The farmer and his plow - When I set out to build a record cutter

[farmersplow]

sounds awesome, congrats on the success!
curious what level you're cutting at?
I wonder if some of the frequency irregularities between each piece come from the overall volume level?
I can hear the drivers compressing and distorting a little differently on each sample

THX

Hello tragwag,

Thank you for your comment. I thought I wouldn't get any response at all. Your hearing is excellent! Unfortunately, it's true that not all frequency ranges are without problems and I can't check why this is (was) the case at the moment. Because the cutting head and the lathe no longer exist in their original form (08/2023).
What I do know for sure is that I cut the pieces in one cutting operation. Before that, I switched on the EQ on the PC so that I could adjust the characteristic curve when I was still working without a feedback coil. However, as the pieces of music are different, clipping was achieved with Billie Eilish, for example (when snipping). As a result, I cut the songs at different volumes.
Billie Eilish a little quieter and Pink Floid a little louder (because there are so many quiet parts).
In each case, the result is too loud an edit because I wanted to achieve a high signal-to-noise ratio (SNR) and I did.

Unfortunately, I can't calculate my PSNR (peak signal-to-noise ratio) and I don't have the necessary instruments to measure it accurately. But I can say that the signal strength (between noise volume and signal volume) is 30% better than all my previous values.

Example:
The values for Silent Cut for this recording were: [SC]
100 Hz: -57dBv
1 kHz: -95dBv
1.5kHz: -95dBv
10kHz: -100dBv
At low frequencies below 150Hz, the drive motor, mains hum and the like are also things that generate high values.

The values for Music Cut were quite high in this recording. The following table should illustrate this.

Bild1.jpg

Bild23.jpg

When looking at the analysis, it can be determined that the individual songs have different maximum volume levels. Of course, not every song has the same maximum volume in all frequency ranges, but I'll ignore that for now.
Pink Floyd was recorded much louder than Green Day. Billie Eilish was recorded by far the quietest!
So I think some of the frequency inconsistencies between the individual tracks are due to the overall volume.
Additionally, after disassembly work (several days and many signal test cuts later), I noticed that the straight bar on the cone of the right channel had loosened somewhat. I can't say whether this already had an influence on the music cuts.

I hope I have been able to answer the question and have not created any more confusion.

By the way, I did not carry out any further tests with the Cutterhead FP-8.
The reason for this: I already decided in July to build a new cutterhead (new housing) and a new completely revised lathe. And the parts for it have arrived. I will report on this shortly.

Thomas

[farmersplow]

Your work and effort is awesome, lot of inspiration. The sound of the cut is well balanced although I can hear some distortion on esses and cymbals in the 4-6kHz range. I have the same problem that I am trying to fight with all sorts of dynamic EQing to get the cut louder, but my head is far from prefect. How loud are your cuts compared to pressed records? In the Green Day sample i can hear some wobbliness is your motor sruggling to keep the speed. I have this problem on the outer edge of 12" disc. That is my main mission now to think about new platter drive.

Keep it up!

Lu.

Hello lulu,

thanks for your review! In fact, when I tried to reduce the silent cut, the rake angle was increased to over 5°. However, this also puts more strain on the turntable drive because the force required is also greatly increased. If a loud cut is also made and low frequencies are cut, this is even worse. Green Day was the first song on the 12" record (then Billie Eilish and then Pink Floyd). As you rightly wrote, the effort is greatest on the outer radius. This can also lead to wobbling. I also think that the contact pressure (cutting depth) was too great.
Regarding the volume: It is as loud as the loudest records I have! So it was a real stress test.

Thomas

[farmersplow]

Next Level - New lathe (FP-9) and new cutter head (CH-9)
08/2023

As already mentioned, I started designing and planning an improved lathe and a new cutter head back in July. There is always something to improve. Always!
But what is it about the current parts that bothers me so much that it is necessary to redo everything?

As far as the lathe is concerned, there are a few things.

For example, the drive that performs the feed. This has several weak points. The spindle has a pitch of 5mm per revolution (0.19685" per revolution). The holder is a ball screw which is "noisy" and requires regular maintenance (greasing). The spindle is driven by a stepper motor with gearbox, which is noisy and transmits motor resonances. With the small movements, these factors are small but still audible.
The guides on the profile rails (linear guides) are also recirculating ball bearing guides. These also make noise and require maintenance.
I would like a maintenance-free and noise-free solution here.

