| sek |
Hi,
this is an unusual request, as I have found nothing about it on the internet.
I want to improve/modify a turntable motor regulation circuit.
I know I could buy good circuits and motors, but the player I want to rescue/improve is something special:
- it's my first player,
- it's as old as me, every part inside was built in the year I was born. ;)
The ugly properties I'm dealing with: a Technics SL-2000, an actual cr@p turntable... :(
I know I could buy 'contemporary' Technics gear for the sake of sticking to direct drives while having an easier start. But hey, it's a DIY project... ;)
Anyway, the SL-2000 has to be improved, if only for the learning experience.
Sorry about the rather long introduction, now to the description of my first problem: the motor control circuit.
I've reverse-engineered the control circuitry up to the point where I'm lacking further documentation that can't even be sourced from Panasonic any more!
About the working principle: the motor appears to be a 'brushless' three-phase 'DC' unit with a built in tacho generator (Matsushita 'MKL-15SI-T'). This unit is enclosed in a servo regulation, based on an integrated motor controller IC 'AN620' (Matsushita). BTW, can anyone supply information on this one, perhaps a modern substitution (does not have to be pin-compatible, but better ;))? I've added a picture of the motor internals, you can see three layers of drive windings and three small pairs of tacho 'pickups'.
The problem with it's current control unit: speed sagging under load, bad regulation and too low torque (even the needle slows the platter...).
Some constraints regarding the final result:
- I want to use lightweight platters (because I want to keep the look-and-feel and can't add infinite torque to the motor).
- And I want to keep it being a direct drive unit with it's oil bearing, of course, as otherwise I could just throw it away and buy better parts...
Some things go without saying:
- it will earn a stiff power supply (the old one sags and humms under load);
- a new base (heavy and unsuspended, because it's easier for now);
- someday a better tonearm, etc.
But how should I approach the improvement of the motor controller?
I have some ideas of course:
- PLL regulation (actually easy to build, but difficult to integrate into the old regulator),
- separate amplifier output stages (to drive with higher power), etc.
Are you still reading? Very good. Read on in the next post, where I'll continue with the controller circuit description. |
|
|
| sek |
I've grouped the interconnects into four connectors (cable colors matching the original setup):
J1 - power supply, 18V DC (upper left conrner);
J2 - speed selection and pitch setting (on the very right);
J3 - 'S' windings, supposedly the speed meter (left);
J4 - 'A' windings, supposedly the actuator (lower right);
In the center of the drawing sits the AN620, which I know nothing about regarding it's internal structure, as I couldn't find any documentation about it.
The circuitry on the upper right (connected between power supply and pitch regulation input) appears to be some filtering and referencing/biasing.
I haven't measured/read the resistors yet, as they are of no particular importance as long as I don't know what the AN620 does whith what value.
D4 (upper right) *might* be a Zener, but it's not marked or printed on.
D1..D3 are of yet unknown type (long glass body, nothing printed on), but definitely connected this way.
What the RC network R5/R10 and C13/C14 (top, between +18V and pin 16) is for? I can't tell without knowing what pin 16 is for...
There is a wire '?' along the 'S' connections, which has no component print and the actual purpose is not yet obvious.
I've studied the schematics of the Technics SP-10MK2 and the SL-1200/1210MK2 for comparison. It's clear how everything is approached (servo loop with frequency-to-voltage conversion, enclosed PLL with quartz reference and phase comparision, higher-torque motor with appropriate drive), but turning a mediocre controller circuit into a better-than-average is not easy for me...
Does anyone have an idea?
Is it possible to increase the output power (haven't measured the motor yet, but a little more should work easily)?
Is it possible to improve the regulation (haven't replaced the capacitors yet, but they appear to be allright)?
