BD's Mechanical Disk Recorder Build

[markrob]

Hi,

Very similar to mine. Still works fine after all of these years.

Mark

[Bahndahn]

With all off my materials in order to get experimenting with my record cutting lathe, I am patiently waiting for my anodized aluminium. I've set up a little corner in my studio for my machine.

For amplification I have a Peavey DPC 750. It is a very efficient amp with 'no slew rate' they say. It's the only thing I have that has enough power so I think it will be fine.

http://assets.peavey.com/literature/manuals/80301435.pdf

For some general experimentation during my first cutting experiences I picked up an Ashly Protea System II 2.24ps which is a 2-channel 12-band parametric EQ, compressor limiter, high & low shelf EQ, etc. processor that may be useful in carving inverse RIAA and open-loop frequency response compensation. It's a digital thing, so saving and recalling settings is possible as well as standalone processing.

http://www.ashly.com/products/discontinued/data-sheets/protea-2.24ps.pdf

Tomorrow is likely the pickup day from the anodizing facility!

[Bahndahn]

Hello folks!

I have been busy setting up and tuning the structural components of my machine, the parts that have been anodized have an intensely rigid surface [compared to the 'raw' aluminium]. It's refreshing not having to worry about nicking the surface etc. A strange reaction occurred during anodizing that left the bottom of my base-plate looking like ripples on the surface of water:

photo (4).JPG

Too bad its not the top!

The machine is largely assembled [along with the first version of the cutterhead], heres a quick photo of the progress, I will post hi-resolution photographs once the assembly process is completed.

photo (5).JPG

While testing, I blew the 'Big Easy Driver' I was planning on using with the stepper motor I have rigged up. I was adjusting the current potentiometer [small thing requiring a screwdriver] and I accidentally touched the screwdriver to a nearby capacitor casing [presumably] shorting something out and frying the chip. I was adjusting the current because the motor was pretty noisy.

I'm having my doubts about the stepper motor. I don't have the money to buy a fancy driver + the power supply it will need and I'm not sure running it with audio amps is working properly [the motor and the amps get really hot yet the torque is poor (perhaps due to using sinewaves)].

In considering my options I realized I had a 8mm film projector sitting around so I tore into it and extracted a very satisfying find:

photo 1 (1).JPG

This is a DC motor with a feedback device on the back end of the spindle:

photo 2 (1).JPG

The feedback apparatus seems to be a magnetized ring [alternating polarity segments or something] that stimulates a 'slotted cage' surrounding it that is connected to a coil.

When I apply 16 volts to the motor, I get an average AC voltage reading of ~3-ish from the feedback. My low-quality smartphone strobe-light tachometer suggests the RPM is around 1250 but that reading is not to be trusted. The torque from this test was pretty good.

When I apply 32 volts to the motor, I get an average reading of ~4.7v from the feedback coil. The motor is spinning way too fast for platter drive without gearing at this point. I didn't take an RPM reading. The torque was high but still malleable by applying resistance to the spindle with a glove on.

Of course, the feedback winding directly outputs the speed changes brought on by applying load to the spindle. This is how this will be useful. I am looking for advice, as I am not very apt with circuitry, on how I can make a motor speed controller with feedback for this motor. It will probably need to be PWM based because even at ~15v the RPM seemed high.

I have been researching a number of voltage regulator based methods of achieving control with feedback but I haven't dabbled in the PWM options. Any advice will be appreciated!

[markrob]

Hi,

A few weeks back I cobbled up a simple DC servo controller for a pancake motor I'm using to drive my turntable. Here is a link to the motor:

http://www.herbach.com/Merchant2/merchant.mv?Screen=PROD&Store_Code=HAR&Product_Code=TM00MTR4434

It is not PWM. Just a pure DC series pass transistor to drive the motor. I used a LM2907 F/V converter and simple PI loop to regulate the speed. I've attached the schematic, PCB layout (1:1), and an assembly print. The pass transistor is located at the PCB edge so that it can be coupled to an external heatsink. You might have to play with the values of U2 to scale for your tach (see the datasheet for the LM2907 for details). The servo gain might also have to be changed to make the closed loop stable. On my build, I ended up removing R6 (the proportional gain adjustment) and ran as a integrator only loop. R4 in combination with C3 sets this integrator gain and R6 in combination with R4 sets the proportional gain. The speed adjust pot, R5, can be located remote from the PCB with no problem. The circuit is ratiometric, so it holds speed in spite of variation in the DC power supply voltage. Feel free to use this if you find it useful.

