Remove LM3886 harsh sound

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Well, the first thing I did was add more capacitance. I added another set of 2200uf caps and a set of 0.1uf caps to the PSU board.

Didn't try it with the B&W's yet, but I can definately say it made a huge difference with the other speakers. It really opened up the sound from the amp. On top of that, it (oddly) literally cooled down the amp. It ran quite warm before when I began pushing it, a little too warm for me. Now it's much cooler, especially considering I now also have it in bridge mode. I guess this is because theres less AC in the power rails the amp has to deal with.

Anyway, like I said, the DRV134 bridge board is complete. I've got two Texas Instruments LM7812 and LM7912 regulators and two DRV134 chips. Both are powered up, but only one chip is driving an amp since only one amp board is completed. This brought about some new problems. Theres a lot of static/hiss being sent into the amp. Although it cancels out in bridge mode, if I try and run it single-ended, you really hear the static/hiss. On top of that, there is a fairly loud buzz being created by the DRV134. Finally, the chips begin clipping the input signals way before the amps are close to clipping. Do you think this is due to not having any capacitors immediately before or after the regulator chips? There are currently no caps on the board. I didn't think it would matter that much, so I didn't put any on. If they are needed I can filter the regulators.

Also, would a circuit such as the following: http://www.diyaudio.com/forums/showthread.php?postid=641075#post641075 work better than the basic no caps no resistors circuit? I wonder because I already have 10k to ground resistors on the amp. Also, what purpose does R3 in that schematic serve? If I used this schematic, the only thing I would add is the R3 resistor. I don't have room for the whole DC offset adjustment circuit.

Thanks again for the great help,
Mike
 
Wow a lot of issues there.

- warning! adding caps usually increases the temp of your chip, not decreases. this is because you have a higher average voltage across the chip as there is a lesser AC drop with more caps. I found this to be true with both increased caps and a regulated supply.

You might want to check your DC values and connections again if you chip is cooler.

- Now lets talk about bridge mode. In bridge mode it should be much hotter than in non-bridge mode since each chip is outputting twice the current.

-OK, maybe you had some type of oscillation before (causing heat) that the bigger caps fixed, so now everything is working as it should be.

- I've used DRV134's with as little as 6V and as high as 15V DC supplies and never had any issue (noise) like you report. It might be grounding or might be the power supplies that you built.


Don't be shy! I'd like to see your schematic (and some pics) and I'm sure others would like to see it too! it would certainly help with these questions.
 
Schematics and pictures still coming, but I've got an update.

First of all, it's too dark to take a picture, as we all know digital cameras suck enough in the daytime, let alone at night. Tomorrow I'll take a couple of pictures.

I'm going to work on the schematic after im done with this post.

Anyway, I gave up on bridging. I decided it was a stupid idea considering the LM3875 amp could only handle 8-ohm and really struggled with the in-walls. Instead, it's paralleled now. This way it's stable into 8/4/2-ohm loads.

So I ditched the DRV134 board and couldn't be happier. The amp is dead silent now and sounds amazing. I connected it to my B&W speaker (only have one channel/2 chips done still) and am extremely satisfied. I can't wait to get the other channel done, add the volume control and finish up the enclosure.
 
Heres the schematic of one channel how it's configured now.
 

Attachments

  • lm3886.jpg
    lm3886.jpg
    39.1 KB · Views: 385
Had a slight setback today...

Being cheap, I used the 35V caps I had rather than 50V caps in my PSU, which runs around +/- 33-34VDC. The small 0.1uf tantalums blew up today and took one of the 2200uf caps. So I tore off the blown caps and one more 2200uf capacitor to balence each rail out, so I'm back to 2200uf until I can get some new capacitors. Until then I'm not going to run the amp too much because I don't need any more capacitors blowing up.

Hopefully I can get some 50V caps, and I'll go for larger values as well, soon.
 
Tantalums suck with any type of voltage over 12V IMO

It's so easy to make them smoke and shoot red glowing stuff. :hot:

Since you are filtering powersupply, use some HV ceramic or plastic caps instead.

And how did the tantalum take out an electrolytic cap? Did the 2200uf cap vent smoke?

I bet even with only 2200 per rail, you can still keep away your oscillation, just add some more 0.1uf caps. It's the High Frequency stuff you have to worry about, and the smaller caps take care of that. Or you can get 0.22uf or larger ceramic or plastic and they work good too.

I'm using 500V 0.18uf ceramic on my amp's rails, they worked good and totally killed any oscillation I had before adding them.
 
I didn't really pay attention to the fireworks. I was more interested in pulling the plug so my house didn't turn to fireworks.

The two tantalums were completely exploded, the tops were burnt and opened. Even the wires directly above them got grilled too. It melted it, but didn't melt through to the wire so its OK.

I don't know how the tantalum blowing destroyed the 2200. One of them right by the 2200s was bulged on the top and the bottom was black and the blue plastic coating was slightly warped. The bottom was very likely caused by heat and black junk from the tantalums blowing, but I didn't like how the top was bulging. So I took it off.

I'm going to be on the lookout for some new caps. When I'm near it I'll look at Amercan Science and Surplus. Thats where I got those 2200uf/35V caps for only 25 cents each. I literally bought out the store (they only had above 20-25 left, but none after I was done there). Hopefully they will have gotten in some 50V or 63V caps.

Also, Nordic, why do I need 63V caps with a 35V power supply? 50V still gives a 15V buffer?

Finally, for the 0.1uf caps, I've got a bunch of brand new surface mount WIMA capacitors. I'll solder some leads to two of the 0.1uf/630V caps and put them into my power supply.
 
Here are the pictures I've been promising.

Cables still need to be neatened up, and will once I'm done with the rest of it.

But here's what I got so far:
An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.