Or is the SL-2000 normally superior to what I describe and mine is broken? ;)
Any comments are appreaciated, thanks for reading! :cool:
Sebastian. |
|
|
| kevinkr |
| From what you say it sounds like the motor and controller are not working properly - I would fix the basic problems first and address upgrades after you are sure the motor itself is ok. It is unlikely at this juncture that you will find an improved motor controller IC for this application that is pin for pin compatible, plus I am sure it can do the job well as long as it is undamaged. Far more likely at this time that some or many of the electrolytics are bad and need replacing. Pots and some resistors may be suspect too. The motor itself is not too likely to respond well to higher current drive, the windings were designed to dissipate only so much power. I would focus on making sure that the current circuit is actually working properly. One area where you might make a big improvement is in replacing the power transformer with a slightly larger one, add more capacitance to the supply, perhaps consider voltage regulation to the servo using something like the LT1085, set to 18V to this I would add hex freds or schottky (if piv is high enough) rectifiers and maybe 10000uF 35V nichicon muse electrolytic after the rectifiers. |
|
|
| sek |
Hi Kevin,
thanks for your reply ;)
Do you know the SL-2000?
| quote: | Originally posted by kevinkr
From what you say it sounds like the motor and controller are not working properly - I would fix the basic problems first and address upgrades after you are sure the motor itself is ok. |
The behaviour didn't change (worsen) since many years. And the motor seems okay. But I plan on progressing slowly and only where I'm sure what I'm doing. So I plan to use the current circuit as long as I don't have a better solution. I could indeed try to 'repair' the old one and hope for a little improvement until the new controller is finished... ;)
| quote: | | It is unlikely at this juncture that you will find an improved motor controller IC for this application that is pin for pin compatible, plus I am sure it can do the job well as long as it is undamaged. |
Well, as I wrote, I'm not lookin for a pin compatible replacement IC (I explicitely said so, because I actually don't expect any pin compatible IC to exist).
But a 3-phase 'DC' control is something very basic, there sure are controllers which implement a 'tighter' servo loop (and have a better reference), but are also suitable for said motor...
| quote: | | Far more likely at this time that some or many of the electrolytics are bad and need replacing. Pots and some resistors may be suspect too. The motor itself is not too likely to respond well to higher current drive, the windings were designed to dissipate only so much power. |
Yep, I suspect that too (as always with old electrolytics). But the motor can handle higher power, because it's "underpowered" due to the poor power supply at the moment. I'not talking about an order of magnitude or so, just about an (as little as neccessary) higher torque for short durations.
| quote: | | I would focus on making sure that the current circuit is actually working properly. |
I honestly don't expect the current circuit to be able to perform good enough, even if I could restore it's original performance. It's a really cheap player, due to the cheap power supply and simplified controller. It's one of the cheapest Technics players ever, known for it's mediocre performance... But the motor is capable of more, as it is used in better players, too.
| quote: | | One area where you might make a big improvement is in replacing the power transformer with a slightly larger one, add more capacitance to the supply, perhaps consider voltage regulation to the servo using something like the LT1085, set to 18V to this I would add hex freds or schottky (if piv is high enough) rectifiers and maybe 10000uF 35V nichicon muse electrolytic after the rectifiers. |
Yep, I agree again, but this is something I'm already planning (as said in the former posts). I actually consider it neccessary, as I wouldn't expect an improvement in performance from the same transformer.
Again, can anyone else recommend an approach to improve or replace the old regulation?
Thanks again,
Sebastian. ;) |
|
|
| kevinkr |
Hey sebastian,
Sounds like you have things pretty under control.. <grin> Here is a thought - you mentioned that this motor was used in better technics tables as well, perhaps you can or already have identified one of those models and from there perhaps you can find a better controller implementation.. (Even salvage one from a damaged TT?)
Philips and National both had lines of motor controller ics at one time that may suit the application. I kind of wonder if anyone is doing much with these servo motor controls these days as the market for these cheap controls has almost dried up with the smallish cheap TT and obsolete cassette deck market.. Might also try JRC and some of the other japanese vendors.
I had an SL-23 25 + yrs ago and thought about getting an SL-2000 to replace it, but ended up going another route to a considerably better TT. As I recall the SL-23 would not win awards for its speed stability either, but the worst I ever owned was a Pink Triangle, (NOT cheap.. lol) even with a new motor and controller it drifted so badly I could not even listen to a single side without having to readjust. (As I recall it used a technics belt drive servo motor with the worst designed controller circuit I have ever seen - it did not really work at all! I disconnected the tachometer and nothing happened, two boards exactly the same.. lol)
Power supply improvement may help a lot as there should be somewhat increased torque. Measure motor currents and make sure IR drops in controller pcb traces are not excessive, bolster with wire point to point as needed. |
|
|
| Havoc |
| You do mention an oil bearing. Are you sure, the oil has not become something approaching sticky glue? |
|
|
| sek |
Hi,
It has still been liquid when I replaced it. ;)
And the replacement oil I applied (a machine bearing lubricant of about the same viscosity) is still in it's original (good) condition...
