Neg Rail on JLH Amp Output

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2005
Paid Member
I have been checking the components on this amp for a couple days now and cannot figure why I have -20 volts on the outputs. I have the JLH Amp from Geoff's design from Skycoral

JLH-2005 – NOTES

Only different thing is I am using the FJL4315 (2SC5200 equiv) for output devices.

I have rails at +/-20V. The two pots (50 ohm and 2K) for Iq bias and DC offset adjustments do nothing. Not current through teh emitter resistors.

I bought these boards from someone here a long while ago and I think I remember them saying there is a fixed trace some where...

Tried 2SA970 / BC559 for Q5 Q6, Q7, and have 2SA970 for Q4 (in correct spot). Q3 is 2SC3421, and Q8 is 2SA1358

Any one?
 

Attachments

  • JLH-2005-AMP-PCB.jpg
    JLH-2005-AMP-PCB.jpg
    65.9 KB · Views: 315
Last edited:
Member
Joined 2005
Paid Member
The positive rail is intact and operational. I see both rails throughout the board.

Only thing I can think of is the Q3 is 2SC3421, and Q8 is 2SA1358 possibly being fakes, as they were cheap and from ebay (China - no offense), as I had problems sourcing them.
 
John, it is always a good idea to say what led up to the problem. For instance poking around with a metal object, etc. Then the guys here does not have to think that there is a design problem to analyze and can often help you to quickly trace the fault if they know it was working moments prior to the problem..
 
Member
Joined 2005
Paid Member
OK, sorry. The amp has not fired up (bias wise) since putting it together.

Only thing I can think of is I put in the 2SA970 (ECB) into the Q4, Q5, Q6, Q7 (CBE). Q4 has the second ECB for the 2SA970, so I moved it there. I removed the 970s in Q5, Q6, Q7 and replaced with new BC559 (with correct CBE pinout)

No smoke, hiss, pops or anything leading me to damaged part... I realized the mistake and made the corrections.

Still has -20v on outputs.
 
Member
Joined 2005
Paid Member
Yes, I had the FJL4315 wired correctly - Board showed different pinout of EBC - I managed to catch that at the beginning.

OK, so the 2SA1358 was toast - shorted on all pins. Still checking the C3421

EDIT - the SC3421 is also toast.
 
Last edited:
Member
Joined 2005
Paid Member
Bingo. Replaced the 1358 and 3421 and all is good. Sounds nice so far.

I have one channel working, and sounds quite clean. Pulled the neg rail fuse and put the amp meter across the tabs and read 2.75 amps, so about 32 or so watts per channel at 4 ohms. The 4 pound heatsink gets nice and toasty hot - I used the Nelson Pass way to check temperature - if I can hold my hand on the sink for 5 seconds, its about 140 degrees.

Incidentally, have the one channel powered by two basic HP laptop power supplies (+20VDC @ 3.2 Amps each for a dual +/-20VDC). When using one of these power supplies on my laptop, I see it pulls 230 watts from the wall, so I biased up the amp up to read 450 watts from the wall to match the load from my laptop. I know a laptop load is somewhat variable, but this works great. These power supplies are clean and quiet, but no one ever believes me...

One more thing, I never put an input cap on any of my amps - Krell KSA-50, Aleph 5, F5, (all my pres have one tho), so why have one here?
 
Last edited:
put the amp meter across the tabs and read 2.75 amps, so about 32 or so watts per channel at 4 ohms.

If you can push or pull 5.5A through 4 ohms, doesn't that mean 60WRMS?
Gonna need big caps to deliver the big current half cycle... 5.5A > 3.2A !
Something more than 2000uF (cutoff of 4 ohms at 20Hz)...

Usually its not so much, JLH tends more to shut one output down than it
turns the other one on. Due to 2n3055 beta droop, you won't see current
double at peak...But you used good 2SC5200, so you might have to worry.

Somehow its only 50WRMS??? if I figure where +/- 20V clips into 4ohms...
Maybe there's some leeway to turn down the amps, if 50W is gonna clip
due to voltage anyways. Don't need to overcurrent those supplies so hard.
 
Last edited:
No, that's 2.5A quiescent current. if you like, the positive rail provides the current, and the negative rail takes it away, if both rails were positive (or negative, and the current went back to the psu via ground -then you'd have 5A.
Look up Kirchoff's current law (probably on wikipedia) - might help explain it.
 
if I can hold my hand on the sink for 5 seconds, its about 140 degrees.
LOL! I read that in shock initially, then realised you meant Faranheit!

I agree with Ken, that's a huge amount of current. I'm not listening at all loud yet, but I'm barely running an Iq at 1A at the moment. Start small and work your way up. 4$ heatsinks sound a bit small for this (unless you found a bargain :) ).
 
Member
Joined 2005
Paid Member
Yes, Fahrenheit...yikes!

Yup, like I said, I always trip that bias calc up - I will just bias the output devices to 30 - 35 watts (using voltage across .2 ohm emitter resistor) and be done with it.

After about 15 minutes at 2.75A, the 1/2 watt Bournes bias pot shorted (Geoff warned of this on his site). One of the laptop power supplies started power pulsing, but still works Fine. These laptop PS are rugged as hell.

I will replace the pot, bias up the amp to 30 - 35 watts/device, adjust DC offset, and then replace both pots with resistors.

Anyone's thoughts on putting a bypass switch across the input caps? Like I said, I don't have them on any other amps. Problem is if ever I would need it, it would already be too late.

EDIT - the 4 pound sinks are test sinks, not the final sinks.

Also, does anyone know the overall gain of the amp?
 
Last edited:
Member
Joined 2005
Paid Member
Here she is...like I said, I need to do some critical listening, as I installed a switch to bypass the input cap...don't know yet if the JLH will unseat the Krell KSA-50 Clone....

Turned out quite nice...
 

Attachments

  • P1010020.JPG
    P1010020.JPG
    725.5 KB · Views: 112
  • P1010013.JPG
    P1010013.JPG
    667.3 KB · Views: 101
  • P1010014.JPG
    P1010014.JPG
    660 KB · Views: 103
  • P1010015.JPG
    P1010015.JPG
    656.1 KB · Views: 99
  • P1010016.JPG
    P1010016.JPG
    651 KB · Views: 101
Member
Joined 2005
Paid Member
I tried. Don't care much on the back, but Acetone won't get it off the front...

I know a local sandblaster tho - $20 and he will sandblast anything...

This amp really surprised me. I gave it to a friend who has high end everything (Levinson, Revel, Bryston, etc), and he thought it was very good too...I am building ESL (with Martin Logan CLS Panels), and can't wait to try this amp on them.

Someone asked about the caps - they are 18,000uf each, 50V from Apex Jr (thanks Steve!)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.