Mark

[Bahndahn]

Mark, This is exactly what I have been looking for! This is invaluable to me, thank you for sharing.

I'm going to piece together this circuit over the weekend, I will report back here soon.

Thanks again!

[markrob]

Hi,

No problem. If you have issues or need help scaling for your motor, let me know and I'll see if I can help.

Mark

[Bahndahn]

I have put together this circuit and am testing it out, trying to stabilize the system. I built it as close to spec as possible, in some places I needed to put a two resistors in parallel to obtain the correct value.
One place needed two capacitors in parallel for this; C6 was obtained with two 470pf caps in parallel rendering that value slightly off.

After assembly, I hooked up the motor and tach, applied 20vdc to the + - terminals and ran a test with the [unloaded] motor. I was surprised to find that it was working as predicted, keeping the rpm [seemingly] stable when I applied random load to the shaft. I began to notice that when relieving the load, the motor would ramp up before stabilizing. This behaviour seemed logical based on momentum or something yet it contributes to my understanding of my setup not being stable yet.

Is this suitable to run at 20vdc?

R9 on the schematic is marked as 0ohms, I first had a wire connecting the corresponding points but now have a little trim pot that ranges from 20 to 50 ohms. I cant tell if thats helping.

I left R6 out too.

The behaviour I am experiencing now when I test it under loaded [turntable belt driven] conditions is as follows:

-when I apply 20vdc [currently I only have 20vdc power supply available] the motor ramps up substantially above the potentiometer-set speed
-it eventually stabilizes but continually oscillates subtly

-the oscillation I am hearing corresponds with the values I am reading [with my multimeter] at the summer/integrator's inverting input [the buffered signal from the frequency to voltage IC]
-the value oscillates around the target value

-it seems to take a [relatively] long time to respond to the changes induced by the potentiometer

I am suspicious that this is due to the values referred to as R1 & C2 on the datasheet for the FtoV [R1,C6 on the provided schematic].

If you have any tips on what I need to look for beyond the advice already given, I'm all ears. One unfortunate factor with my setup is that I don't have a way of measuring shaft speed. I haven't measured tach frequency yet either, but the latter I am capable of.

I will keep trying things out in the mean time!

Here are some photos:

[markrob]

Hi,

Looks like you are in pretty good shape with the build. Just need a bit of tweaking to get things under control.

1. The circuit is fine at 20Vdc. The LM358 op-amp is rated for +36 Vdc and the LM2907 is rated to 28Vdc
2. R9 is just a jumper. It was placed on the schematic to make my PCB layout work well as a single sided design. Adding resistance in place of this will not do anything.
3. There is an error on my schematic. The values of C7 (C1 on the datasheet) and C6 (C2 on the datasheet) are reversed. the 1000pf cap should be at the C7 location (pin 2 on the LM2907). This, in conjunction with R1, gives a voltage output of .002V/Hz with Vcc = 20 Vdc. Try reversing these two caps and see if things get better.

Do you know the frequency output of the motor tach at your max desired speed? This would be helpful to scale the V/F if you are still having trouble. However, I think once you reverse those two caps, you will be in a much better operating range. Sorry about that!

Mark

[Bahndahn]

Alright! Swapped those caps around and it seems to have helped a bit.

Measured the maximum tach frequency based on a allowing the platter to achieve 78rpm+. With a strobe-tachometer printout on the platter and an incandescent bulb, I made sure the platter was clearing 78rpm by a little bit. It was hard to fine tune the high speed with the 20k pot I have hooked up. I measured the frequency of the tach while the platter was at this 78rpm+ speed and got roughly 515Hz. This is the absolute maximum tach frequency I may need to obtain.

The circuit is not yet bearing applied 'load' very well, do you have a recommendation as to a integrator/summer gain calibration method?

Thanks for your help so far, I'm learning a lot doing this.

[markrob]

Hi,

If 515 hz is the max frequency, then you should increase the gain of the F/V converter. You can do that by increasing the value of R1 and/or C7. As it sits now, you should see about 1 Vdc output from the converter at 515 Hz. I would get that up to 5 Vdc at 515 Hz. So increase R1 to 200K (was 100K) and C7 to 2700pf (was 1000pf). That yields a conversion gain of 20 * 200 x 10^3 * 2700 x 10^-12 = .0108 V/hz. At 515hz, you would have an output of 5.56 Vdc. Since you are running at a lower frequency, I would increase the value of C6 from .1uf to maybe .47uf. You may need to experiment with that to get the best trade off between ripple and response time.

