Hi Pjotr,
Overnight I have been thinking some more about your question in post #91, about how much drag is caused by the stylus.
Regrettably, it is close to 50yrs since I was at school, and I have forgotten how to do calculus, for example in relation to calculating the acceleration due to gravity, which is a constantly changing situation, of course.
However, you seem to be quite clued up on theory and calculations, judging by your posts, so maybe the following thought might be possible.
If one could either time or perhaps count the number of revolutions for the TT to come to rest from normal speed, both with the stylus on the record, and without, this should indicate (in some way) how much drag is being caused by the stylus.
My problem here is that as the TT slows down, the stylus drag will decrease proportionately, and this is where I think some calculus might be needed, effectively to 'average out' or 'normalise' the results.
Then with your knowledge of the weight/inertia of the TT etc., could you then do this calculation to determine the overall stylus drag, perhaps?
Clearly in both cases (i.e. also with my earlier crude manual assessment methodology) it would be wise to make several 'passes' when carrying out any measuring/timing/counting revs., or whatever, and then averaging the results.
However, it would appear to be rather quicker and less fussy or subject to error, if one could simply calculate the required results, as opposed to many attempts at pulling strings whilst watching balances and so on.
Regards,
Overnight I have been thinking some more about your question in post #91, about how much drag is caused by the stylus.
Regrettably, it is close to 50yrs since I was at school, and I have forgotten how to do calculus, for example in relation to calculating the acceleration due to gravity, which is a constantly changing situation, of course.
However, you seem to be quite clued up on theory and calculations, judging by your posts, so maybe the following thought might be possible.
If one could either time or perhaps count the number of revolutions for the TT to come to rest from normal speed, both with the stylus on the record, and without, this should indicate (in some way) how much drag is being caused by the stylus.
My problem here is that as the TT slows down, the stylus drag will decrease proportionately, and this is where I think some calculus might be needed, effectively to 'average out' or 'normalise' the results.
Then with your knowledge of the weight/inertia of the TT etc., could you then do this calculation to determine the overall stylus drag, perhaps?
Clearly in both cases (i.e. also with my earlier crude manual assessment methodology) it would be wise to make several 'passes' when carrying out any measuring/timing/counting revs., or whatever, and then averaging the results.
However, it would appear to be rather quicker and less fussy or subject to error, if one could simply calculate the required results, as opposed to many attempts at pulling strings whilst watching balances and so on.
Regards,
maxon REmax speed problems
Boris,
i have the same maxon motor as you. No longterm experience yet as i postponed the replacement of my TT's Pabst GS38/09.
But i did some preliminary experiments with it and i had no speed variation problems at all. Rock-solid stable speed with a standard LM317 circuit as reg, with leader tape as belt and a preliminary pulley with radially oriented set screw. Not even a collet chuck pulley.
My buddy Manfred (who designed the Teres motor controller unit und uses maxon since then) also has an REmax29 in use. Had he had problems, i would have gotten a warning from him.
I see 3 possible causes for your problem.
1)
The pulley has to run smoothly and without the maximum achievable lack of excentricity. Make sure that is the case.
Too much excentricity is not only bad for speed stability in itself, it also causes an oscillating radial load aging the shaft bearing prematurely (that word is a gross understatement )
2)
Your reg is performing some unwanted antics. Tame the thing.
3)
The maxon cannot stand excessive radial loads on the shaft. According to the data sheet the load maximum is 5.5 Newton for the slide bearing and 5.0 for the ball bearing. I presume you had chosen the slide bearing variant.
Now I remember that Manfred reported speed problems with one sample of the Teres maxon motor like you describe it above and he told me it turned out that the shaft bearings were gone. This also is a problems well-known to maxon; their experts warn from excessive radial loading.
So if the motor got speed problems, it is very likely that its shaft bearings are gone (one cause for that could be that the shaft is slightly bent due to too much axial forces when press-fitting the pulley). In this case, order a new sample.
Fitting of the pulley:
the pulley's center hole should slide smoothly and without radial play over the motor shaft. Use either glue or a set screw or a collet chuck mechanism to clamp the pulley on the shaft.
DON'T use axial press-fitting. You may use radial press-fitting (heat up the pulley, slide it over the shaft, let it cool.)
But: you will never get the pulley off the shaft undamaged.
