V1.0 of Bryan's 3D printed head with feedback

[emidisc]

Hi Bryan yes piano wire for torque tube and the soldered links to drivers the chrome tubes are from pen tops (actually propelling pencils)made from very thin brass then they are glued to the plastic discs (they look like brass) that are in turn epoxied to the driver cones,
To facilitate alignment I needed to open up the holes for mounting the drivers and insert spacers as you mentioned in your post,
Using the piano wire for the torque tube as meant it is easy to adjust to suit the final position of the drivers.
I forgot to weigh the components but no doubt it will be stripped down again soon!!
I ran a little low volume audio to the drivers and everything appears to run free and clear
So fingers crossed.......

Emidisc

[EpicenterBryan]

Let's go over how things went this weekend. As you know the driver discs got installed off center, last ditch push rods were installed with flex in them...

After some extra messing around with the Presto to tune down some rumble, some more of the LF noise is gone. I tested the head with nylon, carbon fiber and piano wire torque tube. Frequency response was nearly identical. By a slight margin in the HF was a 0.45mm ( 0.017") piano wire. Below is the response:

BS-1 Piano filter 8D damping io2 input mute.JPG

That big dip followed by the big spike is bad. But I wondered what would happen with those crazy guitar string aluminum push rods... Now I know. For fun I used a parametric EQ with 20 bands per channel and messed around trying to clean the curve up.... just about smoked the drivers with all that extra energy in the HF. Oh, guess I should have set the parametric EQ for -10db output rather than +4db. No damage, but I guess I should lower the fuse ratings.

Will give it another shot with the correct levels and see how we do...

Bryan

[EpicenterBryan]

I should also add a photo of what the setup looked like with piano wire...

IMG_3074[1].JPG

And how bad the over-modulation was when I just about blew the divers.

IMG_3077[1].JPG

Bryan

[dimi751]

Hi Bryan

Great work so far the head is looking nice for a diy project. Have you decided to make this a dynamic head for now or does it have the feedback section as we'll.

I'm slowly trying to make my own diy head now (dynamic) and when I have the confidence
I would like to make a feedback head later on when I get a better understanding.

Also can you recommend a book that explains the inner workings of a feedback cutterhead
Wiring & design etc.

Thankyou for making the time and posting your work.

Best
Dim

[EpicenterBryan]

Have you decided to make this a dynamic head for now or does it have the feedback section as we'll.

For now, I decided to back up and make sure it could work with no feedback. It turns out it can't without a redesign of the push rods, attachment to the driver, stylus holder and torque tube and assembly methods (wow, that's pretty much everything). But all is not lost. Everything that has been learned is about to be implemented in a new non-feedback head. Depending on new lessons learned, I may come back to this design with all the new secret sauce, and then dig into the feedback.

Also can you recommend a book that explains the inner workings of a feedback cutterhead
Wiring & design etc.

Everyone recommends "DISK RECORDING VOL.1: GROOVE GEOMETRY AND THE RECORDING PROCESS edited by Stephen F. Temmer. " available from here:
http://www.aes.org/publications/anthologies/

I don't own it yet. But there is a ton of info on lathetrolls when you use the search function. Also, toss the word "patent" in any google search and lots of info is there.

As far as general feedback concepts goes, I can give you a quick rundown. I am a former analog / digital circuit designer, and did motion control projects as well. It's easiest to explain in a mechanical sense. Say you tell an actuator to move exactly 1.0", and pretend you have a transducer connected to the load to measure exactly how far it moved. And say the load only moved 0.9". That measurement could be subtracted from the commanded position to give an error of 0.1". So to move the actuator to exactly 1.0", you would combine the original command (move to 1.0") with the error signal (0.1 too short"). So next time you tell the actuator to move to 1.1" and you get exactly where you wanted to go ... to 1.0" because of the feedback error.

This example is kind of like in super slow motion. It makes it look like you commanded a movement, measured the result, took the difference, added the error to the original command and issued a new movement based on the last results. In the analog world feedback happens in real time without all the steps.

What we are really talking about is negative feedback. Taking the output (movement), inverting it and adding it to the original signal (or taking the movement error and adding it to the commanded movement)...

The link below will help I think.

http://en.wikipedia.org/wiki/Negative_feedback

Bryan

[EpicenterBryan]

Stay tuned in the next few days for a totally new post about a totally new design for a DYI head.

