KGSSHV Carbon Build Thread

[Pars]

On 12/11/2020 at 12:57 AM, Rodeodave said:

If I recall correctly, they are Molex Eurostyle Terminal Blocks UL certified for 300VAC, which would mean 425V DC peak. I'm not sure if there are terminal blocks rated for 600V with a 5.08mm pitch.

 

Since I'm using the amplifier boards with a ground plane, i used a couple of layers of Kapton tape at the bottom of the terminals and used very little solder in order to avoid the arching that's been reported with these boards.

 

Edit: I just realized that the terminal blocks aren't the Molex that I thought I had installed. I guess I should replace them, which is not going to be easy 😕

Mouser has some phoenix 5.08mm pitch terminals rated for 400V.

What would be a good alternative to the screw terminals? I'm not their biggest fan either.

I had identified Hirose and a TE/AMP parts that were high voltage and used keyed connectors. I haven't installed these yet (and haven't tested the amp or PSUs), but one of these would definitely fit the board. The other seems too big. I'll recheck and let you know which one fits.

571-1-178128-5 TE/AMP 798-DF5A-5S-5C  Hirose

 

 

[Rodeodave]

Goop:

Bolted down:

Fits like a glove:

Good clearance:

And the full monty:

 

[Rodeodave]

Progress is slow during the week, but thanks to home office I can get some soldering done over the lunch break.

For connecting the HV and LV PSUs and the amp boards I'm going with a star ground, held in place by wire wrap and soaked with solder:

Here's how the manifold looks like in circuit:

There's good clearance too:

Although the currents are substantially lower than what I'm used to from solid state, I took care to have equal lengths of fat wire for each of the HV and each of the LV supplies. I hope the grounding scheme  is going to be quiet, I passionately hate ground related hum.

[tkam]

2 hours ago, Rodeodave said:

I passionately hate ground related hum.

I second this passion.

[Rodeodave]

HV and LV wiring is now in place:

Next up is input and output wiring.

[hhobeika]

6 minutes ago, Rodeodave said:

HV and LV wiring is now in place:

Next up is input and output wiring.

Very nicely and cleanly done!  Looking forward to seeing more progress.

[Rodeodave]

Thanks for the encouragement, y'all.

Getting closer now:

[Rodeodave]

Input and output wiring in place:

And I'm happy to report that the amp is working! No smoke, no arching, no shrapnel.

After like an hour of warming up and fiddling with offset and balance, I gave it a listen with my 007mk1. Wow.

I'm treating pin 1 as analogue ground btw, and circuit ground goes to the chassis earth ground point via a CL60 thermistor. Zero noise, zero hum - with my gear at least.

Bias is still around 17mA for all channels, I'll get some more accurate numbers when I set up the optocoupler servo. Not sure if I'll go up to 20mA, this amp is a furnace at 17mA already.

 

[hhobeika]

2 hours ago, Rodeodave said:

Input and output wiring in place:

And I'm happy to report that the amp is working! No smoke, no arching, no shrapnel.

After like an hour of warming up and fiddling with offset and balance, I gave it a listen with my 007mk1. Wow.

I'm treating pin 1 as analogue ground btw, and circuit ground goes to the chassis earth ground point via a CL60 thermistor. Zero noise, zero hum - with my gear at least.

Bias is still around 17mA for all channels, I'll get some more accurate numbers when I set up the optocoupler servo. Not sure if I'll go up to 20mA, this amp is a furnace at 17mA already.

 

Excellently done and very methodical.  It was a pleasure to follow your building steps.  Congratulations!

[Blueman2]

6 hours ago, Rodeodave said:

 Not sure if I'll go up to 20mA, this amp is a furnace at 17mA already.

Yes, this baby will keep you warm in the winter.  I ended up doing the same, starting with 17mA and then increasing to 20mA after it was running for a couple days.  Not sure it really made much difference in either sound or heat production, but just wanted to push the envelope. 

Edited December 19, 2020 by Blueman2

[gepardcv]

Interesting, my 20mA build doesn’t get anywhere hot enough to be described as “a furnace”, and I used smaller heatsinks. OTOH, I also used smaller mounting brackets, so the hot transistors sit closer to the sink fins, which should help somewhat. My on-board heatsink KGSSHV ran hotter than the Carbon, and my Dynahi ran hotter still.

[dingding123]

Congrats on your finished build!

23 hours ago, Rodeodave said:

Not sure if I'll go up to 20mA, this amp is a furnace at 17mA already.