As the pulley i am talking of is a precision component, hence expensive, and as pulley outlive approx. 5 maxon motrs, i do not recommend heat-fitting.
If you glue the pulley, make sure the glue joint can be destroyed at a temperature not yet damaging the pulley. Epoxy e.g. should give up at 160°C. So if you wnat to recycle your precision pulley later, heat it up, unmount it from the motor shaft and keep it at the temperature until you wiped the glue out of the the center hole. Then you can mount the pulley on another motor shaft.
My premium option would be a pulley with a collet chuck inside clamping it on the shaft.
Boris,
i have the same maxon motor as you. No longterm experience yet as i postponed the replacement of my TT's Pabst GS38/09.
But i did some preliminary experiments with it and i had no speed variation problems at all. Rock-solid stable speed with a standard LM317 circuit as reg, with leader tape as belt and a preliminary pulley with radially oriented set screw. Not even a collet chuck pulley.
My buddy Manfred (who designed the Teres motor controller unit und uses maxon since then) also has an REmax29 in use. Had he had problems, i would have gotten a warning from him.
I see 3 possible causes for your problem.
1)
The pulley has to run smoothly and without the maximum achievable lack of excentricity. Make sure that is the case.
Too much excentricity is not only bad for speed stability in itself, it also causes an oscillating radial load aging the shaft bearing prematurely (that word is a gross understatement
)2)
Your reg is performing some unwanted antics. Tame the thing.
3)
The maxon cannot stand excessive radial loads on the shaft. According to the data sheet the load maximum is 5.5 Newton for the slide bearing and 5.0 for the ball bearing. I presume you had chosen the slide bearing variant.
Now I remember that Manfred reported speed problems with one sample of the Teres maxon motor like you describe it above and he told me it turned out that the shaft bearings were gone. This also is a problems well-known to maxon; their experts warn from excessive radial loading.
So if the motor got speed problems, it is very likely that its shaft bearings are gone (one cause for that could be that the shaft is slightly bent due to too much axial forces when press-fitting the pulley). In this case, order a new sample.
Fitting of the pulley:
the pulley's center hole should slide smoothly and without radial play over the motor shaft. Use either glue or a set screw or a collet chuck mechanism to clamp the pulley on the shaft.
DON'T use axial press-fitting. You may use radial press-fitting (heat up the pulley, slide it over the shaft, let it cool.)
But: you will never get the pulley off the shaft undamaged.
As the pulley i am talking of is a precision component, hence expensive, and as pulley outlive approx. 5 maxon motrs, i do not recommend heat-fitting.
If you glue the pulley, make sure the glue joint can be destroyed at a temperature not yet damaging the pulley. Epoxy e.g. should give up at 160°C. So if you wnat to recycle your precision pulley later, heat it up, unmount it from the motor shaft and keep it at the temperature until you wiped the glue out of the the center hole. Then you can mount the pulley on another motor shaft.
My premium option would be a pulley with a collet chuck inside clamping it on the shaft.
Hi Bernhard,
actually I bought this motor-type, after Manfred Huber made some recommendations in the vinyl-lebt-newsgroup some years ago. And I got also the pulley from am member of this group.
Since the pulley is highly precise manufactured and I tested out a few different voltage-regulators, I can rule out the first two points of you.
The motor has slide bearings. When I ordered it, the sales-man of maxon warned me not to use slide-bearing with this low revolution speed and maybe this is why they put a prototype-sticker on the motor…
To avoid too much slipping of the tape, I gave it a more or less high tension. This leads to you third point: I can feel a slight radial clearance of the bearing. So it seems that the bearings are defective.
But I’m not disposed to buy the same type again. The lifespan is to short for an inexpensive motor like this one…
I wonder if anyone ever compared two similar motors whereas one has ball-bearing and the other one has slide-bearing. Is there really an audible difference??
Regards, Boris
actually I bought this motor-type, after Manfred Huber made some recommendations in the vinyl-lebt-newsgroup some years ago. And I got also the pulley from am member of this group.
Since the pulley is highly precise manufactured and I tested out a few different voltage-regulators, I can rule out the first two points of you.