Would increased friction in the bearing really worsen the speed regulation? I can only imagine it to constantly load the platter (and increase regulation delay), but the motor does not keep nominal speed when only slightly loaded (i.e. by the needle, and I assume that means friction by the needle is much higher than overall friction when unloaded).
I could experiment with the bearing (i.e. find out if bearing fiction is angle-dependent or not constant over speed), but that still makes the original SL-2000 a mediocre player with a good motor... ;)
Thanks,
Sebastian. |
|
|
| sek |
Kevin ;)
how did you push your post in there?
I must have overlooked it, as I only answered on the follow-up by Havoc.
Sorry ;)
| quote: | Originally posted by kevinkr
Hey sebastian,
Sounds like you have things pretty under control.. <grin> |
Umm, pun? :cool:
| quote: | | Here is a thought - you mentioned that this motor was used in better technics tables as well |
I've read about that fact on the net. I'm not sure in which models (and can't determine in which not)... unfortunately! There don't seem to be too many people who like to talk about direct drives on the forum, though...
| quote: | | perhaps you can or already have identified one of those models and from there perhaps you can find a better controller implementation.. (Even salvage one from a damaged TT?) |
They all use outdated Matsushita/Panasonic ICs without adequate documentation. But...
| quote: | | Philips and National both had lines of motor controller ics at one time that may suit the application. I kind of wonder if anyone is doing much with these servo motor controls these days as the market for these cheap controls has almost dried up with the smallish cheap TT and obsolete cassette deck market.. Might also try JRC and some of the other japanese vendors. |
Exactly. I have found several suppliers of servo control ICs for three-phase brushless DC motors in the meantime. The keyword here is: BLDC (brush less direct current).
There are modern solutions, and it looks as if there are possibilities to control the motor with just a single IC for servo control and three-phase drive and another IC for the high precision regulation (i.e. PLL). ST Micoelectronics has such ICs, TI makes one, National used to make one, etc.
| quote: | | I had an SL-23 25 + yrs ago and thought about getting an SL-2000 to replace it, but ended up going another route to a considerably better TT. As I recall the SL-23 would not win awards for its speed stability either, but the worst I ever owned was a Pink Triangle, (NOT cheap.. lol) even with a new motor and controller it drifted so badly I could not even listen to a single side without having to readjust. (As I recall it used a technics belt drive servo motor with the worst designed controller circuit I have ever seen - it did not really work at all! I disconnected the tachometer and nothing happened, two boards exactly the same.. lol) |
Odd. Funny, actually. Yes, there are bad motors around, as well as bad Technics turntables. Unfortunately, for many people this means: Technics and DD can't mean anything good :(
At least you've got a good turntable after all ;)
| quote: | | Power supply improvement may help a lot as there should be somewhat increased torque. Measure motor currents and make sure IR drops in controller pcb traces are not excessive, bolster with wire point to point as needed. |
Okay. I'll do excessive measurements before I continue. I have to gather more specific information on the motor before I select a controller driver. Let's see what I come up with ;)
Further input is much appreciated! I know some of you guys are well informed in the direct drive department, ürobably just don't want to come out with knowledge about it (as DD is a no-no for many on the forums)... ;)
Thanks,
Sebastian. |
|
|
| jack daniel |
I am working out the schematics for the SL-23FS and need help with the DIP IC. The only markings on it are "K650 C1003", and I cant find a reference to it anywhere.
I have the CCT boards mapped out and a CCT diagram drawn, just missing the DIP IC info - WHAT IS IT??????
My plan is to build an external power supply with remote mounted controls.
I have 3 if anyone wants spares |
|
|
| MrDough |
Hi, I just found this thread since I am getting ready to put my SL-2000 turntable back in use.
I am the original owner of this device, having bought it in about 1977, and I have to tell you it has always worked very well for me. I have experience none of the problems you report.
One observation, though: I believe the speed adjustment pots get dirty and noisy, just like bad scratchy volume controls of the same era. I have cleaned mine occasionally and it gets the speed back to a nice stable performance level - for a while. The speed control pots are 5K each, I'm thinking of replacing them with 10-turn trimmers since the old pots are probably beyond cleaning.