I did a similar (in terms of size and layout) single chassis build, with the amp mounted in the upper part of the heatsink. Dual 450V and 20mA. At steady state, the temps hover around 46 degrees celsius (ambient 27 deg) on the heatsinks directly behind the amp boards, 43 deg behind the GRHV, and 41 deg at other parts of the heatsink where nothing is mounted behind. 

Not sure about you but I feel it's quite safe to push it to 20mA if you're getting a similar temperature.

Edited December 20, 2020 by dingding123
corrected image links

[Rodeodave]

I have now pulled some numbers from the circuit. And I took more pretty pics, which I‘ll start off with:

 

Now for some numbers:

PSU voltages:

V_HV_pos=402V, V_HV_neg=-403V

V_LV_pos=14.88V, V_LV_neg=14.88V

 

Current sources:

Right channel voltage drop over the RN60C 49R9:

V_ccs_pos=857mV, I_ccs_pos=17.17mA

V_ccs_neg=856mV, I_ccs_pos=17.15mA

 

Left channel voltage drop over the RN60C 49R9:

V_ccs_pos=862mV, I_ccs_pos=17.28mA (adjusted to 857mV/17.17mA)

V_ccs_neg=856mV, I_ccs_pos=17.15mA

 

On the left side the initial B+=20V current source adjustment was the following:

V_ccs_pos=845mV, I_ccs_pos=16.93mA

V_ccs_neg=847mV, I_ccs_pos=16.97mA

So going from a B+ of 20V to full a 402V led to an increase in current of a meager 1%.

 

Next up is gain, THD and clipping. Input is a single ended 0.1V at 1kHz from the Boonton 1121. The – input of the carbon is grounded through a 220R resistor (equivalent the output impedance of my preamp). The output is measured unloaded.

L+: V_out_rms=42.5V, gain=52.56dB, THD=0.3%

L-: V_out_rms=46.0V, gain=53.26dB, THD=0.095%

R+: V_out_rms=41.9V, gain=52.54dB, THD=0.38%

R-: V_out_rms=46.6V, gain=53.37dB, THD=0.09%

The THD is higher than I would have expected (the Boonton‘s baseline is at 0.00085% btw). It‘s mostly 2nd harmonic though.

 

Here’s how I took the gain and THD readings:

And here’s what the THD spectrum looks like (notched out fundamental at 1kHz, peaks at 2kHz and 3kHz, and maybe even 4kHz):

 

Here we see the output just before clipping:

CH2 in purple is the input signal to the amp, 1kHz at like 1.5Vpp (518mVrms):

CH1 in yellow is the output signal. 764Vpp. The probe is rated for 600Vpp at its 10x setting, so that was stretching it a bit…

And here we see clipping with an input signal of 1.9Vpp (668mVrms):

 

I also took temperature readings after the amp has been running for 2hrs or so. Ambient is at 24°C.

Right behind where the SiC Fets are mounted, the heatsink temperature is 45°C (113°F). The corner the farthest away from where the bracket connects to the heatsink is as 41°C, so the heatsink gets utilized reasonably well. Inside the case it read 43°C.

I realize that the brackets are not positioned ideally on the heatsink (too far up), which is why I have used chunky brackets, thinking that the thick material would conduct heat nicely. Seems to work well enough.

Seeing the 21°C rise in temp makes me think I could try upping the current. Calling it a furnace was overstating it a bit perhaps, but the whole enclosure gives off an impressive amount of heat already.

 

I also employed the opto servo btw, going up to 17.5V positive offset before putting in the jumper. The amp now comes on with like 1V offset, which goes down to a few mV after a couple of minutes. I don’t think I could hear any sonic difference between going raw vs. opto servo.

 

So, two questions remain:

Why is THD so high? I would have expected a lot less. Is it the frontend perhaps?

And what's up with the gain imbalance? It’s consistent between polarities (single ended input, - input grounded). Could that be the cause?

 

[simmconn]

The Boonton 1121 should allow you to do fully differential measurement up to 300V. Did you try that?

I measured my unit with SYS2722 and the second harmonic was the only one visible that sits comfortably below -100dB. Too bad I didn't save the test result and the unit is too heavy to move around for retest. The only complaint I had was that 1mV wide-band noise. Compared to the Stax amps this is very good already. I don't remember any single-ended test results. Perhaps I did single-ended input but never single-ended output since that would not make sense.

By the way my PSU is set at +/-450V, bias set at 17mA and all transistors curve-tracer matched where needed.

