KGSSHV Carbon Build Thread

[JoaMat]

Thanks for LTspice files. 

When I open “kgsshv_carbon.asc” I get message  “Couldn't find symbol(s): power_nmos_heat”

[cloudhead]

Please try to update your LTSpice. It should be in your libs

47 minutes ago, JoaMat said:

Thanks for LTspice files. 

When I open “kgsshv_carbon.asc” I get message  “Couldn't find symbol(s): power_nmos_heat”

 

Sorry, for misinformation. Forgot to add it. Please rename to .asy and put it in the same directory as other files.

power_nmos_heat.doc

also make sure you have this setted to yours location

[JoaMat]

Deleted!

Edited April 17, 2020 by JoaMat

[simmconn]

@cloudhead Thank you for putting together the netlist. I appreciate your time. If I may nit-picking a little bit.

If you would like to do harmonic analysis on the differential signal, you could use '.four V(out+,out-)'. Using .four V(out+)-V(out-) will get you the analysis for the single-ended V(out+). You can see the session error log for details.

The HV power supplies starts at full voltage and drops to zero within a few ns at the start of the simulation. This causes a large current spike through the body diode of the SiC devices. This is probably an artifact due to the HV supplies' behavior in the simulation. I don't think this is real, but it does prompt me to look into the power on/off transient behaviors of the circuit as I look for alternatives for the SiCs. They are too expensive, have a positive tempco, and are not strictly DC SoA rated.

[nopants]

How do you guys orient the terminal blocks on the amplifier boards? I have Version 0.5, pretty crowded and I read a little about how people have had issues with the terminal blocks shorting with the tracks near the outputs/rails. I figured you guys with finished builds would have some pointers- it seems like the least headache is to mount them opposite the component side at a glance.

Edited May 15, 2020 by nopants

[jose]

I mounted them at the bottom of the board and used the typical Phoenix terminal .... so far I have had no problems.

PS: About the Phoenix terminal, I have seen some cheap terminals that come with a bit plastic protection that covers the Pin base. At what each terminal costs, these Phoenix guys could add it.

[mwl168]

I also recall the terminal blocks shorting issue also had to do with too much solder being applied.

[hyperma]

Assembling another kgsshv carbon, this time the positive 400V keep droping from first second, don't know why.

Any diode or mosfet leaking? Don't know how to simulate.

 

The tantalum cap or pair ksc5026s area about 95c, or 203F, hmmmmm....

[sorenb]

5 hours ago, hyperma said:

Assembling another kgsshv carbon, this time the positive 400V keep droping from first second, don't know why.

Any diode or mosfet leaking? Don't know how to simulate

 

Check the current source at the front end
should source more than ~1.3mA or so

look here for more 

Edited May 18, 2020 by sorenb

[Superberg]

I had a similar problem with my boards some time ago. The problem in the end end was a layout issue on the board with the optocoupler based psu delay circuit. In my case the -400V worked fine and +400V was not rising. Check what board you have and maybe remove the optocoupler.

Sent from my LG-H930 using Tapatalk

[hyperma]

17 hours ago, sorenb said:

Check the current source at the front end
should source more than ~1.3mA or so

look here for more 

Great great appreciate for your help.

DN2540 is gone,reverse leakage been found,stable now.

[simmconn]

Finally finished the installation of my AliExpress KGSSHV Carbon into a proper enclosure.

I wanted to use a somewhat slim chassis so I skipped those taller ones (100mm+) and ended up with the Breeze Audio SD4309B. It's inexpensive (about $57) and has a pretty solid build. The internal space is just enough for the AliExpress/eBay Carbon PCBs, but a bit shallower than the transformers I have. I had to mill 1mm-deep 'pockets' on the bottom cover for the transformers to fit right below the top cover.

I'm not a big fan of the heat spreaders so those sections of the PCB were cut off. Aluminum angles were used to build a 'cradle' for the transformers that also work as mounting brackets for the extra boards (input switching and LV power supply).

I soon realized that the boards are not so DIY-friendly to work with. The Golden Reference HV PSU board is not designed for center-tapped HV winding which is more popular in surplus. The connectors are somehow placed at the convenience of the PCB routing but not so much for harness routing. Shared terminal pins means that you'll often have to stick two or more wires into the same hole, etc, etc. Not having a mirrored pair of boards seems to make things worse at some places but otherwise not a show-stopper.

I was planning to make this a phase 1 project. The phase 2 will have the input switching and volume pot changed to rotary encoders driving digital pots, which will then make space for phase 3, a tube final stage piggy-backed on top of the carbon PCBs, could be BHSE or Grounded-Grid. There is enough space left, but I'm afraid I won't have the time. Just some ideas for more ambitious builders, I guess.

[JoaMat]

Nice! Interesting build indeed.

 Making generic boards satisfying all builders is not easy. Myself - I do like mirrored amp boards without terminal blocks. 

Until phase 3 – we wait for phase 2.🙂

[powertoold]

Hi all,

I have a KGSSHV (not carbon) 500V that has developed some low-level static / distortion with certain resonant frequencies on the left channel. I'm fairly certain it's caused by the KGSSHV since the static is reduced when I adjust the offset voltage to be more positive (currently at 40V offset; I believe ideal is 0V offset and 0V balance).

Would anyone have any insight into what is causing this static / distortion, especially since it is reduced when I make the offset voltage more positive?

 

Thanks!