The head in this thread is not dead but needs so many changes (wow, what a rhythm there). I'm going to apply the changes to a new head with new drivers. The new head will not have feedback initially, but I have ideas.

These changes / new ideas can also apply to this design and offer many advantages that ease keeping things aligned while assembling. More details in the days to follow.

Now for the mystery photo of a mockup of the new design. In this shot, the push rods are steel. Those will be carbon fiber. Also notice a torque tube screw that is too long at the top, but the bottom one (at the front) also serves to tighten the stylus.

IMG_3096[1].JPG

Now the question before I share some #SecretKnowlage... why in the hell would I have the torque tube connect in the front and the back of the housing?

Also, notice the drivers are offset by 1mm front to back so the push rods to driver connections are easier to align.

In this shot (after a 5 1/2 hour 3D print), I discovered the stylus would not touch the disc, so I milled 5mm off the print on my mill. It will work for now...

More later.

Bryan

[EpicenterBryan]

Now the question before I share some #SecretKnowlage... why in the hell would I have the torque tube connect in the front and the back of the housing?

I should quote this again to poke at Todd, and make sure everyone pays attention to this one little detail. That secret sauce is SO good.... Just like French toast with real maple syrup. Damn.

[opcode66]

Now the question before I share some #SecretKnowlage... why in the hell would I have the torque tube connect in the front and the back of the housing?

I should quote this again to poke at Todd, and make sure everyone pays attention to this one little detail. That secret sauce is SO good.... Just like French toast with real maple syrup. Damn.

Because it will no longer torque. The tube will bow and the stylus won't move appropriately. That is why it isn't tethered at both ends.

And this design still fails to suffice one required design principle for stereo 45/45 paired transducers.

The 9th full build of my cutterhead wouldn't touch stylus to disc either. That was terribly frustrating. But, great learning experience. Because of the arrangements of parts, I had to start all over.

[dimi751]

Hi Bryan

Thanks for the update I'll try and sauce that AES book! Get it sauce !! Hahaha

Thankyou
Dim

[Ciuens]

Hello Bryan, I am working on a similar design using the same transducers. I'm still working on the torque tube, but looks promising.

[opcode66]

Hate to burst that bubble.... Bolts for your tranny linkeage.... You may want to majorly reconsider that.

One of the most important rules I learned working on all of this. Things will look right to you. But, will test horribly. Looking right and working right are two very different things.

What I found to be most helpful is to work on the sections and even the parts in a testbed. Instead of trying to build the whole cutterhead and make it work. Build sections of it and make them work on your bench. Then translate that to your head.

I started with just my transducers. Until they were working well, then I started to work out the head design.

My 2 cents.

What machine are you printing on? What plastic? ABS? Have you done any mods to get better prints? You should compare your body to mine. I've spent about a week total upgrading my makerbot, tweaking print profiles and learning how to trick the machine into giving me very nicely printed parts. That work paid off handsomely now. Parts are coming out straight/flat/nice.

[EpicenterBryan]

Hello Bryan, I am working on a similar design using the same transducers.

Hey there.

Interesting that you sunk the drivers down. That makes the head wider but sure works. The great thing is you have full access to the driver springs so if it turns out we need to dampen them they are accessible.

I'm with Todd about your push rods. Those have to go for sure. I bet you used them just to visualize where things need to align for the stylus holder and torque tube and such.

I know how hard it is to find anything light weight in 5mm to thread into the holes on the drivers, especially in a small town like where I am. I found some aluminum button head screws on line but only in a crazy kit to mount a windshield on a motorcycle. I ordered the kit just for the screws and guess where they are coming from - China. Hope they arrive soon. I also found something that sounds so crazy that I'll save it until they arrive and I can take a photo, again from China. They look like they can be shortened and screw in, and then accept a rod like an aluminum tube or carbon tube - either being glued or using a set screw to connect. Looks interesting and off the shelf, but hard to find. I was just about to buy a lathe to make something like that...

What I did find locally (so I could get on with an initial check with these drivers) after going to Lowes, HomeDepot, Gerry's Home improvement, Eugene fastener and Neilson fastener... was a 10-32 1/2" nylon screw at Eugene Hardware (an ACE hardware store). It's close enough to thread in there. Just cut it short so there are only 3-4 threads left. Then you can drill the screw head to accept a push rod. I drilled the center to 1mm for a carbon fiber rod, and glued the rod in place. When shortened, the screw will be about 0.24 grams. Trimming the head could reduce the weight drastically... which I didn't do. It will work for now. I'm sure I'll have it apart again when the other stuff arrives.

