Lm3875 Help

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

Finished off my second Peter Daniel LM3875 Kit today, and powered it up and nothing happened....Checked and re-checked my wiring and as far as I can tell everything is good, but there is no sound whatsoever:confused:.

I put the boards together a few months ago, but never wired them up until yesterday, so maybe something happened while they were just sitting there? Not sure, is there any way I can find out if the boards are working or not? I'm thinking I might have fried them, whilst being a little over zealous with my fancy new soldering iron....:(

Also, if it turns out I have to take all the wiring apart and take the amp apart is there anything I should watch out for? I know caps can retain a charge for awhile...but not sure how long I should wait, or how to discharge them(if that's the term for it)

Thanks in advance,
Tristan
 
Tristanc1 said:

Also, if it turns out I have to take all the wiring apart and take the amp apart is there anything I should watch out for? I know caps can retain a charge for awhile...but not sure how long I should wait, or how to discharge them(if that's the term for it)

Those caps are pretty small, and won't be particularily dangerous or anything. Waiting won't necessarily work, in some circuits it'll hold a charge for a long time.

to discharge it, I would just grab a big power resistor 10-100 ohms or something, and then touch its pins to the pins of the capacitor for a few seconds. 10 ohms will kill it almost immediately.

It's probably more responsible to do the math and calculate a resistor which will drain the cap over a period of a few seconds. (you should definitely do this for very large caps)

Either way, you're doing a lot better than the people who short them with a screwdriver.


Hopefully you shouldn't need to disassemble it anyways.

first thing to do if it's completely together is probably to check if you've got proper voltages from the power supply. If not, fix that first. ;)

-Nick
 
Thanks for the replies so far, and for the tip on how to discharge caps that will come in handy some day i'm sure :D

I'm not getting any reading on the output voltages with the multi-meter....Well that's not entirely true, it goes up to 12 and then slowly falls down to zero over a period of about a minute or so.. It's weird, it's wired exactly the same as my working one, so one or more of it's parts must not be working. I have a plitron 25V 225va transformer wired for 115VAC, so I should be getting 35V or so rectified right?

So what do you guys think? Should I take it apart and check it over again? If I do take it apart is there a way to test the individual parts to see if they're working properly? could be a bad solder joint or something creating a short somewhere or something, but if that was the case the fuse would've blown correct?

Thanks again!
Tristan
 
Took it apart and put it back together again using a different set of boards, still isn't working but now i'm getting 0.74 on V+ and -1 on V-. Measured the DC offset at the speaker terminals and got 0.

Any Ideas? I know stuff like this is a pain in the *** for everyone to answer, but any help would be much appreciated.

Thanks!
Tristan
 
Did some more tinkering, i'm now getting the correct voltage on both outputs from the rectifier board(stupid newbie mistake I assumed seeing as it's a monoblock that i'd only need to hook up on set of secondaries not both...apparently I was mistaken:D )

Anyways, i'm still not getting sound, every few seconds with my cd player/pre hooked up I get a really quite hiss of distorted music out of my sacrificial speaker. DC offset is 1.4 mv on the + post and 2.7 mv of the -. I unfortunately can't take pictures right now, but I have a feeling it's something really stupid that for some reason i'm not seeing...

Tristan
 
Tristanc1 said:
Did some more tinkering, i'm now getting the correct voltage on both outputs from the rectifier board(stupid newbie mistake I assumed seeing as it's a monoblock that i'd only need to hook up on set of secondaries not both...apparently I was mistaken:D )

Yep.. Still need both rails. :D

Anyways, i'm still not getting sound, every few seconds with my cd player/pre hooked up I get a really quite hiss of distorted music out of my sacrificial speaker. DC offset is 1.4 mv on the + post and 2.7 mv of the -. I unfortunately can't take pictures right now, but I have a feeling it's something really stupid that for some reason i'm not seeing...

Tristan

Offset sounds really good, but I'd suggest checking your wiring yet again.

Maybe play a sine wave into the input. Measure with your voltmeter set to AC. See if you get any significant voltage at the input of the chip. (maybe you've got it wired backwards, or it's shorted to ground or the - in or something)

Just guesses. Good luck ;)
 
Offset sounds really good, but I'd suggest checking your wiring yet again

I keep thinking it's something with the wiring aswell, but i've checked and re-checked a ton of times and i'm 100% sure the wires are in the right places.

