Hypex DSP module(s)

would it be possible to switch between the two USB devices with some relays internally?

This suggestion works perfectly. I bought a cheap 5V 4x relay board (this one), then I wired it up so the DAC would be on the common terminals, the RPi would be on the normally closed (NC) terminals, and an external PC would be on the normally opened (NO) terminals. The power for the relays is looped from the NO connections, so that the board is powered by the external PC, and therefore *only* when the external PC is connected. The relays are triggered when the IN terminals are grounded (this is configurable via jumpers so they could be triggered by the 5V if preferred), so I simply looped the DC- terminal to all 4 IN terminals.

The result is that the relays are normally unpowered, so it acts as a straight through cable between the RPi and the DAC. When an external PC is connected the relays switch over to disconnect the RPi and connect the PC.

I tested it with my RPi (running piCorePlayer) and my Dacmagic. When I plugged in my Macbook the music from the RPi was cut and the Dacmagic showed up as an audio device on the Mac. Unplugging the Macbook reconnected the RPi and the music restarted as if by magic.

So this is the perfect solution from my perspective. I get to have the RPi as the default music source, built into my amplifier enclosure, and with the relays not even drawing any power unless I connect an external PC (or alternative USB music source). And best of all, the Rpi doesn't 'use up' any of the DLCP's inputs - I still have as many as if the RPi wasn't there.

USB%20Relays_1024.jpg
 
Thanks Eirik

The 4 USB wires are on the J3 header, which is a 16-way box header for a ribbon cable, which I believe comes with the DLCP input board (will know for sure tomorrow - just had word that my Hypex order will be delivered tomorrow).

So it should be a simple matter of making a replacement ribbon cable to break out those 4 wires and route them through the relay board (DLCP input board to NO, and DLCP J3 to COM).
 
That sure is a nice way to get some EMI radiation in the case.

Do you mean the relay board? It's not powered unless the external PC is connected, so completely inert when it's just the RPi playing. Even if I want to connect another music source instead of a control PC, that relay is DC powered, and is switching USB data lines, not audio signals, so I can't see how it can affect sound quality.

Or did you mean the RPi itself? Squeezebox is the source for all of my music listening, either from my ripped CD collection or through the various streaming plugins (iPlayer, Spotify), and the only other source I like to run through the amp is the TV (Toslink). So having the RPi inside the case makes a lot of sense for me. My own feeling is that this interference issue is overblown - I'm not expecting it to have any effect.
 
My DLCP hasn't quite arrived yet (waiting at the UPS collection point for me to collect this evening!), so I've had to make do with studying the documentation.
I'm still trying to absorb how all the different inputs and outputs for these boards work together, but as far as I can see there isn't a 12V output. But your amplifier power supply may well have a suitable 12V output.

For instance, the SMPS1200A400 has a regulated 12V output, and in conjunction with the 'supply standby' signal this might do what you require. I can't see whether this 12V output is always live, or only live when the supply is not in standby. If the latter, then that could already be what you require - i.e. so long as you let the DLCP control whether the supply is in standby or not, then powering up the DLCP will make this 12V output go live. But if that 12V supply is live all the time, then you could run it through a relay to provide your trigger output, and control that relay with one of the DLCP signalling lines (e.g. Amp-enable, supply standby). You'd need to pay attention to whether those signalling lines go high or low, and configure your relay accordingly, but that's easily done with these relay boards that are designed to be controlled by Arduino or RPi applications.

PS - I knew I recognised your username - that's a lovely looking enclosure you presented a few pages back.
 
Last edited:
I use the standard DLCP power supply.

Got it, ok - so you're building a pre-amp/processor, and not an integrated amp. My previous answer was coloured by my own plans to build an integrated amp. So the only power supply in your enclosure is the DLCP power supply. In that case I guess your best option is to use one of the DLCP signal lines to trigger a relay that will switch a separate 12V supply on and off. What is the 12V trigger going to do? You could put a little 12V supply board in the DLCP enclosure, or you could feed the signal line out of the enclosure and put the 12V supply inside whatever you're planning to trigger with it.
 
