KGSSHV Carbon Build Thread

[Craig Sawyers]

5 hours ago, sorenb said:

@Craig Sawyers: @Blueman2 is rather asking for any benefit finding a DAC with bal out rather than SE. I'd say it boils down to how the DAC is creating the balanced signal, which might be difficult to determine if no inside pics or schematic is available from the specific DAC.
That said, it might be of higher importance that he likes the sonic signature of the DAC rahter than focus on the DAC being SE og Bal.

Well pardon me - since I have failed to answer the question, read this http://www.analog.com/media/en/training-seminars/tutorials/MT-019.pdf 

 

[Blueman2]

After installing a quad volume control and connecting the XLR inputs to prepare for a balanced output DAC, I noticed something odd.  The LED on the Positive output side of the left channel dims quite a bit after startup.  Initially, all 4 LEDs are about the same brightness, but then after about 1 - 2 seconds after power on, the Left channel positive LED (the one near the O+) dims.  Happens just as the music starts coming from the earspeakers.  Everything still sounds OK.  Only things that I notice is a bit more of a 'pop' about 1-2 seconds after turning the amp off.  The pop comes mostly from the left channel.

Any idea what a dim LED might indicate?  I did a quick check of voltages and they appear OK.  And is it normal to hear a 'pop' when powering down the amp?  It has always done this, but it seems louder now and more focused on the left channel. 

  

[sorenb]

@Blueman2 check the balance offset  if the pot is dialed near one of the extreme,s you might have a minor fault somewhere later in the chain , e.g. bad joint or something similar ... 

[Blueman2]

7 hours ago, sorenb said:

@Blueman2 check the balance offset  if the pot is dialed near one of the extreme,s you might have a minor fault somewhere later in the chain , e.g. bad joint or something similar ... 

Great advice!  I had not checked the output voltages.  Sure enough, left channel shows offset on both sides of -300V (yes, negative).  Right channel is perfect.

I turned off the Servo, which make the LED normal brightness.   Both the O+ and O- of the left channel show voltage of -300V.  Trim pot does have an effect, but not that much.  I can get it down to about -200V or so.   I am still getting 20ma on the CCS current (1V on the test resistor).  Also, still able to balance the O+ and O- perfectly.  It is just the offset that is wrong.  Oh, and it did still play music just fine from an audible standpoint (before I realized how off the offset was!).  

Recommendations on what points to test to narrow what might be wrong?  Main clues I am going off of are:

- Issue is impacting both the O+ and O- sides equally, it must be something that impacts both sides of the board. 

- Since it still makes music, I am guessing the silicon in the music path is still ok?

Edited December 17, 2017 by Blueman2

[sorenb]

dial the trim to around middle, if you have -300V in both O+ and O- and cannot dial in against zero with the offset trimmer, you might raise the tail resistor , if you have 100-120ohms in that position, you might try 150ohms rather, if you already have 150ohms...try 200ohms, you get the drift

Also, check the other channel, if that channel is in balance, you might check if you got all the resistor values right

[Blueman2]

21 minutes ago, sorenb said:

dial the trim to around middle, if you have -300V in both O+ and O- and cannot dial in against zero with the offset trimmer, you might raise the tail resistor , if you have 100-120ohms in that position, you might try 150ohms rather, if you already have 150ohms...try 200ohms, you get the drift

Also, check the other channel, if that channel is in balance, you might check if you got all the resistor values right

I have a 100ohm tail now as I recall.  I think the original spec was 182, right?  But you are right in that this will directionally get me there.  At 150 ohm across the trim pot + tail resistor (trim in middle at 50 ohms + 100 ohm tail), I am at -300V.  When I adjust the trim pot to max (about 195 ohms total across tail and trim) I am down to -220V.  If that is a linear relationship, it is about 1.6V per ohm.  So I would need to increase tail by about 180 ohms, so need about ~280 ohms tail. 

BUT, the odd thing is that this amp ran just fine for several weeks with perfect offset using 100 ohm tail.  I noticed about 2-3 days ago a bit of static intermittently, and an increased 'pop' when powering up or down.  So rather than an incorrect resistor value, I am suspecting something else has caused this.  Some root cause I have not figured out yet. 

Is the OP27A used when no servo is connected?  What else might impact offset so severely yet allow the amp to still basically operate?  

Edited December 17, 2017 by Blueman2

[sorenb]

hard to say what is going on, 100ohms is what I am using for 20mA  ....from what you report, I'd suggest you go through the soldering....if you have been a bit to generous with the solder, it might hide a bad joint or two ....use some wick, and be careful not to use too much solder ...maybe the A42's (just guessing)
OP27A is ony used if you engage the servo

[Blueman2]

Yes, I guess that is the next step.  I did a compare of resistance readings at different test points comparing the working Right channel to the bad Left channel, and so far everything is exactly the same.   I will continue with that route for another hour, and then just pull the board and check solder joints. 

