DIY Purifi Amp builds

[March Audio]

So that the jumper would be on pins 2-3.

I am beginning to think the SMPS is a DOA. It neither light up the LED for the front faceplate from Ghent which I connected to J5 with the Ghent supplied connector, nor the leds on the Purifi. Would it fail to illuminate both sets of those LEDs even if I had misconnected the Vaux jumpers? Even when I entirely disconnect the Purifi harnasses, I get no power to the Ghent LED even though I can hear the SMPS clicking when I turn on the IEC power. Likewise, when I disconnect Ghent's harnass at J5, and leave the J4 and J1 harnesses connected to the Eval 1, the LEDs on the Purifi board do not light.

It seems to me the SMPS should be able to power up something if it was functional even if the jumpers were incorrectly set. And yes, I do have the power jumper to convert to US 115v in place.

I've also checked for continuity on the harnass to J1, and it's perfect. I've checked for continuity from the IEC plug and it's perfect for ground to the chassis screw, hot lead which is from the IEC plug prong to the red ternmination of the Ghent wiring harnass, likewise for neutral. And there is only one was to plug in the harnass, so I can't have neutral and hot reversed. Looking toward the front of the chassis (and toward the head shield) the hot lead is on the right side,and the neutral is on the left. And pins 2-3 on J6-J7 are the two leftmost pins.

OK you need to break this down into simple sections. First test, check the PSU.

Completely disconnect the PSU from the Purifi board and the Ghent connector from J5.
Apply mains power to the PSU. It should default to on. You should here the relay click.
Measure the output volatges are present and correct on J1

 

[TimoJ]

So that the jumper would be on pins 2-3.

I am beginning to think the SMPS is a DOA. It neither light up the LED for the front faceplate from Ghent which I connected to J5 with the Ghent supplied connector, nor the leds on the Purifi. Would it fail to illuminate both sets of those LEDs even if I had misconnected the Vaux jumpers? Even when I entirely disconnect the Purifi harnasses, I get no power to the Ghent LED even though I can hear the SMPS clicking when I turn on the IEC power. Likewise, when I disconnect Ghent's harnass at J5, and leave the J4 and J1 harnesses connected to the Eval 1, the LEDs on the Purifi board do not light.

It seems to me the SMPS should be able to power up something if it was functional even if the jumpers were incorrectly set. And yes, I do have the power jumper to convert to US 115v in place.

I've also checked for continuity on the harnass to J1, and it's perfect. I've checked for continuity from the IEC plug and it's perfect for ground to the chassis screw, hot lead which is from the IEC plug prong to the red ternmination of the Ghent wiring harnass, likewise for neutral. And there is only one was to plug in the harnass, so I can't have neutral and hot reversed. Looking toward the front of the chassis (and toward the head shield) the hot lead is on the right side,and the neutral is on the left. And pins 2-3 on J6-J7 are the two leftmost pins.

Post a picture of PSU connector and EVAL1 connector wirings. How did you remove the switch from the EVAL1 board? Are you sure you didn't damage some SMD parts around it?

 

[tomchr]

My final build. Added guardian-86 speaker protection boards to this build from Neurochrome.

Fantastic! Did you end up taking the power supply for the Guardian-86 from the main ±65 V supply and changing D4 on the Guardian-86 to accommodate the higher voltage, or did you find a different solution?

Tom

 

[starfly]

Fantastic! Did you end up taking the power supply for the Guardian-86 from the main ±65 V supply and changing D4 on the Guardian-86 to accommodate the higher voltage, or did you find a different solution?

Tom

Yeah, I'm tapping power from J6.2 on your board (J1.2 on the Hypex), so the +/- 65V line that comes from the Hypex SMPS. And also replaced the D4 diode to what you recommended, and it works great. No more pops on startup. Also tested with my multimeter on R1 and it takes a few seconds after power-on for it to display ~2.3V.

You can see in the image two red wires coming from the black wire harness on the left input buffer. Not the prettiest solution, in hindsight I would have spliced it a bit differently, but it's soldered and a solid connection, so it works

Thanks again for all the help!

 

[tomchr]

Excellent. I'm glad that worked out for you. I guess I should wait a little before I shrink the heat shrink on the power harnesses in my Purifi build then.

Tom

 

[starfly]

Excellent. I'm glad that worked out for you. I guess I should wait a little before I shrink the heat shrink on the power harnesses in my Purifi build then.

Tom

Ah, you're also doing a Purifi build. Please post pics when it's done.

