DC motor control for belt drive

[Giordano]

Hello,

I'm thinking about to give up with the DC motors, for two reason.
1, yet I could not make an accurate speed measurement. I use a printed 54 segment paper ring, printed in normal (not econo) mode in laser printer, to normal paper. I measure the time of 4 pulses. Unfortunately, even at stedy speed, it jumps up and down sometimes 0.01%. The microcontrollers measures more accurate, if I feed with stable square signal, it measures to at least a decade more accurate. Of course, if I average it out, than it is good. I use a 16 bit counter. On live signal from the platter the error signal jumps up and down about +-300.
2, I do not know how to tune the control loop. I found pure integrative control works the best when the platter is already on speed. If I set the gain too high (0,1-1) than I can hear that it is acting in the background. If I set it to low value (0,005) than the reaction is too slow. The problem is, when I put the pickup on the LP, the speed drops about 0,05%. If I set slow control, it takes 10-20sec to control it out. If I set too high gain, I can hear it working during playback.

This all is applied to a 1.5kg platter. Is this platter light for DC motor drive ?
Do you have idea how to make better control ?

Regards,
JG

[bear]

I'm not a servo expert, or even good...

But a higher mass platter will even out the attempts of the motor to effect its speed IF the torque of the motor is actually low compared to the requirement of 1:1 control or more (overpowered). Or, if the transmission is slightly lossy.

What ur dealing with is "hunting" it sounds like - maybe.

You may need a two step servo time constant, one for start up the other for constant run.

You may not be taking enough samples via ur "ring". I'd think something on the order of 10x or 100x the samples would provide better control...

Not sure that a 0.05% drop in speed is meaningful when translated into change in pitch. I'd expect that it is invisible even to those with "perfect pitch".

How do you "hear it working" during playback? Is it speeding up, overshooting, and then coasting?

_-_-bear

[Giordano]

Hello,
The control already has two phase. First it spins up in PI mode, with relatively high I gain, than it switches to pure I control with very low I gain.
I found it more noticeable at big volume changes when there is single tone during the "silent" period. Not significant at all, but noticeable.
The weight of the platter is the key probably. So far I have seen DC motor drive on heavy platter TTs only, with a few excemptions. One of that is Linn Radikal (I have never heard that).

JG

[planet IX]

Hi Giordano

I ran a DC motor some time ago only with a voltage regulator without problems (LT1085 and L200 [there's a motor speed control-circuit in "A designers guide to the L200 volatage regulator" by SGS]).
I used a somewhat rigid coupling between motor and platter, so that when the platter is slowed down by the movement of the needle, the current in the motor increases and so does the motor-torque. I could not detect any significant wow and flutter caused by this arrangement.
I'm not an motor-expert, but I worked fine even with a 5kg platter. So if you want to use a digital control, maybe you should let it take only very smoth action. And don't use any elastic tape to drive the platter...

[Giordano]

Hi planet IX,
I guess that was voltage controlled, current limited mode. What time it took to settle ? I mean, from 0 RPM to 33 1/3 ?
My pcb can work that way, I can leave the motor on a voltage and has current limit.
The circuit is based on Manfred Huber's circuit what he made for the Teres project, except I use a modern microcontroller from another vendor and a 16bit DAC instead of the 12bit DAC (and it is SMT).
I made a mistake, I wrote I use a 1.5kg platter and yet it is not true. The plexi new platter on my rega is 1.5kg, but on the old deck I used for development I measured, it is only 0.78kg. I'm sure platter inertia is the issue.
I have a spare unpopulated pcb if somebody wants to play with it, it use a PIC18F6722, but I can only recommend it for :
-high mass platter, like the Teres
-high quality motor and platter bearing
Regards,
JG

[planet IX]

Hi Giordano,
yes, it was voltage controlled. It took 3-4 seconds to come to full speed.
Apart from your problem, I think that a light platter and a strong motor is a good solution anyway...
Regards, Boris