PID to control turntable speed? Options

[mccarthyk]

I just got a new (well, used) Technics SL-23 with a Shure cartridge, and I love the sound. The only thing is, it has a little trouble keeping speed. It doesn't change audibly but the little strobe-meter moves by the end of an LP.

I just got the idea of using PID controlled by an arduino or raspberry pi to control the motor speed. All I would need to do is add a color sensor and use the dots already on the platter for the calculation. And it would be easy to use a speed controller connected to the motor.

This seems pretty simple and I think it would really pay off.

Has anyone ever done anything like this?

This post has been edited by mccarthyk: Jan 18 2013, 05:12

[Glenn Gundlach]

Tach pulses and crystal references are the norm in broadcast video tape machines and film cameras. . Technics did have a crystal referenced turntable back in the '70s. Dr. Richard Greiner, ECE professor at UW Madison said it was "an elegant solution to a nonexistent problem". You did say you couldn't hear anything amiss. You might find variable speed a little more 'interesting' with PID - or maybe not. What you could try is a high count tach wheel on the motor and use a monostable to get a constant width and amplitude pulse, integrate it to a DC level and compare to a precision reference Voltage. You might find that to be sufficiently accurate. BUT and this is very important, a tach wheel that is not EXACTLY centered will have once-around phase and consequently speed errors as designed in WOW. Using DSP and a once around tach to always know where it is in the cycle, you could calibrate out the phase errors even if not centered so it's not as trivial as it seems.

But good luck with it.

G²

[mccarthyk]

Tach pulses and crystal references are the norm in broadcast video tape machines and film cameras. . Technics did have a crystal referenced turntable back in the '70s. Dr. Richard Greiner, ECE professor at UW Madison said it was "an elegant solution to a nonexistent problem". You did say you couldn't hear anything amiss. You might find variable speed a little more 'interesting' with PID - or maybe not. What you could try is a high count tach wheel on the motor and use a monostable to get a constant width and amplitude pulse, integrate it to a DC level and compare to a precision reference Voltage. You might find that to be sufficiently accurate. BUT and this is very important, a tach wheel that is not EXACTLY centered will have once-around phase and consequently speed errors as designed in WOW. Using DSP and a once around tach to always know where it is in the cycle, you could calibrate out the phase errors even if not centered so it's not as trivial as it seems.

But good luck with it.

G²

Well, I can't tell if you were actually giving me an alternative, or just giving me something (in my opinion) more complex than implementing PID. And I agree with Prof. Greiner's quote, but I'm an Astro Engineer student, and if I've learned anything: it's that my duty as an engineer is to come up with elegant solutions to non-existent problems!

This post has been edited by mccarthyk: Jan 18 2013, 05:58

[MikeFord]

The first step I would suggest is to examine your player and see why what you are observing is happening. The second would be to look at the existing speed control system and drive components and perhaps give a fresh thought to how it might be regulated. If you give it some thought I believe you will see PID is not well suited for a turntable.

[AndyH-ha]

If the speed is current variable - you can adjust it - then it is, in high probability, cone with a variable resistor, .i.e. a potentiometer. Dirt and corrosion in the potentiometer is a common cause of speed variation. This is often fixable with a squirt of electrical contact cleaner.