To check things out, I would disconnect the motor from the controller and run it with a variable bench supply. Run the motor up to max speed and check the dc voltage at pin 4. First, verify that you are sitting at about 1 Vdc with your present scaling. Then make the change and make sure you are getting about 5 Vdc. You can also measure the ripple by switching the meter to read AC volts and noting the value. As you increase C7, you will see this drop. However, the response to changes in speed will also slow down. You can live with more ripple than you might think because the integrator will average that out.

I assume you are using a 2K speed adjust pot as shown on my schematic along with R8 at 5.1K. This limits the max speed reference voltage to about .3 Vcc or 6 volts with Vcc at 20 Vdc.

With the converter gain scaled better see if you have better dynamic performance. If not, you can tune the closed loop by playing with C3 and/or R4. As you decrease the value of C3 or R4 the integrator ramps more quickly (don't lower R4 too much or you will load down the converter). This will speed up response time. However if you lower this too much you will start to see a tendency to oscillate. In the best case, the control loop will ramp the motor up to speed with little or no overshoot. If you have trouble getting the loop to settle down, try adding in R6. This will improve stability at the expense of a droop from the set speed reference and the converter voltage at pin 3.

Also note that you should make sure that the motor voltage required at full speed and load is lower than the 20 Vdc power supply rail. The combination of the op-amp and series pass transistor cannot swing the motor output too close to Vcc. I'd leave a few volts of headroom there. If you leave too much you will need to burn more power off in the pass transistor. So its a trade off once again.

The servo is not designed to be a high bandwidth fast responding system. Think of it more as an auto pilot. Fast step changes to the load will not be tracked.

Hope this helps.

Mark

[Bahndahn]

With this information I have experimented with a number of values and have achieved a setup that seems to be working well. I went with your tips in a pretty 'blind' way, simply shooting for higher output from the FtoV. With this in order, I went on to tweak the integrator values; by substantially decreasing the size of C3, in combination with the higher voltage from U2, I have obtained a sufficiently stable system as far as I can tell right now.

Thanks you very much for your help, Mark!

With this in hand I have finally cut [perhaps more literally scratched] my first 'records'. Using the bottoms of CD's and some carbide vinyl plotter blades I have engaged in some sort so-called embossing. At this point, I was very happy to discern some semblance of the original recording attempted, but let me assure you the results don't extended beyond semblance.

Now the real experimenting begins for me; time to get things sounding good! Im going to try to grind some carbide plotter blades into shapes conducive to embossing style using my turning lathe's compound slide [for precise angles] and some home made grinding attachments.

Photos and more updates to come.

[markrob]

Hi,

Glad you were able to get things under control. I wanted to give you some additional info. Since you had to increase the integrator cap, you might want to consider increasing the value of R4 instead. As you go with higher values on C3, you could run into problems with leakage current. Also you'll note on the schematic that I specified a polarized capacitor. I did that because its hard to get non-polar non-electrolytic caps in that value range. The cap should really be non-polar, but I get away with it as shown because most of the time the voltage at the output of the op-amp is higher than the the speed reference voltage. However during ramp-up, that is not always the case. The ideal situation would be to lower C3 into a range that you could find a film cap for example:

http://www.digikey.com/product-detail/en/R82DC4100CK60J/399-11516-ND/4833342

Then, use a bigger value for R4 such that you end up with the same RC time constant. Given the large input impedance of the LM358, you could probably go up to a 1 meg resistor with no problem. You could also change the op-amp to a better part and work with even higher values for R4. It should be a rail to rail single supply type with high impedance and low input bias current. Bandwidth is not too important as we are in the 1 Hz close loop region. It must also be able to operate at over 24 Vdc.

Mark

[Bahndahn]

Interesting, I will have to grasp this additional information regarding leakage current at my own pace, however in my next digikey order I'm certainly going to include the recommended capacitor and a rail-to-rail op amp.

I really appreciate this info!

As I mentioned, I have gone about cutting some little drill bits for sharpening as a stylus. My handheld rotary tool happened to have a perfect sized collet for the outer diameter of the drill bit I am using for stock. As a first go at it, I chucked up this stock-bit, and tried my hand at making a cone. It worked well as a first try, I will certainly get better at it.

photo 1 (2).JPG

photo 2 (2).JPG

I came across a thread on this forum where people were discussing lapidary ventures [in 2010] but not a lot of followup was provided. Mark, you were getting nice results forming steel on your homebrew lap. Did you manage to make steel styli capable of cutting polycarbonate at all? How was your experience with that sapphire rod you got?