- Doug |
|
|
| ogeronimo |
I´ve a Lenco L830 DD turntable and the IC AN620 is burned. The motor and the 3 phase controller is the same of Sek´s technics.
Anybody know any circuit for replace the original speed controller?
thanks!! |
|
|
| soundnut |
I know this is along time ago for the reply. Any luck with the sl-2000 Technics turntable? I just bought a used one for $53 usd. to replace my Dual 1219. The 1219 gets honors for being a real piece of ****. All mechanical...the speed switch is stuck, the speed is very slow, the anti-skating and weight balance all failed by just letting the unit not being used in 8 years. The sl-2000 works perfect. I will do some of your suggestions. Such as clean and put in new oil bearing oil. Change pots to 5 or 10 turn. Replace all electrolytics. And beef up the power supply with bigger electrolytics and maybe more if possible I have not opened the unit yet...just got it today.
Mitchell |
|
|
| analog_sa |
| quote: | Originally posted by soundnut
The 1219 gets honors for being a real piece of ****. |
It's probably pretty trivial to fix all that's wrong with the Dual. I bet it will beat the pants out of that Technics sonically. |
|
|
| sek |
I second that, although I don't know this particular Dual model. ;)
| quote: | | Any luck with the sl-2000 Technics turntable? |
Not really. I have to admit that the conditions were difficult to start with. The platter was excentric, the arm bearings were loose, the power suppy was crappy and the acrylic cover was damaged beyond repair. :(
I had stripped the motor assembly from the chassis and experimented with the combination of motor and platter. They could benefit from a hefty laboratory supply (with regulated 18V). My attempt would have been to replace the 33/45 switch and pitch control unit with something more convenient (button switch, single potentiometer) and the stroboscope with LEDs for a sharper image.
But then I got hold of an SL-1710MK2! :devilr:
There's no way back.
I currently use an SL-1200MK2 that I refurbished and repainted (white chassis, white LED illumination, improved power supply, renewed pitch assembly, recabled tonearm, etc.).
Don't get me wrong, I don't dislike the SL-2000 parts for what they are, it's just impossible to improve them beyond what they are capable of - at least not without spending more than for a used MK2.
Cheers,
Sebastian.
PS: I like to keep the SL-2000 parts in the drawer, though. We share the same year of birth. :) So in case you have any questions, don't hesitate. |
|
|
| Steerpike |
We are working on something similar here:
"New Technics SP10 motor controller specification"
http://www.diyaudio.com/forums/show...threadid=121934
Sek, your shematic shows no frequency reference (quartz crystal) so it will not be as accurate as designs like the SP10. (probably because they want pitch control, and pitch control with PLL is (was) not an easy or cheap circuit)
The motor also appears to have no true tacho coil - only the three phase detection windings. These CAN be used for tacho purposes as well, but are not very accurate (from the SP10 motor measuremnts I did, only about 6 pulses per revolution, as compared to 190 for the true tacho pickup). For PLL control you will need that sort of accuracy (190 pulses per rev). You could always add a photo-detector on the rim of the motor, and stick on a laser printed band of stripes for it to read.
To answer some of your questions, I'm guessing the following:
D1-D3 are germanium signal diodes, to demodulate the AM signal from the phase detector.
The white wire at S2 is likely the excitation voltage for the phase detection coils - probably at around 50kHz.
R10/R5 etc. MAY be the lead/lag compensation for the servo loop.
I imagine you could certainly increase the current drive to the motor. - after all, it' only the impulse current to get it up to speed, its normal runing current will be unchanged. It has a long thermal time constant and won't suffer badly from a higher start current. |
|
|
| sek |
Hi Steerpike,
I'm not the one who wants to improve it (anymore). ;)
I'm aware of the things you mention since I dissected a SL-1710MK2 and a couple of SL-1200/1210MK2 for comparison. The SL-2000 incorporates the technology that has also been used in the SL-1200 series (without MK2): no frequency generator, no crystal reference.
