A simple and inexpensive stereo cutter YOU can build!

[grooveguy]

Hey, guys,

Over the past few months I’ve been crafting a simple, easily-built stereo cutterhead. The two heads that came from all this turned out well enough to share with the group.

So I wrote the project up as kind of a standalone DIY article, which is attached here as a PDF. There should be enough info for someone with reasonable mechanical skills and a basic understanding of stereo disc recording to duplicate one of these heads. Web links to MP3 samples are included so you can hear how they sound.

The first part of this write-up appeared in print in one of the audio magazines, but this updated PDF has been enlarged-upon. Please post any questions or comments, and don’t be scared-away by the reference to DSP-based signal processing. That’s actually the easy part. Honest! Bottom line: a head and its driver circuitry for about $200… and many, many hours, of course. And YOU can do it!

Stereo Cutterhead EngrBrief.pdf

[markrob]

Hi,

This is great work! Really well documented and explained.

I've been working on something very similar over the past few months and may be able to add some additional info on the DSP part. Rather than the brute force EQ, I created two DSP blocks that compensate for known characteristics of the head directly. The first is the IRIAA. Below the main resonance, the head is stiffness controlled and exhibits a +6db/oct rising velocity response. This is already defined in the IRIAA curve, so there is no need to compensate the head. You only need to create the boost below 50hz and 2122Hz as defined in the RIAA spec. So I created an IRIAA block with only the 50 hz and 2122 Hz turnover points. It turns out to be a simple second order filter. I calculated the filter coefficients and plugged them into the general 2nd order filter block them that Sigma provides. The second block compensates for the the falling -6db/oct response of the head on the other side of resonance (mass control region). This is just a simple single pole filter set to start just beyond the resonance point. This also uses the same general filter in Sigma. See the attached jpg to get an idea of the frequency response. That leaves you with the need for a notch filters to kill the main resonance and and and secondary resonances present along with some high frequency correction to flatten as needed. I also developed an first order elliptical filter with cutoffs at to 75, 100, and 150hz. This uses standard Sigma building blocks. I developed simple input and output peak detector block that drive LED's using GPIO pins on the Sure board. Finally, I wanted to have some means of monitoring the long term average power that was being fed to the head. This is due to the fact that the head needs very high peak power to accelerate the mass, but must be limited to very low average power to avoid frying the voice coil. To do this, I created a DSP block that calculates the instantaneous power of one channel and averages this via a .1hz low pass filter. The power is calculate by squaring the audio signal of one channel (P = V^2/R). On the Sure board, you can't pass a DC signal so I use this low pass filtered signal to modulate a 1Khz sine wave tone generator. This signal is output via the 3rd DAC on the Sure DSP board. If you connect this to an AC voltmeter (digital or analog), the value represents the long term (~10 second) power to the head. I calibrated this based on the sensitivity of my amplifier such that 1 Vac = 10 Watts. I can post the Sigma project or more details if there is any interest.

Sigma DSP Rersponse.jpg

Head EQ.jpg

[grooveguy]

VERY well done, MarkRob! I'm so glad to see that others have discovered SigmaStudio and are making very good use of it. I can't stress enough how simple the Analog Devices people have made this math-intensive DSP business, and encourage others to use it as no code-writing is required. Obviously you have a good understanding of the elements involved in this and are in a good position to supply much additional tech help to those whom I hope we can inspire to explore this head-making business.

My own background was in magnetic recording, hence my total 'blowing-off' of the IRIAA characteristic. In tape recording the complementary pre-/de-emphasis characteristic is defined as a reproduce curve, making the record part more of a 'whatever-it-takes' to make playback flat. Because of the resonance hills and valleys in the cutterheads, I segmented the equalizer function probably more than was necessary. Doing the IRIAA first and working from there is a very sanitary idea. I also wasn't sure of the cutterheads' turnover frequencies, which I recall from books on the subject had to be factored-in as well.

Please continue to make suggestions and recommendations in this arena. The more of us who undertake these seemingly-difficult chores in our hobby the better for all.

[2bitcomputer]

Oh, very interesting and intriguing!
How would you feel about posting some dimensions of the various parts - just as a kind of "start here" thing?

[grooveguy]

Hi, 2bit,

I'm embarrassed by your suggestion; here's why.

In designing the head, I started with the audio 'exciters,' measuring them carefully with a caliper, etc. and creating their image in... are you ready for this?... Microsoft Publisher! I know, I know, it's not a CAD or even a drafting program, but it generates shapes and angles, allows 'layers,' sort-of, and rotation, and I'm familiar with it through other, more mundane uses for the program.

So I blocked-out the head in a series of 2-dimension drawings until everything fit, and then I tried another terrible-idea. I printed the parts to be made on my el-cheapo HP inkjet printer. I was astounded that the dimensions were dead-nuts-on, as near as I could measure anyway. So, what the heck, I thought, and used 3M "77" spray adhesive to stick the paper templates down the the phenolic or aluminum stock or whatever, and drilled, cut and sanded to the lines. Wonder of wonders, everything lined up and fit and worked. So, you see, there are no dimensions. They could be derived from the Publisher drawings, but with funny angles and all, that's more math than I'm capable of.

I'd be happy to share the Publisher files, if you can run them, or could make PDFs that anyone ought to be able to read. With the PDFs I would think they'd be pretty close, but it would be another generation of digital conversion, so the 'horse's mouth' might be the best. The drawings are not annotated, so you'd be pretty much on your own using them. Also, Publisher does allow objects to be grouped (which many are) and ungrouped, and that doesn't come through on a PDF. If you decide to use the Publisher files, work from a copy of the file then if anything drifts away from what it was attached to you can go back and try again.

[2bitcomputer]

Well, you know, we work with what we know and what we got, and you seem to be getting it done.
So there!
How about using a ruler and just getting 'close enough' measurements of the main parts - the back plate, the blocks, the exciters?
That would be pretty helpful...

[sameal]

This is super awesome! Thank you.

Gives me a lot of ideas for my stereo build when i can get back on it.

Now i just need to figurd out arduino and build this damn pitch computer!

[grooveguy]

Thanks for the kudos, sameal. I will be adding some mechanical drawings, as coerced by 2bit; may take a few days.

Regarding the pitch controller, I am assuming that you are referring to a device that will give you variable pitch on a dynamic basis, as the music dynamics require, not just a stepper controller for preset fixed pitches. Is this right?

The whole concept of dynamic (automatic) variable pitch (historically with an upstream preview head on the master tape playback machine) is great for classical music where you need to fit a couple of movements (or maybe a whole opus) on one side of the disc, and you are dealing with a wide dynamic range. But in reading through the different postings on the group here, it seems to me that far too much emphasis is placed on this function, which is largely unnecessary for the pop/trance/'tronic music in vogue these days.

I went through my own record collection and found very few albums mastered with auto-variable pitch. I have never used it myself, although with my old RCA lathe it was handy to have a selection of pitches. Now my Rek-O-Kut gives me the choice of 120 lpi, still a bit coarse for 45s, a 210 lpi screw, my default, and 240 lpi that I have never had occasion to use. When my forthcoming from-scratch lathe project gets to the details-planning stage, I will certainly provide an option for selecting any pitch that seems right for what's going to be recorded, but I doubt that I'll make any provision to vary it during the recording process... except for spiraling, of course.

I don't know how the rest of you guys feel about this, but I tend to keep things really simple. Since today's music all seems to be highly-compressed, even on CDs where it doesn't need to be, and with artists, producers and engineers all clamoring for loud, 'wall of sound' releases, dynamic variable pitch seems to offer little return on investment. Comments?

[sameal]

Yes sir.

I need some kind of pitch control, as mine was absent from the lathe when i bought it.

Now lately i have been thinking of just using a motor on a pwm controller with a couple preset l.p.i. buttons to start with. Sounds like that might be fine for what i plan to cut (rock/electronic). I might shelve the arduino half of it for a bit just to get cutting records.

I stole soulbears setup he was using, and on the little webster head i had been playing with it seemed ok. But the westrex head i got now is much heavier and both the lpi motor and lift solenoid i developed seem insufficient for that weight of head.

So changes must be made.

[grooveguy]

In my experiments with stepper motors, I think there's enough range with most of them to accommodate a fine microgroove pitch and still do the end spiral-out without shifting any sort of mechanical gears. I'm happy with 1/4" or even 3/16" pitch in the spiral-out and a somewhat finer one between bands. I haven't tried it yet, but by my third-grade-arithmetic-class calculations, the motor is still moving fast enough when cutting music that it won't cog; that is, using a really nice motor controller with micro-step capability, and maybe a rubber belt to the leadscrew to help filter out high frequency noise. Even without an Arduino or even a PIC (which is still beyond my programming capability!), a simple series of pushbutton presets controlling a VCO (with some first-order filtering between to ramp-in and -out of the pitch change) ought to do it. The between-band spiral could do the initial lead-in, and a separate button to spiral out, at the faster speed, would go until a microswitch initiated the lock groove. Auto-lift afterward would be nice, and in my case (a captive lead screw nut) a button to back-up and 'park' the cutterhead in the starting position. All in good time, but "simplicity rules"!

