| Matt Rowland |
Hi folks,
I have read lots of discussion on speed control for TT dc motors. I've decided to try the OPA548 using back EMF sensing as suggested by one contributer. My question is how can I temperature compensate the circuit to avoid long term drift? I could make an "oven" but do not favour this solution. Ideas and circuits welcome!
Matt |
|
|
| hagtech |
You can use either NTC or PTC thermisters. They'll give you reasonable correction over small ranges. But there is more than copper coil temperature that requires compensation. You also have to take into consideration bearing lubricant temperature.
As things get hot, I believe the bearing speeds up (less friction) and the motor slows down (higher resistance).
Anyway, you'd have to attach the temp sensor right to the motor.
jh |
|
|
| gmphadte |
If u go the tachogenerator way like explained in LM2917, u will get better speed control.
Gajanan Phadte |
|
|
| Matt Rowland |
Thanks for the suggestions. I'm trying to avoid the tacho route if possible. I understand the principle of using a thermistor but am unsure of the best way to incorporate it into the circuit or how to choose component values to make it operate correctly. I've attached the proposed circuit in hope of suggestions!
Cheers
Matt |
|
|
| sreten |
This circuit belongs nowhere near a turntable, :) /sreten. |
|
|
| argofanatic |
LMAO
So that's what TT stands for.
:) |
|
|
| Bobken |
Hi Matt,
I tried that particular circuit out a couple of years ago using the recommended OPA547 (low current version) with a Michell DC motor on a Gyrodec.
Regrettably, after many months of extremely determined experimentation and using the highest quality (very low tempco)components available, I was forced to abandon the idea as the speed was just not stable enough.
It was very disappointing, but using an extremely accurate (and expensive!) LED-based independant tachometer, it showed that the speed was continually 'hunting', and this was the cause of the audible problem. It was clearly a low level 'cyclical' problem, which appears to be inherent in this circuit design, and the author does accept that it is not up to the best tacho designs in its operation. Possibly, for larger and less sensitive motors in less critical situations, it may well be a good design, though.
Over the years I have built 4 different supplies, some for AC and this one for DC motors, and a really good-sounding TT supply is more critical than perhaps one might imagine. In my set-up, doing without the feedback arrangement, which was provided by this chip, improved the sound considerably, and I don't have any problems with speed stability now, without any tachometer feedback or temp. compensation for the motor windings.
Interestingly, I spoke to Michell at that time, and it was more or less admitted that their own tacho supply was not considered a noticeable improvement over their similar non-tacho version, but for marketing reasons, they produce the far more expensive tacho version.
I hope this helps. |
|
|
| hagtech |
| quote: | | I've attached the proposed circuit |
You don't need such a fancy circuit unless you need to vary speed over a wide range. If you're just going to use 33 & 45 rpm, then there is no need to worry about back EMF compensation. The circuit you present is to improve the linearity relationship between voltage and speed. You're only working at one or two speeds.
Best to remove the complex feedback paths. Try just a straight linear regulator.
jh |
|
|
| Matt Rowland |
Good information folks!
I have tried the no feedback straight voltage reg option with a premotec motor (I think the original type Origin Live used). I have found it not speed stable and possibly a bit under powered (as well as mechanicaly noisey).
Listening to the experience of everyone else, going dc seems quite a tricky route. Feedback systems are inherently fraught with timing issues- overshoot, hunting etc. Without feedback there are issues with temperature and setting the speed accurately given the platter intertia.
Origin Live use back EMF compensation on their top PSU and it seems to work according to reports! Teres use a sophisticated "soft" tacho sytem, also supposedly good. The Teres is a bit complex to build and tweek for my liking. I would be interested in seeing the Origin Live circuit. In the meantime I am going to buy a better motor to experiment with, probably the Maxon 110189.
Cheers for now
Matt
PS anyone know where to get pulleys accurately machined? |
|
|
| winggo |
I started out using an AC motor from Hurst. You are right, the big problem is getting the drive pulley the exact size, and I mean exact. The ratio of the drive pulley diameter to the platter diameter is so great that very small changes in diameter on the drve pulley makes a big difference on the other end.
My second attempt was with a DC motor/controller package from Oriental motor company. It was the smallest one they make and cost me $240 or so. The controller is way more complex than necessary and it took a while to get it configured right. It runs beautifully at a very constant speed but you can hear it run from across the room. Way too loud.
My next try will be with a Maxon 110184 and a small DC motor speed control I purchased on EBay. The controller has not arrived yet but I will let you know how it works out.
Good luck
George |
|
|
| Bobken |
Hi,
I don't think there should be any problem with precise pulley sizes, if you use a DC motor. I would have thought with any DC motor it is essential to be able to vary the voltage to set the overall speed accurately, so the exact size of pulley is not very important here.
In fact this was one of the main reasons I finally tried out DC motors, after many different attempts with AC synchronous types where the nominal rotational speed is already fixed by the 'mains' frequency. In these AC cases, accuracy of the pulley size is vital, of course as you suggest, unless one does not care about the absolute pitch of any music being played.
When I was experimenting with AC motors, I used a nicely machined alloy pulley (made by Michell Eng.) This was attached to the motor shaft by a grub screw, but, because of the inevitable (although small) shaft/pulley clearance, when the fixing-screw was tightened, the pulley was pulled slightly off-centre. The run-out was only about a thousandth of an inch (or at most 2 thou.) which I could see under magnification, and could just about feel with a finger nail gently resting on the pulley's indented circumference as it rotated.
To correct this, I eventually burnished the pulley with a wooden coctail stick soaked in metal polish (Brasso) and fixed in a small hand vice, and when this was done until the 'bearing' surface of the pulley was running absolutely true and polished, the sound did improve noticeably.
Subsequently going over to using a DC motor, and with no temp. compensation or feedback, like yourself I was initially very concerned about longer term speed stability, especially following the enforced decision to abandon the feedback arrangement like your proposal.
However, this has remained rock solid for about two years now, much to my surprise and great relief. The speed was carefully adjusted initially when the motor was in use for a minute or so, it doesn't vary regardless of summer or winter nor however long the deck is in use, and I have never needed to re-adjust it since setting it up. Also, I can detect no pitch change due to stylus drag, even on heavily modulated passages, which was another initial concern of mine, and is common with many (most?) AC motors. I believe this must be due to the higher torque of the DC motor I used, and perhaps aided by the excellent Michell main bearing.
Merely removing a couple of thousandths of an inch off the diameter of the pulley with the AC motor (when I trued it up) did noticeably upset the pitch of music being played, and the reduction in speed was readily seen by the stand-alone tachometer/stroboscope I checked it out against. It could only have been running fractions of a % slower, but it was immediately audible to me.
Bearing in mind the inertia of a very heavy platter, and the drive in my case being merely through a circular-section neoprene band of only about 2mm cross sectional area, the effect of such tiny discrepancies like the pulley surface being slightly off-centre was surprising, but it clearly indicated just how critical this is and how important the surface finish and true-running abilities of the pulley are, for the best sonic results. It still amazes me that any vibrations or whatever can be transmitted via a thin rubber band to a substantial platter, to the extent which it does.
Incidentally, since having the use of a highly accurate tachometer, I have checked several other AC driven turntables, and not one was running at the correct speed, at any time, and there is very little one can do about this, from a practical point of view!
So, if you do change over to a good DC motor and a suitable variable voltage PS, and pay careful attention to the surface finish and concentricity of the pulley, I think you will find this to be far more important than worrying about any tachometer feedback arrangements, which may or may not be such a good thing, anyway. |
|
|
| Vinyl-Addict |
| quote: | Originally posted by Matt Rowland
In the meantime I am going to buy a better motor to experiment with, probably the Maxon 110189.