The motor has slide bearings. When I ordered it, the sales-man of maxon warned me not to use slide-bearing with this low revolution speed and maybe this is why they put a prototype-sticker on the motor…
To avoid too much slipping of the tape, I gave it a more or less high tension. This leads to you third point: I can feel a slight radial clearance of the bearing. So it seems that the bearings are defective.
But I’m not disposed to buy the same type again. The lifespan is to short for an inexpensive motor like this one…
I wonder if anyone ever compared two similar motors whereas one has ball-bearing and the other one has slide-bearing. Is there really an audible difference??
Regards, Boris
Hi Boris.
It's very simple to check for record slip. Just put a small piece of tape on the edge of the record, and another right next to it on the platter. Play the record a couple of times, and the two pieces of tape should still be exactly aligned. I think it's worth doing this test, just to be sure. It would be silly to spend all this energy on the motor if the problem was slippage. It's always worth re-checking your previous assumptions.
You might also have slippage between the belt and the platter, or the belt and the pulley. I don't know how to test for this, but there's probably a simple way.
I've just acquired a Beuhler motor that uses an LM2817 to control the speed. It's very simple, and controls the speed quite well, though I haven't tried it in my TT as yet. All you need is some means of getting frequency pulses from the platter. This should not be too difficult. It doesn't matter what frequency the pulses are, as the LM2917 converts them to a voltage, and the feedback loop controls this voltage.
Neil
It's very simple to check for record slip. Just put a small piece of tape on the edge of the record, and another right next to it on the platter. Play the record a couple of times, and the two pieces of tape should still be exactly aligned. I think it's worth doing this test, just to be sure. It would be silly to spend all this energy on the motor if the problem was slippage. It's always worth re-checking your previous assumptions.
You might also have slippage between the belt and the platter, or the belt and the pulley. I don't know how to test for this, but there's probably a simple way.
I've just acquired a Beuhler motor that uses an LM2817 to control the speed. It's very simple, and controls the speed quite well, though I haven't tried it in my TT as yet. All you need is some means of getting frequency pulses from the platter. This should not be too difficult. It doesn't matter what frequency the pulses are, as the LM2917 converts them to a voltage, and the feedback loop controls this voltage.
Neil
Hi Neil,
although I still cannot believe that the record could slip, I will do this test. I have some slippage between belt and pulley during the start-up, so probably the slippage happens rather there, than between the record and the pulley.
But since Bernhards posting I’m sure that the motor is defective.
Regards, Boris
although I still cannot believe that the record could slip, I will do this test. I have some slippage between belt and pulley during the start-up, so probably the slippage happens rather there, than between the record and the pulley.
But since Bernhards posting I’m sure that the motor is defective.
Regards, Boris
Hi,
Don’t know if this is an issue but a LM317 regulator needs 100 nF ceramic caps close to its input AND output to gnd for stability reasons, holds for most IC regulators. A simple electrolytic cap of around 10uF – 100 uF at the output is not always sufficient and HF oscillations can occur impairing DC stability.
Just another 2 cents
Don’t know if this is an issue but a LM317 regulator needs 100 nF ceramic caps close to its input AND output to gnd for stability reasons, holds for most IC regulators. A simple electrolytic cap of around 10uF – 100 uF at the output is not always sufficient and HF oscillations can occur impairing DC stability.
Just another 2 cents
Hi folks,
A quick report back on my new motor trials to date.
I have resolved a regulator problem, now the voltage measured at the motor terminals remains constant. However, the speed of the platter consistantly increases slightly (as seen on strobe disc) toward the centre of a record. It seems therefore that the motor is just not load tolerant enough. This may be insufficient torque due to very low motor voltage and rpm (as has been suggested). However, given that the motor has enough torque to start the platter from standing, I would have thought torque was more than adequate (?). I think that instantaneous changes in drag due to groove modulations are smoothed out by the platter intertia and no open or closed loop control system could deal with that anyway, so I think it's a kind of "soft" load compensation system that is required.
If anybody is still interested comments are welcome!
A quick report back on my new motor trials to date.
I have resolved a regulator problem, now the voltage measured at the motor terminals remains constant. However, the speed of the platter consistantly increases slightly (as seen on strobe disc) toward the centre of a record. It seems therefore that the motor is just not load tolerant enough. This may be insufficient torque due to very low motor voltage and rpm (as has been suggested). However, given that the motor has enough torque to start the platter from standing, I would have thought torque was more than adequate (?). I think that instantaneous changes in drag due to groove modulations are smoothed out by the platter intertia and no open or closed loop control system could deal with that anyway, so I think it's a kind of "soft" load compensation system that is required.
