KGSSHV Carbon Build Thread

[audiostar]

18 hours ago, tonyxu said:

Looks like those relays on the input board are used to switch between chinch and XLR inputs. But that soldering job... 

Edited October 21, 2022 by audiostar

[gepardcv]

@tonyxu: Maybe try emailing Bottlehead? It’s local to you. Although it has in-house expertise to do the work, I doubt anyone there will take on this project. That said, the company has been around the Seattle area for many years and might share some contacts. 
 

Meanwhile, for safety, please unplug the amp and close the case. The voltages on some of the wires and traces inside are very dangerous (and who knows if that build is safely grounded at all).

[tonyxu]

40 minutes ago, gepardcv said:

@tonyxu: Maybe try emailing Bottlehead? It’s local to you. Although it has in-house expertise to do the work, I doubt anyone there will take on this project. That said, the company has been around the Seattle area for many years and might share some contacts. 
 

Meanwhile, for safety, please unplug the amp and close the case. The voltages on some of the wires and traces inside are very dangerous (and who knows if that build is safely grounded at all).

Thanks for the advice! I actually have emailed them and Paul there passed on this as he thinks this one could be hard to debug and could easily cause more damage. 

I have unplugged it and closed the case. I'll see if someone would be willing to work on this, or otherwise it'll be very big loss on me unfortunately. 

[Pars]

Wire gauge: I'm wiring up the PSUs to the Carbon boards, but wanted to get an opinion on AWG to use. I had bought a bunch of 22ga teflon jacketed 600V wire from ebay seller skip malley. Since I don't think the HV supplies are sourcing much current, I think this should be fine but want to get any input if you think it is too light of gauge? Usually I've used 18 or 20 for other amps.

Thanks.

[ang728]

6 minutes ago, Pars said:

Wire gauge: I'm wiring up the PSUs to the Carbon boards, but wanted to get an opinion on AWG to use. I had bought a bunch of 22ga teflon jacketed 600V wire from ebay seller skip malley. Since I don't think the HV supplies are sourcing much current, I think this should be fine but want to get any input if you think it is too light of gauge? Usually I've used 18 or 20 for other amps.

Thanks.

I used 22AWG before,they worked fine.

[Kerry]

22awg is good. 

Good luck  

[Pars]

On 11/29/2015 at 5:15 PM, mwl168 said:

More progress today testing my HV Carbon build (V5 board from the GB). I am running 20mA (1.00VDC measured at test point) and have substituted the 182R tail resistor of the offset VR with a 100R one. No problem zero-out the offset and balance.

A few observations at start up (I have all VRs adjusted to mid point of its value before soldering on the board, I also hand measured and picked all passive parts including the LEDs except a few tiny capacitors):

1. the test points measured between 0.71V to 0.76V

2. the balance is around 3.8V

3. the offset is around 38V

I also found the following sequence of procedures worked well for me for adjustment:

1. Adjust the two 100R VR next to the test point to read 1VDC - my desired 20mA output current

2. Adjust the 2K balance VR to as close to 0V as possible (it'll drift a bit)

3. adjust 100R offset VR to as close as 0v as possible (it'll drift a bit and also interact with balance adjustment)

Next will be to wire up the input and output connectors and see if I'll get music. By the way, this thing runs HOT, much hotter than my KGSSHV. My test heatsink got very warm to the touch. 

Here are some photos: 

 

 

 

A couple of questions:

1. For the balance adjustment, should you measure from O+ to O-? Or is this O+ to GND? I'm guessing the former.
2. For the offset adjustment, I'm guessing O+ to GND (and also check O- to GND)?
3. Also, how critical is it to GND out the I+ and I- (I don't have a pot wired in yet, but could jumper these).

It isn't clear in Michael's pics where the blue wire is running to. I bought a 2nd meter (Brymen BM867) so could look at both the +/- offset at once. Thanks!

Edited February 25, 2023 by Pars

[kevin gilmore]

you need the inputs to ground or pot to zero.

balance is o+ to o-

offset is either to ground.

repeat a couple of times.

[Pars]

Fired the first board up after rechecking all connections, and wiring the inputs to ground. I'm getting full voltage (806 Vdc) from O+ to O-. Nothing exploded or seems to be getting hot, though I didn't leave it on very long. Both LEDs are lit evenly on the board. All pots were centered when I built it, and other than the current source pots, which I had adjusted using Soren's method to ~17mA, haven't been touched yet. I know I've seen this problem in the thread somewhere, off to find it.

[kevin gilmore]

this is what happens if the offset pot is out of range.  much easier if you have 2 voltmeters.

