My home made lathe

[sifis1983]

Hi there trolls,

Well, as a lot of people in here, I have a dream... to build my own record lathe.

So, after long reading in this forum and watch many many videos, I have my first try.

You can see the whole machine in the video below.
But here a simple description of what I made:

- A Nema17 motor drives a feedscrew with the help of arduino, with adjustable speed.
- The tonearm base is from a technics 1200 mk2.
- The tonearm is from a paintbrush (!).
- The cutterhead is mounted in this "paintbrush" tip with screws in a piece of plexiglass.
- The cutterhead is very simple. One of these needles (http://www.ebay.co.uk/itm/281255271864) is mounted in a visaton exciter 45s with silicone, in a 22-25 degrees approximately from the vertical.
- I emboss in a blank CD. I polish the plastic with turtle wax original, before heat the plastic with a heating lamp (40W).
- I start to emboss when the temperature of the plastic is about 38-40 celcius. (I have to mention that I wait 5-6 minutes to reach this temperature).

Here is the video with audio results and pics:
https://www.youtube.com/watch?v=y-bvsyihUO8&feature=youtu.be

As you can hear, there is a lot of hum from the motor in the start for the lead in groove, but then is more quiet when the speed of the motor is slower.
Also, you can hear that there are no high frequencies and signal level is not too loud.


Next goal is to achieve better high frequencies.

Now, the first thing that I 'm thinking is to make a better cutterhead.

I will for sure buy a bigger heating lamp to try emboss in higher temperatures.
Also, I want to try liquid lighter fluid for polishing the plastic.
I don't know about the weight of the cutterhead right now, but I'm thinking to make a heavier and more rigid cutterhead.

Do you think the above thoughts, would bring me better frequency response?

So, any suggestions, comments or advices are welcome.

Thank you!

[markrob]

Hi,

Nice job! Clean, simple and elegant design. What are you doing as far as EQ? Can you post some closeups of your head design? Have you made any frequency response measurements? Are you cutting from analog or digital sources?

Mark

[sifis1983]

Thank you, Mark!

Well, the source is digital from my laptop. Just an mp3 using a plugin for inverse RIAA curve (I don't remember where I find this plugin, but is says "Olaf Matthes vacuum sound.de). That's for the EQ.

No measurements for the frequency response. In pics there is a snapshot from a frequency analyser and closeups for the cutterhead.

cutterhead closeup.jpg

tonearm.jpg

ov wright snapshot.png

Can you suggest me a way to have a more precise measurement of my frequency response?

Thank you

[markrob]

Hi,

Is that a piezo driver? I can't see how you coupled to the stylus.

Generate a minute or so a noise signal and cut it using the RIAA plugin. Use white noise if your spectrum analyzer is FFT based and pink if its an RTA (1/3 octave for example). Use WAV or other lossless format, not MP3. Playback and capture the cut on your computer and run the file thru the analyzer. If its able to display in real-time, use a EQ to return the playback spectrum to flat. You may need quite a bit of EQ to pull this off (especially at the high frequencies). Don't go too crazy at the high end (beyond 10 Khz ) to start until you have more experience with the head response. The EQ curve you use to flatten the playback should be saved and used to EQ your source material. You may need to use a much more powerful amp to allow you to cut the highs without clipping your signal. Otherwise, back off on the level until you get a handle on the power required.

One note about the RIAA. If your head is piezo based, the standard RIAA curve is not correct. That's because most cutter heads are velocity based devices and the RIAA curve is specified with this in mind. A piezo head is an amplitude responding device. You still need RIAA, but the curve is different to deal with the response of the piezo based head.

Mark

[sifis1983]

Hi,

Yes, it's a piezo driver. (http://www.visaton.com/en/industrie/koerperschall/ex45s_8.html)
Sorry about the pics. It's the best I can do. It's only this piezo driver with the needle, glue on it with silicone, in angle of 22-25 degrees approximately.
I told you that is extremely simple.

I really really appreciate your advice!

I will follow your procedure, step by step to see what happens.
But how about RIAA for a piezo driver like this?

Thank you again

[grooveguy]

Hey, guys,

Gotta take exception to one statement here, that Visaton transducer is magnetic, not piezo. It's essentially the same as the Dayton Audio transducers others are using.

Having said that, I am interested in any experiments anyone here has done with piezo transducers. The Astatic cutterheads used in most of the 'suitcase' recorders were bimorph crystal heads, and they performed rather well, actually. Seems to me that with the intervening years, improvements should have given us promise for even better results than were had in the '40s. Most of the readily-available piezo transducers are disc-shaped 'buzzers' or used as drivers for tweeters. Those don't have enough displacement to modulate a record groove well, but other types are available, and some respond with enough motion to be used as motors in camera focusing mechanisms, etc. Open to ideas from anyone.

