[JK1dotNET]

Hello Fellow DIYers,
Last night I finally got some more time to work on my Szekeres Amp that I recently built. I've started to notice something that I didn't notice before - I'm not sure if it was always there and I didn't notice or if it has only recently started to appear.

As I turn up the volume, more and more distortion is introduced. I thought maybe it was the drivers in my cans, but I tried them on my CMoy and didn't notice nearly as much at similar levels. I threw a scope across the 15VDC connector points on my perfboard, and at low volume, it looks flat and pretty clean, but as I turn up the volume, I see noise that mirrors the signal that I hear. Loud parts show up even more. The more I turn it up, the more I see.

So, is this most likely the cause of the distortion that I'm hearing?

If so, how do I remedy it? Do I just need some big reservoir caps near the perfboard? I added 1000uF and didn't notice any difference. Do I need a supply that can source more current?

Thanks,
JK1dotNET

 

[Tomo]

Hello,

Post the schematics. I remember you chose different version than the original.

Rail should be kept at constant voltage. You may have mis-wired. Who knows whatelse. I won't be sure until I see the schematics.

Tomo

 

[JK1dotNET]

It's actually pretty standard... I just added your constant current source to the AC Coupled Gain Stage version. Thanks for the help.
Joel

 

[Tomo]

Hello,

This is how you debug your circuit. Pretty standard if you have done some computer programming.

1. Compartmentalize.

In your case, you will separate the amp into the output section and the input gain section. Then you test each module by itself. (That was part of the reason I did not recommend you to soup up Szkeres amp with gain all at once.) Facing all the problems in the amp at once is simply fatal especially when you are not yet confident in your works.

2. Checkpoints

Measure voltage at several points in the circuit and contrast them to the designed specifications or theoretically prausible specifications.

Go ahead and do these. Tell us your results. While you are waiting for our response, you can read up on Common Drain amplifiers.

Tomo

 

[CaptBubba]

Mybe your power supply isn't able to quickly supply current, so the voltage drops when a sudden current is needed to play a loud note, that would sort of explain why you see the noise.

Try adding some caps for a power reserve and see if that helps any (take some semi-large elect. caps and connect one side to the V+ rail and the other to ground) That resulted in improvements in my old amp, but I didn't have any distortion before so I can't tell you if that will fix it.

 

[JK1dotNET]

Well, I found the problem. It was something I never would have suspected. I noticed that when I put a signal generator on one chanel, I got output on both. I searched and searched for a bad solder joint that was shorting the channels together. Finally, I determined that it had to be the headphone jack. I took it off, and sure enough, the channels seperated again. The cheap RadioShack jack had a pretty stupid design that makes it easy to have the right and left plates touch if the solder lug is bent at all. Anyway, that fixed it and it sounds good again. It must have gotten bent somewhere along the lines because I'm pretty sure I didn't hear the distortion to begin with.

While I was debuging the circuit, I found a couple of other problems. The MOSFET gate bias voltage could be a little higher. I have it biased at 7.5V, so the bottom end clips before the top end. This only is a problem if the volume is way up. I guess that biasing at 0.5(Vcc)+0.5(Vgs) would be optimal, but I don't think I'll worry about it.

I also found that at high volume, my Vcc rail moves a little bit. Throwing some caps across helped, but didn't eliminate the movement. However, this doesn't seem to affect the sound at normal listening levels.

Thanks for your help!

 

[Tomo]

Hey,

Your gate voltage should be:

220K
-------------- * (PSU voltage)
220k+100K

So if your PSU voltage is 15V, you should be getting something like 9~11V of Vgs.

Tomo

 

[JK1dotNET]

Nope, my voltage divider is 100k & 100k, so it's pegged at 7.5V. Check my schematic above. That's what I was talking about in my post - I ought to boost that voltage so that the bottom won't clip earlier than the top.