Cheers for now
Matt
PS anyone know where to get pulleys accurately machined? |
I'm frantically looking to get a good motor myself. I spoke with a turntable manufacturer the other day who tried the Maxon 110189 and he said it was not good. Interestingly he uses the same DC drive system as Teres. He did offer to let me try out a Maxon 110945 but I think it's a brush motor.:)
I can machine pulleys in my shop. I posted another thread a few weeks ago for help on a motor build project. I'm willing to trade my machining capabilities for a good working motor, AC with power supply or DC. I just want a good, reliable motor.:)
http://www.diyaudio.com/forums/showthread.php?s=&threadid=70743 |
|
|
| peranders |
Something tells me to avoid switching circuits due to the very weak signals from the cartridge.
One other idea is to use a regular syncron motor and David White's crystal controlled AC source.
http://www.wnaudio.com/turntablepsu.html |
|
|
| Matt Rowland |
Plenty more food for thought guys!
I've decided to try a maxon REMax29, 226756 (about £80!). I 'll try to use the original Linn pulley to start with, but it does mean the drive voltage for the motor will be down in the region of 1.5v for 33.33rpm.
Cheers
Matt |
|
|
| harwoodspark |
Hi Matt, if your going to spend £80 on a dc motor why not just buy the correct dc motor from Origin Live? It comes complete with the correct pulley for your application you just specify.I've had the full OL kit fitted to my Linn for something like four years now and I've never had any problems.It certainly cured the Linn's overblown bass! Mine is the earlier design with the speed control panel mounted in the plinth and the dc supply sits on the floor.The rotary control is in the hole where the original Linn rocker switch used to be.I'm building up another turntable shortly, out of all sorts of spare bits I have but I shall be buying a new OL motor and building the control board myself.The dc supply can be bought from Maplins or ARD for less than a tenner!
Si. |
|
|
| Pjotr |
| quote: | Originally posted by Matt Rowland
Thanks for the suggestions. I'm trying to avoid the tacho route if possible. I understand the principle of using a thermistor but am unsure of the best way to incorporate it into the circuit or how to choose component values to make it operate correctly. I've attached the proposed circuit in hope of suggestions!
Cheers
Matt |
Hi Matt,
The way of speed stabilisation as in that PDF has many drawbacks for a TT. 1st it makes the motor stiffer. The result is that irregularities in the commutation and magnetic circuit of the motor comes out harder as speed irregularities. 2nd it relies on a negative output resistance of the driving circuit to compensate for the DC resistance of the motor. Since the motor resistance is indeed temperature dependant you need to compensate for it. But how do you measure accurately the motor winding temperature, especially when it has an ironless armature?
IMHO the best way to drive a TT with a DC motor is just the opposite. Drive the motor with current so it becomes a torque motor. Now put just enough current in it to keep the platter at the correct speed. That guarantees minimum noise from the motor. For that you need a speed sensor, either a tacho or a pulse encoder,. to control the current and thus the speed. Such a feedback loop is a 1st order loop that has no overshoot or whatever.
Just my 2 cents, cheers ;) |
|
|
| harwoodspark |
| quote: | Originally posted by Pjotr
IMHO the best way to drive a TT with a DC motor is just the opposite. Drive the motor with current so it becomes a torque motor. Now put just enough current in it to keep the platter at the correct speed. That guarantees minimum noise from the motor. For that you need a speed sensor, either a tacho or a pulse encoder,. to control the current and thus the speed. Such a feedback loop is a 1st order loop that has no overshoot or whatever.
|
Hi Pjotr, this is an interesting approach. Could you draw a block diagram, not a circuit diagram, so I can better understand the 1st order feedback loop.My technical abilities only reach so far, I'm an electrical engineer and my abilities are not as good on the electronics side!
Thanks, Si. |
|
|
| Pjotr |
Hi Si,
Can draw a drawing but there are plenty drawings provided in books on control theory.
It is not that difficult. The current through the motor causes a torque inside the motor. That torque is acting on the moment of inertia of the rotating parts with rpm’s (or better Rad/s) as a result. Analogue to an electric circuit you can consider the torque as a current source, the moment of inertia as a capacitor and the rpm as a voltage across the capacitor. So this system behaves as a current source charging a capacitor with a voltage across the capacitor as a result. Such a system is a 1st order system, in fact it is just an integrator (when we ignore losses).
A normal motor driven by a voltage is more or less the same. The current is initially set by the voltage and DC resistance of the motor. When the motor starts turning it is also acting as a generator, generating a back EMF. When the speed is such that the back EMF is near equal to the driving voltage the speed will not go higher because the current then ceases. So there is a kind of speed stabilisation depending on de resistance of the motor. The lower the motor resistance the better it works but also the more the irregularities of the magnetic circuit and armature geometry comes out as speed irregularities.
When driving a motor with current the back EMF becomes inactive and there is no speed control that way, only torque control. So you need a speed sensor (which can be very linear and noise free) to control the current. Now you have also the possibility to tailor the behaviour between the sensor and the current drive with i.e. a PID control or to make the loop very slow.
Cheers ;) |
|
|
| harwoodspark |
| Many thanks for that information Pjotr.The Origin Live speed control I have fitted to my Linn works by sensing the back emf of the motor. When I build my next turntable I will be using an OL motor but I'll try your alternative idea.It does seem the right way to go. Si. |
|
|
| Pjotr |
Hi,
So far so good. But when you connect a rubber belt to the motor to connect it to the platter things becomes more complicated. Now you are adding a spring (the rubber belt) to the motor and the motor system becomes a higher order system. In the electric model in my previous post you are connecting an inductor (the spring action of the belt) between the capacitor of the motor (its inertia) and another capacitor (the inertia of the platter).
But that holds for any motor driving a TT platter, whether it is a DC motor or a synchronous motor. Such a system needs to be well damped, either mechanical or electrical. Mechanical damping is tricky and difficult and can be clumsy. Electrical damping is much easier to accomplish and to tune, but for that you need a control loop anyway. You don’t have that possibility with a simple voltage drive of a DC motor (whether or not with a negative resistance of de voltage drive). The current drive with a speed sensor is the best way to go IMO.
A stiff motor will make the resonance of the belt-platter combo (which is a mass spring system by itself) higher. A weak motor dampens that resonance.
Maybe this is going a bit too far for the moment. The first goal is to keep the noise coming out of the motor itself as small as possible and for that you need a good quality motor in the 1st place.
Cheers ;) |
|
|
| harwoodspark |
| quote: | Originally posted by Pjotr
A stiff motor will make the resonance of the belt-platter combo (which is a mass spring system by itself) higher. A weak motor dampens that resonance.
Maybe this is going a bit too far for the moment. The first goal is to keep the noise coming out of the motor itself as small as possible and for that you need a good quality motor in the 1st place.
Cheers ;) |
Hello again Pjotr, the OL motor I have been using is made by Premotec and is reckoned to be a good quality motor.It is quite small, not very powerful and can be made to run reasonably quietly.The later motors are considerably better in the noise department than the earlier units.I used this motor initially because it was easily available, came with a convex pulley already fitted and the electronics were in the kit with it. I have found another small 12v motor in my spares box today.I think it came out of a good quality cassette deck.It also sports a convex pulley.Also quite small and not very powerful.I'll post some details tomorrow as I've left it at work.If a small motor is needed then this may fit the bill.I may try to put some photos on but I'm not vey PC orientated.I feel I've done well to get this far!
I'm off to bed now, Si. |
|
|
| Pjotr |
Hi Si,
I am drawing an electro-mechanical model of the motor, belt and platter combo and see if I can do some simulations this week to see what it is all about. What Premotec motors are you using exactly (type no, and voltage) and what is the pulley size? Can look up the motor constants then to use.
Cheers |
|
|
| sreten |
Hi all,
I think P is getting his electrical equivalents mixed up.
mass <=> inductance
spring <=> capacitor
force <=> voltage
speed <=> current
So the motors inertia (inductance) is linked by the belt
compliance (capacitor) to the platters inertia (inductance).
And the simplest way to drive a DC motor is voltage drive.
Current drive in a simple form is a very bad idea, as its
ignoring the fundamental principle of the motors operation.