[sifis1983]

Hi,

Well, I'm terribly sorry for my mistake. I had assume that visaton exciter is piezo driver.
But that's better news for me, because as Mark said, now I can use the classic inverse RIAA EQ.

So, I make a cut with the procedure of Mark and I believe that I have better results concerning high frequencies.

Here is what I have done:
I emboss in a CD a white noise for a minute. Then, I captured the result in my laptop.
I try to have the best possible flat response until 10khz, with the help of another EQ.
Then, I load the setting of the above EQ and I put it on a track, with the inverse RIAA, together.

The result was an awful clipped signal with extreme high frequencies. (I almost thought that I could destroy my tweeters in my speakers).
So, I lower the level of my source, (for about -20dB) and then I gave more gain from the export of my mixer (yamaha mg) that goes straight to the cutterhead.

I hope you can listen the result here:

ov wright i'm going home embossed 2nd try (EQ+).mp3

Thanks a lot mark!

Next goal is to have a louder cut with deeper grooves.
Maybe I will try some more weight in my cutterhead to see what happens.

Sifis

[grooveguy]

Hey, Sifis,

That's not a bad result at all, especially for this early in the project. Very commendable. You're on the right track with your EQ, but there are a couple of things to remember about complementary curves as used in disc recording as well as in tape recording.

The curves specified in both tape and disc recording are for playabck only. The inverse of these curves is a logical assumption for a recording characteristic, and not a bad place to start, but you have to account for losses in the process, whether it's magnetic (tape) or mechanical (disc). Thus the recording curve will of course be the inverse of the playback curve, but losses because of gap length or 'fringing' (tape) or stylus geometry and head resonance (disc) mean that you have to put in extra dips and peaks to make up for those and other losses as well.

The best place to start is with your player. Calibration discs, like 'standard' tapes, are no longer readily available, but that's the benchmark. In other words, you need to know what your player is doing before you can EQ the recorder. Here's a response run I made on my old turntable using both CBS Labs and an NAB test record back in the 1970s. The top plot was with a Stanton 881-S cartridge, which I thought I liked the sound of, but which proved really 'peaky' at the top end. I graduated to a Shure Type IV, the middle plot, and use that cartridge to this day. Note that all three test records were pretty close until you get to the top, where the number of playings was probably a factor. Note also that the proper capacitive loading of the Shure cartridge flattened it right out.

Disc playback plots.jpg

Next I made a number of frequency runs on the disc recorder. I was using a Presto 1D head back in those days. With various breadboarded EQ circuits I was able to get pretty flat response. But the equalizer that did the trick was the one shown below. It's similar to the simple bass-and-treble tone controls, but the turnover point has been placed at 500Hz rather than the traditional 1kHz. Note the red diagonal line, which has a 3dB/octave slope. This turned out to be almost ideal for making that Presto 1D head play back flat with an RIAA playback curve.

Disc recording EQ cct.jpg

Here is the final results. The top plot is the playback system once again, the center plot is "almost there," and then the bottom plot shows the overall record/play response with the playback errors (top plot) subtracted from the results.

Disc recorder response.jpg

The other thing you have to keep in mind is that frequency response and power bandwidth are not the same thing, and I think you already found this out. In other words, with systems that employ complementary pre- and de-emphasis, you generally cannot make response tests at full volume. This is true of disc records, tape recordings and FM broadcasts. So you make your response runs 10dB or 20dB below full level. The spectral distribution of speech and music was taken into account when these curves were standardized, and this all works because speech and music do not generally have a lot of power at the top-end, where pre- and de-emphasis are used to reduce noise. So a full-level recording has lots of silky highs, but those overtones are not at high levels.

Hope this all makes sense; keep at it and report your progress!

[markrob]

Hi,

Looks like you are making good progress. A couple of more points. When you cut the white noise, also cut some silence. Make a note of the spectrum of the silent groove noise floor. If your playback of the white noise does not show level above the system noise floor, do not try to make this part of the spectrum flat via EQ. You will just be wasting power to your cutter head for no gain in performance. Also, using the standard RIAA recording curve is probably making you over compensate using your corrective EQ. The standard curve rises from 50 Hz to 500 hz at +6db/oct. The typical open loop response of a moving coil head also has a +6db/oct rising characteristic in this region up to the head's system resonance (you should try to determine where your main resonance it located). It would make more sense to only apply the boost specified at 2122 Hz and forget the low frequency part of the curve. For that matter, it might make sense to forget the RIAA and just do all of your EQ using the procedure I outlined. Grooveguy's advice is spot on, but at this stage, the limiting factor on performance is your head, not the playback system. I think you can assume the playback response it pretty good for now. If you get to the point that you are achieving great results, then it would make sense to check the playback response or look at other ways to characterize the cutting system response (e.g. light band method).