Then there is the cutting head. The cone with the glued tube is a source of error. I would like to replace this with a single part. The aluminum membrane should be improved to prevent cracks or breakages. A new press mold must be created for this. The housing is too big and too heavy and only just manages to cut the end groove. A smaller and lighter housing is needed. The FP-8 also has light, but it is too weak. This should also be improved. In addition, the ease of maintenance is to be improved. As the cutting head is also to be used for blanks that are not so flat, I would like to build the option of a suction tube on the housing (as an alternative). Because I spent so long building the FP-8 to accommodate an external suction tube, I overlooked one thing. I have a VCA (Vertical Cutting Angle) also VTA (Vertival Tracking Angle) of 23°. The target value is 20°. The 3° won't have much influence, but I would like to correct this.

Then there is also the base plate with the superstructure. The current structure cannot be used for changes to the propulsion system. There should also be room for all the electronics in the housing and they should be properly accommodated. This means that the housing has to be modified. All audio and digital connections must also be accommodated and the control board must also be able to be accommodated.

The suspension box for holding the cutterhead also needs to be redesigned. Additional locking and unlocking systems and further electronics need to be adapted.

I would also like to improve the design of the suction container and integrate it in/on the housing. And an improvement to the plate heating, production of PCBs, housing design, extraction system box...

It is important to me to design the construction in such a way that maintenance and updating is easy.

So there is still a lot to do.

And because I won't be slowed down, I'll start right away.


Improvement 1:

CONUS for new Cutterhead CH-9

With the "old" cone of the FP-8, the cone was glued to the rod. As the gluing point is very small and the load on it is high at high frequencies, the gluing point on the right channel had partially come loose.

Picture of the old cone:
.
.

a Konus Old FP8.jpg

.
.
My plan is to make this cone from one part.

To do this, a tool was made first:

Picture of the tool making:
.
.

b Konus FP9.jpg

.
.
Picture of the holding tool:
.
.

c Konus FP9.jpg

.
.
Then the cone was turned from a piece of aircraft aluminum.

Pictures of the cone:
.
.

d Konus FP9.jpg

.
.

e Konus FP9.jpg

.
.
And because it will be a stereo cutting head, two parts are needed.

The production is very difficult and it took a week of work and many failed attempts to get it right. It was a sweaty job to make the funnel. It takes about four hours to make one cone, as only 0.1 mm to 0.05 mm was added at the end.

Finally, both funnels have to be brought to the same weight. This is done by finely sanding the surface.
.
.

f Konus FP9.jpg

.
.
Both parts now have the same weight. The cone weight is 0.41 grams! Another weight saving! At the same time, the new cone is stiffer and more stable than the old glued cone. And I have a good feeling about it.

Pictures of the FP-9 cone:
.
.

g Konus FP9.jpg

.
.

h Konus FP9.jpg

.
.
Greetings from Austria
Thomas

[lulu]

Looks great,

I am really curious how the cone from one piece affects the sound and the resonances.

Lu.

Hello from Prague.

[farmersplow]

Looks great,

I am really curious how the cone from one piece affects the sound and the resonances.

Lu.

Hello from Prague.

Me too

[evildrome]

Regarding ball screws, I considered "upgrading" the half scale VMS 70 to ball screws but abandoned it as an idea due to concerns over noise.

The VMS 70 saddle actually does run on ball bearings (which I did not know when I made the half scale model). If I had known, I suspect I still would not have used ball bearings because the ball bearings used by Neumann were undoubtedly "special" and made for their application. Sourcing such bearings, even if I knew what they actually were would have been difficult and expensive.

The only change I made was I cut the lead screw nut from an "Asonic" plastic bearing material.

You might consider this material as a replacement for the linear guides.

https://www.igus.com/info/linear-guides-low-noise

I discounted the use of stepper motors from the start as I know they're noisy and opted for a brushless EC servo motor & controller from Maxon.

I haven't made a cone yet but I have thought about it & the attachment point of the force transferring rod has always seemed to be an obvious point of failure.

I had considered turning this rod from a larger section but leaving a flared end to A) cause better axial location inside the cone B) prevent separation failure.

http://www.wilsonlogan.com/images/flared-cone.jpg

I like your idea of turning the part as one piece better but recognise how difficult such a thing is to make.

