Deep Grooves Mastering - Feedback Transducer

[opcode66]

I've been working on feedback for a while. I now have some results to post. The work is being done separately from the Bladrunner Project itself. So, I thought I would start a new thread. However, I have a feeling that this is going to be a short one. I'm already achieving good results. And, I know how to increase the current feedback level by threefold with one simple tweak!

Since the first version of Project Bladerunner, I've incorporated the electrical connectivity for feedback. My transducers have always had connections points for the pins. And, the cutterhead body has always had holes for the pin sockets for feedback. The only thing I've had to modify to make a feedback transducer was the bobbin. By adding a secondary feedback coil and feedback magnet, I'm now getting a 4 to 1 ratio of feedback induced by the static field of the feedback magnet vs. feedback induced by the fluctuating electromagnetic field generated by the drive coil. Once I move, I will make the planned tweak and I should be getting something on the order of a 12:1 ratio. That is in excess of what is required to make negative feedback summing work. Video below!

Here is version 2 of my shielded feedback transducer

IMG_6422.JPG

Here is the feedback transducer within my feedback testbed. It is essentially a slice of my cutterhead body.

IMG_6423.JPG

Impeccable drive coil work as always.

IMG_6394.JPG

[Stevie342000]

Hey Todd

Just a couple of quick questions:

Does that ration 4:1 stay the same if you do frequency sweep?

If I understand the principle of Feedback then its purpose is to smooth out any peaks or troughs to give a flatter frequency response?

Is your head with feedback giving results which are the expected or calculated results?

I very much like the concept of buying a non-feedback head and then being able to upgrade to a feedback type by a changing the cutter coils. I wonder if neumann would have got to that point if we had not seen the advent of the little silver disc.

Anyhow good work, as your getting ready to move to California in May there may be a break in communications at that stage? I've been to San Diego, you will love it if that is where you are going, the weather even in winter is dreadful it's like a UK summer, such hardship.

[opcode66]

I'm not sure. I haven't done sweeps yet. I've only been testing with 1K tone as it is the standard for calibration purposes. This is the exact same concept as how a phono cartridge works, by the way. There will be some biases in the response as there are in many different phono carts. Feedback from cutterheads (doesn't matter which one) typically require a bit of eq'ing in order for it to be a flat response. Otherwise, if the feedback is not flat, it would not work to flatten the overall response of the cutterhead in the groves being cut.

Yes, the purpose of feedback is used to flatten the response. Because everything that vibrates has a resonant frequency, it takes very little energy to make that item vibrate when the frequency of vibration is at the RF. Therefore, if you add feedback to a transducer, you can tell when it is vibrating at the RF or multiples of the RF. When you sum the feedback negatively with the input signal before power amplification, it serves to flatten the response at the RF and its multiples. So, feedback is primarily implemented to handle taming the resonant frequency of the cutterhead. Accounting for the physical world electrically. The overall result is very accurate representation of audio.

I have not put a feedback transducer into my cutterhead yet. And, I haven't finished my personal Caruso Pre-Amp build yet. So, I would have no way to test and let you know how all this works in practice. This is all theory and benchtests. It been some tricky work just to get here... Once I get a ratio of 9:1 on the feedback transducer I will be ready to convert Bladerunner 1.C to a feedback head. But, I'll also have to finish my pre-amp too. Give me a bit more time and I'll be reporting on this.

I'm moving to Southern California. I'll still be posting. This represents the last project work I'll do before I move, however.

[Stevie342000]

I thought that is how it worked, having seen the BBC documents for Grampian cutter head with feedback using Modified Presto Cutter amp.

If I have got this right you set the feedback ratio at the usual calibration frequency which you use for setting your cutting level at X cm/s. The ratio would be set by some means such as a potentiometer and then that ratio would be the same for whatever cutting level at X cm/s you were using?

In the BBC amp the feedback forms part of the equalisation chain and works in the Cathode of one of the input tubes, which then changes the bias upon that tube affecting gain and in essence flattening the response of the cutter head across the audio spectrum. Although using the Grampian as my example may not strictly be true but in essence that is how it works.