Mark

[inspector77]

Great progress and great info , Thanks

[sifis1983]

Hi there again,

Well, I reconstruct my tonearm and cutterhead, but I'm deeply disapointed from my new version.

My purpose was to make a more rigid tonearm and a heavier cutterhead, to achieve a louder signal and a better frequency response (especially in high frequencies).

With the help of a 3dprinter, I print lots of plastic parts to make a more rigid, straight tonearm and a heavier cutterhead with the ease of change the stylus. (in the previous cutterhead I just coupled my stylus with silicon glue at the visaton exciter 45s).

The tonearm is good enough. I'm ok with that.

The problem is that the cutterhead doesn't responds as good as the previous version. It seems that the stylus doesn't accelerate as good as I thought.

The new cutterhead use a Dayton audio DAEX58FP, as driver. (http://www.daytonaudio.com/index.php/loudspeaker-components/loudspeaker-drivers-by-series/exciters/dayton-audio-daex58fp-flat-pack-58mm-exciter-25w-8-ohm.html
This is an 25W driver / 8ohm (I assume that it is too much, because I read that even the neumann cutterheads is 10W max).
You can see all the specs in the link above.

Also, I had read that one important thing that I did not pay attention, is the Diaphragm Mass movement.
As Markrob says in a previous post, the CUI driver P/N CMS0401KL-3X (http://www.digikey.com/product-search/en?lang=en&site=us&KeyWords=CMS0401KL-3X) gives good results, because it has a moving mass of about 1 gram.
The Dayton driver indicates : Diaphragm Mass Inc. Airload (Mms) = 9.95g.
The CUI Driver at specs indicates: Total Mass - Mms = 0.5g.

Is this so important? Or I'm searching in a wrong way?

Here is the signal chain:
- The source is a wav file from my laptop with an inversed RIAA plugin, plus an extra EQ with extreme boost at high frequencies till 10khz.
- Then I have a focusrite saffire pro14 for audio interface.
- Then, there is a yamaha mg166 mixer that gives more power through its gain.
- At the end there is a mono Amplifier 18W (https://www.kemo-electronic.de/en/Light-Sound/Amplifier-Splitter/Modules/M033N-Amplifier-18-W-universal.php).

The signal from laptop and my mixer is not clipping. (No red lights)

The Dayton audio driver is coupled with a 3dprinted plastic piece, that holds the stylus.

The audio result is a too low level signal with no bass and highs and sometimes too distorted. (if there is no extra EQ high boost at the signal and have only the reversed RIAA in my source, the result is not distorted and it is at normal level but with no mids and highs).

I will post some audio results in another post.

Any comments or advices are welcome.

IMG_0271.jpg

IMG_0272.jpg

IMG_0265.jpg

IMG_0264.jpg

[markrob]

Hi,

The moving mass is very important. Force =mass x acceleration. If you double the mass, you need double the force to get the same acceleration. That requires twice the current through the head since force is directly proportional to current in a moving coil system. If you consider that power = I^2 x R, then you need 4X as much power to produce 2X force. Even if your driver can handle more power, if it comes at the expense of increased moving mass, you could be in worse shape. 18 watts is way too small for this application. You want to use an amp capable of 100-400 watts to drive your cutter head. That sounds excessive, given its power rating, but you will need this reserve to produce the head current to get the mass moving at high frequencies. You only need this for very short transients or you will fry the head quickly (the average power must still be within the head's power rating). Placing a fast acting fuse in the head circuit is a good idea (but not foolproof). You will find that you need very extreme EQ to make a head like this cut at higher frequencies. First, there is the required RIAA curve that has a boost of almost 14db at 10 Khz. In addition, you will need as much as 30db additional boost at 10 Khz to overcome the natural fall off of the head response due to its mass. That's an overall EQ high frequency boost of over 40db! This assumes that you have no other losses due to secondary resonances poor coupling to the stylus.

Hope that helps.

