Help, new PSU blowing fuses

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I built my BrianGT chip amp tonight, using an old 100VA Hammond EI transformer for testing. It is 25V CT, and after building the PS and wiring it up, I saw 37V+ and 37V- and expected. Fine.

So then I went on wire up the amp boards, taking care to get everything connected correctly. Fired it up with no signal, and immediately got a loud buzz from the transformer and blew a 2A (fast) fuse. Hmm. Checked everything again for shorts on the chassis (boards are each mounted on a single brass standoff screwed into the chassis, but nothing else is touching. Tried it with just one amp connected, and blew a 3A fast fuse. I then tried the PS alone (as I had when testing), and blew 5A slowblow fuse (rather quickly...)

I can't understand how a PS which worked fine and yielded expected DC voltages could just start blowing fuses like this. I've double checked caps and diodes for correct placement and all seems fine.
----
I think I found the problem on the PS. The jumper I have to connect the CT of the transformer appears to have made contact with a diode pin on the bottom side. I trimmed the diode a bit more and the PS now starts without blowing a fuse.

However, wiring one amp back in still produces a buzz at power on - I killed it before the slow blow fuse goes. I can't see any shorts, and I've cleaned the board with alcohol.

Ok, now I'm out of fuses. I give up for tonight anyway. :-(
 
Ok, I think I found my problem here

This describes it exactly. So I need to use the method explained in the older docs for the 3875 kit here.

But the PS boards are a little different in the current 3886 kit I have. Could someone please confirm this method will work with the newer PS boards?

Thanks!
 
I was able to compare the BrianGT old version PS board with the new version, and get it to work with my CT transformer, and no blown fuses! :)

Wired the amps back in and it still works! DC measured 100mv on one side at power on but quickly went down to 15mv after a few seconds. The other side measured 7V (yikes), but again dropped to 15mv in a few seconds.

I wired it up to my speakers, again checking for DC first, and detected some hum. Wired in my 3020 as preamp, and got more hum. Hmm. Decided to have a quick listen anyway, and it sounded not bad. After 4-5 minutes the heat sinks (smallish P3 style) were hot to the touch, so I powered it off.

Next to locate the source of the hum - I know this is a common topic, so I'll search the archives.

That's it for now, I'm getting the evil eye, so that's all I can get away with on Valentine's Day! ;-)
 
They are aluminum sinks from a P3 or Celeron CPU, about 2" square by 1" deep. They were definitely hotter than 40C, I'd say closer to 70C, as they were too hot to touch for more than a few seconds. I plan on trying some earlier P3 slot1 sinks that are about 2.5x the size, as soon as I can find a matched pair.

Still trying to resolve my hum problems before I try it again.

Thanks.
 
Still trying to reduce the hum in my briangt GC amp. I've read several threads about grounding, including this one, and have tried every grounding arrangement I can think of, and still get terrible hum as soon as an RCA cable is connected, even if the cable is not attached to the preamp. Hum is nearly gone with no cable connected. The only combination which gets rid of hum with RCA input connected is to ground the input SG to the chassis ground, which also appears to short out the amps and activate protection, since output is dead until I power off/on the amp to reset.

I've done everything as per BrianGT's stereo instructions, with the exception of using a single CT transformer, which required taking half of the rectifiers out of the PS board (as per earlier 3875 intructions).

I have this ground setup currently:

mains GND -> chassis
CT -> PG+ & PG- on PS board -> chassis
CHG on each amp -> chassis

The above are on a single ground post, in the order shown (mains on bottom). I have tried it without the second line, but this made not difference.

This is the PS setup I have: page 15 of the 3875 briangt board here.

I have tried some of the suggestions in the first thread I referenced above, but they either made no difference, or shorted out the amp outputs, so I am clearly misunderstanding something. Brian's guide makes the grounding sound so simple, so I wonder if my non-standard CT arrangement is the problem?
 
Hi,
disconnect all the audio grounds from the chassis.
maintain the safety ground connection to the chassis.

Create a floating audio ground for all the returns and grounds that the audio side needs. This can be an isolated tag strip bolted to chassis, or a terminal strip , or simply a 3mm nut and set screw for all the 3mm solder tags on the ends of each of your many ground wires.

Keep the RCA ground at the side most remote from the disconnecting network that links to the safety ground.
I place mine in strict order:- signal, speaker, Zobel (if separate from speaker), PCB power, centre tap, PSU 0v, disconnecting network. Some experimentation in the order of attaching each of these to the bolted floating ground may improve the noise performance.
The RCA may go to either the PCB signal ground or to the floating ground but not both.

When you have the floating ground installed but without the disconnecting network check that your multimeter indicates open circuit between safety earth and audio ground.

If you build a multichannel amp then I recommend a separate audio ground for each channel, each with it's own disconnecting network to safety ground. This is different from usual chipamp advice, so more room for experimentation.
 
Andrew,

Thanks for the help, but I still can't figure this out. I think I understand the need to keep SG and PSU grounds separate, but I can't see how this can be achieved with the chipamp.com boards I have. On the amp board, SG, PG and CHG are all tied together on the board traces. So how do I provide separate SG and PG ground leads to the chassis?
 
