ProtoDAC TDA1387 X8 project

[hifinet]

I should mention that the stacking header with pins 2 and 4 removed provides a reversible solution for the Ian Canada reclockers.

For an irreversible solution, simply remove pins 2 and 4 from the ProtoDAC GPIO header (before soldering). That way you always need to apply power to the 5V and GND pads of ProtoDAC. I think the green screw terminals can be soldered there.

[martin155]

The JustBoom Amp (last one from PiShop.us) is also a traditional amplifier that takes balanced or single ended analog inputs, and should work with ProtoDAC.

I would like to make all in one (ProtoDac + amp) solution. Is there somebody using JustBoom amp? How does it compare with other available amps? 

What power supply do you use to power it, do I need clean PS?

[varunach]

I should mention that the stacking header with pins 2 and 4 removed provides a reversible solution for the Ian Canada reclockers.

For an irreversible solution, simply remove pins 2 and 4 from the ProtoDAC GPIO header (before soldering). That way you always need to apply power to the 5V and GND pads of ProtoDAC. I think the green screw terminals can be soldered there.

Hi Mark,

I just built the protodac, hooked it to RPi4 and its glorious. Now I want to incorporate IanCanada's PurePi, Q7 to it. Will this connection schema work?

1. Stack RPi on PurePi. Power the PurePi with a 5v power supply (RPi is therefore powered by PurePi pins underneath and doesn't have its own external power supply)
2. Q7 stacked on RPi GPIO. Q7 is powered from the 3.3V supply of PurePi (J2) using standard jumper wires. 
3. ProtoDAC stacked on the isolated GPIO of Q7 (using a stacking header with pins 2 and 4 removed). Use jumper wires from PurePi 5V terminals (J6) to ProtoDAC 5V and GND pads.

When you say remove stacking header pins 2 and 4, should I just clip it/break it off?

I'm ok with RPI4 and ProtoDac sharing the PurePi 5V power. I'm hoping thats not going to be an issue. 

[hifinet]

Yes, you could share the 5V ultracap supply between the RPi and ProtoDAC. The Q7 goes to extreme lengths to isolate the DAC from the RPi. With an RPi4, there are four screaming cores, and you will be injecting that noise into ProtoDAC, around the Q7 isolation. Consider using the FifoPi MA (no isolation) and a RPi Zero W (single core).

Yes, just clip the leads.

[varunach]

Yes, you could share the 5V ultracap supply between the RPi and ProtoDAC. The Q7 goes to extreme lengths to isolate the DAC from the RPi. With an RPi4, there are four screaming cores, and you will be injecting that noise into ProtoDAC, around the Q7 isolation. Consider using the FifoPi MA (no isolation) and a RPi Zero W (single core).

Yes, just clip the leads.

@hifinet got it. So if I isolate the RPi4 with another GPIO stacking header then I wont have this noise issue. This requires another dedicated power supply for the RPi4. A couple more follow up questions -

1. When you say 'noise' is it the power/EMI leakage noise you are referring to (that the Q7 would have isolated if feeding the ProtoDac)? Because the signal path is still Pi > Q7 > ProtoDac.
2. If I use the FifoPi MA I'll still have the same issue, so not sure I understand the advantage of using FifoPiMa. Unless you are saying I anyways wont use the 'isolation' feature of Q7 so might as well go with FifoPiMa. You had suggested that the entire RPi/FifoPi MA/ProtoDAC stack could be simply powered by shorting J4 1-2 and 3-4, and applying 5V power to either the RPi USB or the FifoPiMA J3. But this is still not clean side power, will this be a problem?

[hifinet]

You only need the one stacking header between the Q7 and ProtoDAC to prevent 5V from ProtoDAC feeding back to the 3.3V of the Q7 GPIO. Q7 isolates the topside isolated GPIO (J7) from the RPi. That is one of the things it does.

The noise is primarily conducted noise that the RPi is imparting on all of the GPIO lines, power, grounds and I2S, the entire GPIO. Q7 isolates J7 from all this. EMI is much less, orders of magnitude less, IMO.

Yes, I am saying that if you are going to circumvent the isolation of Q7, you might as well save yourself some money and go with MA. MA also has two jumpers so that you don't need to use the stacking header. See the previous posts of @michaelagiles . The solution in that case is to use an RPi that emits less noise. The RPi Zero W would have the least of the current RPi in that regard, since it is single core. In my experience, Zero W is less noisy than RPi 2, which is less noisy than RPi 3.

[varunach]

You only need the one stacking header between the Q7 and ProtoDAC to prevent 5V from ProtoDAC feeding back to the 3.3V of the Q7 GPIO. Q7 isolates the topside isolated GPIO (J7) from the RPi. That is one of the things it does.

