Post by Owen Y on May 23, 2020 10:09:19 GMT 12
Your PSU schematic above (29 Apr) shows these 2 x CTapped windings which are Gnd-connected, so yes, continuity.... ?
|
Post by sadface on May 23, 2020 10:54:31 GMT 12
Your PSU schematic above (29 Apr) shows these 2 x CTapped windings which are Gnd-connected, so yes, continuity.... ? Ah but of course, a rookie mistake.
This immediately became obvious after removing the ac connection to the preamp psu. No continuity between secondaries now so that is a relief. The power tranny is still alive.
The 7812 reg on the speaker protection board does not smell happy however.
Investigation continues.....
|
Post by sadface on May 23, 2020 11:51:34 GMT 12
I found the bastard! One of the tracks has failed somehow....
I thought I heard a very faint arcing sound when I powered the amp up last night. I am guessing that was the broken trace arcing across the break.
I have never heard of a PCB trace failing before. Any suggestions as to how this may have come about?
The board was ordered off Ebay preassembled, I have done nothing to the board other than connect wires. The failure is on the underside of the pcb. The board hasn't been removed from its bracket etc since initial testing a month ago so i don't see how I could have damaged it during handling or assembly?
My thinking is this must be some kind of over current situation since its the only way I can see a trace failing. Could this perhaps be a situation of a largish eddy currents caused by the new ground scheme?
I noticed that where it failed is where the trace narrowed from 2.5mm to 1mm to fit between 2 pads.
I think I will now proceed without speaker protection. I mainly wanted the instant turn off rather than having the sound fade out as the PSU discharges.
|
Post by Owen Y on May 23, 2020 12:32:55 GMT 12
I had an amp fault once & the resulting 400V to Gnd overheated (& vapourised) a PCB track! Possibly a fault resulting in excess current?
|
Post by colinf on May 23, 2020 19:54:23 GMT 12
Well spotted. See if you can work out where the track is in the circuit. If the 7812 doesn’t smell happy either there’s probably a fault on the board. If you want to have fast sound turn-off when you turn off the power you could add a small DPDT or DPST relay in-between the preamp output and poweramp input hot signals. You’d connect its coil to a separate small power supply (with its own rectifier diodes and cap running off the preamp transformer winding) that collapses fast when the power goes off.
AMR-iFi R&D
|
Post by sadface on May 24, 2020 14:45:58 GMT 12
The usage case for this amp doesn't really require fast turn off. I do actually have the spare preamp protection board that I made for my backup preamp. It tested and worked flawlessly but I decided I didn't need it. In the backup preamp I wanted it for pops and thumps but with the fancy DC protection in the LME49720 chips it was unnecessary. I could wire the protection board in between two stages but that sounds like another nightmarish rabbit hole of mountings, brackets and rewiring.
I am not certain the 7812 has been effected. My nose is not sufficiently sensitive to tell if I am smelling burnt silicon on the 7812 or simply the fried track on the underside.
Is it possible this was simply an inrush current fault?
From my understanding. Most SS amp failures involve a PSU failure at startup. Things like in rush currents and dead short conditions in components like rectifiers and filter caps as voltages come up.
|
Post by sadface on May 24, 2020 15:01:28 GMT 12
By my maths.
The primary measures 12.6ohms.
Therefore at 230v/12.6R = an inrush current of up to 18A.....
|
Post by colinf on May 24, 2020 20:12:21 GMT 12
Yes they tend to. That’s why most manufacturers now include an inrush current suppressor in series with transformer primaries, especially toroidals that have low resistance windings because of their efficiency.
AMR-iFi R&D
|
Post by sadface on Jun 1, 2020 21:46:35 GMT 12
G'day Guys,
It fired up again today. No changes other than removal of the speaker protection which necessitated a bit of rewiring.
The hum is gone! The amp is completely silent at the speakers. No hiss, buzz or hum except..... 2 fresh issues have reared their heads.
1) when my hand is touching the volume dial I get a buzz in the speakers. It is not 50hz as I am quite used to the frequency now. It could be 100hz but I think it is a higher frequency than that.
2) I get a nasty squeal and a single thump at turn off now.
My suspicion is that the turn off noises are from the NE5532 chip in the preamp stage. I never had an issue with turn off noises with these LM3875 and the NE5532 is known to exhibit turn off noises.
I will have to do some testing but if it is the NE5532 I will have to deploy the unused preamp speaker protection board I built for the backup preamp. Of course this will be a fresh moutning headache.....
Anybody have any ideas regarding the buzz when I touch the volume control?
Bonus kitty shot:
|
Post by colinf on Jun 2, 2020 0:12:30 GMT 12
From the pic it looks like the volume pot casing isn’t grounded. Use an LME49720 in place of the NE5532, it’s better behaved during switch-off. Nice kitty. As he/she is still asleep I guess you no longer have ultrasonic oscillation problems in your amp.