I'm not sure if this would be easier but you might consider metal spinning. I have made some experiments with metal spinning on my lathe and TBF, the results were not good but I was trying to spin parts which required much heavier metal than you would need.

[farmersplow]

Hello evildrome,

two brains - one thought!

Regarding ball screws, I considered "upgrading" the half scale VMS 70 to ball screws but abandoned it as an idea due to concerns over noise.
The VMS 70 saddle actually does run on ball bearings (which I did not know when I made the half scale model). If I had known, I suspect I still would not have used ball bearings because the ball bearings used by Neumann were undoubtedly "special" and made for their application. Sourcing such bearings, even if I knew what they actually were would have been difficult and expensive.
The only change I made was I cut the lead screw nut from an "Asonic" plastic bearing material.
You might consider this material as a replacement for the linear guides.
https://www.igus.com/info/linear-guides-low-noise

At first I also thought about improved ball bearings and then came up with "Asonic" plastic bearing material. It turned out to be an excellent solution (I will write about it in one of my next reports).

I haven't made a cone yet but I have thought about it & the attachment point of the force transferring rod has always seemed to be an obvious point of failure.
I had considered turning this rod from a larger section but leaving a flared end to A) cause better axial location inside the cone B) prevent separation failure.

This cone solution is certainly better for split cone-rod drives. But the weight and the glue must always be taken into account. So don't leave too much flared end.

I'm not sure if this would be easier but you might consider metal spinning. I have made some experiments with metal spinning on my lathe and TBF, the results were not good but I was trying to spin parts which required much heavier metal than you would need.

As I have already written, I have tried it. It was partially possible to bend a cone. But a cone with a rod in one piece is impossible. But it would be possible for spiders. Nevertheless, I built a pressing tool for it, which is more precise. (I will write about this in one of my next reports).

I discounted the use of stepper motors from the start as I know they're noisy and opted for a brushless EC servo motor & controller from Maxon.

As for the stepper motor, I'm still sticking with it, but with an improved system (I'll write about it in one of my next reports).

I like your idea of turning the part as one piece better but recognise how difficult such a thing is to make.

I confirm.


Thanks for your input and thoughts on improvement. It's always nice to see how others think along with a project. That is very motivating!

Greetings from Austria
Thomas

[farmersplow]

Assembling driver coils with spider and cone
09/2023

The new press dies for manufacturing my metal spiders are finally ready. So I can start pressing the aluminum spiders from the new material. The "old" presses were intended for thinner sheet metal and there were always cracks during production with thicker sheet metal.
.
.

01 Spider Tool.jpg

.
.
Now pressing is no longer a problem. The stamps fit the new aluminum sheets perfectly and no problems can be detected under the microscope.

Incidentally, one spider weighs 0.49 grams.
.
.

02 Spider Detail.jpg

.
.

03 Spider Weight.jpg

.
.
So out of sheer joy I made lots of parts and connected the spiders to the driver coils.
.
.

04 Spiders.jpg

.
.

05 Spider with Coil.jpg

.
.
Before I glue the spiders to the speaker housings again, I must not make the same mistake and must prevent the adhesive from coming loose again. To do this, I roughened the gluing surface. This work has to be done very carefully. The tiniest shavings or chip dust would like nothing better than to get to the magnets. This must be prevented at all costs. Mask off cleanly in several layers and then remove all dust with adhesive tape before removing the masking layers. (Remove dust several times)
.
.

06 Prepared LS.jpg

.
.
Once the spider has been glued (and dried) to the speaker housing (as usual), the aluminum cone can be connected. I do this with the 3D milling machine, because it allows me to position it precisely.
.
.

07 Konus to LS 1.jpg

.
.

08 Konus to LS 2.jpg

.
.
Greetings from AustriaThomas

[farmersplow]

THX > 250.000

Thank you for over 250,000 views.
I am pleased that there is a lot of interest in the field of record cutting and also in the process behind it.
You are still welcome to post your ideas, criticisms (or even confirmations).
Inspiration is the basis of every development.

Thank you from Austria
Thomas

[farmersplow]

Improved spindle drive
09/2023

As already mentioned, I would like to improve the spindle drive in three respects. Firstly, it should be quieter. Secondly, it should be even more precise. Thirdly, it should be maintenance-free.

Improvement 2: Spindle

For this reason, I have replaced the ball screw with a pitch of 5mm/revolution (0.196"/revolution) with a trapezoidal screw with 2mm/revolution (0.0787"/revolution).

