Brainstorming Purifi 1et400a amps

Well, in all likelihood I wouldn't, but when mentally considering a Purifi build I would go with the Neurochrome buffers. So, it's mostly understanding whether I sacrifice any potential flexibility going with the prefab route.

Connectivity wise the Neurochrome boards can use the same connections, but guessing the pcb layout would need more work.

You might want to drop an email to Tom at Neurochrome to see if there has been enough interest to produce a kit or pre-assembled version via his TCA brand.
 
You will need a different ribbon cable. Tom can tell you the parts to order. It's a simple build. There will be no matching holes for the standoffs but the connections are tight. Some putty would suffice. The standoffs do not carry any current. Tom's board does not have a voltage regulator unless you ask for one. You will have to set the SMPS to do the regulation. Those NC500 boards do have regulators so the SMPS is set to send unregulated when you get it. The input wire from the xlr's will have to be replaced. A plug is used on the stock input board. Bare wire on Tom's board.
 
I was doing some testing here with my Connex SMPS 800re and Purifi modules and found a problem. While testing with a "brown noise" file I use for break in/testing, the amp modules started intermittantly cutting out, with the indicator LEDs also going out. It appeared that the amp protection was kicking in. I measured the voltage rails while playing (nominally +/- 65 VDC) and saw the voltage jumping all over the place, including to well over 70 VDC.
I contacted Connex and asked them about it, and they suggested it must be due to power supply pumping, and suggested adding 4700 µF more capacitance to each voltage rail, and then to test again. I have not done that yet...

I would suggest that anyone using the Connex supplies with the Purifi modules, take the time to monitor the main power supply rails while driving the amps at a high level to be sure things are OK!

I am going to order a Hypex supply, and test the Connex SMPS 800re with additional capacitance as well. But even with the Hypex supply, I will continue to use the custom linear I built for the +/- 18 VDC supplies, as it is working perfectly.

And, thank you Purifi for having good protection features built into your modules!

@barrows,

What was your final solution to resolve the power supply pumping with the Connex power supply (more capacitance, inverting I/O on one channel, different PS, something else) ?

TIA
 
Add extra capacitors with 15000uF or more per rail and you are good to go.
 

Attachments

  • 7AFBEA16-7E23-415A-940A-1003B2F40BB1.jpeg
    7AFBEA16-7E23-415A-940A-1003B2F40BB1.jpeg
    351.8 KB · Views: 549
well....

@barrows,

What was your final solution to resolve the power supply pumping with the Connex power supply (more capacitance, inverting I/O on one channel, different PS, something else) ?

TIA

I tried adding the extra capacitance as suggested by Connex, but that did not solve the problem. I ended up using a Hypex SMPS 1200 instead which has no issues. I still end up using my linear +/- 17VDC supply for the input stage and modulator, the amp sounds fantastic. Next experiment is to try some Sonic Imagery discrete opamps in place of the eval board's OPA 1612...

Inverting one channel was not a good option for me, as I run a sub from the speaker outputs.
 
input LPF question anyone?

Hi folks, I am considering adding a simple LPF to the input of my Purifi build (using the Eval board). This would be just to roll of the very low bass in order to save the amp from needing to deliver really high currents, and to keep my speaker mid-woofers from hitting their stops with occasional very low frequency pulses (mostly can happen at high volumes with some electronic music, big timpani thwacks, etc, I use a sub for low bass).
So the plan would be to put 4 film caps in series with the inputs. I am just not sure what R value to use for the calculation. The Eval module has a differential input impedance of 10.2 Kohm, would i use then 5.1 Kohm to make my calculation for the -3 dB point? Or, should I measure the resistance to ground of each input (pins 2 and 3) and use that number for the calculation?
I am only looking for a -3 dB point of around 40 Hz.
 
Spectron did something like this with their "direct sense" speaker cables for their class d Musician amplifier some 20 years ago- the speaker wire contained extra conductors that extended the feedback loop to the speakers themselves.

I don't know if there is some sort of cable length constraint requiring the amp to be withing N units close proximity to the speakers.

Some servo plate amps restrict the cable length stating feedback timing delay issues.
 
Last edited: