"In your case, the fuse blows so, remove AMP PCBs and confirm that the PS PCB is good - confirm the unloaded DC rails voltage readings. There should be no voltage drop across those 4 x 5W / 0.47ohm resistors (located on the PS PCB) on each rail side (no current draw)."
You have a short somewhere capable of drawing a lot of current... enough to blow the fuse, so if the PS PCB is okay, and the PS rail voltages are okay... then -> make sure the MOSFETs are electrically isolated from the heatsink... check with the ohmmeter.
You have a short somewhere capable of drawing a lot of current... enough to blow the fuse, so if the PS PCB is okay, and the PS rail voltages are okay... then -> make sure the MOSFETs are electrically isolated from the heatsink... check with the ohmmeter.
Ah, sorry for the misunderstanding. Yes, the 0.47 ohm resistors on the power supply have 0V across them, both while powered off and powered on.
Good idea. I just took some measurements and yes, the MOSFETs are completely isolated from ground and the heatsink.
I switched to a different multimeter and the measurements around the board are settling much more quickly, I think the calibration or maybe the battery in my other meter may be going bad. This will help as I continue to probe around.
I also ordered a transformer with 15V secondaries instead of the one with 18V secondaries that I'm using now. Maybe the issue has something to do with the higher supply voltage? We'll see after the new transformer arrives.
See if you can cancel that transformer order. Here's why:
plus and minus 27V DC rail voltage is perfectly fine with no AMP PCBs connected.
23 - 25V DC is also perfectly fine with AMP PCBs connected.... the exact DC rails' voltage (when the PS is loaded) depends on the amount of quiescent current, which could be as high as 4A (for both AMP PCBs), if you set the bias high.
plus and minus 27V DC rail voltage is perfectly fine with no AMP PCBs connected.
23 - 25V DC is also perfectly fine with AMP PCBs connected.... the exact DC rails' voltage (when the PS is loaded) depends on the amount of quiescent current, which could be as high as 4A (for both AMP PCBs), if you set the bias high.
@corzendonk
Your input wiring looks suspect as well. You've got the inner barrel of the RCA going to IN+, that's correct. However, the outer barrel (ground lug) is going to IN- with no jumper between IN- and GND. I don't see a jumper on either on the back of the board or between screw terminals. Please refer to post #9747 of this thread for correct wiring.
Your input wiring looks suspect as well. You've got the inner barrel of the RCA going to IN+, that's correct. However, the outer barrel (ground lug) is going to IN- with no jumper between IN- and GND. I don't see a jumper on either on the back of the board or between screw terminals. Please refer to post #9747 of this thread for correct wiring.
@KevinHeem Thank you for the feedback, I will add a jumper wire as suggested.
But, I don't think this will solve my issue with shorting, right?
But, I don't think this will solve my issue with shorting, right?
Take the money you saved on purchasing a new transformer and invest it in a variac so you can bring the voltage up slowly at start-up.
@corzendonk
I'm trying to think about what I might do in your shoes.
When I first started up my own Aleph J I had R27 set incorrectly and it started to draw lots of current as measured at R18.
What resistance was R27 set at when you started your amp for the first time?
From this time forward you should disconnect the left board (good channel) from the power supply and troubleshoot with only the right channel connected. You're going to need some kind of current limiting, either a light bulb tester or variac.
I'm trying to think about what I might do in your shoes.
When I first started up my own Aleph J I had R27 set incorrectly and it started to draw lots of current as measured at R18.
What resistance was R27 set at when you started your amp for the first time?
From this time forward you should disconnect the left board (good channel) from the power supply and troubleshoot with only the right channel connected. You're going to need some kind of current limiting, either a light bulb tester or variac.
@KevinHeem Thank you for the advice, I never thought to use a variac but I agree it would be a useful purchase as I intend to build more amps. Would you recommend any particular product? Or is one just as good as any other as long as it meets my voltage/power requirements?
If I tried starting it up again, what would be the "safest" setting to pre-set R27? And for that matter, what about R7 and R8?
To be honest, I was not careful about what I had R27 set to on initial startup, just wasn't something I thought to check. Although, during initial adjustments, I probably turned the pot to both ends of it's range trying to get the offset and bias under control before the fuse blew for the first time. The second time I tried starting up the amp, I was careful to pre-set R27 to the middle of its range, about 50K. And that time resulted in a blown fuse almost immediately.
If I tried starting it up again, what would be the "safest" setting to pre-set R27? And for that matter, what about R7 and R8?
Agreed, I want to protect my "good" board as much as possible until this is resolved.
A variac with a rating of 5A would be sufficient for the FW stuff. If there's a monster amp in your future then 10A rating.
Recommended starting value for R27 = 68k. I installed a fixed resistor at R8 I believe to be 1k, so you could set your 2k pot to half way. R7 could also be set to halfway.
I find it interesting that you didn't smoke any components on that right board. That's a good sign.
Recommended starting value for R27 = 68k. I installed a fixed resistor at R8 I believe to be 1k, so you could set your 2k pot to half way. R7 could also be set to halfway.
Was the left channel connected also at this point?
I find it interesting that you didn't smoke any components on that right board. That's a good sign.
I had the foresight to disconnect the left channel before I did any further testing with the right channel. Actually, I confirmed the left channel was still working with a listening test, with the right (bad) channel disconnected.