There is a new symptom though, the lm3875 is incredibly warm...well actually I shouldn't say warm, it burnt my thumb when I touched it after powering down the amp, and it had only been on for, at the absolute most, 2 minutes. :eek:

So now i'm thinking the fault isn't in the wiring, but something I might have screwed up assembly the board....Any way to check just the amp board for faults? I know the PS is okay, and the Transformer's hooked up correctly so by the process of elimination it must be the board itself.:cool:

Slightly off topic, but i'm kind of glad that I screwed this one up...the last one worked fine, and while that was great I didn't really learn anything, it was more like connecting the dots than anything. I think Peter Daniel, Brian GT etc should purposefully sobotage all their kits when they suspect newbies, such as myself, are involved. That way we'll actually learn about this 'electricity' thing that everyone keeps saying is the bees knees;)

Any other advice, now that I know it's the board itself?

Thanks again!
Tristan
 
You may want to double check that you've got your rails the right way around. Powering things backwards tends to make them very hot.

If not, (or possibly if so) I fear that you may have cooked your amp chip, protection and all :p

You may now be in that dreaded position of having to try another part, potentially blowing that one too, if there is still a problem.

I had a similar experience with a 21" sony monitor. it had several blown parts. Every time I replaced them, I found more, but was never sure if I was finding the true source of the problem, or if stuff was just busting because I kept replacing the other parts that blew first.

I wasted about a hundred bucks before I gave up and tossed it. Fortunately, you've only got a couple parts to blow.

I don't know what you could do to damage the board.

Perhaps you've got the resistors in the wrong places or improperly soldered and are getting severe oscilation?

-Nick
 
I have done the same thing with one of my LM3876T chips. I had it built using the Jaycar 50W design and hooked it up to my sub, had oscillations and the chip's heatsink got very hot!! I fried the chip.

Firstly if yours is the LM3876T (not the LM3876TF) de-solder it from the board, have it sitting on the table and measure the resistance from Pin 1 to the bit of metal on the chip that is attached to the heatsink. It should measure 1-3 Megaohms! The one i fried measured 7.3ohms.

Also measure resistance on pin 3 to the metal back and it should measure 1-3Megaohms. If it measures around 1 ohm the chip is fried.

Finally measure pin 4 to the metal backing and it should measure less than 1 ohm.

If alls ok then check that V+ gets to pin 1 and V- gets to pin 4 when soldered onto the board. Careful not to short between pins. If volts arent getting to the chip check rest of circuit.

Hope this helps.
 
Thanks for the replies guys:)

Measured the LM3875 I got 2.4Megaohms on pins 1 &3 and I couldn't get a reading off of pin 4... But i'm getting -35V on both pins 1 and 4 of the lm3875....:confused: Isn't one of those supposed to be +35? Or is that right? The rails are in the right place, V+ & PG+ goes to V+ & PG+ and the same for negative. It's hooked up like this.

attachment.php


Any more thoughts? With my new found knowledge I think i'll throw the other Monoblock together today and see if I can get that one working.

Fortunately, you've only got a couple parts to blow.

True, and I purposefully bought way more boards than I needed thinking something like this would happen and the parts are cheap too;)

Perhaps you've got the resistors in the wrong places or improperly soldered

That may be, they're all in the right place the only thing i'm suspicous of is the NFB resistor that runs from pin 3 to pin 8 and is soldered directly to the pins on the chip...I remember I had a hell of a time getting that on there for the first board I made....So i'm wondering if this is that board...If I fried that whilst soldering it could that explain things? Because i'm getting distortion out of my speaker it's very quiet and will distort for about 1 second then 2 seconds of nothing and that just repeats, that's oscillation sort of isn't it?

Thanks again!
Tristan
 
You need to make sure that power supply voltages are correct. Use OG point as your ground reference and measure V+ and V- on amp board, it should be approx +/-32V DC.

A quick way to check if the amp works correctly is measuring DC offst without source (or speakers) connected. It should be around 60mV with an input open and drop down substantially if input is grounded. Of course listening test should follow.

Also, you don't really need to solder feedback resistor directly to the pins. There is an alternate spot on board for that resistor marked Rf.

A proper way to install feedback resistor (in a premium kit) was described here: http://www.diyaudio.com/forums/showthread.php?postid=584427#post584427

Info on wiring the amp was posted here: http://www.diyaudio.com/forums/showthread.php?postid=1108820#post1108820 power supply schematic: http://audiosector.com/lm4780 psu.pdf
 
Yes, your power supply voltages definitely sound like the problem.

Follow Peter's test instructions to begin with, but don't bother messing around measuring the offset voltage, since you already know your amp is NOT working properly, and the protection circuitry will likely screw with it.

Double check at the supply board that the polarity is correct (I would expect it is, if this is a board from a kit) If that's fine, then check, yet again, that you haven't crossed the wires up.

Pin 1 should be a Positive voltage from ground, pin 4 should be negative. A measurement between Pin 1 and pin 4 should give you about 70V

You can ignore all that oscilation stuff for now. if you measure 0v between pin 1 and 4, i could make a couple guesses.