I have been considering either making a mini-lapidary grinder or somehow making my turning lathe work as one. The two ways I have thought of with the craftsman618 are:

-use the tool post to hold the stylus stock in a collet that is mounted to a 'jig bar holder' – for indexing the facets, use grinding wheels attached to the lathe spindle [chucked] and move the compound slide to set in various angles and engage in grinding/polishing

or

-chuck the stock stylus into the lathe chuck on the spindle so that the indexing system on the lathe's gears can be used to set the correct facet, the grinder would be mounted to the tool post [my rotary tool has a flexible extension attachment] – the grinding wheels would be moved up to the stationary stylus etc.

If these don't work, I will make a little lapidary system.

A little induction hardening rig could be easily made to harden the steel, has anyone tried this?

Heres a photo from some late night testing:

photo (6).JPG

[markrob]

Hi,

Your build is really coming along. Very impressive.

I had decent results with my home built lapping machine. I was able to grind and lap cutting (not embossing) stylii using drill blanks. I had best results with high speed steel. For some reason, I had poorer results trying to facet carbide blanks. Sapphire rod was harder to work with because it is quite brittle. I'm sure if I spent more time, I would be able to tame that. I used a simple method called template faceting that does not require and expensive machine. Works really well.

https://www.youtube.com/watch?v=1_fmqWBBUdk

I was never able to get a sharp or hard enough edge using HSS to cut anything other than lacquer blanks. I suspect, that if you master faceting sapphire, you would be able to cut some plastics, but with very short life. That's ok, because if you have the means to sharpen in-house, you could live with the short life.

My first idea was much like yours; to use a bench lathe with a wheel in the spindle and a jig to hold the stylus. The problem is that you really need to wet grind with a diamond lap to end up with a mirror finish and the "on lathe" arrangement is not so well suited to that. Also the thought of grinding dust on the lathe bed was not so attractive to me. If you use drill blanks, I don't think you'll need to do any hardening as they are supplied hardened and must be ground to shape.

I do hope to get back to this some day (too many projects and not enough time).

Mark

[petermontg]

Hey BD,

Great work. Long may it serve you well.

Best
Peter

[opcode66]

Post audio!

[Bahndahn]

Thanks for your tips Mark, I best be turning a little lap out for myself on the lathe rather than trying to make it something it's not. I will make a little disk and a brass spindle, power it with a little dc motor etc. Might even make a little arm to adjust stuff, I want to practice machining.

Thanks Peter for the well-wishes!

Todd, as I am sure you know, there is quite a long path ahead of me in getting this machine going. I am learning to cut records for the first time on a machine that has never cut records before. Via my initial tests I have now obtained a trackable cd-record that contains a recognizable track, distorted quite substantially from the original spectra. This is, and will continue to be a 'grass roots' journey of creating the things I need from whatever I can obtain. In due time, I will share some clips from my modest sonic accomplishments.

This week I am studying for a university math final so I will be diverted from my otherwise sole focus of working on this project. More to come!

[Bahndahn]

Attention lathe builders and mechanical types:

I'm looking for input on how to construct a mechanical 'lift mechanism' for my cutterhead apparatus - some sort of lever, some sort of motion reduction to translate more broad movements into more fine ones. Right now I am just lifting it by hand and keeping it up by reattaching a spring..

Any ideas? Here's what i'm working with:

IMG_1708.JPG

The brass bolt is the pivot and the bolt head nearest the cutterhead is the fixed angle-set.

IMG_1709.JPG

Thanks!

[Bahndahn]

Following the leads of others working with the 'surface transducers', I have spent the last few days completely redesigning the spring apparatus to reduce moving-mass and increase compliance. Previously, I separated the stock bobbin from the heavy and rigid springs to fix them to 3D printed springs and cones. This seemed to work pretty well, but I thought I would completely redesign them to acommodate a new spring setup and bobbin/coil.

Most of the parts are 3D printed except for a carbon fiber rod that will pass through the center. I also have an idea to use the magnets already in the transducer for feedback, but I'm not sure it will work; I will get into that in a future update.

For now, heres a look at the new design, currently printing!

Screen Shot 2015-09-08 at 3.51.21 PM.png

[opcode66]

Why not just use two of the same spring. One invested to the other.