Project "Groove Scribe"- A DYI stereo 45/45 head.

[Ciuens]

Hello Bryan, I suspect the problem is in the cartridge. I'm working on my feedback system using coils and am getting good results.

[EpicenterBryan]

Hello Bryan, I suspect the problem is in the cartridge. I'm working on my feedback system using coils and am getting good results.

Hey Ciuens,
share some details on your feedback coils. I would love to see how you are implementing them!

[Ciuens]

That was my first cut test using feedback. Good enough for first test. I cut Polycarbonate, mono without suction. I used a needle 320P MYSHANK

[EpicenterBryan]

That was my first cut test using feedback. Good enough for first test.

Good enough? That's impressive!
Details on your coils and how / where you mounted them?

[markrob]

Hello Bryan, I suspect the problem is in the cartridge. I'm working on my feedback system using coils and am getting good results.

Hi,

I don't think it is just the cartridge. I finally got around to setting up my Kaman eddy current measuring system and getting some plots of the raw driver frequency response:

Driver.jpg

The sensor measures displacement, not velocity. So if the driver had textbook response, you would expect it to be flat with a peak at system resonance and fall off at -12db/oct beyond. If you check out the plot, you can see the response is as expected up to about 4Khz. There you can see two unwanted resonances at 4Khz and 5Khz. Beyond that, it settles back to -12/oct. The system resonance for this driver is located at about 450 Hz. and is not well damped. I think the secondary resonances are coming from the suspension springs as I can kill them by placing a finger on them without any effect on the rest of the response. I'll try to see if some damping can be applied and re-test.

Some notes about the plots. The yellow plot is a full bandwith pink noise source with the RTA set for 1/24th octave. Due to the roll-off of the transducer, the response beyond 6-7 Khz is invalid because it falls into the system noise floor of my sensor electronics. To bring it out of the mud, I would have to heavily overdrive the transducer. So, I created a pink noise noise source high pass filtered starting at 2Khz. With the energy missing below 2Khz, I was able to bring the high frequency info out of the noise floor (see the green plot). I shifted the second plot to align with the wide band plot. Use the yellow up to 2Khz and the green beyond.

Some things become evident here. These drivers have a stiff suspension as compared with a typical tweeter. Probably stiff enough for use as a cutter without the need for much more. With this stiffness, the resonance is only at 450 Hz. The result is that you have about 12 db less output at 10 Khz as you would with the same stiffness and lower mass to move the resonance up to the 1 Khz range. Unless the moving mass could be reduced by a considerable amount, this would put some limits on the drive levels you could achieve. This situation will only get worse as you add in the additional mass of the stylus holder and torque tube. Increasing the stiffness to raise the resonance will not help as you will lower the overall sensitivity in proportion. I didn't measure the BL product to see how it compares wit the driver I used on my head, but that would be a good thing to do.

Mark

[Ciuens]

Mark, my measurements gave a similar result, with the resonance of 350Hz, but I did not know whether the results were real or not.
Bryan, I will post detailed pictures of everything. But, the coils are wound on the rod that print. There are 70 turns of wire, 8ohms impedance. I am dedicating myself much these days.

[EpicenterBryan]

To bring it out of the mud, I would have to heavily overdrive the transducer. So, I created a pink noise noise source high pass filtered starting at 2Khz. With the energy missing below 2Khz, I was able to bring the high frequency info out of the noise floor (see the green plot). I shifted the second plot to align with the wide band plot. Use the yellow up to 2Khz and the green beyond.

Brilliant trick there Mark!

And I'm so glad you have had time to start playing with these drivers. I'm sure I speak for everyone else when I thank you for sharing what you have found so far. It is interesting that the actual response using the green section lines up so well to your predicted -12db/oct line from 6K up...

Are you using the surface of the front facing spring as the target for the eddy current sensor, or do you have something else connecting to the face as a target? Just wondering if you had any issues with those fields near the driver. I always wondered if the carrier frequency of the signal conditioning / modulation / demodulation box was high enough that it really didn't matter.

Mark, you have seen some photos of the driver internals in this thread. Knowing that the spring is normally riveted to another disc that the coil is glued to... Do you think we save enough mass by removing the coil from the non-spring disc and gluing the coil to the spring piece and perhaps lightening it a bit by drilling some extra holes? The problem is it the spring disc has a 5mm hole in the center that still needs some kind of connection adapter for the links. If that darn hole was just smaller in the first place...

I guess another option might be to save what's needed from the spring so the coil can attach directly to it but drilling or punching out as much of the center as possible and gluing a disc / link attachment in the empty center area (perhaps a carbon fiber disc, or carbon fiber strip).