On the first try, I'm using a 1mm solid carbon fiber rod for the push rod. The length I'm using only weighs 0.05 grams.
I just glued everything tonight and am out of time but hope to try it out tomorrow night. Then I'll start a new thread...

Bryan

[EpicenterBryan]

Hi Bryan
Thanks for the update I'll try and sauce that AES book! Get it sauce !! Hahaha

Hey dimi, I just ordered that AES book myself today. Todd speaks highly of it and I think it has some of the #SecretKnowledge.

Bryan

[opcode66]

Sort of. There is actually surprisingly little on cutterheads themselves. It contains information on a wide swathe of topics. There is, however, an image of the internals of a Westrex 3D. That image is what made my brain kick in and understand a few of the principles. It wasn't spelled out in text.

[EpicenterBryan]

That image is what made my brain kick in and understand a few of the principles. It wasn't spelled out in text.

Todd, I'm messing with you. But I should say ... share the #SecretKnowledge, share the #SecretKnowledge, share the #SecretKnowledge!

Bryan

[Ciuens]

The drivers are not fixed properly, only for visualization and measurements for the rods. I'm thinking of using carbon fiber rods, which are light and sufficiently rigid. Stiffness is a very important factor, because the high frequencies very easily dissipate in softer materials. I'm using ABS for prints, but the ABS usually warping during cooling, the ideal would be to use PLA, but I'm a lot ABS here yet to use, . My printer is a RepRap Prusa Mendel

[EpicenterBryan]

No time to test tonight, but here is a photo of a new version that does in fact satisfy the requirements of a 45/45 stereo head!
The glue isn't even dry let!

Yes, the #secretknowledge has been applied and it will be explained in detail in a new thread. And, I have a new code name for the project which I will keep under raps until the new thread.

Here is a quick run down to get you thinking.

Todd has implied the solution several times, again to get people thinking. As a result, I payed attention to every photo I could find of a Westrex and tried to understand the smallest detail and why it might have been done the way it was. I even mocked up a torque tube and V spring assembly using feeler gauges and a threaded rod to demonstrate to myself how a stylus holder / stylus would move if the threaded rod was affixed to the 90 degree spring. This is how I acquired the #secretknowledge Todd is guarding... by experimenting and trying to visualize how it should work.

The goal is to keep the torque tube from rotating in the horizontal axis - ie., not letting the stylus deviate from an exactly perpendicular orientation with the disc surface. If the tube is sufficiently rigid, the V shaped spring in the rear accomplises this task if attached to the tube. The ends of the V spring attach to the body and act as a "torque tube anti rotation device". It is not an "L" shaped metal mounting bracket, it is in fact a spring that moves in all directions relative to it's normal surface, but not in a rotational axis relative to it's normal face. Clear as mud? I'll explain better in the new thread.

Also since the "torque tube anti rotation device" allows the torque tube to move in all directions (except in a rotational axis parallel to the cutting surface), the piano wire at the rear of the tube is required to keep the whole assembly from moving forward as a result of cutting drag.

Enough for now. A full description with diagrams and a frequency sweep in a new thread to follow.

For no, get this... Carbon fiber stylus holder with carbon torque tube extension / piano wire mount tube with carbon fiber collet to spring glue mount... and piano wire to rear housing body mount. Driver links (Push Rods) are 1mm piano wire.

IMG_3118[1].JPG

[opcode66]

The feet to the torque tube should actually bend. That is what drives vertical movement. As long as you have the fixed 3rd attaching point which is piano wire in your case. When the tube gets torqued, it also moves vertically. Now you get true stereo modulation.

You do want the tube and thus the stylus to torque, but, critically, also move vertically. The only way that happens is if the linkage to the torque tube from the transducers is stiff. And, also, you have to double spring the bobbin. Without double springing, the bobbin can teeter instead of moving just in and out. That is the secret. The image of the internals of the Westrex in the red book clearly has a double spring within the transducer. That keeps it from moving any direction other than primarily in and out.

Looking better my man. Let's hear some rips of test cuts!

[chaosbc]

Class act from you to reveal it. Sincerely.
Thanks

[inspector77]

Seems that you are very close to getting a fully functional stereo head , congratulations and thanks for sharing your knowledge.