So the feedback coils form an active gain control in the feedback circuit of the amplifier? Is phase shift an issue? If I think of anymore issues you may run into and I am sure you have thought of most of them already.

The development stage so far has taken longer than I think you would have liked but it is a very important part of the end product, as this way you are more likely to get it right and to be able to replicate the cutter heads in a much shorter time and they are infinitely repairable. It has been most interesting to follow your progress so far. But you need to balance the money coming in from work in order to allow you to further develop any developments in cutting discs.

Well I hope all goes well with the move, I remember the last time I moved over 20 years ago, not looking forward to the next move, if ever there is one. Collected way too much electronics stuff over that time, still de-cluttering and setting myself up in business currently to provide an income.

All going well so far, my aim is then to be able to focus on spending my time doing the things I want to do on my bucket list, by working only when I want and doing what I want to do with the rest of my time allocated to me.

[opcode66]

Input signal from your source goes through a trimmer which then gets negatively combined (either by tube or transistor) with feedback signal from the cutterhead. Feedback is of course amplified and there is a trimmer to control its final level before being negatively summed with input. In the amplification of the feedback there is typically some correction of the low end response of the feedback coil. Also, in preamps which perform this function, there are typically separate trimmers for feedback monitoring level as well. Everything is calibrated. On disc, yes, you use a test tone disc to start from which has 7cm/s 1K tone to calibrate your pickups monitoring. You calibrate drive and feedback levels via a VU meter. 1k and 5k tones are used. You have to do some test cutting of the 1k tone to finish the calibration process.

You ask if phase shift is an issue. Well, it seems logical that you want to be sure that your wiring is correct and that you are maintaining proper phase throughout the system. If you inverted a signal, then negative feedback would not work as expected.... It would be opposite of what you want. It would in fact be amplifying when it should be decreasing level if you inverted signal.

Thanks. I hope the move goes smoothly too.

[Stevie342000]

Thanks for checking my understanding of how feedback works. I have seen the manual/schematics for Fairchild 641 and for the BBC Grampian head.

They both have some kind of equalisation in the feedback circuit and level control in the feedback circuit which to my mind is an active negative feedback system rather than the static kind you see applied across the open loop gain of a regular power amplifier or pre amplifier.

Does it matter where you apply the feedback i.e. in a pre amp or in the power amp or can it be done in the digital domain?. I am more old school and would do it in the power amp around first stage gain but trying to understand all this new fangled solid state and digital technology.

All making sense so far though......

[opcode66]

Most cutting systems I'm familiar with have the feedback summing happening in the pre-amp (signal processing stage). The power amp section typically just does amplification and can be potentially swapped out with other amps.

I suppose you could do the negative summing in the digital realm. I'm not sure anyone is doing that. At bare minimum this would require amplification of the feedback signal before being digitized, eq'd and negatively summed. Dont forget IRIAA applied to the input signal before summing and RIAA applied to the feedback before summing. Each riaa and iriaa has to be switchable so you can remove them for calibration purposes. I think it would be a lot of work to make something like that in software vs the off the shelf preamp solutions that one can find secondhand or build themselves.

Implementing feedback puts my cutterhead squarely into the professional category. It distinguishes finer engineered heads from more economical ones. Right up there with Neumann, Westrex and Vinylium. Of course it requires more to use a feedback head, but the results are well worth it. The clarity and accuracy cant be achieved by using a dynamic head and an eq.

[Stevie342000]

Most cutting systems I'm familiar with have the feedback summing happening in the pre-amp (signal processing stage). The power amp section typically just does amplification and can be potentially swapped out with other amps.

I suppose you could do the negative summing in the digital realm. I'm not sure anyone is doing that. At bare minimum this would require amplification of the feedback signal before being digitized, eq'd and negatively summed. Dont forget IRIAA applied to the input signal before summing and RIAA applied to the feedback before summing. Each riaa and iriaa has to be switchable so you can remove them for calibration purposes. I think it would be a lot of work to make something like that in software vs the off the shelf preamp solutions that one can find secondhand or build themselves.