[mwl168]

40V offset is quite high. Is this the lowest offset you can adjust to?

If your KGSSHV has offset servo, you will need to deactivate the servo before you adjust the offset. Get the offset close to 0V first (it will drift a  bit) and then engage the servo. 

Edited June 2, 2020 by mwl168

[sorenb]

On 6/2/2020 at 12:11 AM, powertoold said:

Hi all,

I have a KGSSHV (not carbon) 500V that has developed some low-level static / distortion with certain resonant frequencies on the left channel. I'm fairly certain it's caused by the KGSSHV since the static is reduced when I adjust the offset voltage to be more positive (currently at 40V offset; I believe ideal is 0V offset and 0V balance).

Would anyone have any insight into what is causing this static / distortion, especially since it is reduced when I make the offset voltage more positive?

 

Thanks!

@mwl168 is right 40V offset doesn't sound right.

The static may origin from leaking resistors or bad solder joints.
Check the 175k's (r10+r11) , r14+r29 and r7+r30 ,  as well as their joints

[simmconn]

On 6/1/2020 at 3:11 PM, powertoold said:

Hi all,

I have a KGSSHV (not carbon) 500V that has developed some low-level static / distortion with certain resonant frequencies on the left channel. I'm fairly certain it's caused by the KGSSHV since the static is reduced when I adjust the offset voltage to be more positive (currently at 40V offset; I believe ideal is 0V offset and 0V balance).

Would anyone have any insight into what is causing this static / distortion, especially since it is reduced when I make the offset voltage more positive?

 

Thanks!

Have you tried swapping the channels and see if the problem follows the left channel amp or the left phone? If some leakage exists in the phone, raising the offset reduces the effective bias and hence you get less arc-overs.

[powertoold]

Thanks all,

I can hear this type of static with both my 007 and LNS, so the problem likely originates from the KGSSHV.

I'll try to check the resistors that sorenb mentioned. I'm assuming this sort of issue isn't likely to originate from the power supply board.

My KGSSHV uses PRP metal film resistors. I don't think they have a good reputation on this board 😅

[Pars]

Some of us do or have used PRP PR9372 resistors in the past. Things I've noticed about them that I don't like include the epoxy or other jacketing used on the resistors is fragile and easily damaged. The 1/4W variants are rated to 300V;  1/2W and above for 500V. I would presume that actual performance meeting this rating would depend on the integrity of the jacketing?

[powertoold]

I tested almost all the resistors on both channel boards, and the resistance on both channels are similar. However, there were some odd measurements:

The 500k resistors were each measuring 1k on both boards.

One set of 175k resistors (sorry I don't have a diagram of the R numbers) start measuring at around 100k and slowly drift to 120k. This was true for both boards.

Overall, the measurements were essentially identical on both boards. I'm using a Klein Tools M400 DMM. At this point, I'm setting the offset on both channels at around 100V and just using the amp that way, since at 100V the static isn't audible so far 😅

I just don't have the tools or know-how to replace those odd-measuring resistors, since it's an off-board KGSSHV with the amp boards glued to the heatsinks.

Edited June 3, 2020 by powertoold
Rationale for not replacing resistors

[Pars]

From the schematic I have, the 500K resistors are from the inputs to ground, so would be in parallel with the volume pot. See if that value changes as you move the pot from 0 volume to full.

That said, measuring resistors in circuit will give you sometimes strange results as you don't know the exact interactions with connected components. Also, ohming out resistors for what is suspected to be a leakage problem won't really tell you anything.

Can you post detailed pics of the amp boards? Are the resistors lifted off of the board, or installed right on the surface? And what is the origin of this KGSSHV? i.e. who built it?

[jose]

What have you used as input devices?, Were they macheted? do you have the transformer near the inputs? ....

[powertoold]

17 hours ago, jose said:

What have you used as input devices?, Were they macheted? do you have the transformer near the inputs? ....

I've used 2 different DACs and also played from several different source files. The static increases with higher volumes and is only audible during "difficult" resonant frequencies (such as reverb from guitar strings).

 

19 hours ago, Pars said:

From the schematic I have, the 500K resistors are from the inputs to ground, so would be in parallel with the volume pot. See if that value changes as you move the pot from 0 volume to full.

That said, measuring resistors in circuit will give you sometimes strange results as you don't know the exact interactions with connected components. Also, ohming out resistors for what is suspected to be a leakage problem won't really tell you anything.

Can you post detailed pics of the amp boards? Are the resistors lifted off of the board, or installed right on the surface? And what is the origin of this KGSSHV? i.e. who built it?

As far as I know, the resistors are installed on the board's surface. This KGSSHV (non-Carbon) was made by Horio / TMoney I believe. Everything seems well-built, and the heatsink is massive.

It'd be great if there was something I could try that doesn't involve replacing components (to troubleshoot or eliminate the static).

 

Edited June 4, 2020 by powertoold
More explanation

[chinsettawong]

I don't have good experience with those red resistors.  On my DIY T2, I had to take a lot of them out.  Your noise could very well come from those.

[powertoold]

This is regarding the KGSSHV (non-Carbon) issue above:

I think I resolved the static / distortion issue. I had essentially doubled the current to ~10mA by adding a parallel 200R resistor to the 180R. I removed the parallel resistor, so the amp is now "stock." After I removed the parallel resistor, static / distortion isn't audible (yet). Thanks all.