I just finished building my gainclone using the LM3875 chip. I am using this design: http://gainphile.blogspot.com/2007/11/47lab-gaincard-based-amp.html with a radioshack aluminum project box. Here is the problem; when I plug it in I hear a very loud sound coming form my speaker; the first harmonic is at 60hz, second at 130, and the third, much lower one at 180. When I plug my mp3 player into it using a Y splitter cable no sound whatsoever comes out besides the buzz. When I first turned it on, it made a similar sound except with the majority of the sound at 60hz, but, I started to smell hot plastic and immediately turned it off. I felt all of the components and it turned out that the capacitors were hot. I checked every connection and some how, a micro line of solder had bridged between the 1 and 3 pin. I quickly fixed the problem by rearranging the solder connection and further fortifying the necessary solitude of everything else. But still, it makes the buzz sound, but without heating anything up. The strange thing though is that I tested that area that the solder bridged right before I turned it on, and the connections were separate...I even moved a piece of paper in between all of the connections to make sure everything was good. I did close the casing when I turned it on...so maybe there is a chance that a spark occurred between the gap, fusing the wires together? I hope I didn't damage the LM3875.....it is arctic silver epoxy glued to the chassis...so I will likely have to scrap the entire case if the chip is dead (I took me a long time to dremmel all of the holes). I read about grounding issues that could cause something like this; how should I have grounded everything? I essentially grounded each necessary connection wherever on the case, is that a problem?
Thanks for your help, I can post pictures of it if you want.
Thanks for your help, I can post pictures of it if you want.
*update*
I just tried connecting the pre-out from my AVR 745 to the gainclone; well...to my surprise the AVR 745 instantly shut off. I thought this could only be because of a back voltage, so I measured the RCA voltage output from the gainclone and it turned out that it outputs .04 volts for 2 seconds once I turn the gianclone on!!!After 2 seconds it returns to 0v. Why????? Isn't this very bad? Did I really nuke the LM3875 that badly before?
I just tried connecting the pre-out from my AVR 745 to the gainclone; well...to my surprise the AVR 745 instantly shut off. I thought this could only be because of a back voltage, so I measured the RCA voltage output from the gainclone and it turned out that it outputs .04 volts for 2 seconds once I turn the gianclone on!!!After 2 seconds it returns to 0v. Why????? Isn't this very bad? Did I really nuke the LM3875 that badly before?
If the capacitors are hot this probably means your bridge rectifier has shorted or is incorrectly wired, and they are being exposed to AC. Check the tops - they are probably now domed outwards. They will most likely be fried and need replacing.
If you bridged V+ and output, you have most likely killed the chip. If you really need to remove it you can probably do it with a chisel.
I would recommend working in stages. First assemble the power supply part with the rectifier. Test it's output - you should have 35V DC and -35V DC, and practically no AC when you measure with a multimeter. The bridge rectifier should not be hot, and the transformer should not be hot or buzzing.
I take it you are doing this without a PCB. I don't recommend this, as it is prone to mistakes. I would recommend the audiosector.com kit.
If you bridged V+ and output, you have most likely killed the chip. If you really need to remove it you can probably do it with a chisel.
I would recommend working in stages. First assemble the power supply part with the rectifier. Test it's output - you should have 35V DC and -35V DC, and practically no AC when you measure with a multimeter. The bridge rectifier should not be hot, and the transformer should not be hot or buzzing.
I take it you are doing this without a PCB. I don't recommend this, as it is prone to mistakes. I would recommend the audiosector.com kit.
My power supply works perfectly...outputs its 24 volts at DC with only up to .2volt variations every 20 seconds. I checked the wiring a million times anyways and everything is in order. I'm using this transformer http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=129-035
with this design http://diyaudioprojects.com/Chip/Nanoo/ of using 2 4.7uf capacitors in series parallel on the rectifier.
S**t.....I think I did kill the chip. Son of a *****! I swear it bridged itself somehow.....maybe I should not have attempted a point to point design where the components are so close together...
with this design http://diyaudioprojects.com/Chip/Nanoo/ of using 2 4.7uf capacitors in series parallel on the rectifier.
S**t.....I think I did kill the chip. Son of a *****! I swear it bridged itself somehow.....maybe I should not have attempted a point to point design where the components are so close together...