To a first order approximation:
the speed of DC motor is dependent on the applied voltage only, the speed
of the armature gives its back emf and this equals the applied voltage.
:) /sreten. |
|
|
| Pjotr |
| quote: | Originally posted by sreten
Hi all,
I think P is getting his electrical equivalents mixed up.
mass <=> inductance
spring <=> capacitor
force <=> voltage
speed <=> current
So the motors inertia (inductance) is linked by the belt
compliance (capacitor) to the platters inertia (inductance). |
Hi sreten,
You can model it either way depending on if you consider toque as a voltage or as a current.
A capacitor is an integrator for current with voltage as output.
An inductor is an integrator for voltage with current as output.
| quote: |
Current drive in a simple form is a very bad idea |
True, and for that you need a control loop with a speed sensor and then the idea is not so bad ;)
Cheers ;) |
|
|
| sreten |
| quote: | Originally posted by Pjotr
True, and for that you need a control loop with a speed sensor and then the idea is not so bad ;)
Cheers ;) |
Hi Pjotr,
its still not good, using feedback to control a current that
effectively ends up setting the right voltage across the motor.
Its a lot more sensible for the control loop to vary output voltage.
:) /sreten. |
|
|
| Pjotr |
Hi Sreten,
Can you explain that more in depth for me? I don’t see your point.
Cheers ;) |
|
|
| 405man |
Has anyone considered using feedback to operate an eddy current brake on the motor or platter. The Garrard 301/401 uses a permanent magnet who’s position is varied to provide increasing drag to slow down the motor this provides very fine control of the load on the motor by using an electromagnet the braking effect would be proportional to the current and this could be used to control the motor speed as a sort of mechanical shunt regulator. Philips used eddy current brake servos in the N1500 video recorders so precise phase control is possible.
Stuart |
|
|
| harwoodspark |
Hello Pjotr, I'll have to remove the baseboard from my turntable to see which Premotec motor is fitted.I know it's a 12v model and it's the usual unit that OL supply for the LP12. The other motor I have and am thinking of using on my home built turntable is made by MPM Singapore. It appears to have some kind of circuit board built in so may need modifying.I will post some pics shortly.
Si. |
|
|
| Pjotr |
Hello Si,
It's no problem, I will pick a 12v - 3W or 5W Premotec motor from the catalogue. OL is not very specific about the exact Premotec motor as far as I can see on the website. That simulation is more to see what it is all about.
Cheers |
|
|
| harwoodspark |
| Pojtr, Ive tried to send some pictures but the resolution must be too great and they aren't being accepted even one at a time.How do I make them smaller?(acceptable). Thanks, Si. |
|
|
| Matt Rowland |
Hi harwoodspark,
I think the motor you have is the 120-18 series as shown in my attachment. I have one branded premotec which I bought from McLennan Servos who supplied the same unit to Origin Live for their early conversion Kits. The idea was first published in an article in a HI-FI World magazine DIY supplement in whch I think a Rega deck was converted. I find this motor rather noisey and slightly under powered. That's why I am going to buy a better quality Maxon.
Matt |
|
|
| Pjotr |
Hi,
Did look a bit deeper in it. When driving the motor with voltage source, the motor resistance works as a damper for the belt-motor inertia resonance. However with a real motor the motor resistance is of such a value that the motor is driven from a current source anyway concerning the dynamic behaviour. From that point there is not much difference between driving the motor from a voltage source or driving the motor from a current source.
To make this damping work effectively you need a much smaller resistance for the motor winding. For a 12V – 3W Premotec motor (120-18105) it should be aprox 1-2 ohms. With a lower than that motor resistance then motor damping becomes inactive again, but now for the belt-platter inertia resonance.
Cheers ;) |
|
|
| Pjotr |
For those interested in, attached the basic electro-mechanical model of a DC motor. The motor parameters are for a Premotec 120-18105 and were taken from the datasheet from www.premotec.nl
Cheers ;) |
|
|
| harwoodspark |
| Hello Matt. Whilst I am still using the early Premotec motor (OL now supply a later unit, supposed to be quieter.I think it's the CL29/3) I've never really been troubled by noise.I admit it does make a sort of rustling noise if you get up real close but at normal distances it's ok.I only bought the OL kit because it was easy. They weren't as expensive then either! I wouldn't consider buying a complete kit again, maybe just the motor. The thing about the OL supplied motor is that it comes with the correct pulley. What will you do for your pulley? Can you buy or will you have to make? Also do you know the cost of the motor direct from Premotec? I'm making my next tt using a Linn subchassis and the platter, hub and bearing from and old Mission tt.I'll post some pics shortly. Si. |
|
|
| harwoodspark |
| Hello again Pojtr, thanks for all the info.I sent you some pics of the motor I pulled from the cassette deck but since then I've managed to get it apart and expose the circuit board.I'll send you another later. Si. |
|
|
| Matt Rowland |
Hi Harwoodspark
I purchased my premotec motor from Mclennan Servo Supplies.
www.mclennan.co.uk it was about £30-40.
I am about to buy an REMax226756 maxon motor rated at 15watts, 36volts, quite pricey @ £97 inc. As for a pulley I have one about 10mm dia from an old deck, but its got a screw fixing so will need to be upgraded to a push fit, close tolerance type, have'nt found anywhere to get one yet. Alternatively I could revert to the Linn 20mm pulley with the motor running at about 1.4v.
Regards
Matt |
|
|
| harwoodspark |
Hi Matt, I think the pulley is going to be the hard part.It really does need to be accurate.That's why I bought the O/L motor because the pulley was already fitted. It definitely does need to be a good push fit onto the motor spindle.Half a thou out here is too much, so a screw fitting is out of the question. I don't know if it's such a good idea dropping the motor's voltage down so low so you can use the Linn pulley either.If the Linn pulley fits the spindle it would be better to machine it smaller so the motor can run faster. Are you modding a Linn turntable or something else?
Si. |
|
|
| Bobken |
Si is quite right here with all of these comments, and see again my post #12.
Once any pulley is firmly fixed on a shaft, it is possible to 'true up' its driving surface by the means I suggested, but this will be very hard work if it runs more than a thou. or two out.
If the shaft/spindle fit is already quite good, try Loctiting the parts together (instead of using the grub-screw) as this will not tend to offset the pulley, in my experience, as the layer of Loctite will try to form evenly around the annulus between the shaft and the pulley.
If using DC motors (where speed corrections can be made electronically), it is still worthwhile 'burnishing' the final arrangement with any pulley 'in situ.' so that the 'bearing surface' on which the belt runs is polished, and is absolutely concentric and true-running. This will always help with the ultimate sound, however well the parts are originally machined or attached to each other, and is something which DIYers can do which is not so practical in production.
Whatever 'tool' is used for this burnishing must be held firmly in a fixed support, so that it only contacts and consequently 'wears away' the high spots on the pulley, of course, until the entire circumference is perfectly polished. Softwood, well soaked in abrasive metal polish (Brasso) works well for this, and the wood can be readily formed to suit the contours of the pulley before commencement.
It won't harm the motor so long as you don't apply too much pressure which would slow the motor down a lot, and it is the only way of ensuring an absolutely true-running surface, as no parts are ever 'perfect' when made, and they always have some tolerance even though it may be very small. Not much pressure is needed here, anyway, but the wooden tool must be supported firmly, perhaps in a small vice or locking pliers to provide some mass to steady the job.
Although DC motors' speeds can be 'adjusted' by varying the voltage, in my experience it is best to keep to somewhere reasonably close to their nominal voltage for several reasons. If run too fast (higher V) too much heat can ensue, which could cause damage to the motor, and if run too slow (reduced V), the motor's torque will suffer. In the latter case, this is bad for speed stability, especially with highly modulated passages on records where the attendant stylus drag can be a problem, but you may also find that the TT will not 'self-start' at all, due to this much reduced torque!
