VMS70 Fixed Pitch Stablization

[opcode66]

I've been performing some interlocking groove cuts (groove in groove) at fixed pitch. I am simply not feeding any preview L+R or program L. However, my pitch is not staying 100% stable. The LPI meter shows wiggling of the needle. And, my two grooves keep running into each other without mannually riding the pitch adjust and microscope.

Any suggestions from anyone regarding this issue? Is there anything that can be tweaked to get more stable fixed pitch on the VMS70? My alternate is to make a separate fixed pitch motor control system that would be far more stable.

[mossboss]

The subject has been bought up here long ago
There are posts on the subject memory tells me about 4 years ago.
Best

[opcode66]

I searched before posting. I can find three thread posts with information that is sort of relevant. But, does not answer my question specifically.

http://lathetrolls.com/viewtopic.php?f=1&t=1403&p=6107&hilit=fixed+pitch#p6107

http://lathetrolls.com/viewtopic.php?f=1&t=1696&p=7835&hilit=fixed+pitch#p7835

http://lathetrolls.com/viewtopic.php?f=1&t=1696&p=7842&hilit=fixed+pitch#p7842

If you can point me to the thread in question that would be very helpful. I've searched the site for "fixed pitch" and read through nearly all the replies. I'm under a time crunch with respect to this so I could really use whatever help anyone can offer.

[mossboss]

I am not in a position to answer you specifically Todd I simply do not know the easy way
However I do know of a vms which has been modified to do an inside out cut with fixed pitch
I will get in touch with the guy and see if he can point out the mods carried out
I think he will be willing to let us know He is a bit of a buff So let me see what comes out of it
I will keep you informed but give it a few days
Best

[markrob]

Hi,

This is one instance where a 6N or similar lathe has a huge advantage over a VMS. Having the feed screw directly coupled to the platter rotation insures that there is no accumulating position error as the head moves inward. In the case of the VMS, there is a DC servo running the feed motor at a constant velocity, but there is no way to to make it track position over time. It boils down to how much accumulated drift occurs of the course of the cut. Given how small the spacing between grooves area at 120 LPI, it does not take much error to cause the second groove to collide. One thing you can do that might improve things is to have a look at the schematic and find the point where the feed servo reverence voltage is developed. There will be a summing junction there that sums the fixed pitch reference voltage with the variable voltage developed by the pitch controller. If you break the variable signal from the summing junction, only the fixed voltage will be present. This will eliminate any noise produced by the pitch controller from causing slight speed command variations. You could also check the stability of the fixed reference to see if noise or drift is adding to you problems. Still, it seem like a tough nut to crack. You would have to develop a servo system that looks at platter revolutions vs. feed screw movement and keep them locked. I've seen this approach used on metal lathes to eliminate the need for a gear box on the feed.

http://autoartisans.com/ELS/

Its been awhile since I followed this project, but it shows how difficult it is to mimic the behavior of a simple geared feed system. Who knows, maybe this could be adapted to the VMS.

Mark

[Greg Reierson]

Dumb question: is it stable without modulation? The pot in mine is a bit dirty and can take a while to settle into the setting I want, but once there I don't notice it moving around.

[opcode66]

Thanks Mark. That is essentially what I was thinking. Either modify the current system to disconnect the variable voltage and only run via the base lpi. But, that might not be 100 stable either. Then, I thought I would make a digitally controlled pwm source. That would end up being fantastically stable. I could use the photosensor to get quater turn timing from the platter.

I will experiment over the weekend.

Greg I think you mean preview. And, yes, that is the only way to achieve "fixed pitch" on a Neumann. With no preview my pitch needle wiggles ever so slightly. As Mark pointed out, that is enough to produce collision.

[markrob]

Thanks Mark. That is essentially what I was thinking. Either modify the current system to disconnect the variable voltage and only run via the base lpi. But, that might not be 100 stable either. Then, I thought I would make a digitally controlled pwm source. That would end up being fantastically stable. I could use the photosensor to get quater turn timing from the platter.

I will experiment over the weekend.

Greg I think you mean preview. And, yes, that is the only way to achieve "fixed pitch" on a Neumann. With no preview my pitch needle wiggles ever so slightly. As Mark pointed out, that is enough to produce collision.

Hi,

PWM is not any better than analog (it could even be worse). At the end of the day the PWM signal is still analog in that its wave shape is averaged to result in analog value. Any variation in the ground level, Vcc level or, time jitter will show up as analog noise. You also have to deal with the time quantization of the processor clock in your ability to set the duty cycle.

You really need a closed loop position system that mimics a gear driven control. If you think about the accuracy required for a velocity driven feed, it gets ugly fast. For example, if you compare two heads moving at a rate to produce 100 LPI with the second head being off by.1% (100.1 LPI), you have a groove spacing of .010" and a nominal speed of .010"/rev or .00555..."/sec at 33.33 RPM. If the second head moves .1% faster its rate is .0056111..."/sec. The speed difference is .000055561"/sec. At that speed difference, you would crash into the adjacent groove after moving .005" in about 90 seconds. Also not taken into account is the speed stability of the feed servo on the VMS.

Look into the ELS project I referenced. Its open source and I think it could easily be adapted to the VMS. It might just be easier to pick up a 6N for these type of projects.

Mark

[opcode66]

I would have to mount my Neumann head on the 6n to get the hi fi quality that is required for this project. Not about to try that.

The Neumann pitch system generates pwm, fyi. Not sure if you were aware if this.