I was also thinking of the A42s, but since this is impacting both + and - sides equally, it would seem odd that both would have equal solder issues.  Any other thoughts of what would impact both sides at the same time??

Thanks again, sorenb, for the advice!

 

EDIT: Bad solder joint, as you suspected sorenb!!!  I have not found it yet, but I decided to use a probe to 'nudge' all the resistors and other components around a bit to stress the joints.  Sure enough, on next power up, the offset was back down to essentially 0 volts.  Of course, now I have to find that damn bad solder joint!!!!

Edited December 17, 2017 by Blueman2

[ang728]

congrats 

yes cold joint will be affected by temperature and cracks very slowly,that's why it occurred few days after you've done your build 

[Blueman2]

19 minutes ago, ang728 said:

congrats 

yes cold joint will be affected by temperature and cracks very slowly,that's why it occurred few days after you've done your build 

Yes, this all makes sense now.  I would feel better if I knew where to focus on reflowing the joints, though.  Given it was only the offset, and it impacted both sides of the board, that should give me a clue.  Just not sure....

[sorenb]

a42's ...would be my first attempt, and the tail resistor + trimmer ...removed the old solder using wick, and re-solder while being careful not to use too much

Edited December 17, 2017 by sorenb

[Blueman2]

OK, I did some rework and am crossing my fingers I got the offending joint.  I did find a few questionable joints, but nothing very obvious.  It is working fine for now.  I will just have to keep an eye on it.  Unfortunately (or perhaps, VERY fortunately) the failure is not immediately obvious.  I will have to watch out for a 'pop' in the earspeakers on startup or shutdown which means the problem is back.  I am sure the instantaneous jump from -300 to +250 (which is what was happening on shutdown with the issue) is not good for my earspeakers!  But at least it was both O+ and O- at the same time. Otherwise, boy, that would be nasty. 

 

Thanks again for the help!

Edited December 17, 2017 by Blueman2

[JoaMat]

Is the jump up to +250 volts not normal behavior? That's what my Carbon does, 500V rails and my multimeter catches +330V before slowly down to 0.

[Blueman2]

 

21 hours ago, JoaMat said:

Is the jump up to +250 volts not normal behavior? That's what my Carbon does, 500V rails and my multimeter catches +330V before slowly down to 0.

Yes, exactly.  But usually it is from 0 to around +300 then drift down.  Going from -300 to +300 is more concerning.  Especially since the membrane bias is at +590V on my setup.  +590V right next to -300V is asking for trouble.  

On 12/17/2017 at 11:19 AM, ang728 said:

congrats 

yes cold joint will be affected by temperature and cracks very slowly,that's why it occurred few days after you've done your build 

Yes, especially given the high temperatures of the Carbon.  The thermal cycling will stress anything less than a really good joint.  No kidding, I have been thinking about this every time I turn the system off after use.  I just did not think it would happen so early.  I guess infant mortality for the poorest solder joints.  

EDIT: I accidentally left the amp on all night last night, so about 14 hours.  Still working fine.  Of course, it is not the heat that kills solder joints, but rather the cycling from hot to cold to hot to cold that does it.  Only time will tell.  

Edited December 18, 2017 by Blueman2

[bbest]

Hello all!

Successfully run one board of GRHV (boards 1.7 version). Wanted to turn it on 450V. Could I use three resistors instead of two R8, R9 ? Is not it criminal?

For example 390K + 390K + 100K = 880 = 2x 442 .

[sorenb]

5 hours ago, bbest said:

Hello all!

Successfully run one board of GRHV (boards 1.7 version). Wanted to turn it on 450V. Could I use three resistors instead of two R8, R9 ? Is not it criminal?

For example 390K + 390K + 100K = 880 = 2x 442 .

only reason to use more resistors is spreading the heat dissipation 

[Blueman2]

I was about to pull the trigger on another set of GRHV boards for another project, but see that the C2M1000170D SIC MOSFETs are out of stock at Mouser and Digikey.  Would this part work as a replacement?  755-SCT2H12NZGC11.  Biggest difference is the much slower fall time (74ns vs the 9ns for the Cree).  Thoughts?  

Alternatively, I have a couple FQFP8N80C's on hand as well if those would work.  Again, this is for the GRHV PSU, not for the amp itself.  

 

Edited January 7, 2018 by Blueman2

[Tinkerer]

Those Crees will be back in stock in a month or two. They tend to run out a couple times a year from what I've seen. Looking at Mouser, it says they expect the next big shipment around the end of February.

