i'm on a roll... the kgsshv

[mwl168]

What's supply rail voltage and the bias current of your KGSSHV? 

It's probably not healthy for the active devices on the amp longterm if the heatsinks are so hot that you can barely touch it. May want to consider either reduce the voltage and/or the bias current.

Edited April 28, 2020 by mwl168

[bwck2000]

15 hours ago, mwl168 said:

What's supply rail voltage and the bias current of your KGSSHV? 

It's probably not healthy for the active devices on the amp longterm if the heatsinks are so hot that you can barely touch it. May want to consider either reduce the voltage and/or the bias current.

I am using a1968 so its 450v config. Unlike the carbon, it has no adjustable ccs current so I didnt measure it.

Maybe I need to order a pair of new heatsinks with longer fins. Its also a fail with no ventilation openings at the bottom but on the top only.

[mwl168]

R5 and R6 set the bias current of the C4686A output devices. The voltage drop across them divided by their resistance is the bias current. 

IIRC, the position of these two resistors varies depending on the version of the board. You’ll need to trace the board to locate them.

Edited April 29, 2020 by mwl168

[powertoold]

Hi everyone,

I originally posted about this issue in the Carbon thread, but it's more appropriate here.

After 2 years, my 500V off-board kgsshv started developing left-channel distortion (sounds like sand in a cardboard box) with certain frequencies (only audible when music is playing and more audible with higher volume). I believe this is a result of increased current (100R in R5/R6) and overheating (for a short time, I had kgsshv sideways with the left channel board on top).

The distortion is reduced with any of the following steps:

1) Increase left-channel offset bias voltage to ~70V or higher

2) Change R5/R6 back to 180R

The distortion is present immediately at power on (cold), and doesn't get much worse with warm amp.

My issue now is that the higher current sounds much better (microdynamics and "thickness") to me (100R at R5/R6), so I'd like to figure out how to fix the distortion by replacing components. I can't live with the stock current, and this kgsshv has sentimental value to me and some others on this forum

My question is

1) What are the most likely components causing the distortion? I was thinking the resistors or the IXYS mosfets.

2) When the amp is warmed up, I notice that the heatsinks near the right channel are warmer (2x by feel) than the left channel. Is it possible that the thermal interface between the left-channel mosfets and heatsink is compromised?

Thanks all!

 

Edited June 13, 2020 by powertoold
Describe distortion

[mwl168]

I've asked this question before - what's the lowest offset voltage you can adjust to with the higher bias current of the output devices (lower R5, R6 values)?

I would first get the offset and balance to reasonable range and see if that affects the distortion issue you are experiencing. If the amp is set up properly you should be able to get both to below 1V easily. 

If the amp was originally set for lower bias, increasing the bias current may cause the amp to run out of offset adjustment range without making change to the offset adjustment circuit. 

  

[powertoold]

12 hours ago, mwl168 said:

I've asked this question before - what's the lowest offset voltage you can adjust to with the higher bias current of the output devices (lower R5, R6 values)?

I would first get the offset and balance to reasonable range and see if that affects the distortion issue you are experiencing. If the amp is set up properly you should be able to get both to below 1V easily. 

If the amp was originally set for lower bias, increasing the bias current may cause the amp to run out of offset adjustment range without making change to the offset adjustment circuit. 

  

With 100R, the lowest offset voltage I can adjust to is +70V. At this voltage, the distortion is barely audible. With 180R, I can set the offset and balance as close to zero as possible, and there’s no audible distortion. Using a trimpot adjustment tool, I’ve adjusted the offset voltage while music is playing and didn’t hear any static while adjusting the pot, so it doesn’t seem like the distortion originates from a noisy pot?

The distortion is also reduced when I adjust the balance voltage (although I forget if more positive or negative).

I’m willing to replace whatever parts would be likely causing this problem. I hope to narrow it down, so I don’t have to replace everything on the amp board 😅

[mwl168]

Ok, that makes sense. 

When you increase the output device bias current, you need to tweak the offset adjustment chain (R26 and VR1 - 2K and 1K respectively in the schematic) so you can adjust the offset close to 0V.  The easiest thing to do is to decrease R26 but make sure the voltage drop is somewhat even between R26 and VR1 so they are not stressed.