Even if it is less than entirely accurate, I can ride the pitch adjust knob while watching on the scope. As long as it is more accurate than the stock system.

Worn out tachometer might be part of my issue.

[markrob]

Hi,

I was not aware. I would bet its an analog PWM not digital.

You might be able to make that work using the scope. Sounds tedious. But whatever gets the job done.

Mark

[opcode66]

Precisely. Which is why I was thinking a digital counterpart that is switchable might be a more accurate arrangement. Thank you Markrob.

Still taking input from anyone who's done this on a vms.

[jtransition]

I assume that you have tried replacing the belt?

[opcode66]

Within the last year, yes.

[opcode66]

The red box highlights the area of the diagram that represents the Control Voltage to Analog PWM conversion. The line that has the Yellow circle represents the connection to the AS-L1 card that controls the pitch motor. Pin number 31.

I see two possible solutions to my dilemma. Markrob, your suggestion represents a third that I will try if these two ideas fail to work. Since the lyrec motors are relatively stable, I'm not sure I need to tie the pitch motor to the platter as tightly as you are suggesting. As long as I supply a more stable fixed pitch than the stock solution I should be golden.

1. I could simply make the connections in the green box switchable. Thereby cutting off the input from the pitch computer and from the tachometer. The pitch motor will run with no compensation at that point. And, it may be stable.

2. I could make the connection in the yellow circle switchable. I could then inject my own digitally controlled PWM to replace the analog pwm. No input whatsoever from the tachometer. Just consistent digital PWM. I'm not sure what voltage I need to pulse however. My guess is that it is between 5V and 14V. I can use my test cables to run the AS-L1 card outside of the lathe and measure the PWM voltage between pin 31 and that card's ground pin. Then implement a very simple arduino based circuit with potentiometer to vary the PWM being generated. The circuit would have a transistor and a power supply properly regulated to mimic the analog PWM. I feel this solution will work just fine for my particular application.

[markrob]

Hi,

Looks like a good plan. It easy to try and I don't see any harm. I think you'll need to keep the tack feedback otherwise, the motor speed will change over time and load. The tach will compensate for that. But you'll find out fast if you start to play.

Mark

[opcode66]

I understand. I fear, however, that one of the main issues with the on-board system is that the tachometers are worn out and no longer give smooth feedback. Ultimately, I'd like to upgrade the tachometer to an optical rotary encoder with a translation circuit to mimic the built in motor tachometer. Starting with that, I would know I was getting accurate feedback. I'm guessing the on-board system would magically start working much better for fixed cuts (and variable).

Either way, I still feel that digital with no feedback will actually be more stable than how much it currently drifts...

Thanks for your ongoing input Mark.

[smithadamm]

I think you'll need to keep the tack feedback otherwise, the motor speed will change over time and load. The tach will compensate for that.

I could then inject my own digitally controlled PWM to replace the analog pwm. No input whatsoever from the tachometer. Just consistent digital PWM. I'm not sure what voltage I need to pulse however. My guess is that it is between 5V and 14V. I can use my test cables to run the AS-L1 card outside of the lathe and measure the PWM voltage between pin 31 and that card's ground pin. Then implement a very simple arduino based circuit with potentiometer to vary the PWM being generated. The circuit would have a transistor and a power supply properly regulated to mimic the analog PWM. I feel this solution will work just fine for my particular application.

I ran my pitch motor for a while with a simple PWM signal from an arduino through a power transistor and it would drift some under load. I upgraded to using DC voltage coming from the motor for feedback. I had to stick a cap across it to get DC smooth enough for the microprocessor to work effectively. The whole rig worked best using a PID algorithm to process the feedback signal. I still have the arduino PID code I copied from the internet and modified if you need it. It was stable and the pitch was pretty much bang on, the only drawback was motor noise from the PWM signal. I tried several different things to get rid of it, without any success.

[opcode66]

Please share. I would be very thankful. I'm still a little fearful that the tachometer on my pitch motor is worn out and therefore would not be the best thing to use for feedback. However, I could be wrong.

Also, it should be noted that I determined from the I card schematics that the voltage that is pulsed for PWM is rated at 12V in case anyone was interested.

I'm creating a hand drawn schematic for this idea with no feedback right now. Stealing power from the L card for 5V for arduino. And, using the 12V from the I card for PWM.

[smithadamm]

Please share. I would be very thankful. I'm still a little fearful that the tachometer on my pitch motor is worn out and therefore would not be the best thing to use for feedback. However, I could be wrong.

Also, it should be noted that I determined from the I card schematics that the voltage that is pulsed for PWM is rated at 12V in case anyone was interested.

I'm creating a hand drawn schematic for this idea with no feedback right now. Stealing power from the L card for 5V for arduino. And, using the 12V from the I card for PWM.

An old trick I used on MCI tape decks with worn out tachometers was to strap a bi-polar cap across the motor tachs to stabilize the phase locked loop system, you might give that a go.

I'm attaching all of my notes on the project and the labs I found online that I used to get my feet under me. I ran mine at 12VDC, and, if I remember right, needed a ton of current. If you have a second, I'd love to see how Neumann was able to use a PWM system to drive their pitch motor and keep it quiet. My Haeco pitch computer drifts from time to time and I'm always kicking around the idea of going back to the PWM system.

I cut on a Scully BTW, so there are some differences to be sure.

[opcode66]

I'm planning to use the existing 12v to drive the PWM. Just rerouting from the stock system to mine.

Thanks!