 

[Pars]

Thanks Kevin. I'll check it out tomorrow (having a beer, not conducive to HV stuff). On your production schematic (kgsshvcarbonproductionv5.pdf), the offset pot is RV4 and the balance pot is RV2 I take it? I'm pretty sure I have either a 120R or 100R ohm resistor in instead of the 182R.

And yes, my cheapie Harbor Freight expendable meter didn't like the HV (read 400V for the bias when testing the PSUs; Fluke disagreed at 580), so decided I needed another good meter.

Edited February 26, 2023 by Pars

[mwl168]

9 hours ago, Pars said:

Fired the first board up after rechecking all connections, and wiring the inputs to ground. I'm getting full voltage (806 Vdc) from O+ to O-. Nothing exploded or seems to be getting hot, though I didn't leave it on very long. Both LEDs are lit evenly on the board. All pots were centered when I built it, and other than the current source pots, which I had adjusted using Soren's method to ~17mA, haven't been touched yet. I know I've seen this problem in the thread somewhere, off to find it.

IIRC, I encountered a similar “issue” when I was adjusting my Blue Hawaii first time I powered it up. As Kevin already said, it turned out my offset pot was way off range. Once I got the offset adjusted to within reasonable range, the balance started to fall in place as well.

I need to find the schematic but sometimes it’s necessary to change the resistor value that’s in series with the offset pot to be able to adjust the offset to 0V.

Good luck!

[Pars]

Went down and started working on this again. The offset pot measured 54.5 ohms in circuit. No adjustment I made to this made any difference and stayed pegged on 404Vdc on the O+->GND. The only thing I've noticed is the right side current adjust TPs show 0.11V. The left side is 0.85V, which is the 17mA i had adjusted to using Soren's procedure.

Ohming out the larger devices shows no differences. When you say do a diode check on them, you mean use the diode function on the DMM? I also assume to check them both ways (flip the leads)?

From the schematic, the DN2540 or 10M90s could be the culprit on that side. I didn't try adjusting the trimpot, but know they were both set to 17mA. I have a couple of DN2540s, but don't think I have any more 10M90s.

I'm going to hook the other channel up and see how it behaves.

I verified that the tail resistor for the offset is 120R and not 182R.

Edited February 26, 2023 by Pars

[Pars]

Update: fired up the other channel and it works fine. Bias was right where I set it at 17mA. Offset and balance look good. Initial offset was around 32V, adjusted easily down to 0. Let it run a few minutes, it had drifted around 10V, adjusted back down (no servos yet). The heatsink (Modu 2U 300mm) was starting to get warm on that side. The one on the bad channel not getting warm.

I did attempt to adjust the bias on the bad channel right side, but the pot had no effect, so guessing a bad device somewhere. I'll probably start with the DN2540 as I have a couple of those. I do not have any more 10M90s. I of course have some C2M1000170D but they are back in stock at Mouser now, and probably not one of those which is bad. I'll attempt to use the good channel to troubleshoot the bad one.

[Pars]

I started troubleshooting this today (I'm working, so limited time). After disconnecting from heatsink/wiring, I decided to rerun Soren's CCS test using a bench supply. Both CCS's adjust fine on the bench, so I guess that tells me the 10M90s and DN2540 are fine?

I guess the SIC and the PZTA42 are next up. I'll look the board over for solder joints, etc, and run diode tests on those devices. I do have some SICs and I should have extra PZTAs.

I did have the opamp and the opto in on this board (no jumpers in). Hope this didn't destroy either of those. I don't think these (if dead) could be causing this issue, but will remove them for the time being.

[Pars]

Posting some board pics, hopefully of good enough quality. #1 is the bad board, #2 is the good one.

[Pars]

Spotted a solder bridge on 2 pins of the TO-71 socket. It was definitely shorted, now removed. I'll try to test it later. If that is it, that was easy

[kevin gilmore]

a short in that place is definitely going to cause trouble.

[Pars]

Put it back together and that was the problem. Offset and balance adjusted fine. I was a bit concerned when I first turned it on as it started to come up but blew the fuse (3.15A Slow blow). Replaced the fuse and it came up fine. No soft start or anything. Lights didn't dim or anything. Not sure if this is a problem or not. I'll see how it starts up in further starts to adjust and implement the opto servo. Bias is ~17mA; heatsinks don't seem to be warming much, but I didn't run it much (5 minutes). Might turn it up to 20mA bias, need to read the thread more (again) to see user comments on that.

