|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
203.214.90.154
I'm happy to say that after several months I seem to have reached my desired goal of being able to lock the speed of an external DC motor to a stable frequency reference.You will forgive me for not revealing all the details on how it works, I don't wish to compromise my IP in this. Basically I used some techniques borrowed from analogue computers to derive a frequency from the motor voltage and current. I then wrapped this into a frequency locked loop with a reference frequency provided by an oven stabilised crystal oscillator. The loop locks the motor speed to the oscillator frequency.
I need to do more testing (and listening) but it actually seems quieter than the motor with my standard DC controller was. Not sure what the speed stability is yet, it's beyond my measurement range. I suspect it's about 1 part per million which is about 100 times better than a standard quartz PLL and about 1000 times better than a mains synchronous motor.
First application will be as a replacement motor and drive for the Garrard 301, so I now have to make it work on three speeds. Since everything is slaved from the oscillator this shouldn't be hard, i'll keep you posted on progress.
Follow Ups:
Mark,if/when you get around to production will you make a single 33.3 speed as well as a three speed motor?
I will eventually replace the logic circuits with a software routine running on an Atmel AVR. This will mean that the speed selection is a software subroutine so there would be no or very minimal cost difference between single speed and three speed.Since producing two different versions would cost far more than simply not using two of the three speeds a single speed version is very unlikely.
Congrats on this, Mark. I predict you'll really like how it works on the Garrard. If you are interested in seeing schematics of another circuit that has accomplished this, utilized via a belt-idler hybrid drive, e-mail me with your snail mail address and I'll be pleased to send them to you.
to what's actually turning, namely the LP (platter) not the motor? So if I understand your concept the motor/oscillator does not actually "know" how fast the platter is spinning?Just curious.
Garth - many designers have gone down the path you suggest but they all end up with the same problem. It basically all depends on whether the platter speed is strictly dependent on the motor speed. If it is (tight coupling), controlling one controls the other and the control can be very tight indeed. This is the case with direct drive and is closely approached in idler drive.If it is not (loose coupling) the looseness creates a lossy term which also occurs inside the feedback loop. This means that the loop response must be slowed down so that the loss does not cause oscillation. The problem is analogous to applying global feedback around an amplifier with significant phase shift in the open loop - see any electronics text on using a Bode diagram to understand this. Basically the loop response must fall to less than unity before the phase shift hits 180 degrees.
This is why belt drive TTs with platter feedback use very slow loops - a good example is the controller on the Teres decks. Unfortunately such a loop cannot do anything about short term fluctuation such as that induced by modulation drag so it gives the classic belt drive wow.
This problem can only be overcome by compensating for the loss term and if that could be done accurately we would be back at tight coupling.
In my opinion it is better to use a tight loop on the motor and then worry about the transmission loss once the motor speed is right. I'm pretty sure I can solve the problem of belt drive transmission loss but since i'm concetrating on idlers at the moment you'll have to wait.
I'd be interested in a motor/controller that could be retrofitted to the likes of Thorens and Linn suspended decks. (TD150 - LP12) Platter feedback could be arranged by the addition of a strobe ring attached to a subplatter and perhaps an optical pick-up. I think this would be the biggest challenge for a couple of reasons: 1) elastic belt is very lossy 2) relatively light weight platter system and lower moment of enertia. But if you do work in this area.........With regard to your assessment of the Teres motor and controller I think one also needs to consider platter mass-moment of enertia at the rim and bearing drag into the whole of it. Like other large-platter turntables, once you get that platter up to speed it tends to keep spinning there, taking longer than a minute or two to coast down. An extreme example might be the Walker with its 70 lb air bearing platter system. Lloyd's not particular about what drive belt you put on it because once that platter gets up to speed, he can cut the belt and it just keeps on turning and turning and..... Anyway back to the Teres situation. Wow isn't something that is going to be heard on one of these when all is working properly. Perhaps you can measure it by instrument, but the human ear is going to think things are very pacey indeed. Or perhaps I'm just predisposed to think that. There's always room for improvement. An improved motor/controller for the Teres would be welcomed by me fwiw.
-Steve
Someone else will have to do that. For reasons outlined above I do not like feedback around the drive system, far preferring to get the motor speed right and keep tight coupling to the platter.In a sense it's like the difference in amplifier design strategies. On one hand there are those who say it's best to get it right so you don't have to fix it (all triode, no feedback) and those who let it drift then correct it back (whatever device with lots of feedback). I'm a triode man myself.
Yes, I think you are right belt drives with only the motor speed controlled don't seem to work well, not the ones I have tried at least.The Teres system (actually developed by Manfred Huber) works extremely well. No wow here.
What about tape (mylar) instead of rubber belts?
... but they are very different in the way they couple platter and motor, and I gather a good DC motor is needed with mylar.In earlier discussions people have said bass and transient definition is better(?) with a mylar belt, and even that strobe markings were sharper, implying less flutter.
This post is made possible by the generous support of people like you and our sponsors: