Hello everyone! I own an English analog Soundcraft console and now that my favorite tape recorder has passed away, I decided to build my own converter for recording to digital. I'll say right away that the options on the market (ready-made) don't suit me for several reasons: price, functionality, and external layout. My console has 56 mono channels, 4 stereo channels, and 8 more groups, so I'd like to understand how the DAC/ADC works and build a custom one. I came across various schematics from Behringer for their DAC/ADC, but I didn't like the chip used in the circuit. Does anyone have experience in creating converters? Can you share your skills and schematics of ready-made single-channel converters (so that I can then install them in one case). It's also interesting to read about connecting to a PC. Installing cards into the computer itself is not an option (there are no free slots), and via USB is a good option, but if you remember the above idea of using single-channel converters, the question of combining them and transmitting information via one USB comes up. I will be glad to receive any useful information. Thank you.
Before your deep dive into DIYing 56 ADC channels, can you tell us what "options on the market" you have looked at or considered before deeming them un-suitable?
Of course! I looked at options from Avid, Universal Audio, PreSonus. Some are too complicated and some of their functions are useless, some are too empty in terms of functionality. And besides, I am interested in assembling something with my own hands. I have already assembled compressors, equalizers, preamps and they still work in the studio, and in the end they cost much less than exactly the same but factory assembled, although I did not take into account my labor costs.
56 mono, 8 groups and 4 stereo over USB? Have you looked at RME? They have 64 channel convertors and then a madi-to-usb for 64 (in and out) channels. Won't be cheap but it will work. RME is good stuff. I'm using a Digi9652 + ADI-8 AE since years, it sits now in its third pc.
Doing this yourself isn't going to be easy. You are recording so you will likely want to have bit-synchronisation over all channels. And then you will need a driver in your pc to handle all of it.
Doing this yourself isn't going to be easy. You are recording so you will likely want to have bit-synchronisation over all channels. And then you will need a driver in your pc to handle all of it.
It's not quite as simple as designing and building an analog amplifier, compressor or the like. Given the channel count required you would probably need a dante or similar interface that supports a large number of channels.
Kaamos Tech in the UK makes ADC modules based on AK5572EN converters, I am using a pair of mono configured ADC boards with their clock generator and I2S to SPDIF/AES transmitter board. It is very transparent. Getting I2S into a format you can record will be a big challenge.
https://kaamostech.com/ (I have no affiliation)
Kaamos Tech in the UK makes ADC modules based on AK5572EN converters, I am using a pair of mono configured ADC boards with their clock generator and I2S to SPDIF/AES transmitter board. It is very transparent. Getting I2S into a format you can record will be a big challenge.
https://kaamostech.com/ (I have no affiliation)
Writing that driver software will require a very significant learning curve if you haven't been doing low level programming already, even more than making the circuit design and low noise circuit board layout, IMO. This kind of thing isn't usually something done by a team of 1. But big congrats if you can pull it off!
And what if we use 56 absolutely independent converters, which will already support current drivers for recording? Or should we still consider options for rehabilitating my old tape recorder or replacing it with something from Revox or Studer?
Depends on what your goals are, but if you are interested in rehabilitating the current recorder that is not a bad option provided the heads are OK. What is it? ReVox was prosumer and I am unaware of anything beyond 2 track stereo, Studer made 24 track machines, probably cost prohibitive. Otari made some good 24 track recorders (MTR series) which might be a little more cost effective.
What if I make 56 independent converters and combine them in my DAW? It should work out. You can recommend single-channel adc/dac converters for this. Ideally, of course, with ASIO support or No latency options. Maybe there are already ready-made chips for such purposes?
With a sufficiently fast clock, you maybe able to use an input multiplexer to convert many channels using a single ADC. However, this technique requires some embedded programming to control the sequences and format the converted data correctly.
how about this method? I know that in MacOS you can take several devices and combine them into one, then the recording program will think that this is one aggregate device. But will this work in Windows or is it just that the converters don't care how many of them will work with one PC?
Maybe helpful to tell us about the Soundcraft console. Is it all analog, with an analog output for each channel? How does it sound? What kind of condition is it in?