I don't know anything about Premotec nominal voltages, but if it is near to the more usual 9-12 volt range, I am sure you will have problems running at the suggested 1.4v, unfortunately. |
|
|
| harwoodspark |
| You echo my thoughts Bob. This is why I bought the Premotec motor and pulley package from Origin Live.Whilst I accept that there are better motors available, at least I knew this would work fairly well.And it does too! Si. |
|
|
| Matt Rowland |
Hi folks,
I hear what you say about the low voltage, but the OL kits must run their motors at low voltages given the pulley/sub-platter ratios involved. Looking at the torque/speed curves for the maxon motor it appears that they maintain torque at low speeds (ie low voltage). I think the Teres kit uses a 36volt motor run at a much lower voltage also. If you use a low nominal rated voltage motor the problem of low drive voltage is worse as the speed at rated voltage is several thousand rpm, so you would be running the motor at fractions of a volt to keep the speed down.
TTFN
Matt |
|
|
| harwoodspark |
| Hi Matt. my O/L motor aka Premotek is set at 7.5 volts ie around two thirds of nominal voltage.It turns at 3000 rpm @ 12 volts.It drives the inner hub of my turntable ( Linn LP 12 ). Si. |
|
|
| Pjotr |
| quote: | Originally posted by Bobken
If the shaft/spindle fit is already quite good, try Loctiting the parts together (instead of using the grub-screw) as this will not tend to offset the pulley, in my experience, as the layer of Loctite will try to form evenly around the annulus between the shaft and the pulley.
|
Hi,
That is a good way to go. But be careful, Loctite comes in many flavours. Never use the thin fast curing one. It will for sure creep into the motor bearing and then you have a big problem. Use the thick slow curing variety and one with a not so permanent bond, so you can remove the pulley when needed. But even with the thick variety use extreme care the glue does not reach the motor bearing, so keep the motor always upside and the pulley downside when assembling this way.
Cheers |
|
|
| Bobken |
Hi Matt,
Not trying to rain on anyone's parade, but offering (hopefully) some cautious advice here.
At a quick glance, I don't see anything in the ref. you show which indicates that torque is maintained (or otherwise) with any lower than nominal voltages. Maybe I need to look at the specs some more, but no time just now.
Are you certain you are reading that graph correctly, or is there some more info elsewhere, perhaps, which I haven't seen yet?
The lowest voltage motor shown in the specs seems to be about nominal 7-ish V, and reducing this to 1.4 V is a huge reduction.
Running at 2/3rds nominal V should be fine, but running at 20% or less (7/1.4) is another matter altogether.
I am only going from general experiences with DC motors here, and maybe the manufacturer of your motor has come up with something different, but, for your own sake, please make certain. |
|
|
| Matt Rowland |
Hi Harwoodspark
This is rather confusing! I think I have the same premotec motor running my Linn with a 10mm pulley. My regulator has to be set for about 1.4v to give 33.33 platter speed. Looking at the data for the motor it has a figure of 3.16volts per 1000rpm. Given that the diameter of the sub-platter is 160mm and the pulley is 10mm, thats 1:16 which gives about 533 rpm for the pulley (@33.33rpm).
So I calculate that the motor voltage should be about 1.7v which close to what I've found in practice. Is your 7.5 volts at the motor terminals or what's going into the regulator?
Matt |
|
|
| Matt Rowland |
Hi Bobken
Maybe I'm not understanding the graph correctly, but the red portion is labelled as recommended operating range. However the torque at low rpm is not clear, so you may be right. However as I said before others seem to be using higher volatge motors at lower speeds with success.
Matt |
|
|
| Matt Rowland |
Hello again!
Looking again at the maxon motor data (REMax226756), the speed constant is 185rpm/v which gives an operating voltage of 2.88v for 33.33rpm and the torque gradient is 33.1rpm/mNm which gives about 16mNm at this voltage. So the question is, is this sufficient torque to maintain constant speed?
Matt |
|
|
| Bobken |
Hi Matt,
I'm afraid I don't know the definitive answers to these questions. You would need to check with the motor manufacturers about this, as I am going by personal experience of other motors and intuition here. So far, I haven't seen much to indicate the torque of these motors at lower voltages in the info you mention, though.
What I do know is that now you have specified the motor you have in mind, the concerns I have already advised in this respect have increased dramatically.
Now you make it clear that the proposed motor is nominally 36v, quite frankly I would be astounded to learn that it would drive a TT adequately at a voltage somewhere in the region of merely 5% of its nominal rating.
I know of others (and indeed I do it myself to help reduce motor vibrations etc.) who run at somewhat reduced voltages, but I don't know of anyone who runs them at merely a few % of 'nominal'. The starting torque will be much higher than that needed to maintain motion, and this is what particularly concerns me, as already mentioned.
If you are looking at the red portion on the graph you referred to, I think (I regret I don't know for sure, though, and you do need to check with the makers) that this merely refers to the 'safe operating area', where no harm should come to the motor providing you remain within these guidelines.
I tend to read the comment as "*maximum* recommended operating range" here, I regret to say, but again I could be wrong.
This looks to me to be very similar to the 'SOA' graphs for transistors, which only indicate *maximum ratings* for safety ( i.e. voltage & current), but do not indicate that the transistors will work at their best (or even at all!) in the lower ranges.
I have not played about much with Linn TTs (except to set them up for others) and therefore don't know what their 'running' voltages are, but many years ago I believe their AC motors were powered somewhere in the region of 12V.
I will certainly be amazed to hear that any TT is powered with a voltage as low as the 1.4 or 1.7 volts you mention, but I am always happy to learn something which is new to me. |
|
|
| Bobken |
Pjotr is quite right to advise caution if you use this method to secure pulleys, but it does solve the offsetting by a grub-screw problem quite well.
Don't use Engineering Grades of Loctite (Studlock or its equivalent) here unless you know you will not need to remove the pulley ever. This grade of adhesive can be overcome with applied heat, but this is not ideal for motors, of course.
Of the weaker grades of Loctite, Screwlock is the least strong, and this, or Nutlock (which is the next up in strength) will be more than strong enough.
It is a good idea to leave the parts upside down for a while after assembly, also as Pjotr states, but more importantly, make sure you oil the end bearing well before assembly, as this will prevent any possible creepage into the bearing.
Also, don't apply the adhesive to the shaft, itself, as when you push the pulley onto the shaft, this will force the adhesive down towards the bearing, because the pulley scrapes it off the shaft on its way down into position.
Instead, apply the adhesive (perhaps with a coctail stick, or whatever suits the hole size) all over the inside bore of the pulley, only. Then, when the pulley slides down the shaft, any excess adhesive is merely pushed out of the hole in the top of the pulley, and can be cleaned up with some tissue etc. |
|
|
| Pjotr |
Hi,
Here I posted a model of a DC motor. Speed is set by the difference of the back EMF voltage (as a result of the speed and the voltage constant) and the driving voltage. This difference causes a current through the DC resistance of the motor and depends on the torque needed. So if the requested torque is virtually constant so the speed is. The stabilization of the speed depends on the size of the DC resistance and the torque constant. IMO running the motor at low speed is no problem when driven from a low voltage if the torque requested is low compared to the available torque.
The Maxon Remax motors require very low current at no load which is an indication that the mechanical motor losses are quite low. That 15W Remax motor has plenty torque available to keep the platter spinning at the correct speed I think, even when driven from a simple LM317 adjustable regulator.
Cheers ;) |
|
|
| Bobken |
Hi Pjotr,
This is interesting, and I have no qualms over using generally reduced voltages here, as it can help to reduce motor vibrations etc.
However, Matt is talking about merely using (less than) 5% of this motor's rated voltage, and I question the sense in going as low as this.
An acquaintance brought his DIY TT set-up to listen to in my system some years ago. I don't recall the precise figures, but I believe he was running at about 20/25 % of the motor's rated voltage, and this would not even 'start up' without being assisted by turning the platter by hand.
Also, I could clearly hear pitch changes during heavily modulated passages, caused by stylus-drag induced motor-braking, although the owner said he was not so sure about this, himself!