So this would be one instance where applying negative feedback from the cutter head is better done in the analogue domain. It's basically simpler to apply, I see no reason why it could not be done in the digital domain. It's just a whole different set of issues to solve....

[opcode66]

Analog would have nearly zero latency. Digital processing would have latency at several points: adc, digital eq'ing and summing, and dac. All that latency would seriously and dramatically reduce the overall effectiveness of the system. I dont think it would actually work because of the latency in digital. You need the feedback to be instantantly prepared and summed with the input. Not some number of milliseconds later. That wouldnt work. Electricity travels at near the speed of light...

[electronrancher]

So this would be one instance where applying negative feedback from the cutter head is better done in the analogue domain. It's basically simpler to apply, I see no reason why it could not be done in the digital domain. It's just a whole different set of issues to solve....

I tend to agree with opcode66 on this one - it would be expensive and somewhat complex to digitize the feedback, difference it with a delayed version of the input, and send the resulting correction to the power amp. But if what you want is digital programmability, you could still switch in gain resistors, etc using digital means if the feedback loop were analog.

[opcode66]

I tend to agree with opcode66 on this one - it would be expensive and somewhat complex to digitize the feedback, difference it with a delayed version of the input, and send the resulting correction to the power amp. But if what you want is digital programmability, you could still switch in gain resistors, etc using digital means if the feedback loop were analog.

Actually you would need future feedback signal. Not delayed input. Which is of course impossible. If you mixed with delayed input and sent to power amp, it would no longer be in sync with the real time signal. The issue of latency in this model could not be overcome by mixing with a delayed input signal. That would be cart before the horse. The signal you are sending out creating feedback can not then be mixed with a delayed signal summed with latent feedback and have everything sound correct.

[EpicenterBryan]

I'll chime in on this....
Doing this in the analog world is the only practical way to go. It's real time, with real measurements and adjustments done in real time. However, correction could be done in the digital world but only by developing a model - a stimulus and response model accounting for many (if not infinite) conditions. As we all know by now, it's not just about frequency response under a specific condition like one drive level, or one cutting force. It's a multi point, multi conditional response - and in the digital world it would be an attempt to capture a real life response to generate a simulation of the real world. With enough data under enough conditions it should be theoretically possible.

Antares has used this kind of "holistic" modeling method to develop models of many microphones for use in their Microphone Modelers. I own two of them. The result is you can select the microphone you used as the source, and select the microphone that better fits the situation (or you could never afford). Does the result sound different? Absolutely. Is it an absolute match of the Telefunken U47 I selected? I can't tell you - I have never owned one. But does the response change as level changes ... Yes. Is it correct? Not sure. Does the simulation change when I say the distance is 31 inches compared to 6 feet from the source? Yes. Is the simulation correct? Again, I can't tell you.

I guess a model could be developed for my home-made head - and if I ever make a second one with a hole drilled 0.01mm off, a new model could be developed. I just don't see it as being practical.

If a guy were to make 100 at a time on an automated production line, yes.... I think it's possible.

Bryan

[opcode66]

Maybe with nanotube technology or graphene processors or some other future voodoo, latency will approach analog. But, until then, feedback is best accomplished via analog.

[Stevie342000]

Maybe with nanotube technology or graphene processors or some other future voodoo, latency will approach analog. But, until then, feedback is best accomplished via analog.

The question had to be asked as I may not be the only one to think about could this be achieved in the Digital domain.

Would one way of achieving this be to apply the feedback i.e. process the signal in the digital domain as if feedback was being used?

That would be in real-time to process the signal to correct the cutter head response and iron out any of those peaks and troughs to make its response more flat?

I prefer the analogue method just thinking about options here but given that your feedback stereo cutter head is going to be reasonably priced I do think the ideas I am thinking here are a mute point.