I will eventually add a 3rd channel to this build, at which point I'll redo some of the wiring to clean it up a bit. But I'll need to buy a center speaker first

 

[tomchr]

Ah, you're also doing a Purifi build. Please post pics when it's done.

Will do.

Yeah. I was getting quite a few questions around how to use my Purifi/Hypex Input Buffer to make a Purifi 1ET400A amp. It seems that with this buffer I've started tapping into a segment of DIYers who have good mechanical skills but who don't have as much electrical experience, thus, writing an article on how to put the puzzle together seems like a good idea. It'll also make it possible for me to answer the, "but what does it sound like?" questions. And, who knows... Maybe it'll be like my Modulus-686 prototype build where I built four amps, mostly for my own use, and started getting requests for fully assembled amps. I just sold the last one. Mind you, this was done with no advertising or even a "hey, I have this" on my website.

Tom

 

[kukocz]

Currently I’m using Eval1 kit based amp with DX3 Pro as a preamp.
I noticed that when I bypass input buffers, sound is somehow more flat, less dynamic and bass is less powerful.
As the opposite the stereo imaging tends to be better when buffers are omitted but this is probably caused by decreased amount of bass.
All this is very subjective and even I’m biased it should be opposite feeling as purify without buffers has better SINAD...
Output impedance of DX3 Pro is 100 ohm on RCAs (as far as I know). Eval1 with buffers has 10kOhm input impedance (balanced/unbalanced) and 4.4/2.2 kOhm without buffers respectively.
What am I missing? Impedance matching ratio is 20x in the worst case so there shouldn’t be any difference at all.

 

[tomchr]

You don't have to worry about impedance matching at audio frequencies. In fact, you do not want impedance matching as that would mean the amp loads the preamp/source to the point where it only produces half its no-load output voltage.

Tom

 

[kukocz]

You don't have to worry about impedance matching at audio frequencies. In fact, you do not want impedance matching as that would mean the amp loads the preamp/source to the point where it only produces half its no-load output voltage.

Tom

Agree. Indeed it’s not RF.
By saying “matching ratio” I meant that amp impedance is high enough to not cause any difficulties to be driven by DAC.
In the nearest future I’m going to pair it with Soncoz LA-QXD1 then I’ll see if pure XLR makes a difference in that matter.

 

[kukocz]

Hi,
For those who have newer batch of EVAL1 board and also have issues to get READY signal from J3.2.
Kim from Purify helped me to identify the root cause. It turned out that black smd elements with white strip, I mentioned in my previous post, are diodes (R6_A1 and R6_A2).
They were wrongly placed and they have to be rotated by 180 degrees.
It looks like newer batch will have corrected design.

BTW. An updated revision of documentation should be available soon (v1.24).

 

[jae]

Is anyone using any passive preamp/attenuator in their projects? Any suggestions?

 

[Casey Leedom]

Hi,
For those who have newer batch of EVAL1 board and also have issues to get READY signal from J3.2.
Kim from Purify helped me to identify the root cause. It turned out that black smd elements with white strip, I mentioned in my previous post, are diodes (R6_A1 and R6_A2).
They were wrongly placed and they have to be rotated by 180 degrees.
It looks like newer batch will have corrected design.
...

Is this on the 1ET400A Amplifier module or the Buffer Board? I'ver order an NAD M28 Amplifier which is I presume made up of 7 1ET400A Amplifier Modules, their own Buffer Modules, and some moderately large Switching Power supply ala the Hypex SMPS3KA400 or similar.

Casey

 

[kukocz]

Is this on the 1ET400A Amplifier module or the Buffer Board? I'ver order an NAD M28 Amplifier which is I presume made up of 7 1ET400A Amplifier Modules, their own Buffer Modules, and some moderately large Switching Power supply ala the Hypex SMPS3KA400 or similar.

Casey

This is about EVAL1 board which is acting as buffers stage.

 

[phoenixdogfan]

Is the power supply outputting the right voltages from the various pins?

I tested the output voltage off the power supply from J1 by attaching my wiring harnass to the SMPS, then I inserted the probes of the multimeter, and turning on line level power to the SMPS after disconnecting the Eval 1 and the little LED Ghent supplied to connect to J5.


Voltages were as follows

For J1.1 (Red probe) and J1.4 VDR (black probe)-- .8 volt constant voltage.

For J1.2 (Red probe) and J1.5 HV (black probe)---starts off at 115v and begins dropping until it reaches .8 v.