Mark

[grooveguy]

Good observations, Mark, and a lesson I had to learn the hard way many years ago. In that case I took Presto's figure for cutting an average level, which was only a watt or two, but they were referring to sinewave tones around 500Hz. Funny, but I didn't even wonder at the time why their own amp was rated at more than 50 watts. So, yes, my 10W amp cut a nice loud groove, and at the normal low levels that highs have in typical music it sounded okay. But hard rock and spitty vocals sounded terrible. Using the same EQ but bumping the power up to 50W made all the difference in the world. And with a more massive armature 'structure' in Sifis' head, you are talking about the need for quite a lot of power. Once you pass the resonance of the moving mass, efficiency plummets.

One question for Sifis from his photo: is that leadscrew mechanism a ready-made buyout item, and it it pretty solid? It looks like a good choice for the job, can you give us some details?

[sifis1983]

Thank you markrob and grooveguy!

Yes, your reply really helps me. But as I understand, I NEED to built another head with different driver and amplifier...
Maybe the dayton audio driver is a bad choice for a driver...
Or I need to have a bigger amplifier (100-400W) for my head and get a fuse for the head circuit. (I don't know how to do that... I'm not an expert at electronics... anyway, I have to learn more about this).

As I understand, for example, that's why souri suggests 250W till 400W amplifier for his head... I don't know what kind of drivers he uses though.

But let me ask a silly question. As far as I can gain my source through my audio interface and my yamaha mixer... Can I reach the power I want for my driver?

One more thing in general. Can you suggest a combination of an amp and driver that really works?

Anyway, I just emboss a logarithmic sweep from 220hz to 10khz (40 seconds duration) with only inversed RIAA.
Next is a white noise with only inverse RIAA again and much more gain from my mixer compared with the log sweep.
At the end there is a music track that I have inversed RIAA plus +45db from 1.6khz to 8kHz and the master level is at -41.9db for avoid clipping in the digital signal (see image for the exact setting).

My led meter of my mixer shows a green signal that reaches -6db.

I have to mention that, when I run the logarithmic sweep, I can listen my cutterhead resonances in specific frequencies! (that's scary enough)!

Also I have to mention that the recording/embossing has a wow/flutter thing. I assume that's because of my cutterhead is too heavy? The downforce weight is about 180grams, so the plastic doesn't have a smooth movement while embossing. Is the 180grams too much?

(You can listen the mp3 if you want :
https://www.sendspace.com/file/0pzff6)

Grooveguy the leadscrew mechanism is in the link above.
And yes, it's really solid and it works very nice!
http://www.ebay.co.uk/itm/195-mm-Q-module-Linear-Guide-Lead-screw-system-1-set-/161819402614?hash=item25ad302176:g:2rkAAOSwPhdVTceu

sweep grooves.jpg

Screen Shot 2016-03-22 at 11.54.48 PM.png

[grooveguy]

Thanks for sharing that, Sifis; very interesting. My initial observation is that you emboss a nice quiet groove. Having said that, the level of the signal seems very low, and when you get into the music portion of your recording, something sounds very nonlinear there. I hear much louder modulation coming in on the signal peaks, but near the 'baseline' of the waveform, the level remains very low. It's as though something is holding your armature tightly and letting-go when the peaks come along. The opposite of clipping, as if low level signals are not being recorded properly.

The 'wow' you talk about is very apparent. It's too fast to be a 'once-around' of the turntable, assuming that you are recording at 33-1/3 r.p.m.; it even seems too rapid for 78 r.p.m. Look for something that links to a mechanical rotation and figure why it 'hangs up' just at one point in the revolution. Does this correspond, for example, to the increment that you are turning your leadscrew? If that's it, maybe you need either to 'microstep' the leadscrew or use a larger pulley ratio between the motor and the leadscrew. Try rotating the leadscrew slowly by hand... is the wow the same?

I'm afraid that I can't advise you on amplifiers and transducers. I, myself, am experimenting with "moving iron" and "moving magnet" transducers to reduce the mass that has to vibrate. I did see one encouraging YouTube video with a guy using a small loudspeaker, who seemed to get really good results in a very simple manner: https://www.youtube.com/watch?v=pcCsAzHoSpQ

[kabuye]

Hey Sifis,
I love your project! And honestly I listened to your MP3 a couple of times and really like it. Thats the kind of sound I would like to cut myself.
I can't watch your video for some studid reason (I love german laws!).

Why did you change the driver? Was the first "Visaton EX 45s" so limited that you couldn't push it for better results?