AMR-iFi R&D
|
Post by sadface on Jun 2, 2020 17:56:09 GMT 12
That's a smart suggestion to just use an LME49720. Much easier that trying to shoehorn the preamp protection board into there. I happen to have them in stock so I can try this after dinner. You are correct that the metal case on the volume pot is not grounded. I haven't encountered this issue before. I used the same 50k jaycar pot on the gainclone boombox, also mounted into wood and also ungrounded but it never presented the issue. Is this a standard issue with such pots or something that can come up occasionally?
PS: I am starting to get excited.
JLCpcb fabbed my pcbs within 48 hours despite quoting 3 days for black coloured pcbs.
Zee Germans (DHL) just picked them up in Shezhen, China. I will be quite interested to see what the shipping time is like. During lockdown I ordered some parts from Mouser via DHL and they arrived in the usual week or so.
From what I heave heard, shipping out of China is taking a bit longer than usual. However my hope is that since DHL have their own planes I might get lucky.
|
Post by sadface on Jun 5, 2020 22:07:48 GMT 12
I changed the NE5532 to an LME49720 chip.
The squeal is gone but there is still a bit of a pop. Probably best described as a fart.
Is this actually likely to do any damage to a speaker driver or is it simply an inconvenience?
Regarding the volume pot: Would it be a good idea to solder a ground wire to the metal body and take it back to the preamp ground?
|
Post by Owen Y on Jun 6, 2020 10:38:38 GMT 12
Vol pot - is usually fixed through a metal chassis, so maybe a Gnd wire attached to a washer, or under a washer, at the bushing end of the pot?
|
Post by colinf on Jun 6, 2020 19:17:26 GMT 12
Ground the casing of the pot any way you can! Owen’s method is a nicer way to do it than to solder to the casing. Some casings are hard to solder to and heating the pot might melt the plastic wiper former parts inside. The sound of it turning off should be ok, as long as the speaker driver isn’t trying to go beyond its full travel. It just doesn’t sound that nice. Remember on an LM3876 / 3886 there is a mute pin. Connecting it to B- power rail through a resistor makes the chip operate, and disconnecting it makes it mute. You need to source over 1mA from pin 8 to get it to operate (refer to the data sheet). Perhaps you could use a switch so that before you switch it off you mute the chip, or hook it up to the protection board relay contacts. From memory the Jaycar boards have provision for a mute switch.
AMR-iFi R&D
|
Post by sadface on Jun 14, 2020 13:49:23 GMT 12
Well dag namit!
I decided to try and put the Corona amp through some tortue testing today to see how hot the heatsinks get.
Instead I find a dead channel. I've traced the issue to the right channel on the preamp board. Changing the OP amp didn't help so I can only assume an issue has developed to do with the PCB.
Queue another round of disassembly to find what I am guessing is either a cold joint or an issue with a trace.
I am beginning to think this project is cursed!
|
Post by colinf on Jun 14, 2020 18:17:57 GMT 12
Ha! Clue in the name? Check power supply on the pins of the opamp first, then check for cold solder joints. It’s a simple circuit so you should find it pretty soon. What does kitty think?
AMR-iFi R&D
|
Post by sadface on Jun 14, 2020 20:47:11 GMT 12
Kitty reigns from the royal bed demanding to be served fresh bird.
Its a single OP amp for both channels in this one so I know the opamp works. I've checked continuity all around the right channel and everything beeps except the output capacitor which I can't get to without pulling the board out.
I think it might be time to shake the boss down for some parts money. A salvaged home theatre amp chassis would make things much much easier.
I've been toying with the idea of rewinding the secondary side of the power transformer from 35-0-35 to 25-0-25.
|
Post by Owen Y on Jun 15, 2020 11:03:41 GMT 12
I've been toying with the idea of rewinding the secondary side of the power transformer from 35-0-35 to 25-0-25. I've thought about doing that task too, now & then, ie unwinding of adding turns to a PTX sec. I guess it depends on the type of TX. Toroid (unpotted) should be easier, perhaps, than EI (bolted, clamped).
|
Post by sadface on Jun 15, 2020 17:33:38 GMT 12
The easiest approach I've seen with an EI core TX is to simply cut off the offending secondaries with a hack saw and rewind rather than try and remove a certain number of turns. It does of course assume that one is very careful not to compromise the primary.
I just won an auction for a box of transformers on trademe so I will run the experiment on one of the smaller and less useful ones.
I've got a large sized tube preamp project brewing quietly in the background that requires separate transformers for each channel. I cannot find any transformers with the right combination of HV and LV secondaries at a decent price internationally. I've emailed Brian Smith in Foxton for a quote on some custom jobbies which come out at about $100 each. I could use separate toroidal TXs which would be relatively cheap but of course take up a fair amount of space. The other option I've been considering is buying a pair of ebay R core transformers with the right HV secondary and then rewinding the LV. It would be about half the price as an estimate. From what I have seen. R core transformers are very simple to rewind as the secondaries are on a nice little bobbin far away from the primary.
But of course I would like to keep local little guy in business so we have the old situation of the wallet vs the heart.
|
Post by Owen Y on Jun 15, 2020 21:59:27 GMT 12
I have, in the past, had a local TX winder to rewind PTXs for me - eg, when one has failed. If he is willing to, it means that he doesn't have to supply the lams & other hardware that can be reused, rewound, reassembled & then re-varnished.
|