Can't believe I hadn't noticed that solder bridge before, but glad I don't need to try to source any unobtainium parts right now

[n_maher]

Nice job @Pars - there are few things as satisfying as finding and fixing a problem that plagues you like that on a DIY build.  

[nopants]

I run mine at 20ma/400V and I think it gets a little strident in the highs- more noticeable on the non-007 models.

Simmconn did some measurements that may or may not back this up:

Depends on what power supply you're using, iirc 450V required either taller or stacked caps which is why I didn't go that route

I've been meaning to turn down the output current but I need to root around for the adjustment documentation/tips

Edited March 4, 2023 by nopants

[mwl168]

On 3/24/2022 at 7:08 PM, simmconn said:

 

On 3/22/2022 at 3:41 PM, simmconn said:

In a previous post I mentioned the less-than-ideal performance when Carbon is powered by a ±400V supply, and I suspected that the lower Vce on the PZTA42 is the culprit. Now it's been proven. The PZTA42 being a high voltage transistor, has a non-linear region at low Vce, as the slanted curves you can see on the upper left side. With 407V on the negative rail, the transistors on my board works at Vce=8.4V and Ic=20mA, right around the knee. The global negative feedback would have a hard time correcting that non-linearity. It also explains why some people prefer setting the Carbon at a lower current when powered with ±400V supply, as it also improves the linearity of the PZTA42, albeit to a lesser degree.

I guess Kevin chose the high voltage PZTA42 to deal with the power-on transients. I have a quick and dirty fix. Just bias the SiC MOSFET a little higher to give the PZTA42 more headroom. The SiC MOSFETs are biased by two 175k and an 20k at the gate. Reducing either 175k or increasing the 20k would do. The goal is to move the PZTA42 operating point to the right, well into the constant-current region (parallel lines).  I would use Vce=14 to 15V. Pushing it even higher would increase the power dissipation on the PZTA42, eat into the max output voltage swing and have diminishing return. What I did was to put a 260k resistor in parallel with one of the 175k resistors. YMMV because it has to do with the operating point of the PZTA42 in your circuit, the Vgs(th) of your SiC MOSFET, etc.

After the quick fix, one of the channels now measures as good as with the ±450V supply. We can see that the max output voltage is slightly less compared to with ±450V supply. The difference is subtle with the log scale, though.

Now I'm continue to work on the other channel and see if I can find something else.

 

 

Today I decided to make this mod to my Carbon I built in 2015 and parallel a resistor to one of the 175K to raise the Vce of the PZTA42. My Carbon is running on +/-406VDC rails and 20mA bias.

To my surprise, instead of the 175K + 175K bias resistor chain, I had already replaced them with 150K + 100K at some point. As such, I measured Vce  = 10V for PZTA42 and 21.8V drop across the 20K resistor. According to the graph above, this should put the PZTA42 in its linear operating region. Some wise man must have suggested this mod to me (I suspect SorenB) very early on. 

However, I measured Vgs  of the C2M1000170D to be only 3.2V. Based on how I understand simmconn's posts, I was expecting 20V or higher. What gives?     

 

 

[MLA]

If I read simmconn's post right, simmconn measured G to B- (over the 20K resistor) for those 20+ volts, not Vgs directly. So your 21.8V is that measurement and thus indicates all is well  

[mwl168]

Thanks MLA. That was what I was thinking too. Then I don't understand why Simmconn was worried about exceeding the C2M1000170D's max Vgs rating?

Here is an interesting thing - I went back to my PM's and could not find anything that discussed this mod to the 175K+175K bias resistor chain. However, I came across an early correspondent with Kevin about my plan to use 200K+150K instead (values I  had on hand at the time). Now I wonder if the one that looks like a 100K may be in fact a 200K. It's hard to tell visually. 

I can desolder the resistor in question and measure it. But part of me, the lazy part, is thinking if all is well why not leave it along? Maybe between the +/- 406VDC rails and the 20mA bias current somehow puts the PZTA42 Vce in the desired region?    

[simmconn]

The long tail resistor (offset pot plus the one in series) should drop about 8.xV so

Vce(PZTA42)+V(offset res)+Vgs(SiC)=Vg(SiC)-V(B-) //10+8.6+3.2=21.8

For PZTA42, Vce=10V and Ic=20mA is still not out of the woods. Just look at the DC current gain curves in the datasheet below.   

The non-linear region (due to Early Effect?) is not apparent in the SPICE model from the manufacturer. That's probably why it didn't catch the amp designer's attention.

On the other hand, I was probably over worried about exceeding the C2M1000170D's max Vgs rating.