I am very interested in that driver but I would like to know everything before I spend my very little student income. Any help is apreciated.

kabuye

[sifis1983]

Thanks for your kind words kabuye,

Well, I change the visaton exciter 45s because I was thinking that there in not enough power from my driver.
But I will set it up again, to push it for better results.

As I understand there is no reason for more power in watts to have better acceleration of the stylus.
So, too much mass to move with the Dayton driver. However, visaton exciter doesn't have any specific measurement for the diaphragm mass. But I suppose is less than the 9.95g of the dayton. (Maybe I will contact visaton for full specification of the product).

Grooveguy, thanks for your reply,

I don't think that there is problem with the leadscrew. I suppose that the downforce of the stylus is bigger than the rotation of my disc in 33 1/3. I must achieve better coupling of the plastic disc, to the platter. I just have a rubber mat, but maybe I should try a weight clamp to my turntable. That's why there is "wow".


Anyway, I will post some of my new results in a few days.

[sifis1983]

Hi,

So, I set up my previous driver, the visaton 45s - 8ohm. The stylus is glued to the driver with silicone in an angle of 25degrees approximately.

I use Ronsonol lighter fluid to soften the plastic and the temperature of cut is about 35degrees Celcius.

After a lot of tries, I can't make it work ... at least, as good as my first results in my first setup.
Now, I understand that the whole box that supports the driver, creates lots of resonances in my stylus.

Here is an mp3, that I emboss with only the RIAA EQ from a plug-in. A white noise first, a sine sweep from 50Hz to 5000Hz, and finally some test tones (A2, A3, A4, A5, A6, A7, A8).

emboss test noise + sweep + tones.mp3

The noise part, is good, as my first try, I think. In the sweep part, the reproducing stylus "jumps" in around 440hz. (Lot of resonance here I think).

(If you listen it, BE CAREFUL WITH YOUR SPEAKERS AND YOUR EARS IN THE SWEEP PART) !!!
Here is a jpg, of the grooves in the sweep part, when the stylus "jumps".

sweep grooves.jpg

It seems that the level is higher, than my first try. I don't know if downforce weight matters. It is about 126-128grams. (see pics).

IMG_0338.jpg

IMG_0340.jpg

I try to emboss the above, with RIAA plus a corrective eq, as markrob had told me.
The result is maybe better concerning the high frequencies, BUT the level of the signal is extremely low, because I lost gain, to avoid clipping the signal. Then, I try to push the gain hard from my preamp but the result is awful, with lot of clipping in my driver and also it was very annoying in my ears!!!

Well, three queries here.

First, is the downforce weight matters a lot? Is the 128 grams enough? Do I need more or less? (In the manual of embossing cones, it says : "Put as much weight as possible to the cutterhead". (In my first project, the downforce weight was about 90grams)

My second query is about suspension in vertical. In my stylus there is no suspension to the vertical. (it is only glued to the surface of the visaton exciter).I built it like this, because I was thinking that I need only horizontal movement of the stylus for a mono cut. So, is mono cutterheads has suspension to the vertical? and if yes, how much is this affects the embossing?

Third query: What should I do to avoid loss of gain when I try to emboss with the RIAA + plus corrective EQ? Is there a simple way to compensate the energy that goes to the driver?

Any comments or advices are welcome.

Regards,
Sifis

[grooveguy]

Hi, Sifis,

Looking at the way you have attached your embossing stylus to the speaker, I am wondering if you are actually inscribing lateral (monaural) modulation, or if your recordings are more vertical (hill-and-dale)?

Take a look at the way this one experimenter (from YouTube) attached his loudspeaker-based cutterhead to the stylus to ensure a lateral cut. I have seen others use this same technique. One advantage with the leverage obtained this way is that the speaker cone is allowed to move over a much greater distance than the modulation of the embossed groove. This 'mechanical advantage' might actually help the whole situation, as a loudspeaker was really never meant to move much more than very compliant air. This way the cone moves with less load since the force to the stylus is multiplied by the leverage.

But the main thing here is that the stylus movement is forced to be lateral.

Snap1.jpg

[sifis1983]

Hi,

Thank you for reply grooveguy.
The truth is that your reply, really confuse me. In my driver the movement is lateral. (or maybe i'm wrong?)
The visaton driver is moving back and forth and the stylus is glued to the mounting plate.
It's like this one:
https://www.youtube.com/watch?v=Hm7fD59zPa4

visaton.png

I had seen this video in youtube. I can see in your pic (from this experimenter) that maybe there is a verical suspension with a metal rod.
Is that thing critical ?

and what about the downforce weight?

Regards,
Sifis