Oh, and I recognized your thread and followed it from the beginning. ;)
In other forums, I discussed the option of retrofitting the motor assembly with a frequency generator coil (i.e. from a superseded MK2 series player or a DIY one) and a reference generator circuit. But the amount of work is similar to your approach of a redesign...
| quote: | | D1-D3 are germanium signal diodes, to demodulate the AM signal from the phase detector. The white wire at S2 is likely the excitation voltage for the phase detection coils - probably at around 50kHz. R10/R5 etc. MAY be the lead/lag compensation for the servo loop. |
Thanks for your description.
It's even more severe: I made a mishap in my circuit description and accidentally took the drive coil connections for the position detector connections and vice versa. It's the other way 'round. :rolleyes:
The circuit can be improved as you mention - by increasing drive current and adjusting the lead/lag network in the feedback loop to the servo. Measured frequency is lower than 50kHz, btw...
But as the controller IC is not documented anywhere, the risk of increasing the current through the IC is too high for me. At least, it incorporates a metal plate for mounting a heat sink. External driver transistors are a possible solution, but too complicated to fit into the original housing...
Cheers,
Sebastian. |
|
|
| soundnut |
Heres what I am going to do. I will upgrade the power supply. It is 18 volts. I will use a LM7818 T220 chip instead of the existing supply. The LM 7818 can supply 1 amp constant and 2.2 amps peak (sounds good). Will use existing heatsink. However I will need a bigger transformer, I will try and find one that just fits inside. I need one that is 24vac (thats what I measured, 20-22 volts is ok too). Capacitor after the bridge rectifier will be 700 t0 1200uf. Capacitor at output of regulator will be 330-500uf. The bridge rectifier will be changed to 1 amp.
This is a little overkill since fuse for existing power supply is 350ma, I might change fuse to 500ma if I get a bigger transformer only if fuse blows once in a while. Existing power supply has only one capacitor at 330 uf. Using a LM7818 is better than using discrete componets. Existing transformer looks like it is 6 watts, hope to replace it with a 10 watt. In reality I need a beefed up power supply that will supply 1/2 amp instantly:). |
|
|
| soundnut |
| Oh I will also change the speed pots if I can find ones that will fit and hopefully get 5 turn ones or 10 turn ones but 10 turns ones are overkill. |
|
|
| Steerpike |
Is anyone still watching this thread?
A couple of weeks ago I got an wrecked Marantz TT, that upon disassembly proved to contain a Technics motor - not *quite* the same as the SL2000, but *very* similar. I'm still trying to figure out which Technics it comes from - the motor ID is PM2330009, with the ubiquitous AN620 IC in it. And no FG!
So I might look into upgrading this motor too - but the SP10 one comes first. |
|
|
| sek |
You could try to take some pictures from the stator and rotor assemblies, as well as the bearing and of course the circuit board. ;)
Cheers,
Sebastian. |
|
|
| Steerpike |
I'll take a pic later today!
That AN620 is a very elusive chip - no datasheets or pinout data to be found anywhere, as you said. |
|
|
| Steerpike |
This is the motor from the Marantz 6170 turntable, but it looks *very* Matsushita to me. Maybe never used in a Technics TT, but only sold to OEM manufacturers? The armature looks like that from the SL1200-1.
There is little on the PCB but for the AN620 power amp IC, a few diodes, resistors and caps and the two speed trimpots.
And there is a thing that looks like three photodiodes - not shown in the pic is a white stripe on the bottom edge of the rotor, that passes over the 3 holes, once per rev. |
|
|
| sek |
The only differences to the SL-2000 motor so far seem to be the more precise stator windings and the presumably optical position sensing (where the SL-2000 had a set of three pickup windings). The complexity of the board looks alike.
Cheers,
Sebastian. |
|
|
| Steerpike |
What I thought were photodiodes are just boring old magnetic pickups - not Hall sensors though.
I've been analysing this thing's behaviour, and traced out the circuit - the component number are my own invention - none are marked on the PCB. Very similar - but not identical - to yours Sek.
The bulk of an 'almost SP10 control system' is inside this IC, in a simplified form.
The same principles of operation apply, so we can improve its performance.
There is no push-pull drive on the motor coils - they are switched 'on' only with current in a single direction. I'm not sure I'd bother to change that.
The torque isn't particularly high: the normal operating current is around 60mA, yet the startup current is around 250mA, indicating that when running normally, the system has quite a good current headroom.
Trying to get more current through the motor might damage the IC, and wouldn't improve the torque substantially for reasonable currents.
Beefing up C25/26/27 might be advantageous though.