Obviously it will depend on many factors here, including:
The motor's torque, how this torque is affected at lower voltages, what the rotating mass of the platter is, and how sensitive the listener's ears are to pitch changes, etc.
However, if you are entirely satisfied that this chosen (36v) motor will perform satisfactorily with merely 1.7v (less than 5%), I wouldn't argue with you.
I am rather puzzled, though, by your comment "plenty of torque"....."when driven from a simple LM317 adjustable regulator", as *how* the voltage is regulated doesn't have much to do with the torque available from any motor, loaded in a particular way, and with a specified voltage being fed to it.
Regards, |
|
|
| Pjotr |
Hi Bob,
The torque gradient is set by (torque constant)*(driving voltage)/Rdc at zero speed and at the other end by the no load speed at the driving voltage. As long as the torque gradient of the load, that is from the belt, platter and stylus/groove drag, crosses the torque gradient close to the no-load speed (say between 3/4 no load speed and full no load speed) I don’t see any problems.
No load speed in this case is the speed set by voltage constant and the driving voltage. NOT the nominal no load speed as stated in the data sheet.
Note that most commercial TT’s are driven from a 3W motor and deliver enough torque then, so that 15W Remax will do anyway me think.
Cheers ;) |
|
|
| harwoodspark |
Hello Matt. You are correct in your assumption. I am feeding the regulator with 7.5 volts not the motor itself. I've dug out the O/L paperwork for my dc motor conversion and it states that for the O/L turntable which has an inner hub of 164 mm. the dc voltage at the motor will be 1.98v for 33 rpm and 2.6v for 45 rpm. So the voltages you give for your O/L motor on your Linn must be correct.
I apologise for my mistake. Si. |
|
|
| patriz |
| quote: | Originally posted by Bobken
Hi Matt,
.
Over the years I have built 4 different supplies, some for AC and this one for DC motors, and a really good-sounding TT supply is more critical than perhaps one might imagine. In my set-up, doing without the feedback arrangement, which was provided by this chip, improved the sound considerably, and I don't have any problems with speed stability now, without any tachometer feedback or temp. compensation for the motor windings.
Interestingly, I spoke to Michell at that time, and it was more or less admitted that their own tacho supply was not considered a noticeable improvement over their similar non-tacho version, but for marketing reasons, they produce the far more expensive tacho version.
I hope this helps. |
Bobken
do you have a link or schematic of a open loop dc motor speed controller with no feedback? |
|
|
| Bobken |
Hi patriz,
I have only just seen your query.
Unfortunately I don't have any links which I know would help you here, but I have seen several schematics over the years on the Internet, so I think you will need to search for yourself.
I am not at liberty to post my own design here, which has been the result of several years in development, as I sold the rights to a TT manufacturer a while ago.
What I can say is that it is a straightforward two-stage voltage regulator, starting off with an over-specified toroidal t'former, soft recovery diodes, and non-polarised Blackgate smoothing caps.
This leads on to the pre-regulator, which is a 3 terminal LT device, again using Blackgates and Vishay bulk foils for further smoothing and output voltage setting etc. before the main regulator, so that the voltage 'dropout' in the main reg. can be tightly controlled.
This then feeds a very well-smoothed, low-noise discrete regulator of my own design (using an LM329 voltage reference as the control element) and, again, using all Vishay & Blackgate components, with a Vishay bulk foil 20 turn preset for accurately setting the final voltage used for the motor.
Whilst under development, I initially used 'lesser' components in several iterations, and none of these performed as well as the final version. Therefore, I have no doubts that its excellent performance in final form is due to using these very low noise/low tempco etc parts, which initially might seem like overkill.
I have never needed to re-adjust the output voltage in the prototype since initially setting it up, and I know this is mainly due to the lack of drift in specs of these carefully chosen parts.
Personally, it is my belief that using a very good non-feedback design which produces a clean and stable voltage at all times with an adequate current reserve, is more important for good sound than using any tacho style arrangements.
I appreciate that this is a contentious subject, but as with power amps etc., no feedback arrangement can ever be truly instantaneous, and this can lead to 'flutter' occurring, which I don't care for. Also, as time goes by, I have enjoyed better results throughout with my own electronic circuits, the more I have managed to simplify them.
What happens in reality with feedback circuits is, the TT begins to slow down as a result of varying stylus drag, caused by differing modulations in the recordings. This then needs to be 'sensed' via (in my case) an elastic belt by the motor (either through its back EMF, or maybe a separate tacho coil) and this braking effect then needs to be transmitted via a cable to the PS sensing circuit, before any reaction can even be contemplated by the PS electronics.
Having sensed this additional drag, the PS circuit reacts to increase the voltage accordingly, this is then transmitted back down the cable to the motor, which then needs to overcome its inherent inertia and speed up, and then transmit this speed increase via the same elastic band back to the TT, which again needs to overcome its own inertia (often quite high) before regaining its correct speed.
Although the electronics will react quickly (but certainly not instantaneously!) my experiences suggest that the other variables do not, and this inevitable 'time-lag' is a problem which all such circuits will suffer from, to some degree or other.
In other words, the correcting 'signal' will always be too late to be fully effective, and hence the hunting effects I have seen and measured do occur.
The only way of successfully avoiding this potential, is if one could 'read ahead' somehow, and anticipate any changes in TT speeds *before they actually occur*, and then applying the correcting forces at the right time so there is no net change in TT speed at any time. However, although this procedure might be easy in digital cases, I simply haven't come up with any effective way of doing this in the analogue domain, unfortunately.
Incidentally, even the quality of wiring used between the motor and TT will affect the overall sonic results, so, as I have said before, this area is more sensitive than many people appear to appreciate.
I hope this helps. |
|
|
| Pjotr |
Hi Bobken,
Keep in mind that the motor is kind of a feedback system by itself. In the motor the back EMF is in fact a speed sensor that is subtracted from the driving voltage and as such giving an error signal to set the motor current. This feedback is not instantaneous, but delayed by the inductance of the motor winding.
IMO the only optimum way to control the speed of the platter is by using a controlled braking mechanism at the platter itself like an controlled eddy current brake as suggested somewhere else in one of these threads.
But I agree that at good simple voltage drive can give very good results with a high quality motor.
Cheers ;) |
|
|
| pixpop |
| quote: | | What happens in reality with feedback circuits is, the TT begins to slow down as a result of varying stylus drag, caused by differing modulations in the recordings. |
Bob,
do you have a sense of the magnitude of this drag? Is it meaningful to express it as a torque, with an RMS value, for example?
Of course, I understand it will vary depending on stylus geometry and material, as well as the composition of the vinyl.
Neil |
|
|
| harwoodspark |
| I have read bobken's post with great interest and have to agree with what he says. As I see it, which ever method of speed regulation you choose, there will always be some delay in speed correction.And what about the belt? This will be expanding and contracting in sympathy with the change in loads on the platter and making matters worse. Maybe a direct drive TT is not such a bad idea after all. Therefore the only way to deal with this problem is to have a platter of sufficient mass that stylus drag is not an issue, use a motor with just enough torque to keep that mass revolving at the required speed and use a psu that maintains a rock steady output without any other external interference. Does this make sense? Si. |
|
|
| Bobken |
Hi Pjotr,
Yes, these are some more sensible comments, although I have not really found the need to consider this suggested "controlled braking system" you mention, so far.
By whatever means the desirable accelerating or braking effects needed to maintain absolute speed control are carried out (to overcome the inevitably varying drag) it seems to me that there will always be some inherent delays in its operation.
Until the platter actually slows down, and then allowance also needs to be made for the stretchy driving band to 'catch up' and similarly affect the motor's speed (which initially will attempt to maintain speed due to its own inertia, anyway), the 'sensing element' will not know that some correction is needed.
This is so even if this sensing is done directly at the motor, itself, of course.