No smoke or visible damage on either channel that I can tell, caps aren't bulged or shorted, resistors look good and measure consistently with their rated value. So by process of elimination, my most likely suspect is one (or more) of the transistors, or maybe something going on with the pot settings. Although, my basic resistance measurements haven't been perceptive to any discrepancies so far.
I went to the hardware store earlier and scrapped together a dim bulb tester, got bulbs that will limit current to about 0.5A and 1.5A (as measured with an ammeter). Now I should be able to test with the amp powered on and current limited, but at this point I'm not sure what to look for. Any recommendations? Perhaps the measurements as shown by Extreme_Boky in post #9,757?
@corzendonk were you measuring the resistance between the middle pin of each of the MOSFETs and the chassis? I think you did already from how I read your posts, but just in case.
Good news and bad news, everyone. The bad news is that I am a tremendous idiot that overlooked a very important detail. The good news is, I now have a seem to have a functional Aleph J, and a bunch of new test equipment!
During initial power up and bias/offset adjustment, I thought I was being a smart engineer by hooking up my oscilloscope to measure voltage across R18. It was the only voltmeter in my house with alligator leads. I didn't realize that the grounding clip on the probe wasn't just for voltage reference, it was actually tied to earth ground. So naturally, every time I tried to power up, the -V rail was dumping current through R18 and into my oscope ground. Stupid mistake.
More good news, shorting R18 to ground is a very direct path to the -V rail, so I think it's very unlikely that I damaged any components during this whole debacle.
The amp is warming up for final offset/bias adjustments, I'll report back after my listening test if all goes well, but so far this board is behaving just as well as its sibling.
The lingering mystery is, why didn't the fuse blow when I was adjusting the other board?? I used the same oscilloscope...
During initial power up and bias/offset adjustment, I thought I was being a smart engineer by hooking up my oscilloscope to measure voltage across R18. It was the only voltmeter in my house with alligator leads. I didn't realize that the grounding clip on the probe wasn't just for voltage reference, it was actually tied to earth ground. So naturally, every time I tried to power up, the -V rail was dumping current through R18 and into my oscope ground. Stupid mistake.
More good news, shorting R18 to ground is a very direct path to the -V rail, so I think it's very unlikely that I damaged any components during this whole debacle.
The amp is warming up for final offset/bias adjustments, I'll report back after my listening test if all goes well, but so far this board is behaving just as well as its sibling.
The lingering mystery is, why didn't the fuse blow when I was adjusting the other board?? I used the same oscilloscope...
It passes the listening test with flying colors. Thanks to all who took your valuable time to help me! Extreme_Bocky, KevinHeem, and beedlo.
As far as I can remember, the measurement setup was exactly the same for both boards. Maybe the "good" board wasn't connected properly to ground? I'm content with never knowing.
The grounded (negative) alligator clip was placed after (inclusive of) the 5W resistor... or, the amplifier IEC ground lug was not connected to the chassis (oscilloscope ground)
Also, search for the fuse fatigue... which may or may not apply to the type of fuse you used.
I was on holiday for a few weeks, so I had a day here and there I could spend on my hobby. So, logically, I looked at the Aleph J schematics again and considered what else I could do to improve the sound even further. The things I've done so far were a bypass of C1 with a copper link, removing a short circuit protection transistor and associated voltage divider resistors, and the work I put into decoupling power supply rails with polypropylene film caps and my favourite Black Gate N-type caps.
So, I kept looking at the schematics diagram hoping the Muse would strike... and then I realised that 2k trimpot that sets DC offset should be treated the same way as I/V reisistor in current-out DACs. What I had there before the mod, was a Bourns 2k trimpot, but I decided to find a suitable discrete resistor replacement to see what would happen with the sound. The more I looked at it, the more I wondered how the hell I didn't realise the importance of that 2k resistor before...?
So, I ended up using my favourite graphite film resistors, 1.2K each, 0.25W. And the result was mind-boggling. My kids thought that I bought a brand-new amplifier. What stood out was the speed, extension at both ends... and in particular the top-end extension in "all directions".
The resistor choice (carbon, metal film, Rhopoint (could be inductive!), TX2352, Holco...) would depend on your personal listening preferences and definitely on the rest of your system. I remember achieving very good I/V results with TX2352s, and TDA1541 / DDDACs... but in the case of Aleph J, the carbon films worked best for me.
So, if you do not have better things to do this 2024, put the above mod at the top of your list (of improvements).
So, I kept looking at the schematics diagram hoping the Muse would strike... and then I realised that 2k trimpot that sets DC offset should be treated the same way as I/V reisistor in current-out DACs. What I had there before the mod, was a Bourns 2k trimpot, but I decided to find a suitable discrete resistor replacement to see what would happen with the sound. The more I looked at it, the more I wondered how the hell I didn't realise the importance of that 2k resistor before...?
So, I ended up using my favourite graphite film resistors, 1.2K each, 0.25W. And the result was mind-boggling. My kids thought that I bought a brand-new amplifier. What stood out was the speed, extension at both ends... and in particular the top-end extension in "all directions".
The resistor choice (carbon, metal film, Rhopoint (could be inductive!), TX2352, Holco...) would depend on your personal listening preferences and definitely on the rest of your system. I remember achieving very good I/V results with TX2352s, and TDA1541 / DDDACs... but in the case of Aleph J, the carbon films worked best for me.
So, if you do not have better things to do this 2024, put the above mod at the top of your list (of improvements).