For J.1 (Red) to J1.3 (black) VDR to Ground -- .8v constant

For J1.2 (red probe) to J1. 6 (black) HV to Ground-- Starts at 55 v and drops until it reaches .4 volt.

Is this normal behavior for a fully functional SMPS? Was my test procedure correct? What should correct voltages be off the HV and VDR outputs on J1?


I have also checked end to end continuity for the wiring harnasses connecting J1 on the Purifi to J1 on the SMPS and the aux power going from J3 on the Purifi to J4 on the SMPS. I used the 3M plug Jim recommended, and it does have continuity for all 5 wires, and I also rechecked the locations where each wire was plugged to be certain it was correct on both of the wiring harnasses. Also checked that the aux jumper were set to pins 2 and 3, and my 115 v jumper at j3 is also in place, and my wiring of the IEC plug receptacle on the Ghent case was double checked against his diagram on the website.

 

[phoenixdogfan]

A picture of the Vaux jumper setting at J6 and J7.

As can be seen, they are both set for pins 2-3.

 

[phoenixdogfan]

And also a photo of the mains connection at J2 and the 115 v jumper in place at J3.

Also, as I've already pointed out the Ghent LED connected at J5 does not light even when the Eval 1 harnasses are disconnected. And the Leds on the Eval 1 doe not light, and of course no sound when music is played.

The one thing I do see is the back of the Eval 1 is very close if not touching the back panel, which it of course needs to be for the Xlr and bannana receptacles to protrude through the back plate. Could anything on the Eval 1 cause a short by virtue of being so close to the back plate?

I also applied thermal paste to the bottom of the SMPS. That could not possibly be causing any kind of failure, could it?

 

[phoenixdogfan]

OK you need to break this down into simple sections. First test, check the PSU.

Completely disconnect the PSU from the Purifi board and the Ghent connector from J5.
Apply mains power to the PSU. It should default to on. You should here the relay click.
Measure the output volatges are present and correct on J1

I measured the voltages with everything disconnected (the Purifi power and aux cables, and the Ghent cable out of J5). The procedure used and results are as follows:

I tested the output voltage coming off the power supply from J1 by attaching one end of my wiring harnass to the SMPS, then I inserted the probes of the multimeter into the opening in the housing at the other end which would normally plug into J1 on the Purifi. The I turned on line level power to the SMPS (with both the Eval 1 and the little LED Ghent supplied LED at J5 fuly disconnected).


Voltages were as follows

For J1.1 (Red probe) and J1.4 VDR (black probe)-- .8 volt constant voltage.

For J1.2 (Red probe) and J1.5 HV (black probe)---starts off at 115v and begins dropping until it reaches .8 v.

For J.1 (Red) to J1.3 (black) VDR to Ground -- .8v constant

For J1.2 (red probe) to J1. 6 (black) HV to Ground-- Starts at 55 v and drops until it reaches .4 volt.

Is this normal behavior for a fully functional SMPS? Was my test procedure correct? What should correct voltages be off the HV and VDR outputs on J1?

 

[March Audio]

I measured the voltages with everything disconnected (the Purifi power and aux cables, and the Ghent cable out of J5). The procedure used and results are as follows:

I tested the output voltage coming off the power supply from J1 by attaching one end of my wiring harnass to the SMPS, then I inserted the probes of the multimeter into the opening in the housing at the other end which would normally plug into J1 on the Purifi. The I turned on line level power to the SMPS (with both the Eval 1 and the little LED Ghent supplied LED at J5 fuly disconnected).


Voltages were as follows

For J1.1 (Red probe) and J1.4 VDR (black probe)-- .8 volt constant voltage.

For J1.2 (Red probe) and J1.5 HV (black probe)---starts off at 115v and begins dropping until it reaches .8 v.

For J.1 (Red) to J1.3 (black) VDR to Ground -- .8v constant

For J1.2 (red probe) to J1. 6 (black) HV to Ground-- Starts at 55 v and drops until it reaches .4 volt.

Is this normal behavior for a fully functional SMPS? Was my test procedure correct? What should correct voltages be off the HV and VDR outputs on J1?

It isnt normal but lets make sure. I will try and make a video tomorrow to show exactly what to do and where to measure.

 

[JimB]

Regarding Ghent cases, a couple weeks ago, I heard they had sent design plans out to some major customers for comment, and that they received some feedback and will now be moving ahead to the next stage. I hope we hear more, soon.