If anybody is still interested comments are welcome!
It's hard to believe that the motor can see this tiny load variation.
However, this is the kind of variation that a feedback circuit would eliminate.
The original Burr Brown circuit you posted indicated that the voltage would need to change when the load changes, in order to maintain constant speed.
You will have to sense the speed somehow, if you want to compensate. You'll either need to sense the current through the motor, or look at some strobe pulses from the platter, using the 2917 device.
Presumably, what you'll find is that the current reduces when playing the inner tracks, because the load is less.
Have you actually tried the Burr Brown circuit?
However, this is the kind of variation that a feedback circuit would eliminate.
The original Burr Brown circuit you posted indicated that the voltage would need to change when the load changes, in order to maintain constant speed.
You will have to sense the speed somehow, if you want to compensate. You'll either need to sense the current through the motor, or look at some strobe pulses from the platter, using the 2917 device.
Presumably, what you'll find is that the current reduces when playing the inner tracks, because the load is less.
Have you actually tried the Burr Brown circuit?
Matt Rowland said:
Bobken said:
Pjotr said:Hi,
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
Pjotr is correct, you need to consider combination of back emf, drive voltage and torque when calculating the speed of the motor. For instance in Matt's calculation, the motor speed at the stated torque would be zero as all the voltage would go to driving the torque current across the winding resistance. I think the trick that you have missed in relation to DC supplies is that they are ideally not constant voltage supplies. The drive output impedance must also be included in the analysis of motor speed: The speed is given by
S = Ks. (V supply – (T load / Kt ) . (Rw + Rd ) )
so with constant voltage supply like a "perfect" regulator the motor has a constant negative speed torque slope given by
Delta S / Delta T load = Ks (Rw + Rd ) / Kt .
For the motor given this is 34.2 rpm / mNm. This is relatively good, but since load torque for a turntable is <B>not</B> constant it will still give problems with speed variation. The torque required from the motor is affected by stylus drag, the torque equivalent of which varies with stylus position. I calculate a speed variation of 1.2% if that motor is used without compensation in a standard belt drive set-up. Anyone who claims stable speed with a simple regulator fed DC motor is not measuring speed properly.
Compensation consists of using a transimpedance amplifier to reduce the output impedance of the power supply to below zero. If the sum of the output impedane of the supply and the motor's winding resistance is made close to zero, the speed torque slope becomes nearly zero so the problem disappears. At a combined drive / motor resistance of 1 ohm the speed torque slope for the motor above is 3.5 rpm /mNm, ten times better, but with a 226764 it can be brought down to 0.3 rpm / mNm at 1 ohm, which is a hundred times better. It is easy to achieve 1 ohm combined impedance, I used a compensation circuit giving less than this in the motor drive referred to by Vinyl Addict. I have just completed the PCB design for the fourth prototype, I will be publishing this in Bas Hornemann's on-line DIY mag.
Differing opinions
Hi,
Quote:
"Anyone who claims stable speed with a simple regulator fed DC motor is not measuring properly."
That is an interesting overstatement, if ever I heard one!
I have been playing about with TT supplies for maybe 40 years now, and until quite recently, always relied on test records for ultimate speed checking.
These recordings were plentiful in the heyday of vinyl, and certain tests were specifically intended to be used in conjunction with wow and flutter meters, where errors of a fraction of a % were significant in these measurements. They were therefore extremely accurately recorded and carefully pressed, and in their time were considered to be industry standards.
On one test recording I have, the entire side is taken up with a 'constant' 3150Hz @ nominal 10cm/sec. sine wave, and on another, there is a 3kHz signal @ nom. 5cm/sec. Both signals being fairly average, middle-of-the-road choices here so as not to unwittingly bias any results.
For years I used these in evaluating TT speeds etc., amongst other parameters, and this was usually done by feeding the ouput into an accurate frequency counter, although inspecting the results on a 'scope was also instructive.
Using both of these recordings, I could not reliably see any repeatable speed variations when using my current P.S. when comparing the results at the start or end (or anywhere in between) of these tests. In fact, as nothing is absolutely perfect in life, nor in any engineering either, I saw greater frequency variations when simply deliberately blowing across the TT with my mouth!
I have never experienced any noticeable variations in speed over the entire side of a record of anything like the levels talked about recently (even 1% is collossal here), but not being entirely satisfied with these results since my interest had moved on to the effects caused by highly modulated passages on recordings, I sought some other means of speed checking.
This eventually resulted in the purchase of what is claimed to be the worlds most accurate means of TT speed checking 'on the fly', which tests, of course, are normally precluded when using test records with their otherwise constant-velocity recorded tracks.
Whilst this new arrival (a Lys Gyrascope) did then enable me to see better what happened in real time whilst listening to heavily modulated conventional records, it also confirmed (without a shadow of any doubt) that my TT's speed was perfectly constant, no matter what the circumstances.
I don't know how you can explain these results, bearing in mind your sweeping statements, but maybe you can point me to some better way of TT speed checking, since as your chosen quote suggests "I am always happy to learn something which is new to me" and you have categorically stated that (I am) "not measuring properly."
Could you kindly tell us how you verify your own results empirically, and hopefully I will learn something new from this?
For what it is worth, (IMHO) using a high torque motor together with a robust high current supply which is extremely quiet and capable of being controlled (regulated) very accurately (which precludes using these suggested ultra low voltages, anyway, for reasons I have already advised) this will give a far better subjective result than any attempt at complications like servos, or whatever.
In my opinion, to satisfactorily explain this, it must be one of those situations where even if the theory is sound, in practice the results (and these are what I listen to) simply depart from this theory or any 'calculated' expectation.
I simply cannot imagine why there is an apparent predilection with such low voltages, anyway, especially when used with vastly (nominally) higher-rated motors, as I have found that this is just making things harder here for one's self than need be. Presently I am using approx. 9-13 volts DC in my set-up, as do some other TT manufacturers, and except where one cannot obtain the correct speeds otherwise because maybe the available pulleys dictate this, why create more problems/complications in a P.S. than one needs?
However, as I have not found any need to consider neg. output impedance DC supplies, myself, I have no idea what this will do to the resultant sound, and I await with interest seeing your own design.
Regards.
Hi,
Quote:
"Anyone who claims stable speed with a simple regulator fed DC motor is not measuring properly."
That is an interesting overstatement, if ever I heard one!
I have been playing about with TT supplies for maybe 40 years now, and until quite recently, always relied on test records for ultimate speed checking.
These recordings were plentiful in the heyday of vinyl, and certain tests were specifically intended to be used in conjunction with wow and flutter meters, where errors of a fraction of a % were significant in these measurements. They were therefore extremely accurately recorded and carefully pressed, and in their time were considered to be industry standards.
On one test recording I have, the entire side is taken up with a 'constant' 3150Hz @ nominal 10cm/sec. sine wave, and on another, there is a 3kHz signal @ nom. 5cm/sec. Both signals being fairly average, middle-of-the-road choices here so as not to unwittingly bias any results.
For years I used these in evaluating TT speeds etc., amongst other parameters, and this was usually done by feeding the ouput into an accurate frequency counter, although inspecting the results on a 'scope was also instructive.
Using both of these recordings, I could not reliably see any repeatable speed variations when using my current P.S. when comparing the results at the start or end (or anywhere in between) of these tests. In fact, as nothing is absolutely perfect in life, nor in any engineering either, I saw greater frequency variations when simply deliberately blowing across the TT with my mouth!
I have never experienced any noticeable variations in speed over the entire side of a record of anything like the levels talked about recently (even 1% is collossal here), but not being entirely satisfied with these results since my interest had moved on to the effects caused by highly modulated passages on recordings, I sought some other means of speed checking.
This eventually resulted in the purchase of what is claimed to be the worlds most accurate means of TT speed checking 'on the fly', which tests, of course, are normally precluded when using test records with their otherwise constant-velocity recorded tracks.
Whilst this new arrival (a Lys Gyrascope) did then enable me to see better what happened in real time whilst listening to heavily modulated conventional records, it also confirmed (without a shadow of any doubt) that my TT's speed was perfectly constant, no matter what the circumstances.