Yes, once I spotted my screw up on the RCA input (misread the board labels), most of the hum is gone, I will continue to work on getting it gone altogether.
Now I need to deal with the heat issue. Using a pentium3 heatsink of about 2" square by 1" deep, the sink gets way too hot after 5 minutes of use. I suspect the amp is oscillating, but am not clear what I need to do the prevent this, even after reading many threads mentioning oscillation. It appears the chipamp.com design uses a zobel on the output, so that is covered I think.
So is there a simple test to check for oscillation, and a proven way to prevent it with the design I am using?
 
Too much heat?

I think my amp may still be oscillating, because it gets very hot when driven - to the point where one channel cuts out after 3 minutes of normal listening levels, so I assume it is reaching thermal shutoff because it can't shed heat fast enough. At this point, heat sink temp is about 40C, but this is up from 20C in only 3 minutes, which seems way too fast to me. Sink I am using now is from an early Pentium 2 designed for use without a fan, so it has 3x the area and mass of previous one. 5"x2.5"x1.5" (but I am now using both chips on this one sink). I am trying to keep to this size heat sink to avoid changing case design, and this seems to fit with other sinks I've seen used.

Testing today at idle with no input, only a loose RCA cable plugged in, and test speakers attached, the sink reaches 37C after 1 hour.

Is there a method for determining oscillation without a scope? If I were to borrow a scope, how do I test for it?

I am running 37V +/- as my supply - is this voltage to blame for the heat? My speakers are 8ohm nominal, with a minimum 6ohms according to specs. DCR including cables is 9-10ohms.
 
A bit more on the heat issue. I now have tried one larger heat sink per amp, and the idle temp after 1 hour is a few degrees lower at 35C. Running the amp for 15 minutes on my den speakers (6ohm) from PC source generated only moderate heat, about 45C at the sinks. But running on my main speakers, the right amp runs considerably hotter - 20C - than the left. I changed right side speaker cables (16ga zip cable), but no difference in temps.

Note the right side has always been the one cutting out on earlier tests as it heated up quickly. However the catch here is that I swapped amps - this is now the *other* amp heating up on the right side speaker and input. Both speakers are original PSBs I've had since new (except for re-foamed woofers) and have no problems driving them with my NAD amps.

I've also swapped heatsinks side to side, as they are a bit different. My digicam is dead right now, but one looks identical to these, and the other is very similar, with different fins.

So it appears either one side of my pre-amp, or one speaker, is causing one amp to run hotter. Any ideas?
 
I've now re-wired everything in the amp to include a better layout, with all power pairs twisted, and a different IEC plug with EMI filter and fuse. The EI transformer fits nicely on the aluminum power supply chassis which came with the cable modem chassis I am using, and includes a rubber mat.

My heat problems have not gone away however. After 2 hours of low level listening, the amp heatsinks seemed to stabilize at about 58C (right channel still heats faster, but they both seem to reach 55-60C after 1 hr or so). Of perhaps greater concern is the transformer, which is nearly 50C after 2 hours.
It is an old Hammond EI 100VA 25V CT. At moderate levels, I would think 100VA should be plenty, and the amp seems to have lots of punch and decent bass (better than both my NADs).

However, it puts out way more heat than either of the NADs do at similar levels. I also notice a fairly consistent hum from the transformer, which from this thread seems to indicate I may have DC on the line power. I do have a dimmer controlling a woodstove fan, so this is likely.

The amp is fairly quiet now, with very little hum at the speakers. I do hear a high pitched squeal at the speakers, perhaps 10khz or so. I just switched back to using my NAD 3020 as the amp to let the chipamp cool down, and I notice I have no hum, but the *same* high pitched noise with the 3020!

So perhaps the pre-amp section of the 3020 is causing the HF noise? Could this HF noise lead to oscillation of the amp (perhaps explaining the heat output?) I may be able to borrow a scope, but am not really sure how to test for oscillation, so will start searching for that next...

ps: I am listening to Chopin, and having switched back to the NAD mid-cd, it sounds like someone stuffed pillows in the piano!
 
NAD 3020 causing oscillation?

After a week without any further testing, today I decided to attempt to isolate the source of my oscillation. (after reading many more threads here, I am convinced this is what is causing my amp to heat up).

Here is what I've found:

1. Amp connected to NAD 3020 pre outs and Rotel CD or NAD tuner source will reach 60C after 1 hour or less.

2. Amp connected to PC sound card runs much cooler - 30 to 40C after 1 hour. I have tested with two different PCs, and two different sets of speakers.

I originally thought my Rotel CD could be the problem, as there is a low level HF sound from the speakers when it is connected, and the sound is present with the chipamp, and both of my NAD amps. But I then used the NAD tuner as a source, and saw the temp climb from 30C to 40C within 5 mins, after running at a stable 30C with the PC input.

So it seems my NAD pre amp section could be to blame. It is possible some of the mods I have done recently are the cause (bypassed tone controls, replaced some but not ALL caps). The 3020 runs fine when using it as an preamp/amp combo, with very little heat output, so I assume the amp section is better protected from oscillation.

Not sure where to go next, but I will go over the mods I did on the 3020 and try to spot any mistakes.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.