No, 4700uf capacitors.
I'm not sure what a virtual ground is. I noticed that it does not have a ground wire....so I thought that just by having the bottom of it contact the chassis compensates for it in some way. The cases are 4'' apart. The wire distance is around a foot.
Well....lol, I just built another one quickly using a spare LM3875 that I bought along with all of the spare resistors, with the same capacitors. I turned it on, and no smoke!!! But still a humming sound with no music but at a different frequency than before. This one is star grounded, but with the negative speaker line an inch or so away from the center of them (due to a few spacing issues).
What is up with this? Am I just an imbecile? Nothing would make me happier right now than to get this working.
BTW, my receiver does not shut off now when I use the pre-amp output.
I would really prefer to build this independently, but if I absolutely have to, I will buy a circuit board.
What do you mean by input polarity?
I'm not sure what a virtual ground is. I noticed that it does not have a ground wire....so I thought that just by having the bottom of it contact the chassis compensates for it in some way. The cases are 4'' apart. The wire distance is around a foot.
Well....lol, I just built another one quickly using a spare LM3875 that I bought along with all of the spare resistors, with the same capacitors. I turned it on, and no smoke!!! But still a humming sound with no music but at a different frequency than before. This one is star grounded, but with the negative speaker line an inch or so away from the center of them (due to a few spacing issues).
What is up with this? Am I just an imbecile? Nothing would make me happier right now than to get this working.
BTW, my receiver does not shut off now when I use the pre-amp output.
I would really prefer to build this independently, but if I absolutely have to, I will buy a circuit board.
What do you mean by input polarity?
Ok, I am seriously starting to doubt the usability of my transformer with an amplifier. Looking at everything else, like this: http://www.partsexpress.com/pe/show...605&vReviewShow=1&vReviewRand=9766190#reviews mine seems to be completely inadequate. What do you guys think?
Apparently the filter capacitors make sure that the DC is not pulsing; I feel as if this could also be the problem. I have the capacitors wired perfectly, but maybe there is a particular technique involved?
Apparently the filter capacitors make sure that the DC is not pulsing; I feel as if this could also be the problem. I have the capacitors wired perfectly, but maybe there is a particular technique involved?
hi Synesthesia,
Post a picture. I think you are doing more than one thing wrong.
Both those tranformers you posted are not really suitable. You need at least 100VA, most people use 160VA, 225VA or 300VA per chip. The transformers you mention are 50VA and 30VA. 12.5-0-12.5 VAC will work, but not on 4 ohm speakers.
Is the LM3875 a T or TF version? Is it isolated from the chassis?
If your caps got hot you have a wiring problem or the caps are reversed.
I hope you are testing on a crappy test speakers. Test with a cheap source and a cheap speaker. Always measure the DC offset after changing something.
regards
Post a picture. I think you are doing more than one thing wrong.
Both those tranformers you posted are not really suitable. You need at least 100VA, most people use 160VA, 225VA or 300VA per chip. The transformers you mention are 50VA and 30VA. 12.5-0-12.5 VAC will work, but not on 4 ohm speakers.
Is the LM3875 a T or TF version? Is it isolated from the chassis?
If your caps got hot you have a wiring problem or the caps are reversed.
I hope you are testing on a crappy test speakers. Test with a cheap source and a cheap speaker. Always measure the DC offset after changing something.
regards
Ok, I found the problem. I hooked a speaker directly up to the power supply, and what do you know, the speaker shot in one direction (DC) and produced the same loud hum. I will trouble shoot this tomorrow (can't believe I stayed up this late) and likely pinpoint the cause. If I do not, I will let you guys know 
TF
It is isolated.
Nothing gets hot anymore, it was only that one time.
Can't post any pics, I already destroyed it in a rage (didn't hurt any of the parts), and my current makeshift one isn't pretty + its so chaotic you wouldn't be able to tell much anyways.
TF
It is isolated.
Nothing gets hot anymore, it was only that one time.
Can't post any pics, I already destroyed it in a rage
(didn't hurt any of the parts), and my current makeshift one isn't pretty + its so chaotic you wouldn't be able to tell much anyways.