By this time, it is already too late, unfortunately, and this completely disregards the additional 'reverse' delays with the need for the motor to subsequently react (against its present established inertia, commensurate with its then slightly reduced speed) and via the same stretchy belt speed up the platter again, and overcome the platter's then established inertia, due to its also then lower speed.
It appears to me to be a "no-win" situation, unless, as I said, one could somehow ideally 'read ahead' and anticipate these speed changes, which would be another matter, altogether.
It is much the same with the amplifier analogy I mentioned, except that, with purely electronics circuits, their 'reaction' time is many orders of magnitude faster than that of quite high mass rotating physical bodies, with their also inherently quite high inertial effects.
Perhaps I have misunderstood you here, but I am guessing you refer to some kind of controlled 'magnetic influence' working against the (possibly speed-varying) magnetic forces of the motor, itself, which are probably caused by its back EMF.
I have certainly heard of magnetic 'brakes' to control TT speeds, but if I recall correctly, these are 'static' rather than 'active' devices, simply used for overall speed control.
On a slightly different note, I well recall some of the early Bang & Olufsen decks which I looked at for repair/modification, over 30 Yrs ago.
Their speed was adjusted/controlled by a very crude arrangement of a felt pad which was forced against a ring which rotated along with the platter, and by way of a knob for adjustment, the felt pad was pressed harder or softer against this ring to control the speed!
I imagine they might have described this in their impressive and colourful brochures at the time, as a "Controlled Braking System", but this is rather different from what you have in mind. |
|
|
| harwoodspark |
| Does the Garrard 301/401 not use an aluminium disc mounted under the platter as an eddy current brake or whatever you wish to call it? Si. |
|
|
| jo-briggs |
| Just an idle question, do any of you listen to the music, or do you just sit there watching rev counters? Absolute speed is not usually an issue unless you have perfect pitch, and few do; speed variation is the bigger worry. |
|
|
| Bobken |
Hi Neil,
No, unfortunately I don't have any way of 'quantifying' this effect, or at least so far I have not attempted to do this.
However, two points I raised on the Forum recently are that when I polished off no more than a thousandth or two of an inch from a pulley which was (IIRC) about 3/8" to 1/2" in diameter, I could clearly hear the change in pitch with the same synchronous motor I was then using.
This simply confirms my comments that this area of interest is incredibly sensitive to even quite minute differences, as the nominal speed of this motor was, of course, governed by the mains frequency and shouldn't have changed.
The other point was that a few years ago I purchased an extremely accurate self-contained stroboscope/tachometer, which did not rely on mains frequency, or whatever, for its operation.
It has a series of maybe 50+ leds around its circumference, which light up (flash) at pre-determined frequencies, controlled by an accurate quartz crystal clock, and this is very revealing when used whilst records are being played.
It then became easy to 'see' what effect the heavily modulated passages have on the TT's speed, and to watch how things change, and it was only after obtaining this device that I made any further real progress with my PS designs.
This finally confirmed precisely the effects I was earlier hearing, but previously the reasons for these abberations were not quite so clear, and still left some doubts in my mind, as the degree of these effects were not initially at all intuitive to me.
Maybe a good analogy here would be with viewing signals on a 'scope, where it is infinitely more clear what is happening to a signal, which subjectively does not sound as good as it might.
I have ended up with a fairly high mass platter, with a lot of the mass concentrated at the outer periphery, to deliberately encourage a flywheel effect, and this tends to 'plough through' any short-term attempts for the stylus drag to reduce the overall speed.
With this arrangement, allied to a high torque motor fed with a very smooth and clean DC voltage, which is also extremely accurately regulated, I have achieved the best results so far.
I still have serious reservations over using very low DC voltages here, as expressed earlier in this thread, as I found that keeping extremely low voltages so well regulated is less easy than higher voltages, all else being equal.
Hi Harwoodspark,
I hadn't read your post when I last contributed in response to Pjotr, but you will now see that I agree entirely with your comments on the relatively elastic nature of flexible belts, and their own attendant problems.
I have also experimented with duplex belts, in an effort to reduce any 'stretching' abberations, and this does help, somewhat.
However, nothing in life is free, and although this gave some benefits, I generally preferred the single belt, overall, in listening tests.
As they say, the jury is still out on this one, though, and I will try some more with this arrangement at a future time, as it is one way of reducing the otherwise (albeit minor) inevitable belt-stretching effects. I cannot hear any pitch changing effects with my current set-up due to any belt stretching, but common-sense tells me that on longer term heavily modulated passages, this must happen to some degree, and I would like to eliminate this possibility.
Presently, I am inclined to try out some other belt materials first, of a less stretchy variety like nylon cords, but joining these is not the easiest thing to do I have found, without ending up with a lump somewhere on the belt where it is joined.
This, unfortunately, tends to give the platter a minute 'kick' every time it passes around the quite small motor pulley.
As you will also have seen, your comments about using a TT with "sufficient mass", and a "rock steady output" from the DC supply make a lot of sense, and to me these are far more important attributes to the sound than any feedback or servo attempts.
Incidentally, as with preferring not using very low voltages, I don't see any reason to restrict the motor's torque, as in my book, the more torque you have (within sensible bounds), the less any stylus drag will affect the resultant speed.
I now need to pinch my TT's belt (when it is running between my fingers) quite hard before there is any noticeable speed reduction, and this is entirely down to the higher torque motors I have progressed to.
Any such pinching needed to affect the TT's speed is many times (orders of magnitude, I suspect) greater than any stylus-induced drag, I believe. |
|
|
| pinkmouse |
| quote: | Originally posted by Bobken
The other point was that a few years ago I purchased an extremely accurate self-contained stroboscope/tachometer, which did not rely on mains frequency, or whatever, for its operation.
It has a series of maybe 50+ leds around its circumference, which light up (flash) at pre-determined frequencies, controlled by an accurate quartz crystal clock, and this is very revealing when used whilst records are being played. |
Sounds interesting, do you have a link? |
|
|
| Bobken |
Hi Jo,
Its a relief to see that you have been paying such close attention, as it makes the effort all worthwhile!
"Speed variation", mainly due to stylus drag, and trying to eliminate it, is precisely what most of this has been about for several days.
Personally, I haven't been able to enjoy listening to a record for many years due to all the flashing leds I need to keep a close eye on!
Regards, ;) |
|
|
| pixpop |
Bob,
with your LED strobe, do the bars on the strobe disc have sharp edges?
I've been using a traditional neon strobe, synced to the AC line. You can tell if the bars are moving left or right, but you cannot use it for detecting flutter, as there is too much jitter in the flash of light, which makes the edges of the bars appear blurred. I've been toying with the idea of making a strobe with LEDs, maybe now is the time. I notice a few such DIY projects on the web. |
|
|
| pinkmouse |
| Bob, your link's broken! ;) |
|
|
| pixpop |
The link worked for me.
I see that your device doesn't even use the strobe disc. You just look directly at the LEDs. Very clever. |
|
|
| Bobken |
Hi pixpop,
You will see the device I use in the thread above.
I know exactly what you mean here, and I gave up a long while ago with any mains controlled devices, as the mains is not sufficiently accurate quite a lot of the time.
In the UK this is especially so late at night (which is my preferred listening-time) when the Electricity Authorities make up any 'lost time' (during the daytime) so that clocks still keep good time, overall!
Copying a Gyrascope would not be an easy DIY effort, I regret to say, and the blue led version I have cost about £350 ($600+), several years ago.
This cost was only justifiable as I was involved in a commercial development at that time.
Edit for Pinkmouse. It works OK for me, as I just tried it out immediately after I posted it. |
|
|
| Bobken |
Hi PM,
Have you cracked it OK now? I just tried the link again, and it seems fine to me.
Hi Arne,
Yes, that is certainly a lot less costly, as you say. I wouldn't like to speculate whether or not the abberations I was concerned with could be readily seen on this tacho/strobe, although I am sure it would enable accurate overall speed settings to be made.
I wasn't aware of the Hagtech device when I bought the Gyrascope, but I was under pressure to finalise a design, and I went for the most accurate device I could find, anywhere.