There is no PLL, but this is not a terribly difficult thing to add.
Poking at pins on the IC, it can be provoked to run the motor a low as 16 rpm, and it will also go up to beyond 78 rpm. Current consumption is only marginally higher - at around 68mA.
So it can easily be converted to a 4 speed turntable - which I will certainly be doing.
Adding a Quartz PLL is not hard either. There is no FG winding, but the platter is already machined with 4 strobe tracks, 33 & 45 rpm for 50Hz and 60Hz.
Any one of these can be watched with a LED/photodiode combination, to derive a tacho signal.
If only 33 and 45 rpm are to be PLL locked, the tacho signal can be read by two opto-pickups (one for 33,one for 45) which means no divider switching is necessary - and a easy crystal frequency like 4.000MHz can be used.
Otherwise a single opto-pickup is fine, but the divider ratios become rather inconvenient. Still, if a microprocessor is being used, the 'nasty' divider ratios are much easier to cope with.
Once a useful phase error signal has been derived, it can be injected into the junction of R4 and R5, adding it to the velocity error signal.
More to follow! |
|
|
| sek |
How do I have to imagine phase sense coils on the PCB? Can you take a picture of these? I'm asking because that's a true difference to the SL-2000.
A true phase sense would be a great thing to have, as a PLL could reference this to a clock. That's also pretty much how the MK2 series does it; split the tacho signal into magnitude and phase, then inject both of them into separate loops (using integration and differentiation), then recombine the resulting signals into a single error voltage and feed that into the drive IC's error amplifier.
On another note, the drive currents in a SL-1200MK2 motor are 100mA to 180mA continuous (load dependent) and 600mA maximum. Just for comparison. :)
Cheers,
Sebastian. |
|
|
| Steerpike |
| quote: | | A true phase sense would be a great thing to have, |
It would have to be done by means of an added tacho stripe. The resolution inherent in my motor and the early SL1200 is too low - its about 20 pulses per rev I think. Machines like the SP10 have a tacho resolution of 190 pulses per rev.
I'll get a photo of mine. Its esentially just a pcb mounted module with 3 sensors aligned in an arc, about 5mm apart.
This is an extra sensor, not part of the motor itself.
On your circuit, the pink wire of J4, can you see what that does to? I'm guessing it might be a speed sensor.
| quote: | | drive currents in a SL-1200MK2 motor are 100mA |
Do you know the voltage? |
|
|
| sek |
| quote: | | On your circuit, the pink wire of J4, can you see what that does to? I'm guessing it might be a speed sensor. |
Now you made me think! :D
Take another look at the picture I attached in the first post. You see three pairs of pickup coils arranged 120 degree apart. Additionally, there are three diodes in the upper right mounting area (note that this PCB is not the control PCB, just the one that mounts the coils).
It is easily possible that the pairs of pickups are an arrangement of what you call velocity sense and phase sense coils, respectively. This would then result in a control circuit frighteningly similar to yours. :)
Please also note that some things in my schematic drawup from back then shouldn't be taken too seriously, particularly the connections labeled with a question mark and the orientation of polarized components like diodes and capacitors.
| quote: | | Do you know the voltage? |
4Vpp (peak to peak) as per the service manual, with about 11Vdc bias. I haven't measured it under load, though. Do you have the MK2 service manual handy?
| quote: | | It would have to be done by means of an added tacho stripe. |
The voltage derived from the three detector coils (i.e. your L1 through L3) is rectified (i.e. by your D1 through D3) and filtered (i.e. by your R1 and C?), thus injecting a DC voltage into the AN620's pin 15 (pitch).
One idea is to replace this whole arrangement by a quartz referenced tacho stripe DC error signal, just as you noted (PLL injection point) in your schematic.
Building (or reusing) a tacho stripe isn't too difficult, I would expect. The MK2 drive unit has it arranged on the outer side of the rotor's permanent magnet. It gets filtered and amplified to a level of about 0.6Vpp by a simple bipolar transistor circuit and delivers a sinusoidal signal with a period of 20ms for 33rpm, 15ms for 45rpm. Inside the MK2's AN6680 control IC, this signal is then fed to an arrangement of a (presumably integrating/differentiating) amplifier buffer and Schmidt trigger.
Cheers,
Sebastian. |
|
|
|
