Thanks Morten for letting us discuss your design, I think this should be listed as one the features of your preamps
Arduino was an example, I think once you put a more powerful system this is good as long as it is Linux. But it has drawbacks, as you say, it adds complexity. Our preamp should boot and periodically be updated when a new version of the OS or when a new driver is available, as it happens for our NAS. Turning off becomes a shutdown, which takes some time. At the end, the LDR preamp is going to become a sophisticated machine on the SW side, which is good but I am afraid of the complexity. I prefer Occam's razor principle, and go for the simplest way to get the job done (and I have to manage a lot of complexity in my daily work). Everyone has experienced the computer getting stuck with a blue screen, very often because of drivers: what if the preamp got stuck with a blue screen and the volume jumps to maximum? With a micro controller, things are more under control, I would use it in an elevator, not a Raspberry. So it seems to me an overkill to use a Linux system to control the preamp, unless we want Spotify directly on your preamp, so it becomes a network player, another beast
As for the noise, I would not underestimate its impact as people spend good money on isolating that noise in computer-based high-end audio. Sonically speaking, it is not a hum but its effect is that it flattens soundstage and makes mid-range dull.
I see that there many requests for many inputs, it makes sense. I prefer the shortest path, and maybe it suffices having extra solder pads close to the LDRs.
With the removable box, one could have connectors already soldered on it, so the signal path would be all isolated from the rest, I like this idea of yours
Note that the box is still plugged in as you say, but in the control signal path, not in the analog audio signal path, which is different to my eyes:
- current analog audio signal path (my guess, assuming no input switch): connector (XLR/RCA) -> soldering -> wire -> soldering -> trace on the board -> soldering -> socket -> plug -> soldering -> LDR -> soldering -> plug -> socket -> soldering -> trace on board -> soldering -> wire -> soldering -> connector (XLR/RCA)
- analog audio signal path in the box with connectors: connector (XLR/RCA) -> soldering -> trace on the board -> soldering -> LDR -> soldering -> trace on board -> soldering -> connector (XLR/RCA)
As you can see, the number of steps is half in the signal path. Increasing the steps in the control path is less detrimental to my eyes.
Maybe I am too extreme, but you know, this is our hobby: a preamp is a preamp, a computer is a computer
Cheers
-Roberto
Hi Roberto. Good to hear from you and thank you for your thoughtful input.
There's little to distinguish an Arduino from our controller board aside from the fact that the Arduino's are mass produced. Both use standard available microcontrollers and are programmed at the basic hardware level. We would have to mate any Arduino with our own custom interface board same as we would for the Raspberry. I prefer the Microchip family of microcontrollers whereas the Arduino uses Atmel although I've worked with both. Our earliest prototype LDR board used an Arduino. Personally, I'm not very keen on the Arduino programming language and prefer working in either C or Pascal. I currently use MirkroElectronika's development hardware and compilers.
The Raspberry Pi is totally different animal wherein we'd be programming on top of a full Linux operating system. A lot more complexity under the hood and we'd be insulated from the microprocessor at the hardware level. Fortunately there are sufficient drivers available for everything we need - and then some - to handle IR receiver and SPI communications, USB, wireless etc. not to mention high level display functionality. Concerns over noise and shielding may be valid but then again our current boards are already running 16 MHZ microcontrollers with shared power supply driving the LDRs and no shielding all without discernible downside.
I'd say it's around 50/50 in terms of comments I receive on the topic of number of inputs. Half are asking for more than 3 and the other half can't fathom why more than 2, perhaps 3 at the most...and some say why not just 1. There's no perfect one size fits all solution here. The current prototype board design comes close insofar as it has the flexibility to only have 1 input or have up to 6. It's only a matter of adding the LDR modules which are plug in.
It would be a simple matter to minimize trace lengths by eliminating J1L and J1R as such and simply having individual solder pads as close as possible to each input LDR switch. This would have the added benefit of shortening the board up by perhaps another 0.5 inch. However, as a practical matter this doesn't really shorten the signal path. You still have to run wires from each input RCA/XLR jack to the board. The traces may be shorter but the wires to each trace/pad would have to be longer. Pick your poison.
One of the more interesting philosophical issues around design is whether one is designing around a real or imaginary problem. Is noise really a dominant problem here that needs a solution? In my view that remains an interesting question as it relates to the bigger topic of perceived benefit of balanced audio within short run stereo audio...to which I say... I don't know. I can say that with some conviction that noise has not been an issue with any of our LDR preamps aside from the your run-of-the-mill ground loop hum issues which are really a separate noise category best solved through good grounding practices. RF/EM noise has not been a problem as far as I'm aware. Therefore taking special precautions such as shielded boxes and/or cables internally within the preamp do not seem warranted.
Also I doubt anything is gained by having to replace a whole set of LDRs within a removeable shielded box vs. individual LDRs modules. If it's removeable it's still plugged in and not soldered in. Tubes live quite effectively within a plugged in reality, so my thought is LDRs should be able to as well. We use gold plated pins and sockets so corrosion shouldn't be an issue.
One design architecture that we've explored quite a lot is to move the input switching out to the individual RCA/XLR input jack. While this would indeed minimize the signal path of each individual input, you still end up with common signal(s) that have to run back to the LDR board. Plus you'd end up with numerous switching control signal cables running from the LDR board to each input jack which frankly would end up looking really awful.
After going around the track numerous times with various alternatives we arrived at the design you see below in the recent post. By locating the LDR board as close to the rear panel as practical you minimize the signal wire runs to/from the LDR board and ultimately the total signal run distance between the rear panel inputs/outputs and the LDRs. At least that's the thought behind it. 
There's one additional step we could take but have resisted. The LDR board could be custom designed to mate up directly with the RCA or XLR jacks. This eliminates any additional wiring between the jacks and the board. The issue with that is you end up with a different custom board for each and every model of preamp. Many designs use this approach since it has the added benefit of reducing the manhours needed to build a preamp.