I don't know how you can explain these results, bearing in mind your sweeping statements, but maybe you can point me to some better way of TT speed checking, since as your chosen quote suggests "I am always happy to learn something which is new to me" and you have categorically stated that (I am) "not measuring properly."
Could you kindly tell us how you verify your own results empirically, and hopefully I will learn something new from this?
For what it is worth, (IMHO) using a high torque motor together with a robust high current supply which is extremely quiet and capable of being controlled (regulated) very accurately (which precludes using these suggested ultra low voltages, anyway, for reasons I have already advised) this will give a far better subjective result than any attempt at complications like servos, or whatever.
In my opinion, to satisfactorily explain this, it must be one of those situations where even if the theory is sound, in practice the results (and these are what I listen to) simply depart from this theory or any 'calculated' expectation.
I simply cannot imagine why there is an apparent predilection with such low voltages, anyway, especially when used with vastly (nominally) higher-rated motors, as I have found that this is just making things harder here for one's self than need be. Presently I am using approx. 9-13 volts DC in my set-up, as do some other TT manufacturers, and except where one cannot obtain the correct speeds otherwise because maybe the available pulleys dictate this, why create more problems/complications in a P.S. than one needs?
However, as I have not found any need to consider neg. output impedance DC supplies, myself, I have no idea what this will do to the resultant sound, and I await with interest seeing your own design.
Regards.
Why would the load change between the outer and inner tracks? Since the downforce remains constant, the coulomb friction should not change with velocity.
Is there something special about stylus drag that makes it not operate like ordinary sliding friction?
(I'm a little out of my depth here, trying to understand.
Is there something special about stylus drag that makes it not operate like ordinary sliding friction?
(I'm a little out of my depth here, trying to understand.
Hi Bobken,
The reason I'm using low voltages is not to make life more difficult but because I'm stuck with the diameter of the Linn sub-platter and the original 20mm pulley. Even with an 8mm pulley the voltage will only be about 3v.
Hi Mark Kelly,
Your circuit sounds interesting can you give details of where to see it ie Bas Hornemann's on-line DIY mag?
Thanks to everybody for the continued input.
Cheers
Matt
The reason I'm using low voltages is not to make life more difficult but because I'm stuck with the diameter of the Linn sub-platter and the original 20mm pulley. Even with an 8mm pulley the voltage will only be about 3v.
Hi Mark Kelly,
Your circuit sounds interesting can you give details of where to see it ie Bas Hornemann's on-line DIY mag?
Thanks to everybody for the continued input.
Cheers
Matt
Never mind, I figured it out.
Stylus drag acts as a torque, and hence is proportional to radius.
If Matt increased the load on the motor, e.g. by adding constant friction to the platter, would the speed variation be reduced? It seems that the stylus drag would be a smaller percentage of the total drag.
Stylus drag acts as a torque, and hence is proportional to radius.
If Matt increased the load on the motor, e.g. by adding constant friction to the platter, would the speed variation be reduced? It seems that the stylus drag would be a smaller percentage of the total drag.
Hi Matt,
No offence intended here as for some time I realised that you were stuck with this pulley size, and that is why I referred to that restriction of yours in my last posting.
However, for most others it is simply adding to the problems in my opinion, and I don't understand the attraction here. I have found that reducing voltages to maybe say 50% at most can be beneficial to reduce some motor vibrations, but going further than this is not wise in my experience, as I have also mentioned many posts ago.
Pixpop has just figured out one of the most important points here, which I have obviously failed to convey previously, in spite of my efforts.
Consider a much exaggerated situation where say the motor is a 100 HP unit driving the TT. You could load this (equivalent to a medium car engine!) with a dozen pickups and maybe see a speed reduction of 0.000001%, or whatever!
Not much to worry about here, and other factors will dominate, anyway.
This is why I have mentioned several times about high torque in the first place, and higher voltages too, as any errors and noise in the DC circuit (of which there will always be some, no matter how hard we try to avoid this) will be related to the output voltage, also as a percentage, as I mentioned in an earlier thread.
All of these undesirable effects will have far more serious consequences if the motor is say merely a 'flea-powered' affair, and fed by a very low voltage, as they represent a much higher percentage of the power used to keep the TT spinning at normal speeds.