[markrob]

Analog would have nearly zero latency. Digital processing would have latency at several points: adc, digital eq'ing and summing, and dac. All that latency would seriously and dramatically reduce the overall effectiveness of the system. I dont think it would actually work because of the latency in digital. You need the feedback to be instantantly prepared and summed with the input. Not some number of milliseconds later. That wouldnt work. Electricity travels at near the speed of light...

Realtime digital feedback control is done all the time. You do need to set the system sample rate properly, pick a proper ADC/DAC and have enough processing power to do the calculations on a sample by sample basis (or very small block size), but this is possible with modern DSP chips and converters. One downside is that you are no longer totally in the analog domain. If you are an analog purist, that would be a big issue. I suspect it would be more costly as compared with an analog approach. There are several possible advantages:

1. Totally digital signal path including the power amp.
2. Self tuning
3. Better stability (no component drift).
4. Better loop compensation (higher order filters).
5. Possibly better driver protection.

This approach certainly would not have been possible in the 60's and 70's. But it is possible now.

Mark

[Stevie342000]

Analog would have nearly zero latency. Digital processing would have latency at several points: adc, digital eq'ing and summing, and dac. All that latency would seriously and dramatically reduce the overall effectiveness of the system. I dont think it would actually work because of the latency in digital. You need the feedback to be instantantly prepared and summed with the input. Not some number of milliseconds later. That wouldnt work. Electricity travels at near the speed of light...

Realtime digital feedback control is done all the time. You do need to set the system sample rate properly, pick a proper ADC/DAC and have enough processing power to do the calculations on a sample by sample basis (or very small block size), but this is possible with modern DSP chips and converters. One downside is that you are no longer totally in the analog domain. If you are an analog purist, that would be a big issue. I suspect it would be more costly as compared with an analog approach. There are several possible advantages:

1. Totally digital signal path including the power amp.
2. Self tuning
3. Better stability (no component drift).
4. Better loop compensation (higher order filters).
5. Possibly better driver protection.

This approach certainly would not have been possible in the 60's and 70's. But it is possible now.

Mark

Afraid I am an analogue purist but I thought it would be possible given your last statement. What was not possible in the past may now very well be possible. It would of course be much easier to do in mono but can not see why the digital processing could not be done in real time for stereo. The digital equivalent of feedback may be more than possible in this scenario.

[opcode66]

I disagree with you Mark. From my perspective, the amount of latency introduced by processing feedback for the purposes of cutting audio would make the results inaccurate to the point of having no considerable beneft over simply using a well constructed digital EQ curve. I understand feedback is processed digitally all the time. But, not necessarily for the purposes of audio transport. Innaccuracies and timing lags become very apparent very quickly. Additionally, you would not achieve as flat of a response as you would analog with any amount of latency. No way around that.

[Sillitoe]

Is this system posted by Bahndahn a similar example to what is being discussed here?

https://cycling74.com/2015/02/04/transforming-room-sound-with-a-max-patch/

[markrob]

I disagree with you Mark. From my perspective, the amount of latency introduced by processing feedback for the purposes of cutting audio would make the results inaccurate to the point of having no considerable beneft over simply using a well constructed digital EQ curve. I understand feedback is processed digitally all the time. But, not necessarily for the purposes of audio transport. Innaccuracies and timing lags become very apparent very quickly. Additionally, you would not achieve as flat of a response as you would analog with any amount of latency. No way around that.

I don't get why you think there is any significant latency involved in the processing. Not all A/D converters are delta-sigma types with large FIR decimation filters. Not all processors need to work on large blocks of samples to have high throughput. Your thinking is based on how audio processing is done on a PC with a CISC processor and a non-realtime OS running. DSP hardware is much different. Disk drive position digital servos sample at rates in the 10Khz and up range. They would fail if there were latency in the 1ms range as you alluded to in a earlier post. Have you ever done any digital servo design work?

Mark

[opcode66]

Ok, we'll see when you make sych a system. Until then, I'll maintain my opinion that you cant reach as flat of a response via dsp. Until I'm proven otherwise imperically, this is all theory and opinion.