Synesthesia said:
I tell you the trick to know. Power supply voltage.
Looking in the LM3875 datasheet, we can see a lot of measurments & curves.
This is always the case for any chip ( even small op-amps ).
When manufacturer make these curves, these Very good data,
they will use some power supply voltage & set up that this chip will like the most.
If they did not use 'the best' voltage, they wouldn't get those great data.
They would a bit less impressing numbers...
In the case of LM3875, I would say a good supply should be +- 35 Volt (=70 Volt total).
And to find out one transformer thatwould give this voltage after rectifier diodes/recitifier bridge
we use the formual:
VAC ~ VDC / 1.4
(where VAC is the rating of trasnformer)
VAC ~ 35 / 1.4 gives ~ 25 Volt
--------------------------------------------------------------
So result is we should buy one transformer: like 2 x 25 VAC
Now I would rather go down, than up, if no such 25 VAC.
look for to buy: 2 X 22 VAC .. or 2 X 24 VAC
And you will get a power supply close to what National Semiconductor used
in their Audio Labs to get all these great data for LM3875.
Regards
Lineup Audio Transformer Labs
http://lineupaudio.freehostia.com/forum/index.php
... why don't you join us .. cost nothing at all
... we are a bunch of DIYers gathering at Lineup forum .. making good Audio Amplifiers, Projects & Schematics
Yeah, I noticed a 10v drop between the AC and DC outputs.
Thanks for the info on that.
So I took everything apart and determined that the switch was not the problem, the bridge rectifier was not (I tested it with many other ones, all obtaining the same result) and neither was the particular transformer; all transformers (I ordered 3 of them for other projects) create a humming noise. Does this mean that all of them are faulty? What can you use a transformer with this much noise for? The only other thing that might be messed up are the 4.7uf filter caps which I have no real means of testing. Should I order new filter caps to check that out first before I lay down the cash needed for a decent transformer?
Is it possible that the buzz messed up my chip? It has only been exposed to a total of 5 or so seconds of it. If the chip still works, I may yet be able to use my case! (Trust me, there is no way in hell to get the LM3875 off the the chassis now)
Thanks for the info on that.
So I took everything apart and determined that the switch was not the problem, the bridge rectifier was not (I tested it with many other ones, all obtaining the same result) and neither was the particular transformer; all transformers (I ordered 3 of them for other projects) create a humming noise. Does this mean that all of them are faulty? What can you use a transformer with this much noise for? The only other thing that might be messed up are the 4.7uf filter caps which I have no real means of testing. Should I order new filter caps to check that out first before I lay down the cash needed for a decent transformer?
Is it possible that the buzz messed up my chip? It has only been exposed to a total of 5 or so seconds of it. If the chip still works, I may yet be able to use my case! (Trust me, there is no way in hell to get the LM3875 off the the chassis now)
Synesthesia said:
Is that a typo? If you are really only using 4,7µF filter caps they are way to small. You should aim for 10000µF to 20000µF per rail. You can use less with small speakers that don't reproduce deep bass (1000µF and more) or for the upper ways of active speakers (several 100µF).
With 4,7µF only the supply voltage will have a significant ripple that might be the source for your hum.
And concerning hum there is always the grounding to analyse. You'll find links in the Favourite Post section.
Success! I found a 2700uf capacitor and wired it where the 4.7's were; the hum was reduced by 5X! Looking at most circuits, most people do not seem to wire 2 capacitors in series parallel on the bridge rectifier, and only have large ones with the second terminal grounded and the first connected to the +- and the power input/output pin. I'm assuming then by doing that you are accomplishing the same effect as shown here: http://en.wikipedia.org/wiki/Reservoir_capacitor ,am I correct? If this is so, then my capacitors must be blown. I'll head down to radioshack tomorrow...er, today, and get some spare 4700's.
Turns out the capacitors are not blown. Both of them have the same effect as the 2700uf one...except with less noise. This must mean that you need large filter capacitors along with the two that have one terminal grounded. This is strange though considering I've never seen a psu diagram with large capacitors being used as filters. Like in this: http://gainphile.blogspot.com/2007/11/47lab-gaincard-based-amp.html
He is only using the grounded ones.
He is only using the grounded ones.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.