There are some problems in using servos or feedback arrangements, as I have found, but they do have some merits too. However (so far) my choice, having tried both, is that subjectively the non-servo type is simply sonically preferable to my ears.
As I also said many posts before, I have found throughout in electronics (as time has gone by) that simpler circuits are generally preferable overall, all other things being equal, and many other engineers are beginning to realise this too.
There are always downsides to any engineering solutions as nothing is absolutely perfect, and it is a question of obtaining the best compromise overall, which for me, and for now, I have found.
Any categoric statements saying that this choice is "no good", though, or maybe "the only way" to solve any problem, I tend to view with some caution.
Never let it be forgotten with any audio components, that how it *sounds* is what matters most, not how it measures, although this often gives one a good indication, and this is what has 'driven' me for very many years in my own trials.
Regards,
No offence intended here as for some time I realised that you were stuck with this pulley size, and that is why I referred to that restriction of yours in my last posting.
However, for most others it is simply adding to the problems in my opinion, and I don't understand the attraction here. I have found that reducing voltages to maybe say 50% at most can be beneficial to reduce some motor vibrations, but going further than this is not wise in my experience, as I have also mentioned many posts ago.
Pixpop has just figured out one of the most important points here, which I have obviously failed to convey previously, in spite of my efforts.
Consider a much exaggerated situation where say the motor is a 100 HP unit driving the TT. You could load this (equivalent to a medium car engine!) with a dozen pickups and maybe see a speed reduction of 0.000001%, or whatever!
Not much to worry about here, and other factors will dominate, anyway.
This is why I have mentioned several times about high torque in the first place, and higher voltages too, as any errors and noise in the DC circuit (of which there will always be some, no matter how hard we try to avoid this) will be related to the output voltage, also as a percentage, as I mentioned in an earlier thread.
All of these undesirable effects will have far more serious consequences if the motor is say merely a 'flea-powered' affair, and fed by a very low voltage, as they represent a much higher percentage of the power used to keep the TT spinning at normal speeds.
There are some problems in using servos or feedback arrangements, as I have found, but they do have some merits too. However (so far) my choice, having tried both, is that subjectively the non-servo type is simply sonically preferable to my ears.
As I also said many posts before, I have found throughout in electronics (as time has gone by) that simpler circuits are generally preferable overall, all other things being equal, and many other engineers are beginning to realise this too.
There are always downsides to any engineering solutions as nothing is absolutely perfect, and it is a question of obtaining the best compromise overall, which for me, and for now, I have found.
Any categoric statements saying that this choice is "no good", though, or maybe "the only way" to solve any problem, I tend to view with some caution.
Never let it be forgotten with any audio components, that how it *sounds* is what matters most, not how it measures, although this often gives one a good indication, and this is what has 'driven' me for very many years in my own trials.
Regards,
Thanks for the reply Bob,
I am begining to wonder whether the dc motor route is appropriate with a deck such as the Linn. Maybe a well designed supply for the original ac motor would have been the best idea. However I remain curious as to how Origin Live manage to make a dc motor conversion kit for the Linn, with a similar motor to mine, that sells very well and apparently works OK (at low voltage).
Regards
Matt
I am begining to wonder whether the dc motor route is appropriate with a deck such as the Linn. Maybe a well designed supply for the original ac motor would have been the best idea. However I remain curious as to how Origin Live manage to make a dc motor conversion kit for the Linn, with a similar motor to mine, that sells very well and apparently works OK (at low voltage).
Regards
Matt
Re: Differing opinions
I cannot explain your results. A DC motor with a fixed voltage supply changes speed with changing torque for the reasons I explained. This can be reduced by using a motor with a lower speed / torque ratio but it cannot be eliminated. I'm not making this up, it's there in the manufacturer's data sheets as speed / torque constant.
We agree that effective load torque due to stylus drag changes across the record plus there is a (smaller) effect from modulation drag. I use a very simple load / no load test - get the thing running on speed with the stylus off the record, cue the stylus then remeasure the speed. Using some simple assumptions regarding torque step up ratios this will give around 1% speed variation for a motor speed / torque constant around 22rpm / mNm, 0.1% for 2.2 rpm / mNm and 0.01% (100 ppm) for 0.22 rpm / mNm. The only DC motors I know of with ratings that low are big high torque affairs - 100 watt and up. Such big motors are noisy, for the simple reason that a given amount of runout in the mechanical components generates a much larger amount of kinetic energy.