It did the job well! |
|
|
| Cobra2 |
I guess the Gyrascope also function as a clamp/weight, it looks good, and would be an interesting option for older (non-quartz) DC-motorized TT's, like MicroSeiki...
Arne K
For speed-variation checking, you will find wow & flutter-meters on ebay for "next to nothing" |
|
|
| Bobken |
Hi Arne,
The last time I used one of these meters, it was necessary to also use a test record with accurately recorded sine waves on it.
Otherwise, I don't know how you can get any meaningfull readings from a TT set-up to measure with a wow & flutter meter, but perhaps you can enlighten me.
Irrespective of the levels and frequencies of these test signals, they are recorded effectively as 'steady-state' signals, and therefore their modulation does not vary during each track.
The problem with this arrangement is that it precludes you from carrying out the kind of tests I am interested in and have been discussing, which are mainly how the TT speed is affected *during heavily modulated passages* on 'normal' recordings.
I have looked at waveforms etc. on 'scopes, and tried several other ideas, but so far, the Gyrascope has proved to be the most illuminating tool I have used in this connection.
Regards, |
|
|
| pinkmouse |
Got it now, don't know what happened there.
It's certainly an impressive piece of kit, but the Hagtech one looks a little more feasable for home construction. :) |
|
|
| Bobken |
Hi Al,
Glad you made it!
Actually, I tried the link again this AM, and I had difficulty navigating the site, so it looks like they may be having some intermittent problems.
As you say, it is a very well-made piece of kit, but so it should be at the price!
It does, however, exactly as Lys claim "takes into account the drag of the stylus attached to the cartridge", and I have not (yet!) found anything better for this purpose.
Its a silly cost for DIY'ers, and maybe the Hagtech device would be worth experimenting with. It all depends on how clearly the strobe effects can be seen, as Neil previously discovered.
Also, with any unit which has quite heavy batteries etc. within, it is important to ensure that the final 'set-up' is statically balanced as well as possible. Otherwise, when it is rotating on the platter, out of balance rotational/centrifugal forces will tend to induce wow on its own, and this will very likely adversely affect the results.
Regards, |
|
|
| Matt Rowland |
Hi folks
I've just caught up with the latest postings which interestingly reflect my recent experiences- stylus drag etc.
I just got my new motor, REMax 226756. Its quieter than the premotec which is a good start! As I still haven't got a good quality 10mm pulley I'm using the Linn 20mm one and running the motor at 1.6v. I am using a 317 reg mounted right at the motor terminals. The sound is good, however, the speed is not constant when viewed using a strobe disc. I estimate that there is a increase in speed of about 1.5-2% from the first to last track of the record as reflected in a voltage increase at the motor terminals of this magnetude. I was suprised as I expected the regulator to easily cope with any current the motor was likely to draw, also I thought that given the heavy platter of the Linn stylus drag would be smoothed out. I'll try a better regulator, as the 317 is down near its lower operating limit (1.2v).
Cheers
Matt |
|
|
| pixpop |
Matt,
did you measure the speed difference with the stylus in contact with the record, and not in contact?
Also, do you know the mass of the rotating parts? |
|
|
| Bobken |
Hi Matt,
Sorry to hear of these speed drift problems, but I am not entirely surprised.
Inconvenient though it may be, I am still quite certain that you will have better results with a smaller pulley and using higher voltages.
As I suggested in post #62, such low voltages are not so easy to control accurately in my experience. Also, and more importantly, whatever regulator you end up with, it will not be perfect (nothing in life ever is!) and it will have some inherent degree of accuracy in maintaining a set speed etc, and a 'noise' level, neither of which are good for the best results.
Suppose for argument's sake, the output accuracy of the regulator will allow a drift of 10mV, and you are running at 1volt nominal for the motor. This is a 1% variation here.
However, with an identical 10mV drift with the same regulator, if one is using 10 volts to drive the motor, this inaccuracy reduces by a factor of 10, down to merely 0.1%.
Have you tried out deliberately braking the platter gently by hand, either by pinching the belt (if it is accessible when the TT is in use) or perhaps with a finger on the platter's edge?
This is crude, but should give you some idea of how resistant the arrangement is to drag effects generally, together with some rough idea of how much torque you have currently. |
|
|
| Pjotr |
| quote: | Originally posted by Bobken
Hi Matt,
Sorry to hear of these speed drift problems, but I am not entirely surprised.
Inconvenient though it may be, I am still quite certain that you will have better results with a smaller pulley and using higher voltages. |
Hi Bob,
Yes a smaller pulley will help a lot here. The smaller the better, I would opt for a 8 mm one.
The LM317 regulators are pretty good with quite low noise. Long term stability is not that an issue when properly heat sinked. What is important when driving a motor is that these regulators do not “starve of current”. That can be a result of motor irregularities especially due to the back EMF. When they starve of current they present a high impedance to the motor and electric motor damping is lost. It is best to load the regulator with a bleeder resistor (in parallel with the motor) that draws a current of app. 1/2 the steady motor current.
Cheers ;) |
|
|
| Bobken |
Hi Pjotr,
Yes, that's a good point about pre-loading (any) regulators with a resistor at the output, which I omitted to mention, myself, even though I use them and find them beneficial.
I confess to being puzzled by some of Matt's comments though, as an increase in RPM of 1.5 to 2% simply across a record is horrendous in my opinion, and suggests almost negligible torque from the motor, or something similar. I have never encountered anything like that in my many years of experimentation. The surface velocity at the inner grooves will, of course, be less than at the outer rim (smaller circumference here but the same RPM) and consequently the stylus drag would be marginally less here.
However, for this to noticeably affect the speed at all would cause me to think that something is very wrong, somewhere. I even wonder if I have understood what Matt said correctly here.
I would guess that any changes in drag would be far more prevalent during heavily modulated passages, than any overall differences in drag across the record, so goodness what this will be doing to the pitch of music when replaying 'enthusiastic' passages!
Unless Matt is using a 317*L* (100mA max.) the other suffix 317s should provide somewhere between 500mA, and 1.5A, which should be plenty of current available for such motors.
Off the top of my head, my present motor takes only a few tens of mA during normal running, although it does draw a substantially higher current momentarily during start-up, as I mentioned as a caveat earlier on in this thread.
Regards, |
|
|
| Matt Rowland |
Hi Chaps
Thanks for the info. The increase in speed is visible on the strobe from outer to inner part of the groove. This is matched by an increase of voltage at the motor terminals of 1-2%.
I will try to get a smaller pulley as I can see how this will improve matters. I am busy with other stuff at the moment so will probably not be able to report back anything new for a while.
Cheers for now
Matt |
|
|
| planet IX |
Hi all
I’m experimenting with a Maxon RE-max29 (22W/48V) for about two years now and I was not able to find a satisfying setup so far! I have a very good quality 10mm pulley. I tried all possible combinations of belt/strings/tapes and voltage regulators (LM317, LT1083 and L200 (L200 circuit involving motor current)).
Next I will try a much weaker motor (A-max with 4W). This is an advice of the guy who gave the pulley to me and he is very pleased with his TT-drive…
I’m really tiered of experimenting with motors and if the 4w-Type won’t work as I want it, I will disassemble my old rega P3 and use the premotec motor (driven by a frequency generator)
Regards, Boris |
|
|
| Bobken |
Hi planet IX,
You don't say what your problems are here, unfortunately, merely that you are not (yet) satisfied with the results.
Would you care to elaborate on this dissatisfaction, perhaps, and suggest whether this is due to speed stability difficulties, poor resultant sound, or whatever?
Otherwise, is it not possible to make any valid comments about this.
Regards, |
|
|
| planet IX |
Hi Boken,
sorry, but I thought that it was clear that we're talking about the speed stability of the platter :rolleyes:
... and they were audible in all my setups that I tested (concerning the Maxon motor). I always checked the speed with a strobo-disc, but most of time that wasn't even necessary- on some records I could hear the speed-fluctuation immediately
...