Improvement 3: Flange threaded nut

This material has finally arrived!
I replaced the ball screw with a flanged threaded nut made of iglidur®. This is quiet, highly precise and maintenance-free.
.
.

01 spindel.jpg

.
.
So that the spindle also fits, I first had to make it fit on the lathe.
.
.

02 spindel.jpg

.
.

03 spindel.jpg

.
.
For the new flanged threaded nut made of iglidur® material, I also need a new retaining block. This is CNC milled and now fits perfectly. In comparison, it looks like this:
.
.

04 spindel.jpg

.
.
Improvement 4: Bushings
I also took this opportunity to replace the bushings. Here too, recirculating ball bearings were replaced by slide bearings made of iglidur® material.
.
.

05 spindel.jpg

.
.
The disadvantage of the conversion is that the transmission ratios no longer match. The geared motor and drive must therefore be replaced.

Greetings from Austria
Thomas

[farmersplow]

Improved feed drive
09/2023

To ensure that the drive is not only quieter due to the bushings and the spindle, the drive for the feed of my turntable should also be improved. Changing the pitch of the spindle makes it necessary to adapt the entire drive motor structure.

Improvement 5: Stepper motor
In order to improve the precision of the stepper motor, it is necessary to install a larger and even more powerful stepper motor. This is because an increase in microsteps per revolution (for example from 800 steps to 1600 microsteps per motor revolution) also causes a loss of torque.
.
.

20230926_095926.jpg

.
.
Improvement 6: Gear ratio of the drive
The gearbox on the "old" stepper motor also caused (minor) noise.
I therefore decided to replace this gearbox with a toothed belt drive. This turned out to be more difficult than originally thought. The difficulties arose not only in the manufacture of the flanges between the drive shafts and drive wheels. The suspension of the stepper motor also had to be designed to reduce vibration and still allow the belt to be tensioned even when installed.

But the main problem arose due to the larger space requirement. This meant that I had to rebuild my entire cutting machine accordingly and relocate the turntable drive under the base plate.
This again revealed a principle that ran through my entire development work. Every solution to a problem has an impact on its surroundings and (usually) causes a new problem.
.
.

20230926_105228.jpg

.
.
The good thing is that I had already thought about this in July and redesigned all the parts, including those of the overall structure of the lathe. And these parts have also arrived. So I can continue right away.

Greetings from Austria
Thomas

[farmersplow]

New FP-9 cutting head (CNC)
09/2023

Before I start assembling the new parts of the complete lathe, I have prepared a look at the CNC milling of the new cutting head.
Why am I building a new cutting head again? The reason is not because I can. It's because I have to. An inner voice tells me that the current head is no good. (And my wife's voice saying that "Big Mac" is too big). It's not the sound that bothers me so much. That can certainly be improved. But I can also make the improvements in a new case. It's the weight, the size and the construction. And I want to make the obvious improvements right away before I start working on the details to improve the sound.

And I was very happy when I saw the first CNC cuts of the FP-9 in front of my eyes.
.
.

01 CH9 CNC.jpg

.
.

02 CH9 CNC.jpg

.
.

03 CH9 CNC.jpg

.
.

04 CH9 CNC.jpg

.
.
I hope my design is better this time. I will report in more detail about the head when I have all the parts.

Greetings from Austria
Thomas

[farmersplow]

New complete lathe FP-9 (CNC)
09/2023

In order to be able to install the new drive in the lathe, I have redesigned the entire device and am now assembling the CNC parts.

First, the central base frame is produced. I use aluminum profiles for this:
.
.

20230920_175913.jpg

.
.
Then the new base plate is adapted and attached.
.
.

20230920_180151.jpg

.
.

20230920_181540.jpg

.
.
In this way, the drive unit now fits under the base plate.
.
.

20230926_105234.jpg

.
.
And even if nobody will see the base plate in the end, because it is completely concealed by the drive carriage and housing. Nevertheless, it is important to me that the areas that are not visible have the same quality of workmanship as the obvious parts. So I have also subjected the base plate to a high quality of workmanship.
.
.

20230920_181602.jpg

.
.

20230920_181607.jpg

.
.
By the way: I love it when things are beautiful!
What do you think?

Greetings from Austria
Thomas

[Fela Borbone]

That looks realy nice, Thomas,
Well done.