The negative impedance supply can achieve effective speed / torque ratios in this region quite easily, using a small and therefore quiet motor. I'm not claimimg any credit for this, it is a well known principle (Maxon calls it I x R compensation). All I did was figure out a simple way to add it to my extremely simple shunt regulator design.
I am not willing to divulge how I measure motor speed it is the subject of a patent application. I measure platter speed with a strobe disc and an LED strobe with a part shade so I can count strobe sectors as they move across the division. No visible movement in 3.3 minutes is 100 ppm. I am toying with the idea of building a wow and flutter / speed meter, haven't gotten around to it yet.
BTW I don't actually use the DC drive at the moment, I installed a three phase EC (brushless) motor with a dedicated three phase AC drive in my Garrard.
I built the DC drive for Vinyl Addict. He measured about 0.1% drift but that was before we built the thermal drift compensation in, I expect it to be better now.
Bobken said:
I cannot explain your results. A DC motor with a fixed voltage supply changes speed with changing torque for the reasons I explained. This can be reduced by using a motor with a lower speed / torque ratio but it cannot be eliminated. I'm not making this up, it's there in the manufacturer's data sheets as speed / torque constant.
We agree that effective load torque due to stylus drag changes across the record plus there is a (smaller) effect from modulation drag. I use a very simple load / no load test - get the thing running on speed with the stylus off the record, cue the stylus then remeasure the speed. Using some simple assumptions regarding torque step up ratios this will give around 1% speed variation for a motor speed / torque constant around 22rpm / mNm, 0.1% for 2.2 rpm / mNm and 0.01% (100 ppm) for 0.22 rpm / mNm. The only DC motors I know of with ratings that low are big high torque affairs - 100 watt and up. Such big motors are noisy, for the simple reason that a given amount of runout in the mechanical components generates a much larger amount of kinetic energy.
The negative impedance supply can achieve effective speed / torque ratios in this region quite easily, using a small and therefore quiet motor. I'm not claimimg any credit for this, it is a well known principle (Maxon calls it I x R compensation). All I did was figure out a simple way to add it to my extremely simple shunt regulator design.
I am not willing to divulge how I measure motor speed it is the subject of a patent application. I measure platter speed with a strobe disc and an LED strobe with a part shade so I can count strobe sectors as they move across the division. No visible movement in 3.3 minutes is 100 ppm. I am toying with the idea of building a wow and flutter / speed meter, haven't gotten around to it yet.
BTW I don't actually use the DC drive at the moment, I installed a three phase EC (brushless) motor with a dedicated three phase AC drive in my Garrard.
I built the DC drive for Vinyl Addict. He measured about 0.1% drift but that was before we built the thermal drift compensation in, I expect it to be better now.
Matt Rowland said:
Origin Live use a negative impedance supply which they call "current compensation".
If you want to build a drive for the Linn, there is a design for a quadrature oscillator in the article in Bas's magazine. I sent the first rough draft to him yesterday, I don't know the publication date yet.
The design for the quadrature drive is on my website. Cost of the drive is about UKP 40 without power supply or enclosure, for a low power motor like the linn these wold be pretty cheap as well.
Quadrature drive page
pixpop said:
Doesn't work that way. The drag removes a certain amount of energy from the system and dissipates it as heat. For the speed to remain constant this energy must be replaced. Since the only source of energy is the motor, the motor must supply extra torque, which slows the motor down and the sytem comes back into equilibrium.
Adding extra sources of viscous drag will damp the response but not change its equilibrium point. Adding extra moment of inertia will reduce the rate of change of speed (eg the rate of deceleration) but also not change the equilibrium point. "No free lunch" is known in physics as the law of conservation of energy.
I understand. But what I meant was, apply a constant friction, and then set the motor speed. At that point, the motor will be producing much more torque, to balance the extra friction. Now, the additional load of the stylus would be seen as a smaller percentage of the steady state torque, wouldn't it?
I mean, surely, if there's a frictional load, you must get a different speed/torque curve from the motor. If the speed of the motor is constant, then the torque it is producing is equal to the total friction. If you apply more load, then you can increase the torque (by increasing armature current) to keep the speed constant.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.