LM3875 and Distortion

[newbie1]

Hi All,

I have just built my first gainclone using an LM3875 and it works! Well kind of...when I first power up the amp it sounds great but after about 45 seconds distortion begins to set in. The distortion sounds kind of like an intermitent crackle which gets worse the longer the amp plays. All the wiring is pretty much P2P and I have the 1500uF caps as close to the chip as possible.

I am using an variable SMPS 30V 3A supply for testing, could this be the root of my problems? I have the parts to build the dedicated PS but I wanted to make sure I could get the amp working before starting on the PS. I have the voltage set for 30.1V and the LCD on the SMPS shows the chip drawing about 0.07A max (this seems a little low?). I am using a single channel (one RCA connector) from a diskman as the input source/preamp.

I don't think this is related to the temperature of the chip, I have measured it numerous times with a thermocouple and it was never hotter than 88 degrees F. It's kind of strange that everything sounds great for about a minute then there is all this distortion. I really don't even know where to begin troubleshooting. I looked at the connections over and over and everything appears to be connected properly (no shorts or bad solder joints). I saw a simillar thread about distortion with the LM3886 but that appeared to be ground related. All my ground connections are connected to a single point on the perf board and are as short as I could possibly make them. I even ran a connection from the star ground to the SMPS earth ground terminal but it didn't make a difference.

Anyone have any suggestions? I am using Doovieman's LM3875 gainclone schematic here:

http://members.aol.com/skimanii/schematic.jpg

Thanks.

[cm961]

Do you mean you're providing the circuit +30V and -30V as V+ and V- or do you mean you're putting the 30V to V+ and the 0V to V-? Or do you mean +15V and -15V for the rails?

Pete

[newbie1]

Sorry about that, I am applying +/- 30V. Another thing I noticed is that if I disconnect the source unit RCA then reconnect it the distortion goes away for a minute or two then returns. The same thing happens when I cycle the power. Plays fine for a minute or two then starts to mildly distort. I applied a 0.47uF cap in series to the source path to see if the distortion was coming from the HU (Sony Diskman) but it didn't filter anything out, after a couple minutes the distortion returned. I suspect it's the PS...?

[uvodee]

I described on the forum 'another gc project.

exactly the same problem I faced, music sounds good for about 30 secs to one minute and then starts to crackle and make a serious distorted sound....

I think you have the same problem I had... grounding ! Does the ground really go to ground???? If not, connect the wire and your probles will go to eternal fields forever!

J-P

I tought I had checked everything untill I found out that the IEC receptacle I used, had a faulty ground pin!!!!!!! it did not let the ground signal come through! ofcourse I always think that this only happens to me!

[newbie1]

Originally posted by uvodee
I described on the forum 'another gc project.

exactly the same problem I faced, music sounds good for about 30 secs to one minute and then starts to crackle and make a serious distorted sound....

I think you have the same problem I had... grounding ! Does the ground really go to ground???? If not, connect the wire and your probles will go to eternal fields forever!

J-P

I tought I had checked everything untill I found out that the IEC receptacle I used, had a faulty ground pin!!!!!!! it did not let the ground signal come through! ofcourse I always think that this only happens to me!

Interesting...so what exactly are you saying here? Does the ground need to be connected to "earth" ground? The LM3875 has no GND pin on the chip so it seems as if the chip just floats. I tried connecting the circuit star ground to the PS "earth" ground but it didn't make a difference. But yes, I am experiencing the same symptoms you were having with your amp. So were you able to fix it?

[uvodee]

I was able to fix it by connecting the ground star to the 0 volt and from there to the ground of the wall plug.....

problem was solved immediately

I hope that what I did was ok for the power supply company, haven;t had any complaints yet.....


J-P

[newbie1]

Well I tired troubleshooting some more. I thought that the 47uF electrolytic in the feedback loop should be changed to a 22uF ceramic instead to be more compliant with NS schematics for the chip. Well, it didn't solve the distortion problem it just decreased the low frequency response for the speaker. So then I got rid of the 1K resistor and the 22uF cap in the feedback loop (bad idea). Apparently the LM3875 needs the 1K resistor to achive unity gain but says the 22uF cap is optional. Well without the cap there is continuous distortion coming out of the chip so I put the cap back in.

If this was a ground problem why wouldn't the distortion be present all the time, and not only after a minute or two of use? It seems kind of like a capacitor maybe slowly charging somewhere.

[uvodee]

but let me tell you, before I fixed the groundloop and connected it to the ground via the wallplug I had tried everything as you mention
it is to say removing components, decreasing values, increasing values adding extra caps etc.....
it is the ground that is bringing you the problem, ofcourse imho!

it is too much of a exact copy of situation I faced.


J-P

and it is also a fact that I tried to change, alter remove, increase value and decrease value of the whole circuit. at the end it was the ground that did not go to the wall that did it!

[runebivrin]

On the face of it, it sounds as if the PSU 0 volt pin isn't connected to loudspeaker and signal ground. This floats ground with the 1500uF decoupling caps as a funky voltage divider.

Mains earth doesn't really matter in this case, although for security it should be connected as well.

Rune

[newbie1]

Originally posted by uvodee
but let me tell you, before I fixed the groundloop and connected it to the ground via the wallplug I had tried everything as you mention
it is to say removing components, decreasing values, increasing values adding extra caps etc.....
it is the ground that is bringing you the problem, ofcourse imho!

it is too much of a exact copy of situation I faced.


J-P

and it is also a fact that I tried to change, alter remove, increase value and decrease value of the whole circuit. at the end it was the ground that did not go to the wall that did it!

Well I fixed it. Turns out it the .33uF cap on the input source was the culprit. It seems as if the cap was charging and after a couple minutes the charge was enough to be audiable. So I removed it .33uF cap and the 22K resistor to GND and the amp sounds great now! Originally I thought it was a problem with my ground wiring too...but guess not. Anyway, the amp sounds really sweet now, gonna build another channel and PS for it.

Anyone know of any good preamp opamps?