diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Chip Amps (http://www.diyaudio.com/forums/chip-amps/)
-   -   Diagnosing hum??? (http://www.diyaudio.com/forums/chip-amps/171166-diagnosing-hum.html)

jstan 1st August 2010 10:19 PM

Diagnosing hum???
 
I recently completed a tube buffered chip amp using Joe Rasmussen's design. It sounds great except for the hum. I'm getting a constant hum (probably 60hz, but I don't have a scope) that is way above what could be considered normal or acceptable to even the least demanding listener (I'm getting about 9-10 mVAC with no input). The hum does not vary with changes in volume, connecting or disconnecting sources, shorting the inputs, or touching the chassis. The heater is supplied with DC and the power supply is a separate unit attached via a 3 foot umbilical. Any ideas as to what could be going on here or things to try?

jstan 2nd August 2010 01:41 AM

Here's an interesting update...I've been poking around with my meter trying to find AC in the amp. I had my meter in frequency mode to see if I could get a reading between ground and V+ for the power amp section. When I ran the meter between V+ and ground, the hum magically disappeared from that channel. I repeated on the opposite channel and got the same result. My first instinct was that I was shorting the power to ground through the meter and basically just de-powering the amp, but when I tried this with music playing, the volume remained. What does this mean?
-Jeff

Gary Diamond 2nd August 2010 01:49 AM

Quote:

Originally Posted by jstan (Post 2260203)
I recently completed a tube buffered chip amp using Joe Rasmussen's design. It sounds great except for the hum. I'm getting a constant hum (probably 60hz, but I don't have a scope) that is way above what could be considered normal or acceptable to even the least demanding listener (I'm getting about 9-10 mVAC with no input). The hum does not vary with changes in volume, connecting or disconnecting sources, shorting the inputs, or touching the chassis. The heater is supplied with DC and the power supply is a separate unit attached via a 3 foot umbilical. Any ideas as to what could be going on here or things to try?

No scope needed just throw another filter cap from ground to B+ if it gets a little better it's coming from your DC supply also just try a .o1 cap

Also how clean is the power supply for the DC heaters, try a filter cap there also

AndrewT 2nd August 2010 10:00 AM

Quote:

Originally Posted by Gary Diamond (Post 2260335)
try a .o1 cap

what??
add a leading zero to avoid not "seeing" the decimal point.
use 0 not o for numerals.
add mF, uF, nF, pF to define your recommended value.

are you saying 0.01uF? = 10nF

jstan 3rd August 2010 05:32 AM

So after trying snubber caps across the diodes and additional filter caps closer to the load with no improvement, I began what I'll call 'desperation experiments': random rearrangements, pixie dust, magic incantations and the like. To my surprise, one of these experiments solved the problem and the amp is absolutely quiet now. I had all of the grounds connected in a star in the ps, running through a single wire in the umbilical, and spreading out again via a star in the amp chassis. When I separated the tube ground from the ss ground, the hum disappeared. The arrangement now has the tube ground (grounds for heater and tube power) running through its own wire in the umbilical, and the ss grounds (grounds for the chip amp power) connected to the chassis and running through the braided shield of the umbilical. I have no idea why this matters, but I won't complain. Still, if anyone can explain why this eliminated the hum, I would appreciate the education. Thanks for all the suggestions.
-Jeff

Gary Diamond 3rd August 2010 05:38 AM

Quote:

Originally Posted by jstan (Post 2261377)
So after trying snubber caps across the diodes and additional filter caps closer to the load with no improvement, I began what I'll call 'desperation experiments': random rearrangements, pixie dust, magic incantations and the like. To my surprise, one of these experiments solved the problem and the amp is absolutely quiet now. I had all of the grounds connected in a star in the ps, running through a single wire in the umbilical, and spreading out again via a star in the amp chassis. When I separated the tube ground from the ss ground, the hum disappeared. The arrangement now has the tube ground (grounds for heater and tube power) running through its own wire in the umbilical, and the ss grounds (grounds for the chip amp power) connected to the chassis and running through the braided shield of the umbilical. I have no idea why this matters, but I won't complain. Still, if anyone can explain why this eliminated the hum, I would appreciate the education. Thanks for all the suggestions.
-Jeff

Ground loop

Redshift187 3rd August 2010 12:47 PM

Are you tube and SS grounds still connected in the power supply?

jstan 3rd August 2010 05:34 PM

Tyler, the arrangement now has the tube grounds and ss grounds separated in the PS (it's like there are seperate star ground for the ss and tube sections, with only the ss grounded to the chassis). In the amp, though, all of the grounds come together in a star.

