So here's my strange update on my 9300:
My plan was to replace the output fets to see if it would cure the problem. I started reassembling the circuit, replacing back the resistors I had removed. Just for the fun of it I decided to take measurements along the way. Every time replaced another resistor it seemed the measurements got better. By the time the last resistor was replaced and I added the fuses back the measurements on right side (the "bad" side) were almost identical to the left side--as though nothing were wrong. The DC offset now measured ~8mV. I could not resist so I connected the right side of the amp to one of my old Paradigm reference (first release) which I'm planning to sell. It sounded very nice from what I could tell. But after a few minutes I got silence again--both fuses on the right side had blown. I once again measured a large DC offset (~8V). But this time I replaced the fuses and everything went back to normal, with good measurements everywhere.
I should mention that the fuses originally in the amplifier were 7A, but I replaced them with 5A fuses, which were handy.
I'm not used to blowing fuses in my amp--would 5A vs. 7A explain blowing the fuses, or do I (as I suspect) still have a substantial problem with this amp? Do FET's ever fail intermittently? I know diodes can fail this way. The amp sounded good enough to me during my short listening test that I really would like to get it working for more than 5 minutes.
Thanks again for any help.
Ron
I should also mention that prior to putting it back together I replaced all of the 100uF caps, (C207, C208, C4, C104, C21, C121). I had planned to replace all the electrolytics, but I only got as far as these. Right now it's still idling fine, with both sides heating up as expected. Previously the right side never got warm.
Last edited:
Its not due to the slightly lower amperage of the fuse. Even a 5A fuse will pass a substantial amount of energy.
Until this problem is corrected I would not put anything larger than a 1 or 1.5A fast blow fuse into the DC rails on the channel that's failing and I'd keep it on a DBT with a larger bulb if needed. You could dial the bias down a bit to allow the amp to idle and operate (play music @ low level) while still on the DBT.
Until this problem is corrected I would not put anything larger than a 1 or 1.5A fast blow fuse into the DC rails on the channel that's failing and I'd keep it on a DBT with a larger bulb if needed. You could dial the bias down a bit to allow the amp to idle and operate (play music @ low level) while still on the DBT.
Hi Ron:
I just saw this thread (I should probably check "Solid State" forum more often).
A few years ago I was working on a Hafler 9300 which had a faulty LV power supply and high DC offset. The installed LM337 & LM317 and the precision 221 & 3.92K resistors were replaced to get it back to spec at +/- 24.3V with the LM337/LM317 just warm to touch.
Hafler Transnova 9300: DC offset & power supply puzzle
I attributed the high DC offset to (presumably) component sourcing issue at the shop floor for the hard to get jFETs. The correct complement of 2SK163 is 2SJ44 and not 2SJ74 which were installed in my amplifier. 2SK163/2SJ44 are 40V devices while 2SJ74 is the lower 25V device. I replaced the SK163/2SJ74 with matched pairs of 2SK163/2SJ44. With the faulty (high) power supply, I believe the 2SJ74s failed.
Furthermore, as already mentioned here, the amplifier had MPSA06/MPSA56 transistor pairs which were operating extremely hot. As a result, the copper tracks on the double sided PCB had hair line burn cracks due to the heat. These were carefully repaired and the MPSA06/MPSA56 replaced with their 1W MPSW06/MPSW56 siblings. With these changes I could get the DC offset to a respectable 0.3mV level.
I had also noticed there were a few typos on the schematic. Transistors Q6 & Q7 are MPS6521/MPS6523 and not 2N6521/2N6523 as marked on the schematic and parts list. Also, C9 & C109 are 1000uF NP electrolytics.
Good to see KC here!
Rgds
Mayank
I just saw this thread (I should probably check "Solid State" forum more often).
A few years ago I was working on a Hafler 9300 which had a faulty LV power supply and high DC offset. The installed LM337 & LM317 and the precision 221 & 3.92K resistors were replaced to get it back to spec at +/- 24.3V with the LM337/LM317 just warm to touch.
Hafler Transnova 9300: DC offset & power supply puzzle
I attributed the high DC offset to (presumably) component sourcing issue at the shop floor for the hard to get jFETs. The correct complement of 2SK163 is 2SJ44 and not 2SJ74 which were installed in my amplifier. 2SK163/2SJ44 are 40V devices while 2SJ74 is the lower 25V device. I replaced the SK163/2SJ74 with matched pairs of 2SK163/2SJ44. With the faulty (high) power supply, I believe the 2SJ74s failed.
Furthermore, as already mentioned here, the amplifier had MPSA06/MPSA56 transistor pairs which were operating extremely hot. As a result, the copper tracks on the double sided PCB had hair line burn cracks due to the heat. These were carefully repaired and the MPSA06/MPSA56 replaced with their 1W MPSW06/MPSW56 siblings. With these changes I could get the DC offset to a respectable 0.3mV level.
I had also noticed there were a few typos on the schematic. Transistors Q6 & Q7 are MPS6521/MPS6523 and not 2N6521/2N6523 as marked on the schematic and parts list. Also, C9 & C109 are 1000uF NP electrolytics.
Good to see KC here!
Rgds
Mayank
@Mayank - After 7 pages of this topic and everyone who wanted to help,luckily someone appeared who is well versed in the matter and knows the basic data.
I went to check and I found that in the original schematic circuit diagram, it's truly jFETs transistor pair SK163/2SJ74.It means that with 25v 2SJ74 works on the edge.Similiar with 2N6521/2N6523
which are found in the original shematic diagram,instead of MPS6521/MPS6523.
Now we are still waiting for a happy ending.Great job.
Regards
Tomislav
I went to check and I found that in the original schematic circuit diagram, it's truly jFETs transistor pair SK163/2SJ74.It means that with 25v 2SJ74 works on the edge.Similiar with 2N6521/2N6523
which are found in the original shematic diagram,instead of MPS6521/MPS6523.
Now we are still waiting for a happy ending.Great job.
Regards
Tomislav
Last edited:
Having just fixed a Hafler 9505, here is what I learned:
Remove the output MOSFETs. Test them. Discard any fried ones.
Meanwhile, get the FET input/driver stage working on its own. Best way is 2 temporary 47K resistors from each PNP/NPN driver stage to where the feedback resistor normally connects to the FET pair.
The input/driver stage should show symmetrical voltages across the mirrored topology. If not, continue to remove parts and troubleshoot.
Once the input/driver stage is ticking, put output MOSFETs back in and remove the temporary feedback resistors.
Normally all will come back up and work.
Good luck.
D.
Remove the output MOSFETs. Test them. Discard any fried ones.
Meanwhile, get the FET input/driver stage working on its own. Best way is 2 temporary 47K resistors from each PNP/NPN driver stage to where the feedback resistor normally connects to the FET pair.
The input/driver stage should show symmetrical voltages across the mirrored topology. If not, continue to remove parts and troubleshoot.
Once the input/driver stage is ticking, put output MOSFETs back in and remove the temporary feedback resistors.
Normally all will come back up and work.
Good luck.
D.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.