Regards, Boris |
|
|
| Bobken |
Hi Boris,
There's absolutely no need for any apologies, and you are quite right that (most of us) have been talking about speed stability issues here.
However, whilst I (and most other posters, I think!) have been more concerned with short-term variations caused by stylus drag during heavily modulated passages on records, Matt has just surprised me with some different concerns related to speed changes across the entire record.
I wondered if your problems were similar to Matt's, as this is not a phenomenon I have come across before, and suggests to me to be due to a more serious deficiency in the motor and/or PS set-up. :bawling:
Regards,:) |
|
|
| planet IX |
Hi Bobken
No, I have never had any problems with speed stability over the side of a record, like Matt reported. This sounds very curious…
My anger are the short-term variations. But funnily enough I could not say that they appear only at heavily modulated passages… They seem to appear coincidental, regardless what record is playing and if the passage is quiet or loud or so.
The output-voltage of the LT1083 is absolutely constant and I believe that a LM317 will do the job just as well.
Boris |
|
|
| Pjotr |
Hi Boris,
Point is that virtually all motors exhibit some torque ripple, the one more than the other. Problem is to find one with the lowest ripple at the speed needed. The properties of the belt are crucial to dampen the ripple. Not only should the belt have a proper springiness but also a proper damping to absorb the short time speed irregularities. The more poles a motor has, the higher the ripple frequency and as such it is better absorbed by the belt and platter inertia. Has anyone ever looked at the PML pancake motors?
http://www.pmlflightlink.com/motors/GP.html
And all for all a synchronous motor with its 50/60 Hz ripple frequency is still a pretty good one but lacks simple speedcontrol.
Cheers ;) |
|
|
| Bobken |
Hi Boris,
Now I see.
Certainly the LT1083 should be adequate for this job, and these LTs are generally better than LM regs., so this is a real puzzle, especially as you suggest it is not consistent.
It still suggests to me that there is either a lack of adequate torque here, or is it possible there is some 'outside' influence, perhaps?
Does the TT (without any belt to the motor) spin freely and run on for a long time if you rotate it by hand?
Also, have you tried 'braking' the TT by hand (gently) when it is running to see how much drag needs to be applied to replicate this effect? This might indicate how much drag is occurring (if it is drag which is the cause) because if a lot of braking effect is necessary, it would look more like a PS problem to me.
The amount of drag caused by any stylus, should be very little in reality, and in my experience it will tend to increase consistently when playing heavily modulated passages.
I don't know what else to suggest right now, but I will give it some more thought.
Regards |
|
|
| planet IX |
Hi Pjotr,
I read this thing with the low torque ripple in a german dissertation (!) about turntables once. But no motor-manufacturer mentions its value in the motor-datasheets…
Furthermore I read in this dissertation, that one way to lower torque-ripple is to drive a motor with a very low voltage.
As I said, I tried out a non-slippery and stiff tape instead of a rubber-belt. The idea behind it is (as I was told), that a dc-motor has some kind of self regulation: when the platter is slowed down caused by the stylus drag, the motor-current will increase coevally, what in turn increases the torque. In this way the platter-speed should be constant when using a tape. But in my case this didn’t work…
These PML-motors look very good. Probably I will try to get one, if they fit to my pulley.
Boris |
|
|
| planet IX |
Hi Bobken,
the motor I use has plenty of torque and the platter spins for minutes if I rotate it by hand without the belt (It’s a Scheu Platter and bearing).
If I break down the platter by hand it speeds up relatively quick and if I do it to heavy, the tape or belt slips over the pulley… So far everything as it should be. |
|
|
| Pjotr |
Hi Bob,
Do you have any idea how much the drag of the stylus is? I did some simulation with a 4 kg platter and an ideal running motor (that is, with a rock solid speed). The temporarily speed deviation due to the inertia and the belt will be then app 0.1% at a change of the drag of 1 gram (at the outside of the record where the effect is strongest). With a heavier platter the speed change will be lower but the duration will also be longer. To halve the speed change you will need a 4 times heavier platter, this relation goes with the square root of the mass of the platter.
@Boris,
Yes, that pancake motors look very interesting but these are quite expensive. They are available from Farnell and will cost around 130 eu
Regards ;) |
|
|
| pixpop |
Boris,
do you see rapid speed variations (flutter) or slow speed variations (wow) or both?
Do you also see the same speed variations when no record is playing, or only when a record is playing?
Can you try driving the motor directly from some batteries, to eliminate the regulator as a problem?
Neil |
|
|
| calucci |
| will someone post a decent DC motor controller PCB? Pick a decent brushless motor, somewhat strong to drive a 3+ kg platter and accompany it with a good controller. Let's build it then and then comment on it... we can improve it later. you need 500 euros to get something decent from the market. its ridiculous... |
|
|
| planet IX |
Hi Pjotr, if I’d know the problem eliminated with this motor, I would buy one. But at least I have no pulley for the axis of this size…
What you say about platter-mass and speed deviations might lead us to lightweight platter in combination with a strong motor, doesn’t it? But then the torque-ripple of the motor becomes more a problem, if I understand right…
Hi Pixpop, particularly I hear these speed variations. So far I never listened to the music and simultaneously observed the strobe-disc.
There are no speed variations without a needle on the platter (possibly they are there, but then much lower).
The same is for battery-supply. I’ve tried really everything possible.
But now... finishing time for me today and time for a beer:cool: |
|
|
| Bobken |
Hi Pjota,
No, unfortunately I don't, or at least I haven't measured it for myself (yet!).
Actually I have been thinking about this issue for a day or so, and I wonder if the following would be worthwhile trying out.
Fortunately, I don't have any of these problems, and right now I have rather too many other matters to attend to, to try out such an experiment, myself.
However, what I would suggest is to obtain (or maybe make!) a spring letter-balance, the kind which used to be in use for weighing individual letters before posting. I haven't seen one for some years, but basically it consists of a suitably weak coil-spring with a hook at the ends, and one end is fixed relative to a slide along which the spring extends.
This slide has graduations marked along it, and according to Hooks Law, the extension of the spring (within sensible boundaries) will be in proprtion to the force (the weight of the letter) acting upon it. Something similar could quite readily be made by a DIYer, I think, and calibrated using gravity by reference to some known weights.
I would then attach a long piece of thin nylon fishing line/sewing nylon, or whatever, to the rim of the TT with some adhesive tape, and wind this several times around the outer rim of the TT, possibly where the belt rides normally.
Then, by attaching the balance to the other end of the nylon, one could rotate the platter by pulling on the nylon, by way of the balance, to see how much 'pull' is needed (read off the scale) to rotate the TT with the cartridge in place.
This is rather crude and probably not very accurate, but I still suspect it would be quite revealing, and the biggest problem to overcome is the inertia of the set-up until the TT gains speed to 33.3 RPM.
Accordingly, it would possibly need maybe ten or so windings of nylon to enable the TT to be got up to speed (check this with a strobe) and a helper to keep an eye on things to avoid too many tangles! Clearly some practice would be needed too.
The tester would need to walk steadily away from the set-up with the balance held carefully in his hand and watching the scale, while the helper reads the strobe, and gives instructions about the TT's speed.
Depending on how this initial exercise turns out, it might also be possible to make a comparison without any stylus pressure as well, and this could provide a 'difference' which would indicate the true stylus drag, itself. However, this may be expecting rather too much of this basic idea.
Before you laugh too much at all this, it would be (almost?) free to try, except for needing a lot of patience, and I have had some success with several similar situations in the past, believe it or not. After all, we are talking about DIY, aren't we?
Doubtless I will try it myself at some future date, but if anyone else does so in the meantime, I would be interested to learn if it was successful.
Regards,:bigeyes: |
|
|
| pixpop |
| quote: | | There are no speed variations without a needle on the platter (possibly they are there, but then much lower). |
Boris, I think you should verify that the record is not slipping on the platter/mat. Perhaps try a different mat, or add a weight on top of the record. | | | |