[Rodeodave]

With differential measurement - which better corresponds to what the membrane sees if I'm not mistaken - there should be some cancellation, right. Guess I'll be breaking out the analyzer over the holidays again, currently too busy listening when time permits

Edited December 22, 2020 by Rodeodave

[simmconn]

The Boonton should be measuring THD+n instead of purely THD given its vintage and price range. So if your one half has a noise problem it will also show up in the “Distortion” reading. Might be easier to figure out if your unit had the 400Hz high-pass filter.

[Pars]

On 12/13/2020 at 3:05 PM, Pars said:

I had identified Hirose and a TE/AMP parts that were high voltage and used keyed connectors. I haven't installed these yet (and haven't tested the amp or PSUs), but one of these would definitely fit the board. The other seems too big. I'll recheck and let you know which one fits.

571-1-178128-5 TE/AMP 798-DF5A-5S-5C  Hirose

 

 

It is the Hirose series which will fit, and are 500Vac rated (https://www.hirose.com/en/product/p/CL0676-0019-7-35).

Just to note, the Phoenix connectors shown on the BOM I had, as well as many of the BOMs in this thread (651-1729047) are only rated for 250V (data sheet claims 400V)? I haven't looked into the III/2 and II/2 spec to see what that means.

I thinking of replacing with these, which are 400V: 651-1800090, though it uses the same III/2 rating. These do say 400V nominal however.

I am using Kapton tape under these, and would guess that many have built theirs using the 651-1729047 parts?

[simmconn]

Google “Overvoltage category” and “Degree of pollution”

[Pars]

Thanks. Not quite sure how to apply that here though, as the datasheets are inconsistent. I suppose the 651-1729047 should be OK, though Mouser calls these a 250V part. The 651-1800090 is shown as a 400V part. I need to order the CPC1117N and resistors for the soft start anyway (universal PSU; not for the Carbon), so I guess I'll get the 651-1800090 and use them.

[simmconn]

On 2/2/2021 at 12:04 PM, simmconn said:

Google “Overvoltage category” and “Degree of pollution”

That's the III/2 and II/2 spec you haven't looked into. II/III is the Overvoltage category, and 2 is the Degree of pollution. Once you find their definitions you can match them with your application and determine the proper voltage rating that applies.

[Pars]

OK. Actually, I had looked at them but was having trouble discerning which OV cat was most applicable since they all looked to be more building power related.

– Overvoltage categoriesThe standard has divided the possible overvoltages into four categories. The three categories which relate to connectors are shortly described below:

Overvoltage category I  Equipment (e.g. connectors) intended for the use in applications or parts of installations in which no overvoltage can occur.Examples are low voltage equipments.

Overvoltage category II  Equipment (e.g. connectors) intended for the use in installations or parts of it, in which lightning overvoltages do not need to be considered, however switching overvoltages generated by the equipment. Examples are household appliances.

Overvoltage category III  Equipment (e.g. connectors) intended for the use in installations or parts of it in which lightning overvoltages do not need to be considered, however switching overvoltages generated by the equipment, and for cases where the reliability and the availability of the equipment (e.g. connectors) or its dependent circuits are subject to special requirements. Examples are protecting means, switches and sockets.

From this, it would seem that category II is the most applicable. Pollution degrees of 1 or 2 would also be applicable since the amp shouldn't have any condensation present, even briefly (degree 2) so no conductive pollution should be expected (briefly is allowed for degree 2). So the ii/2 spec would be applicable, but for safety, the iii/2 rating would be what I would look at.

 

[starcat]

Phoenix Contact indeed makes 400V rated terminal blocks where the usual green colored ones are only 250V rated. 

[Rodeodave]

According to their website, part nr. 5452258 has the following electrical parameters:

Nominal current 13.5 AN

Nom voltage 400 V

Rated voltage (III/3) 250 V

Rated voltage (III/2) 400 V

Rated voltage (II/2) 630 V

Rated surge voltage (III/3) 4 kV

Rated surge voltage (III/2) 4 kV

Rated surge voltage (II/2) 4 kV

Edit: It looks like part nr. 1729131 has the same properties listed.

 

Edited March 1, 2021 by Rodeodave

[JoaMat]

Datasheets are nice.

[Sebtdi]

Hello

I am looking for two spare Carbon amp boards from the last GB, either Version 5 or 6.

I already have the GRLV & HV boards

Please send me a private message 

Thanks