[grooveguy]

Fellow Trolls,

A few days ago, 2bitcomputer suggested that some drawings of the head would be helpful. I didn't have any dimensioned drawings, but pulled-up the MS Publisher (shudder) sheet I'd been using as a free CAD program and added some dimensions. When I made the parts, I simply printed-out the drawing and used it as a template. As long as you print the original on a decent printer, things are pretty darned accurate.

Unfortunately this newsgroup setup doesn't allow a .pub document to be attached. So I've made a .pdf of it and have attached that instead. It still may serve as a template, even with another layer of digital conversion, but I'd be glad to email the original .pub document to anyone who wants it. Just drop a PM.

Cutterhead 2.pdf

[2bitcomputer]

This is great grooveguy - thanks!

[markrob]

Hi,

Thanks for posting that info. That should be more than enough info to duplicate your work!

One question I have is the basis for your stylus and head geometry. In your design, you mount the stylus at and angle and tilt the head to bring it back to 90 degrees with respect to the disc surface. In a Westrex torque tube type of design, I was under the impression that the VTA is present due the fact that the motion of the stylus mounted in the torque tube is an arc rather than truly vertical. That motion, coupled with the length that the stylus sticks out from the torque tube, results in the tip of the stylus moving at an angle in reference to the surface of the disc. The value of the angle depends on the distance from the stylus mounting hole to the torque tube pivot point and the length that the stylus sticks out from the holder. I may be wrong, but I think that the original Westrex head produced the 23 degree VTA due to its dimensions, not by design. I don't think any head tilt or stylus tilt was employed. Am I correct about this?Below is a simple diagram I drew to illustrate what I am talking about.

VTA.jpg

In the case of the Ortofon rocker style head, the is no vertical arc in the motion of the stylus and thus no VTA. The stylus motion is truly vertical in this case. Since it came after the fact, they had to tilt the head and stylus holder to match the existing standard. In other heads (e.g. Neumann), I'm not sure if some tilt is employed to match the standard even though it is similar to the Westrex design (e.g. different dimensions). Can anybody add to this? Am I off base here?

In addition to the VTA caused by the head geometry, there are additional factors such as stylus shank deflection and lacquer spring back, but the bulk should be due to the dimensions.

Mark

[grooveguy]

Hi, Mark, and good food for thought in your comments; thanks for bringing this up. My own understanding of the VTA is quite sketchy and clouded, and based for the most part on the old Noel Keywood article, which I have attached here. I am certainly open to any thoughts on this subject and would really like to understand it better.

My understanding was that pickup cartridges necessarily had to have an angled cantilever because you couldn't put coils or magnets at disc surface level. That meant that vertical groove modulation did not displace the cartridge EMF-generator on a 1:1 basis with vertical groove displacement, thus inferring an automatic limitation in separation (and maybe even in vertical even-order distortion). This could be compensated with a complementary transfer function in the L-R signal domain, but simply tilting the cutterhead by the same amount would suggest either: 1) perfect compensation for the effect, or 2) making the effect twice as bad! Having no foundation in the physics and higher mathematics of this issue puts me at an embarrassing disadvantage.

If we temporarily assume my premise here, I don't see the difference between the traditional 45/45 cutterhead and the Ortofon "rocker" in this regard. It would seem to me that, regardless of how the vertical action is generated, the distortion that Keywood speaks of in the (pirated!) article would be the same.

One advantage to tipping the head, as I did in my design, is in gaining some working room for the mechanics. Just like the playback cartridge, it's difficult to get the stylus 'anchor point' down at the disc surface plane.

One bit of design that snows me is Jerry Minter's "Hydrofeed" lathe write-up, where he puts the vertical pivot for the cutterhead back behind the turntable so it can be right at the disc surface. I have long held that there was sort of a magic relationship between the vertical pivot for the head and the distance between that pivot point and a line through the axis of the stylus (how far back the pivot was). Oliver Read in his book, The Recording and Reproduction of Sound, makes note of this with respect to vertical cutterhead oscillation. That's what prompted my request to Trolls members on where the cutter was pivoted on their lathes. My Rek-O-Kut is about 1-1/2" off the disc surface, and I think the Presto 6N is about the same. Anyway, I've never had tendencies toward vertical chatter and thus never used a dashpot. Of course all this goes back to mono recording, but the vertical 'hill-and-dale' cutterheads used by World Broadcast Services and some others were mounted on these same lathes and made some really nice-sounding recordings. I dunno!

N. Keywood VTA Article.pdf

[markrob]

Hi,

Great article. Thanks for posting it.

Figure two is basically the same as my diagram. I think there may be some confusion on the location of the pivot points. If I understand you correctly, you are talking about the suspension pivot point for the head. In my case and in the article, the pivot is the location referenced to the cutter head torque tube and pickup stylus suspension. If the vertical arc motion in figure 2 is traced by the cutting stylus, you would like the pickup to reproduce the same motion on playback. If the angles match, that will be the case. If, for example, the cutter stylus moved perpendicular to the record surface for vertical modulation and the playback stylus moved in an arc as in figure 2, there would be a tracing error. That would be the case if the cutterhead was a rocker style with no tilt.

Just guessing here, but I suspect that if the rocker head design came first, stereo pickups might have been developed without an angled stylus to work with a VTA of 0. As you point out, that may not have been feasible and perhaps that would have forced the head designers to tilt the head anyway.

I see from your article that there were some differences in US and European standards that added confusion to this. I wonder if that is due to the differences in the Westex vs. Neumann torque tube and stylus dimensions? Its also interesting how far off some of the pickups measured. As was stated in the paper, the real world effects may not be as bad as the numbers would suggest due to typical program material. Maybe not a big deal in a DIY cutterhead design other than that gross errors may reduce the separation.

I thought that it was a cleaver idea in your design to boost the L-R signal path to improve the cutter separation. My only concern is if this increased the peak depth of the cut to the point of causing tracking problems on playback. I didn't notice if you implemented some sort of elliptical filter to move low frequencies to the center or lateral channel. Going this route might make my concerns moot as most of the vertical excursion would be due to out of phase or hard panned low frequencies.

Mark

[grooveguy]

Good points, Mark. Yes, the US 'leans' toward 15° (if you'll pardon the pun or allusion, or whatever), and the Europeans like 20°. Not sure why, but I figured 18° would be a decent compromise.

Here's another article on the subject by Geoff Husband: http://www.tnt-audio.com/sorgenti/vta_e.html, which may contradict the first, or may shed additional light if you are able to read it all the way through (Zzzzzz... on a rainy Sunday afternoon).

Diddling the L-R gain is meant to be just a trim adjustment, and I found that in practice with the head I built, 6dB gain there is all you need. More than that doesn't help separation, and separation is frequency-dependent to a large degree. What I did was to tweak that adjustment while cutting monaural material and then listen to the L-R (vertical) component on playback, the idea being to null the 'vertical residual' as much as possible. The extra L-R gain only compensates for mechanical (stiffness) losses, so you may push a few more milliwatts into the cutterhead but the actual vertical displacement doesn't get bigger than what's optimum for best separation, if you see what I mean.

So take a look at that Geoff Husband article and see if it helps us get to the bottom of this.

Jim

[markrob]

Hi,

I did some more reading of my archives over the weekend and I think your approach may be the right one. I found a paper and associated patent by Bauer at CBS that talks about the issues we are dealing with here. In particular, he indicates that although the Westrex cutter has a built in 23 degree VTA, in reality due to spring back effects in the lacquer, the real cut VTA is near 0 degrees! One of his proposed solutions includes almost exactly your method of tilting the head and drilling the stylus mounting hole at an angle. See the attached PDF's.

Was this change in mounting ever adopted by Westrex head users? Can any pro cutters out there comment on this? We know the Ortofon was designed to be tilted. Are the Neumann heads mounted in this way as well?

This is also dependent on the properties of the medium that is being cut. Given many DIY cutters are not cutting lacquers, maybe all this needs to be re-examined.

Mark

[grooveguy]

Hey, Mark,

Thanks for those references, which I have added to my own stash! Ol' Ben Bauer was a very prolific inventor and theorist, best known for his development of Shure's original Unidyne microphone, the "Fat-Boy Elvis Mic" as it's commonly known.

The concept of 'springback' is a mysterious one, and deserves further study to determine that same quality in the materials we are being forced to cut nowadays, given the limited availability and cost of lacquer blanks. As an aside, I think that someone ought to look into getting into production on an experimenters' grade lacquer-coated disc. From what I'm told, Apollo/Transco and a very small Japanese firm are the only game in town now.

Not long ago I read the biography of Cecil Watts, the Brit who was deep into developing recording lathes and lacquer blanks between the wars. An entertaining read that, to me, somewhat demystified the making of a lacquer blank, although wide-groove 78s were more forgiving of imperfections than microgroove recordings. Cecil used a commonly-available lacquer (paint, I guess) and added some ingredients of his own. Every now and then, in the course of doing chores around the house, I come across paints and varnishes that suggest themselves to our line of work. My only warning to someone experimenting in this area is to remember how volatile most of these things are, and from vapor-high/health and explosion/fire standpoints, one would want to be very careful.

[tragwag]

ok this is amazing THANK YOU! a local friend with manufacturing experience has been soaking up the cutting based literature I've sent him, can't wait to drop this bomb on him and get going with a head!

[xave2000]

Hello
have you a link in ebay for buy some Aluminum “perfume funnels”
i bought the same excisters but i can't find the good funnels.
thanks