The noise is primarily conducted noise that the RPi is imparting on all of the GPIO lines, power, grounds and I2S, the entire GPIO. Q7 isolates J7 from all this. EMI is much less, orders of magnitude less, IMO.

Yes, I am saying that if you are going to circumvent the isolation of Q7, you might as well save yourself some money and go with MA. MA also has two jumpers so that you don't need to use the stacking header. See the previous posts of @michaelagiles . The solution in that case is to use an RPi that emits less noise. The RPi Zero W would have the least of the current RPi in that regard, since it is single core. In my experience, Zero W is less noisy than RPi 2, which is less noisy than RPi 3.

@hifinet  So when you isolate the ProtoDAC from the GPIO (via stacking header with pins 2 & 4 disabled) you are only cutting off the power part of that chain, right? In other words the other isolation benefits of Q7 for the I2S/audio signal should be intact? Or is the 'isolation' benefit of a Q7 entirely hinges on the power flowing through it?

[wolf7250]

You only need the one stacking header between the Q7 and ProtoDAC to prevent 5V from ProtoDAC feeding back to the 3.3V of the Q7 GPIO. Q7 isolates the topside isolated GPIO (J7) from the RPi. That is one of the things it does.

The noise is primarily conducted noise that the RPi is imparting on all of the GPIO lines, power, grounds and I2S, the entire GPIO. Q7 isolates J7 from all this. EMI is much less, orders of magnitude less, IMO.

Yes, I am saying that if you are going to circumvent the isolation of Q7, you might as well save yourself some money and go with MA. MA also has two jumpers so that you don't need to use the stacking header. See the previous posts of @michaelagiles . The solution in that case is to use an RPi that emits less noise. The RPi Zero W would have the least of the current RPi in that regard, since it is single core. In my experience, Zero W is less noisy than RPi 2, which is less noisy than RPi 3.

@hifinet I recently ordered one of your kits on EBAY and was intending to use it with the FiFoPi MA and a Raspberry Pi 3B+. Would you have any advice as to the optimum way to power that setup to reduce noise? I was thinking of powering the entire stack using the 5V J3 jumper on the MA and then shorting J4 1-2 3-4 (not exactly sure how to do that yet...).

But reading through the last few pages (and your post above) on this thread this will pass through RPi noise to the ProtoDAC correct?
If I'm wanting to use the MA, do you suggest I use a Zero W to minimise noise? I see from your signature that you use a Zero W. Do you find that WiFi interferes with/creates noise at all?

[varunach]

You only need the one stacking header between the Q7 and ProtoDAC to prevent 5V from ProtoDAC feeding back to the 3.3V of the Q7 GPIO. Q7 isolates the topside isolated GPIO (J7) from the RPi. That is one of the things it does.

The noise is primarily conducted noise that the RPi is imparting on all of the GPIO lines, power, grounds and I2S, the entire GPIO. Q7 isolates J7 from all this. EMI is much less, orders of magnitude less, IMO.

Yes, I am saying that if you are going to circumvent the isolation of Q7, you might as well save yourself some money and go with MA. MA also has two jumpers so that you don't need to use the stacking header. See the previous posts of @michaelagiles . The solution in that case is to use an RPi that emits less noise. The RPi Zero W would have the least of the current RPi in that regard, since it is single core. In my experience, Zero W is less noisy than RPi 2, which is less noisy than RPi 3.

@hifinet I recently ordered one of your kits on EBAY and was intending to use it with the FiFoPi MA and a Raspberry Pi 3B+. Would you have any advice as to the optimum way to power that setup to reduce noise? I was thinking of powering the entire stack using the 5V J3 jumper on the MA and then shorting J4 1-2 3-4 (not exactly sure how to do that yet...).

But reading through the last few pages (and your post above) on this thread this will pass through RPi noise to the ProtoDAC correct?
If I'm wanting to use the MA, do you suggest I use a Zero W to minimise noise? I see from your signature that you use a Zero W. Do you find that WiFi interferes with/creates noise at all?

@wolf7250 yes thats what @hifinet is suggesting, to use ZeroW. Otherwise you'll have to have PurePi, lift the RPi3 a bit higher so that it doesnt use the ultracapacitors to power it (and use a separate 5V external power supply). On the shorting thing, you should get 3 small square black caps that you insert on top of the pins 1-2, 3-4 to short. Its shown in the FifiMa manual.

[hifinet]

Yes, the other benefits of the Q7 are intact. It is best to provide ProtoDAC with a clean 5V. So you can take the 5V output from the terminal of the PurePi. That will be "clean" as long as you don't contaminate it with noise from the RPi.