Gary, how does this eliminate a ground loop? I feel like this would only make a larger ground loop. One day I may begin to understand this stuff!

Gary Diamond 3rd August 2010 07:32 PM

Quote:

Originally Posted by jstan (Post 2261890)
Tyler, the arrangement now has the tube grounds and ss grounds separated in the PS (it's like there are seperate star ground for the ss and tube sections, with only the ss grounded to the chassis). In the amp, though, all of the grounds come together in a star.

Gary, how does this eliminate a ground loop? I feel like this would only make a larger ground loop. One day I may begin to understand this stuff!

Ground loops are hard to understand for me too, trial and error

gootee 4th August 2010 01:40 AM

Here is an explanation of one way that it might have eliminated the hum:

When a ground-return current flows through a ground conductor, it induces a distributed voltage across the distributed inductance and resistance in the conductor. The distributed incremental voltages add together and appear back at the non-ground end of the conductor, as what is sometimes called a "bouncing ground" voltage.

Do we care? How bouncy is it?

That all depends on the current in the ground conductor, and its resistance and inductance. The effect of the resistance is easy to grasp. A larger current through a resistance induces a larger voltage across it, in proportion to the value of the resistance. And changing currents will induce proportional changing voltages. But what about the inductance? That can be worse, because the voltage across an inductor, due to a current through it, is proportional to the RATE OF CHANGE of the current. So that means that even small-amplitude ground-return currents that have fast-changing components in them will be able to induce relatively LARGE changing voltages, back at the non-ground end of the conductor! That's "a bad thing".

So, what happens when you let ALL of your ground returns share a single conductor?

Well, first imagine the resistor that goes to "ground" from the input pin of your amplification stage, or chipamp, or whatever. Its grounded end is THE "ground reference" for your amplifier's input signal (unless it's a differential [i.e. "balanced"] amp). That means that the amplifier's input "sees" the voltage difference between the input signal conductor's voltage and the voltage at the grounded end of the input resistor.

Now imagine all of the ground-return currents from everything in your amplifier, all running through a single conductor. It's a nasty mixture, with some higher-frequency junk and lots of large-amplitude changes from the power supply ripple (bingo!), which your bypass capacitors dutifully put there. And there is probably some sort of distorted image of the output signal there, too, especially if your local bypass caps are too small to meet the demand for large, fast current draws by the amp (which the power supply caps can't quite satisfy because of the inductance in the supply lines, which is even worse with longer lines). But I digress.

All of those nasty ground-return currents create voltages, back at the non-ground ends of EVERY ground-return conductor that branches into the single one. And that INCLUDES the one that connects to your signal-input resistor's "ground reference" end. Can you say "bouncing ground"? So your amplifier's input now "sees" the difference between the input signal conductor's voltage and the bouncing ground voltage at the other end of the resistor! Essentially, the bouncing ground voltage is arithmetically SUMMED with your input signal voltage. Run the grounds separately and that won't happen.

You might not have to run ALL of the grounds separately. But at least the nastiest ones, which have either very large or very dynamic currents, need to be separated. (Note that others that aren't separated might not create noticable hum. But they might add a little bit of distortion.)

---------

Is it all starting to make sense?

Next in the series: "Faraday's Law", or, "How separating natural pairs of conductors [or current paths] (e.g. signal & ground, power & ground, AC & AC, in and out, etc) and making "loop area" between them turns them into antennas for stray time-varying electromagnetic fields (AC and RF), causing HUM and possibly other undesirable effects". Some transmit and some receive! Or both! Avoid it!

<grin>

Cheers,

Tom

"Current does NOT take the path of least resistance. It takes ALL paths, in inverse proportion to their resistances."


All times are GMT. The time now is 12:29 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2