LM3875 kit puzzling problem: thoughts appreciated
I've completed my first GainClone! Yee-hah! It is based on Peter Daniel's LM3875 kit found on audiosector.com, but despite what I hoped it doesn't really work well yet.
I used a conventional aluminium U-profile, mounting everything on a wooden board, much like the excellent tutorial that came with the kit (although not nearly as neat as the tutorial, but this is just the demo -does it work- version). A picture says more than thousand words, so here it is:
One transformer, one PSU board, powering two amplifier boards.
At first inspection it seems to work. The sound produced sounds reasonably well, although there tends to be some hard-edged-ness to it.
However, what is really bugging me now, is that there seems to be some sound pressure/problem beyond the range of human hearing which results in a feeling of pressure on the ears, it even starts to get a little painful after a while.
After taking measurements with a scope this afternoon, I can at least conclude the following:
- There is no noticable DC-offset at the output
- The chips stay wonderfully cool
- The power supply looks fine, only when turning up the volume really loud it starts to show
- There is no oscillation
Does anyone care to share their ideas as to what might be causing the ear pressure problem? I don't really know how better to describe it, besides there seeming to be a unheard sound that irritates the ears.
The extra puzzling thing is that it almost sounds ok, and it is hard to pinpoint what is wrong with the sound. I'm not an expert at this. It just sounds hard edged somehow.
Something I did notice, using a test CD with a freq. sweep track, is that low frequencies (50 Hz and below) are amplified more than the rest. So my next step would be to add high pass filter capacitor.
My guess is it is related to wrong grounding of some sort. I wired it similar to the kit instructions with one twist:
- I connected PG- and PG+ on the PSU board to each other with a solid piece of wire
- To this wire I connected wires to CHG on each amp board
- And a wire to the chassis (which is grounded)
- Input ground comes together from the input RCA and potmeter on the amp boards.
I assume this is enough, although I am a bit worried I also should have explicitly connected PG+ and PG- from the amp board to the rectifier board.
Any helpful comment is highly appreciated.
Perhaps I can market it as a Mosquito (to buzz away teenagers) :)
i might be wrong, but if you have hi speed rectifiers, and the filtering caps are on the amp boards, the hi frequency from the switching would go through the 10 cm of wire and ground between the boards. I hope you haven't ignored the decoupling capacitors.
and now that i looked up the kit, i can see that you have to use two wires, pg+ and pg-, so you don't have a ground loop like you do right now.
The crucial earthing information for using one rectifier board with two amps is on p.29 of the user guide. I don't think you have followed these instructions, but whether that is causing your problem I couldn't say.
Why not ask in the Audiosector section of the Forum? - the man himself will very likely reply. :)
A high-pass filter is a good idea nevertheless, but not with a corner frequency of 50 Hz. Start with the one in the feedback loop. If the DC offset was not as unnoticable as you thought, you may get rid of the harshness by that.
Ground loop buzz is at 100 Hz (2 x 50). You would hear that quite clearly, not as inaudible pressure on the ears.
What speakers do you use that are able to produce subsonic content at sound pressure levels which bother your ears?
I also had to apply star grounding before the sound was good in this design however 'ear pressure' as you describe it does not sound familiar.
I having the same 'ear pressure' like you said.
but it is from my pc sound card, tested with headphone.
with shielded pc sound card, I can't feel the pressure.
once take the shield off, I can feel the pressure and kinda headache.
cannot listen to any longer than 10seconds.
I feel like going to deaf after remove headphone.
Please do not quote entire posts in this fashion where it performs no useful purpose and eats up bandwidth.
I've listened to Simon's Gainclone and have helped him with taking some measurements.
As I see it, there are two separate problems:
1) A 'pressure' sensation that feels as if there is a very high frequency noise present. I can't hear it, but there is definately something there. Weird thing is that the feeling persists, even after the Gainclone is turned off. Measuring the output did not show any oscillation, however.
2) We measured an increasing amplitude below 50 Hz. A frequency sweep clearly showed the output to rise significantly below 50 Hz and even more below 30 Hz. These frequencies are approximate, as they are measured on an oscilloscope and not verified by a counter. Even so, output is almost doubled below 30 Hz compared to, say, 1 kHz.
This 'bass-boost' has been observed both with speakers attached as well as with a dummyload. The effect of the speaker can therefore be ruled out.
Apart from these two issues, the GC actually sounds allright, there is no obvious flaw in the soundscape.
Anyway, just thought I'd add some more detail and clarification.
|All times are GMT. The time now is 01:51 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2013 DragonByte Technologies Ltd.
Copyright ©1999-2013 diyAudio