Do you mean the relay board? It's not powered unless the external PC is connected, so completely inert when it's just the RPi playing. Even if I want to connect another music source instead of a control PC, that relay is DC powered, and is switching USB data lines, not audio signals, so I can't see how it can affect sound quality.

Or did you mean the RPi itself? Squeezebox is the source for all of my music listening, either from my ripped CD collection or through the various streaming plugins (iPlayer, Spotify), and the only other source I like to run through the amp is the TV (Toslink). So having the RPi inside the case makes a lot of sense for me. My own feeling is that this interference issue is overblown - I'm not expecting it to have any effect.

USB data lines are high speed and are routed deferentially for a reason: https://product.tdk.com/en/products/emc/guidebook/eemc_practice_02.pdf

What you have there is a USB aerial. :-D
 
USB data lines are high speed and are routed deferentially for a reason: https://product.tdk.com/en/products/emc/guidebook/eemc_practice_02.pdf

What you have there is a USB aerial. :-D

Ah I see - I jumped to the wrong conclusion and assumed you were talking about differential audio signals.

So what's important then? Is it the length of the cables, the shielding, the connection types? Note that the Hypex DLCP input panel routes the USB wires along adjacent strands of a ribbon cable, so no shielding on that short section. I had already been thinking of replacing the screw-in terminals on my relay board (which aren't going to be ideal for strands of a ribbon cable) with cable-to-board connectors (e.g. header pins) to make the connection easier. I guess I should minimise the differences between the data line path lengths, but how much difference is acceptable?
 
I just thought it was funny given the conversation previously about EMI in cases.

I'm not a layout expert so stand to be corrected, but I think it's the trace length matching and the trace spacing that is key. Any deviation and the signals are out of phase which results in radiated noise.

A ribbon cable is better, in the scheme of things, but doesn't seem ideal. I wonder if they have to do any EMI compliance testing or whether the DIY status of the product means they can bypass it.

Either way, as Julf said, it shows the robustness of the interface but if you're streaming audio in isochronous mode, then there's no error checking. Use it as it is and if you have the time, money and interest, then building a PCB might be an interesting project to see if you can hear a difference.
 
Use it as it is and if you have the time, money and interest, then building a PCB might be an interesting project to see if you can hear a difference.

^^This is exactly why I like DIY Audio. Designing PCBs is new to me, but I've had great fun designing one to carry some Raspberry Pi buttons and an IR receiver in precisely the same configuration as the DLCP control panel. 10 samples of my first attempt are due to be delivered from China on Monday!

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


So a bespoke PCB to carry some low voltage, low current relays, with matched path lengths and convenient wire-to-board connectors is something I've already considered. Thanks for the heads-up about the data transmission issue - it gives me a good reason not to just go with the easy pre-built option. :)
 

Good find, thank you. So switching between computers is an acceptable thing to do. But that particular device requires a button press, and I plan to put mine inside the enclosure where it will do it's job automatically.

I had a quick look at some relays, and as far as I can tell, most of the other gubbins on my pre-built 4x relay board must be for lighting up LEDs, regulating power, handling the switching logic (high versus low) etc. If I simply want a device to switch 4 USB lines, using the 'hot' external PC to power it with it's own regulated 5V, and it will be inside the enclosure where LEDs won't be seen, then I believe I just need a couple of something like this DPDT relay.

So I envisage a tiny little PCB, with nothing more than two of those relays and three 4-way pin headers. I'll put the data lines through one relay and the power lines through the other, and I'll match the wire lengths. I may try to test this out this weekend.
 
Good find, thank you. So switching between computers is an acceptable thing to do. But that particular device requires a button press, and I plan to put mine inside the enclosure where it will do it's job automatically.

Keep searching then :)
That was first result from a search for "USB switch"

USB was never designed to be switched by relays. USB was never designed for the individual signal wires to be split by centimetres.

If that little Amazon box exists, then if a remote control variant doesn't already exist, then there will be a USB switching IC somewhere which allows the USB signals to be switched correctly whilst maintaining full USB spec