But then again a whole new spring / attachment surface may be the way to go. It kind of looks like Ciuens has gone this way...

Just for fun, how about a totally minimalist spring attachment plate.

The idea really comes from one of my favorite sources of inspiration. The Humble Pizza Box Center Support.

Years ago there was a wild flurry of development and improvement. Each time I ordered a pizza there seamed to be a new an improved version. Eventually, I had examples of a dozen of the most interesting changes on my desk - as a reminder that it's the simple stuff...

UNTITLED.jpg

Here are some highlights:
The first one (not shown) was a Barbie table with four legs... Then everything started to change.
1). Why 4 legs when 3 do the job and use less plastic?
2). The next guy made the disc smaller so they could stack for easier shipping.
3). The next guy did the same but eliminated some material.
4). On the bottom row, a rebel said "why does it need to be round"?
5). The next guy eliminated the center and changed to "T" beams.
6). The guy that won reconfigured with the correct strength for the application and least amount of ... wait for it ... mass too!

So I'm going to share one crazy photo that might get people thinking about a minimalist spring change. This goes back to the other head I was working on with the other drivers that were hollow in the center... I will be re-naming it BS-1 if this idea works. The big problem (beside my total lack of understanding about how the 45/45 internals needed to work) was that the drivers could not handle any kind of off axis load (like cutting a record) without having the voice coils go out of alignment and rubbing. In this photo you see some "Xs" as I started to play with the idea again just before going on vacation and I haven't touched it since. I'll post some more details about this on the other thread when I have some solid info...

IMG_3559[1].JPG

So pretend something like this "X" gizmo was the new spring for these Groove Scribe drivers, with "T" beams in the center to be very rigid at the link point, something rigid enough near the spring connections, and 4 points where the coil can attach (not a circular connection), I think you see we can eliminate a bunch of material (and mass)... The actual connection points to the driver body could be curved but the point is to remove as much material (and mass as possible).

So chime in people!

New springs with less mass? Do you think we should look into it?

Bryan

[markrob]

Hi,

Here is a photo of my test stand with the Kaman sensor:

Test Stand.JPG

I was using the provided threaded "pusher" as the target for my measurements. The part represents about 2.23g of mass. The plots include this.

All of those suggestions might improve the situation. I'm going to look at damping methods to try and kill the secondary resonances and tame the Q of the main peak.

The Kaman carrier is up at 1 Mhz, so its very far outside the audio band. I've not seen any interaction from the fields generated by the driver. Fairchild used a very similar approach to their feedback cutterhead system.

Mark

[EpicenterBryan]

Fairchild used a very similar approach to their feedback cutterhead system.

Nice.
Are you thinking about using sensors like this for feedback (perhaps bent pencil style)?

[emidisc]

Hi Bryan it's nice to see you revisit the old set of drivers I still think they could be useful ?
But I still can't get my head around plastic springs! Apart from the ease of making them,
Going back to the more recent transducers you used, could you not simply remove some material from the original spring in order to get the desired effect for example if the spring has four legs remove two of them or perhaps reduce the cross sectional area of specific areas of the spring ? I hope this makes sense ?

Enjoying your progress

Emidisc

[EpicenterBryan]

could you not simply remove some material from the original spring in order to get the desired effect

I'm going to give that a shot this weekend. I'll drill 4 new holes in the existing spring, as big as I can between the center and the edge. I might also be able to enlarge the 4 existing holes. I hope to trim 1/3 of the mass.

Now that I have coils detached, I'm beginning to see that the entire circle is useful in eyeballing the center for gluing the coil back on. I'm going to have to think about it... I think drilling 4 holes will be easier than trying to cut 4 sections of pie out. This is pretty tough stuff and am unsure how my Dremmel will do cutting pie.

I guess I could also make a jig to make gluing easier... Things to think about.

This is fun!

B

[EpicenterBryan]

I decided to try removing the coils tonight, and for anyone interested in trying it - it's not too bad.

The glue seams to be something like a rubber cement. When dry, it bonds like crazy and has a slight give if you try to scrape it off. The only practical way to remove it is to soften it.

The first thing I tried worked. I used Acetone. It did not dissolve it, but it did soften it. There may be a better solvent but this works so I am happy.

IMG_3771.JPG

IMG_3772.JPG

Acetone will eat through plastic, so you want to use an old plate or something.
Guess what I used? An aluminum Pizza pan! Can you tell I love pizza?

You want to tape off your wires so they are not near the bottom of the pan. With the spring facing down, put some Acetone inside the coil. I found using a syringe works great for directing the acetone. Yes, some will leak out, but you just want to make sure it is in contact with the black glue all the way around. Let it sit. If it starts evaporating put more acetone in. You want the glue moist the entire time.