Your positive voltage isn't connected at some point and the whole chip is trying to float high.

You have a short circuit between your rails, and something on the + rail has blown open. Maybe a trace on the board, since blown diodes would kill the voltage at the power supply output.

Your chip is just cooked (Always leave this until last if you've got other guesses)

Backwards capacitor? I'd expect it to explode, but you could check it in 1/10 of a second, so it's worth checking.

-Nick
 
So i'm getting voltage in the right places, I get just over 35 on V+ and the same on V- from pin one to pin 4 I get a little over 70 give or take( I have trouble holding the leads still enough and don't want to short anything...) i'm now getting music, but it cuts in and out every second. It was distorted, as I mentioned before, but while I was measuring the voltage on the pins music came on all of a sudden:confused: and it's been that way since yesterday.

What have you got connected to the output?

Just an older crap(but working) speaker.

The lm3875 is still VERY hot, so I think, as Andrew suggested maybe it has something to do with overheating...Would heatsinking help this? I have some smaller heatsinks that I'd planned on using, but i'm thinking I may need something slightly more substantial. I got a reading of 35 C about 30seconds after powering up, with nothing connected to the amp.

Thanks again for all the advice guys, it is very much appreciated!
Tristan
 
It's a heat issue...I just tried it with a pulled Pentium 3 heatsink(a little bigger than the ones I had planned on using) and the distortion, although still very much there, was signifigantly less, and I didn't have any interuptions in the music, whereas before it would cut in and out every few seconds. So i'm still going to need a bigger heatsink I think, but it's good to know I didn't wire it up wrong....*knocks on wood*
:D

That said, even with a 225VA 25V toroid should the chip be getting this hot?

Tristan
 
Tristanc1 said:
The lm3875 is still VERY hot, so I think, as Andrew suggested maybe it has something to do with overheating...Would heatsinking help this? I have some smaller heatsinks that I'd planned on using, but i'm thinking I may need something slightly more substantial. I got a reading of 35 C about 30seconds after powering up, with nothing connected to the amp.

Are you running it without any heatsink at all?! You really do need a heatsink.

EDIT: Missed your latest post.

I would have thought a PIII/PII heatsink would be OK if you can find one. One of my channels is using an old PII heatsink. One of those big slot 1 ones, but without a fan. It works very well indeed.

The other is slightly larger, but the fins are smaller and closer together. It gets really quite hot, but despite this the protection has never kicked in yet ( and no it's not oscillating either :) ).

Attached a picture so you can see what I mean... hopefully! Do excuse the mess of wires :)

An externally hosted image should be here but it was not working when we last tested it.
 
New heatsink works great, doesn't even get warm. Thanks for the help everyone, it sounds really good, no hum, dead quiet, pretty good sound overall i'd say...can't wait to get it hooked up to a decent system instead of the sacrificial one i whipped together....

MarkieMrboo, what all do you have going on inside that chassis? Looks pretty elaborate! And just out of curiosity, how to you go about hooking up 2 transformers in the same chassis? And how does that work? I'll try a search aswell.

Thanks again everyone!
Tristan
 
Tristanc1 said:
New heatsink works great, doesn't even get warm. Thanks for the help everyone, it sounds really good, no hum, dead quiet, pretty good sound overall i'd say...can't wait to get it hooked up to a decent system instead of the sacrificial one i whipped together....


Out of curiousity, roughly how big were the old heatsinks compared to your new ones?

I actually still have trouble understanding how everyone says their chipamps barely get warm :) Mine get pretty hot with what I would consider quite large heatsinks!

I have experienced oscillation before (in another thread somewhere) and fixed it, so I dont think it's oscillation, but it's still pretty hot... I haven't heard SPiKe kick in yet (heard that when the oscillation was there), and it doesn't distort or anything... so I guess it's OK?

MarkieMrboo, what all do you have going on inside that chassis? Looks pretty elaborate!

Nah it's pretty simple really. I think it's just the messy wiring that makes it look complicated :D

It's a separate transformer per channel ( 120va's 25VAC, they don't seem to struggle either! ), so they have separate power supply capacitors. The two littler boards with littler heatsinks are regulators, regulating from post-cap-filtering 36VDC to 28VDC, then the boards with the [bigger] heatsinks are the LM3886 amp boards.

It's pretty messy :) I intend to replace that cat5 wiring with something a little thicker some time. It's working OK for now ;)

And just out of curiosity, how to you go about hooking up 2 transformers in the same chassis? And how does that work? I'll try a search aswell.

I'm not terribly sure what you mean. I have just wired the two transformer primaries in parallel to the mains... and... well... yeah :) It is basically a 'dual mono' setup in a single case rather than split in to two separate cases, and the transformer primaries wired to the same switch / power socket.

Good to hear you got it working anyway though!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.