[PLD]

What do I think? I think it looks rather wonderful...

[zdenek]

Hello ,
I don't know how you achieve a frequency range close to 9kHz using these speakers and I gave up on them a long time ago, Chinese car tweeters for $2 each turned out to be much better. I checked the small vibrators that you used in your 4OHM 20W head and the maximum bandwidth what I cut was 5kHz and the speaker coils were burning up, so I decided they were completely useless for this project and not suitable for the head, so after burning several sets, I gave up on them permanently. Your recordings indicate, as you say, that you did not use feedback in your head, but after the recordings you can hear that there is no echo that occurs when there is no feedback from the head or after turning it off, so it's very doubtful for me...
I also don't think it's a good idea with a moving platter [such a system was used in the early years where variable travel pitch was not used because during variable travel, the turntable platter retains its ballast like a gyroscope and can jump out or fall out of synchronization and quartz stabilization, you can check it by connecting an oscilloscope to the turntable motor encoder, each jerk of the turntable motor causes displacement modulation and after cutting the disc, the signal with a frequency above 10kHz is vibratto, i.e. additionally modulated [unnecessary when recording]. Just compare the signal from the turntable encoder and the time base signal of the turntable PLL quartz synchronization generator. ,control system and there will be great differences in signal purity. Any vibration or jerk of the moving platter disrupts the smooth operation of the platter.
I've now gone in a completely different direction with my project, but that's a different story...

[farmersplow]

That looks realy nice, Thomas,
Well done.

Thank You

[farmersplow]

What do I think? I think it looks rather wonderful...

Thank You

[farmersplow]

Thank you for the questions and doubts are always justified. I would like to respond briefly.

Hello ,
I don't know how you achieve a frequency range close to 9kHz using these speakers ...

Thanks to matching spiders and extreme weight savings.

I checked the small vibrators that you used in your 4OHM 20W head and the maximum bandwidth what I cut was 5kHz and the speaker coils were burning up, so I decided they were completely useless for this project and not suitable for the head, so after burning several sets, I gave up on them permanently.

If you show me a picture of your speakers (with which you could not get above 5kHz) I might be able to tell why you could not get to a higher frequency. I think that high frequencies have nothing to do with the coil or the magnet. It has to do with the coil suspension and the moving weight. With my drivers, the coil weight is hardly higher than with smaller coils, but the total weight reduction of all moving components (as reported) then makes up for it. Incidentally, none of these coils have burned out yet.

Your recordings indicate, as you say, that you did not use feedback in your head, but after the recordings you can hear that there is no echo that occurs when there is no feedback from the head or after turning it off, so it's very doubtful for me...

The recordings were made without feedback compensation. Only with EQ. Why should an echo occur? What is doubtful?

I also don't think it's a good idea with a moving platter [such a system was used in the early years where variable travel pitch was not used because during variable travel, the turntable platter retains its ballast like a gyroscope and can jump out or fall out of synchronization and quartz stabilization, you can check it by connecting an oscilloscope to the turntable motor encoder, each jerk of the turntable motor causes displacement modulation and after cutting the disc, the signal with a frequency above 10kHz is vibratto, i.e. additionally modulated [unnecessary when recording]. Just compare the signal from the turntable encoder and the time base signal of the turntable PLL quartz synchronization generator. ,control system and there will be great differences in signal purity. Any vibration or jerk of the moving platter disrupts the smooth operation of the platter.

This also happens when the cutting head is moving and the platter is stationary. I think that both systems have advantages and disadvantages. When the cutting head moves, the suspension box, the calibration unit, damping, the suction hose, the cables, etc. must also be moved. The bottom line is that the needle moves towards the plate.

... [such a system was used in the early years where variable travel pitch was not used because during variable travel, the turntable platter retains its ballast like a gyroscope and can jump out or fall out of synchronization and quartz stabilization...

History: It was used in the early years because there were no microphones yet and an incredibly large sound funnel was mounted on the cutting head (mechanical transmission like the Gramophone). The musicians and the singer stood in front of the funnel and played and sang at full power. It was not possible to move the cutting needle with the large funnel in a controlled manner and therefore moved the plate.

I've now gone in a completely different direction with my project, but that's a different story...

It would be interesting to hear or see something about it.

[farmersplow]

THX

Thank You for more than 1.000.000 views. I am delighted that so many people are interested in the development work.

Greetings from Austria
Thomas