What kind of SQ do you want from ADCs? Do you want to mix their outputs in the box (digitally)? Or, how do you want to mix?
Or maybe you are thinking of mostly gutting the electronics in the Soundcraft console, and then make it into a fully digital console with a digital mix bus?
Once we understand some basic things better, then we can drill down more deeply into the exact area where you want to go.
What kind of SQ do you want from ADCs? Do you want to mix their outputs in the box (digitally)? Or, how do you want to mix?
Or maybe you are thinking of mostly gutting the electronics in the Soundcraft console, and then make it into a fully digital console with a digital mix bus?
Once we understand some basic things better, then we can drill down more deeply into the exact area where you want to go.
The problem in that case could be making the various input devices synchronous with each other. Otherwise you get clock drift problems between different input devices. Whether or not that can be managed/compensated properly depends on how they are interfaced going into the computer. If they all go into one digital interface device via AES, the ASRC on each channel can sync them up. However, the would be some (possibly minor) SQ penalty for going that way since ASRC are in some ways analog devices. Thus they are sensitive to power supply noise, board layout imperfections, bypass design, etc.
The console is completely analog and in perfect condition (technically, there are some flaws externally, but that's not the point). Each channel has an input and a direct output. When I had a tape recorder working, the sound sources and outputs from the tape recorder were plugged into the patchbay. From the patchbay, the sound went into the console input. I mixed it and when recording, the sound from the direct output went to the tape recorder for recording. Now I'm trying to do the same without it, only instead of a tape recorder there should be a computer. In a word, I would like to replace the tape recorder (since I don’t have time to deal with it and it’s unlikely that it will be possible to repair it, only to buy a new one, and the prices for vintage equipment... in a word, three or more zeros in dollars) with a computer, it’s easier to mix because 10 plugins for every 56 channels is somehow more compact in digital than in analog.
Okay. To get 56 each of, say, maybe 24/96 inputs and an equal number of outputs, is probably going to exceed the bandwidth of USB2. Also, so far as I know there is not a standard USB3 audio interface defined yet. Once you need multiple USB2 interfaces, keeping them in time sync with each other is likely going to be significant issue. In that case there would be a strong reason to consider a professional interface standard, such as Dante.
EDIT: as far as there being a way to interface multiple external hardware devices to recording software, probably the external devices are time calibrated in some way to account for processing delay (such as by sending a pulsed tone over the interface and measuring the round trip time delay in samples, then adjusting the recording software to account for the sample delay). However, that sort of thing can't compensate for ongoing clock drift in external converters.
EDIT 2: If you want superb SQ from this digital tape recorder, that's a different requirement than emphasizing very low cost as a specification. For one example, for the best SQ Mobile Fidelity could find, they selected a Merging Technologies DSD256 Horus system. Probably that's about right for excellent SQ, but not ideal for every budget.
EDIT: as far as there being a way to interface multiple external hardware devices to recording software, probably the external devices are time calibrated in some way to account for processing delay (such as by sending a pulsed tone over the interface and measuring the round trip time delay in samples, then adjusting the recording software to account for the sample delay). However, that sort of thing can't compensate for ongoing clock drift in external converters.
EDIT 2: If you want superb SQ from this digital tape recorder, that's a different requirement than emphasizing very low cost as a specification. For one example, for the best SQ Mobile Fidelity could find, they selected a Merging Technologies DSD256 Horus system. Probably that's about right for excellent SQ, but not ideal for every budget.
Last edited:
This does not work. All your channels need to work in absolute sync.
I'd buy two RME M32 AD units, use them in sync with a RME MADI card, with this you could record 64 channels in 48kHz (on a capable PC, that is).
That would be one of the most cost-effective solutions and more importantly, it will work as expected.
I certainly don't want to spoil the party but in general, ceating this kind of really elaborated large-scale stuff yourself as a DIY project has to remain a pipe dream. You'll easily loose much more money than for any off-the-shelf solution and even if you master all the problems and issues (looks like you are starting from scratch in terms of skills and knowldedge) it will take years... realistically, the whole project will likely end in nothing else than endless frustration and lot's of burned money.