I posted earlier in this thread of a similar effort I made on my build: https://www.head-case.org/forums/topic/8327-im-on-a-roll-the-kgsshv/page/187/

The goal is to get the offset and balance close to 0V as possible. They will drift a bit and that's normal. 

 

Edited June 14, 2020 by mwl168

[powertoold]

41 minutes ago, mwl168 said:

Ok, that makes sense. 

When you increase the output device bias current, you need to tweak the offset adjustment chain (R26 and VR1 - 2K and 1K respectively in the schematic) so you can adjust the offset close to 0V.  The easiest thing to do is to decrease R26 but make sure the voltage drop is somewhat even between R26 and VR1 so they are not stressed.

I posted earlier in this thread of a similar effort I made on my build: https://www.head-case.org/forums/topic/8327-im-on-a-roll-the-kgsshv/page/187/

The goal is to get the offset and balance close to 0V as possible. They will drift a bit and that's normal. 

 

Perhaps there is some misunderstanding here. With 100R on R5/R6, I am able to adjust my balance and offset close to 0V. However, if I do so, there is some distortion in the 20-200hz low frequency region (only in the left channel, right channel is perfect). Also, the amp sounded great for 2 years with 100R and only started developing this distortion in the last month.

To reduce this distortion, I would have to adjust the offset voltage to around +70V on the left channel.

Edited June 14, 2020 by powertoold
Duration of issue

[mwl168]

Hmm, what’s the voltage drop and dissipation on R26 and VR1 when you adjusted the offset to 0V?

[powertoold]

11 hours ago, mwl168 said:

Hmm, what’s the voltage drop and dissipation on R26 and VR1 when you adjusted the offset to 0V?

I'm not really comfortable working with a live circuit right now, as I don't have any experience building amps 😅

I have the tools necessary to replace parts, so I'm trying to go that route first. I'm fairly certain that my issue is related to a failed or overheated component, as the amp was working fine until a month ago. Also, my right channel has no issues.

As for R26, are you implying that it's necessary to make changes or check on R26 when changing R5/R6? I've searched this thread about changing R5/R6, and other than your post, I didn't see anyone else talking about R26.

[mwl168]

IIRC, significant voltage is dropped in the chain of R26 and VR1 and I am wondering, depending on where VR1 value is set, if one of them is overly stressed beyond its rating and causing some issues.

  

[powertoold]

On 6/15/2020 at 9:06 AM, mwl168 said:

IIRC, significant voltage is dropped in the chain of R26 and VR1 and I am wondering, depending on where VR1 value is set, if one of them is overly stressed beyond its rating and causing some issues.

  

Just wanted to update everyone on the amp.

About 2 weeks ago, I started trying to replace the MOSFETs and c4686 transistors on the left channel. I also bought a new 1K pot and 2 ixys chips from mouser.

When I initially removed the amp board from the case and looked at the existing c4686 transistors, it seems like the heat conducting material (looks like Arctic silver) was lightly applied to the back of the transistors and was therefore not dissipating the heat well. This led me to speculate that overheated c4686 transistors were the culprit for my left channel distortion.

I also bought some new c4686 transistors off dalbani on eBay. I don't know the authenticity of these transistors. When I received the transistors, they had different fonts on some of the transistors, so I wasn't very confident.

When I started replacing the parts on the left channel, I replaced both the MOSFETs and all of the c4686 transistors along with the 1K pot. Unfortunately, after doing this, the left channel offset bias couldn't get under 70 or 80 volts no matter how much I adjusted the 1K pot. This led me to believe that maybe the transistors I got off of eBay were fakes.

Long story short, I made all manner of component replacements and configurations in order to troubleshoot this issue. After about seven or eight attempts at replacing various components with new versus old parts, I finally found that the IXYS MOSFETs I got from mooser were the culprit for the bias offset voltage problem. This was very bad luck because I hadn't expected that I would be getting a bad part from mouser.

I was lucky in that I was about to give up on the amplifier because having to solder and desolder off-board transistors is a PITA.