[bbest]

sorenb, good news, so I just add resistors for required voltage.

Blueman2, wait some time and C2M1000170D will be sell soon. I also wait a couple week when assemble my BOM.

[Blueman2]

 

On 1/6/2018 at 9:07 PM, Tinkerer said:

Those Crees will be back in stock in a month or two. They tend to run out a couple times a year from what I've seen. Looking at Mouser, it says they expect the next big shipment around the end of February.

 

On 1/6/2018 at 9:21 PM, bbest said:

Blueman2, wait some time and C2M1000170D will be sell soon. I also wait a couple week when assemble my BOM.

 

<Despair> noooooooooooooooooooooooooooooooooooooooooooo! </Despair>.  Wait many days and many hours to build my 4th amplifier, of which I only really need one?  Clearly you have overestimated my maturity and patience.  

Seriously though.  For the PSU, I wonder if the  755-SCT2H12NZGC11 would be an acceptable substitute.  Just not sure how much of an issue the lower current rating (3.7A vs 4.9A) and slower fall time will be.  It figure it is always nice to have a backup source part for when the primary part goes OOS.  

Here is the parts comparison:

EDIT (Aug 4, 2018): now running this amp for over 6 months using the 755-SCT2H12NZGC11.  Working fine for me as a direct replacement for the C2M1000170D.  

 

Edited August 4, 2018 by Blueman2

[spritzer]

That one should be just fine if the pinout matches.  Couldn't be bothered to test it...   

[Blueman2]

On 1/7/2018 at 10:23 AM, spritzer said:

That one should be just fine if the pinout matches.  Couldn't be bothered to test it...   

Pinouts match, and TO-3P appears to have same pin spacing as TO-247.   I am going to be a guinea pig and give it a try.  What could possibly go wrong?   

EDIT: Well, sorenb was correct!  The part did work just fine and is now powering my Blue Hawaii BJT build.  It appears to operate just like the C2M1000170D.  Does not get very warm at all, either. The packaging is a bit different in that the hole is not are far down from the top as the C2M.  This means that it can go into the same hole spacing as a normal TO220, rather than the hole that is lower down for the C2M (notice the boards have 2 holes that almost overlap).   

Edited January 21, 2018 by Blueman2

[wink]

This, perhaps....?

 

[Blueman2]

If you read about Silicon Valley blowing up or a major California power outage, you will know why.  

[bui501]

I finally received the last few components for my carbon build and finished populating the boards this weekend.  (carbon board v5 + GRHV fat, split boards)

I'm using the split fat board version of the golden reference high voltage power supply boards and a SumR toroidal transformer with 2x360, 2x16 secondaries.  Actual unloaded measurements from the transformer's secondaries at 220VAC input were 387/387VAC, 17.8VAC.

When I hooked up the ps boards to a variac and cranked up the input line voltage to the transformer, the high voltage outputs from the boards measured +400/-400V .  

The LEDs on both amp boards come on and stay on.

Problem 1:  The low voltage outputs from the GRHV boards goes up to only +/-10V even at 220VAC input on the transformer (17.8VAC unloaded measurement from the transformer's output, low voltage winding).

Problem 2:  I hooked up the power supply boards to the carbon boards anyways and started to play with the current adjustment pots.  At the test point FURTHEST away from the servo 2 jumpers, I can get the current to 20mA (started out at 18mA).  However, at the test point/pot closest to the servo 2 jumpers, the current started out "normal" at close to 20mA, but after 10 or 20 seconds, it rose to 36mA -- I cannot/cannot get the current to go much lower at that test point.  This is the same for BOTH amp boards, so I'm assuming it's a problem with the PSU boards.

- I've double checked all the solder joints three or four times with a 20X loupe.   (FYI, instead of using the phoenix terminals that everybody else has been using, I decided to use molex mini fit jr. cable connectors that are rated for 500V because I was paranoid about messing up and not being able to examine the solder joints underneath the phoenix terminals.)

- I've gone through and cleaned up all stray flux and debris several times on both sides of each board.

I've read through this forum and have tried some of the suggested troubleshooting suggestions.  E.g., measured the resistance between HV+ and GND and HV- and GND leads on the amp boards as suggested by SORENB to INSANITY last sep in this thread, and the resistances measured as expected in the ballpark of "250kohm-ish" (actually, around 280Kohm).  

Any suggestions on next steps for troubleshooting?  I've attached photos of the power supply boards and one of the amp boards (unpowered).  I haven't attached the ceramic pads on the 10M90S'es yet -- those are coming in the post from mouser along with the thermal grease for the pads.

 

 

 

Edited January 15, 2018 by bui501