Then use a straight dental pick and make the tip go around the edge right at the connection. Apply more acetone - it will get into the groove. After about 15 minutes you will notice the glue starts pealing off of the spring as you run the pick around the perimeter. Then you are ready to pull it off. This is also the time to clean the goop from inside the coil. You have to do this while it is still wet. Don't bother with a rag, just use your fingers to roll it off, and be sure to clean the edge of the coil where it will need to make contact when you glue it back.

One thing to note: Do not touch the coil wires if your fingers have acetone on them. It will eat the coating on the coil wires!


Now that everything is apart, I decided to weigh things. Below is a shot with weights for the spring, the round backing plate, and also the spring / plate and rivets.

You can see that if we ditch the inner disc and the rivets we go from 5.6g to 3g.
And if I can ditch another gram by drilling holes we may get down to 2g.

weights.jpg

Bryan

[EpicenterBryan]

Bryan, I will post detailed pictures of everything. But, the coils are wound on the rod that print. There are 70 turns of wire, 8ohms impedance. I am dedicating myself much these days.

I understand you are busy. We will look forward to photos when you have time. I'm really interested in what you are up to. That test cut was really impressive!

[markrob]

Hi Bryan,

Nice work getting the parts separated. You should be able to make some really good improvements. Looking forward to seeing what you come up with. Here are some more measurements I made on the driver along with those on the CUI driver I used on my DIY head:

Parameters for Parts Express Driver BCT-2

Re = 3.28 ohms
Le = 388 uH
Fs = 529 Hz
Mms = 8.31 g (8.31 e-03 Kg)
Cms = .0109 mm/N (10.9 e-06 m/N)
BL = 4.32 N/A (Tm)
Impedance at 1Khz = 4 ohms
Max power 10 W rms

CUI Driver CMS0401KL-3X

Re = 7.1 ohms
Le = 410 uH
Fs = 464 Hz
Mms = .78 g (780 e-06 Kg)
Cms = .15 mm/N (150 e-06 m/N)
BL = 2.51 N/A (Tm)
Impedance at 1Khz = 7.6 ohms
Max power 5 W rms

Some things jump out.

The mass of the BCT-2 vs. the CUI driver is over 10:1. They both have about the same resonance, but the CUI driver suspension is 10X more compliant (less stiff). If you decrease the compliance of the CUI driver to the same level as the BCT-2, your system resonance would rise to 1.7Khz. This makes it easy to get a decent trade off between system stiffness and resonance.

On the plus side the BCT-2 can handle 2X (not verified) the power of the CUI driver and is a stronger force generator with a BL product about 1.7X that of the CUI. The ease of interfacing to the BCT-2 is also a very big plus. I also love the hackablity of this driver as compared to the CUI.

If you are able to remove some mass and increase the compliance a bit (not too much) the net result may be a wash over the CUI. It will probably require more power for the same cutting level, but you have margin over the CUI and I can see some ways to improve the power handling (forced air cooling, heatsinking, ferrofluid). Some work will also have to be done to address the secondary resonances I'm seeing.

[Ciuens]

bryan, this is my coil feedback. yes, it's ugly, but that's what got the result posted


Enviado do meu iPhone usando Tapatalk

[EpicenterBryan]

If you are able to remove some mass and increase the compliance a bit (not too much)

4 holes, 8mm each got us down to this:

IMG_3776.JPG

So Mark, I'm thinking the easiest thing to do would be to trim the outer curve on all four spring arms with a fine grinding wheel. I'm also finding that the width of the springs are not identical. I've measured anywhere from 1.85mm to 2.17mm.

I'm up for giving a try. How much do you think I should try to trim? 20%?

[EpicenterBryan]

bryan, this is my coil feedback. yes, it's ugly, but that's what got the result posted

Nice job there Ciuens!
What gauge wire did you use?

It also looks like you have some kind of shield at the driver. Is that MuMetal?

[Ciuens]

I used the wire from an old speaker, then I do not know the gauge. I used Mu metal to shield, but need higher magnetic isolation.

[Bahndahn]

Great experiments!

One way you could reduce mass and rigidity/increase uniformity in thickness could be to mill the surface of the spring. Passing across the entire spring could yield the right results. An obvious challenge would be 'holding the workpiece'. The remedy I can think of is to hold the spring to a machined surface that is electromagnetically charged – make a high power electromagnet and use it as a special-purpose vice!

[Ciuens]

https://vimeo.com/131547513

new test. Using MyShank black blank. I am very impressed with the quality of the cut.