Tldr: mouser sent me some bad ixys MOSFETs, leading me to try all sorts of component replacements in order to resolve the problem. The irony is that I actually liked the semi-distorted sound more than the clean sound I have now.

[kevin gilmore]

if the thermal grease really is the color of artic silver, that stuff is electrically conductive. causes all sorts of problems.

[powertoold]

13 hours ago, kevin gilmore said:

if the thermal grease really is the color of artic silver, that stuff is electrically conductive. causes all sorts of problems.

I took a picture of it here. You can see the contact isn't that great either.

[Pars]

Lots of peeps have had problems with Dalbani selling fake shit. I'm not so sure I would be so quick to think the Mouser IXYS parts are your problem... possible, but not likely.

[powertoold]

9 hours ago, Pars said:

Lots of peeps have had problems with Dalbani selling fake shit. I'm not so sure I would be so quick to think the Mouser IXYS parts are your problem... possible, but not likely.

Thanks. Yes, what I ended up doing was mix and match older looking C4686s with "real" looking Dalbani C4686s, lol.

The reason I think the IXYSes were the culprit was that I tried old C4686s + old pot + new IXYSs and wasn't able to get the offset bias below 60V. Only after using the old IXYS was I able to, regardless of old or new C4686s.

[powertoold]

Just to update anyone who happens on my issue in the future:

I wasn't satisfied with the new C4686s from eBay (sound seemed a little more "metallic" and the transistors have questionable authenticity), so I took a risk and switched back to the old C4686s, and the initial distortion problem is now gone. So I'm not exactly sure what was the cause. The only new part on the board is the Offset Trim Pot.

Now, the sound is back to its former glory, and I'm very happy.

Edited July 23, 2020 by powertoold

[waltzingbear]

Is there a current source of earspeaker jax? I am aware of the Ampenol (WDI) parts, but are there any of the custom teflon ones available anywhere? I couldn't find them.

 

cheapest source in US https://vacuumtubesinc.com/index.php/sockets-adapters-and-parts/5-and-6-pin-tube-sockets/a5-5285.html

 

as a side note, contacted Richard Summer, he is moving his shop and won't be available for a (unspecified) period of time.

[spritzer]

I have my own custom made teflon ones in stock but not listed on my site. 

[hhobeika]

After 5 years from the time I acquired the PCBs and parts, I finally took advantage of the free time created by the pandemic and completed my KGSSHV.  This is the offboard version with +/-500V supply.  I know many have since moved on to the Carbon but I am very happy with the results and plan on building myself a Carbon in the near future.  Hopefully it won't take me that long this time.

Thanks to Kevin for the great design, Spritzer for his support and providing some of the parts and this thread from which I learned a lot.

The case is from Aliexpress but I got a new front panel made at the local machine shop.

[hhobeika]

Here is a picture of the inside.  There is a limit on the file size I can upload so I had to dramatically shrink it and compress it to be able to post it.

[Blueman2]

@hhobeika, great build!  Very impressed with your details such as keeping the signal output wires parallel and flat just like in the headphone cable.  I love those little details!

[hhobeika]

17 hours ago, Blueman2 said:

@hhobeika, great build!  Very impressed with your details such as keeping the signal output wires parallel and flat just like in the headphone cable.  I love those little details!

Thank you.  I had the advantage of being late in the game and building on the collective experience in this thread.

[Pars]

Stupid question: where do you get multiple solid colors of hookup wire such as that use in hhobeika's build. Rated for 600V, teflon jacketed? Stranded or solid? Thanks!

[Laowei]

20 minutes ago, Pars said:

Stupid question: where do you get multiple solid colors of hookup wire such as that use in hhobeika's build. Rated for 600V, teflon jacketed? Stranded or solid? Thanks!

One source I used. Great to deal with.

https://www.ebay.com/itm/PTFE-wire-Alpha-5854-7-or-equiv-20-ga-Mil-16878-4-silver-plated-10-feet/292416116301

Many other wire gauges, solid or stranded available by seller. Do a search of his other items.

Edited November 3, 2020 by Laowei