disconnect PGND and measure the resistance to PROVE that this is your only ground connection.
Hi koldby and AndrewT,
Initially I did want to have a connection from the amp boards to safety ground and even ordered custom made double insulated transformer for this. THis would exclude ground loop issue entirely. Unfortunately the buzz issue is there and I do not know how to get read of it without PGND connection to chassis. I'll twist the transformer leads tighter, but I am not sure it will help. I'll report back if I manage to solve the buzz issue. In principle this forum provide enough info for possible cause/solution, so I'll just use brute force approach and try it all. It may not be the smartest way of solving issues but its a hobby
I'll report back if I find a solution.
PS: half way to the solution of any problem is to understand where the problem really is. At the moment I do not fully understand this part
Initially I did want to have a connection from the amp boards to safety ground and even ordered custom made double insulated transformer for this. THis would exclude ground loop issue entirely. Unfortunately the buzz issue is there and I do not know how to get read of it without PGND connection to chassis. I'll twist the transformer leads tighter, but I am not sure it will help. I'll report back if I manage to solve the buzz issue. In principle this forum provide enough info for possible cause/solution, so I'll just use brute force approach and try it all. It may not be the smartest way of solving issues but its a hobby
I'll report back if I find a solution. PS: half way to the solution of any problem is to understand where the problem really is. At the moment I do not fully understand this part
Start by getting ONE amplifier working with ZERO measured HUM and Noise on the output when the Input is shorted.
Copy that successful solution on the second amplifier.
Then combine the two amplifiers into a stereo Chassis.
This last stage introduces extra loops that usually create more hum and sometimes Hum and Buzz.
If you omit the first and second stages and go straight to stage three (stereo) you can have two, three, or more, sources of loop errors. That is extremely difficult to diagnose.
Keep it simple. Do one stage at a time.
Copy that successful solution on the second amplifier.
Then combine the two amplifiers into a stereo Chassis.
This last stage introduces extra loops that usually create more hum and sometimes Hum and Buzz.
If you omit the first and second stages and go straight to stage three (stereo) you can have two, three, or more, sources of loop errors. That is extremely difficult to diagnose.
Keep it simple. Do one stage at a time.
OlegSH, there is an alternative that you probably would like to avoid, but it is actually what I do first with every amp I build. Eliminate the chassis - attach all the components on a piece of plywood. You would need a couple pieces of metal if you don't have some heatsinks like those in the photo for the LM3886 chips. I make little brackets for the RCAs and posts, but they can hang free. As Andrew says, get one side working and free of hum - then the other. It's a lot of messing around, but it does work and will let you see if something with the chassis is causing the hum. (V1.3 pictured)
Attachments
instead remove all cable connections to the second amplifier.
That is input cables, output cables, power cables and grounding cables.
Leave the PE connected to Chassis.
Remove all amp cables to the Chassis.
This last leaves the isolated secondaries separate from Chassis with lot's of metal work around parts.
Be careful !!!!!!
Now you can debug ONE channel. Take measurements, see what the scope has to tell you.
Expect 1mVac to 10mVac if you have a hum loop.
Expect <0.1mVac if you have eliminated the hum loop.
Test with a 1k substituted for the input short to get a feel for how much noise is included in the hum measurement.
Then swap to a 100k for lots more noise.
the shorted input gives just the intrinsic noise of the amplifier. 0.1mVac is very low nV/rtHz
the 1k0 is equivalent to 4nV/rtHz.
You can use that to guess at what the amplifier is adding to the resistor noise.
100k is 40nV/rtHz, just in case the amp noise is obscured with the 1k0 measurement.
On the scope the peak to peak voltages look much bigger than the Vac reading that you get from your DMM. Firstly because of the noise added by the scope and secondly because Vpp is roughly 6times Vac.
A 100k hum + noise reading of 4mVac on your DMM will be very approximately 24mVpp on the scope, a 5mV/div setting will show 24mV = 48mm, up and down the scope screen, or 24mm above the line and 24mm below the line.
That is input cables, output cables, power cables and grounding cables.
Leave the PE connected to Chassis.
Remove all amp cables to the Chassis.
This last leaves the isolated secondaries separate from Chassis with lot's of metal work around parts.
Be careful !!!!!!
Now you can debug ONE channel. Take measurements, see what the scope has to tell you.
Expect 1mVac to 10mVac if you have a hum loop.
Expect <0.1mVac if you have eliminated the hum loop.
Test with a 1k substituted for the input short to get a feel for how much noise is included in the hum measurement.
Then swap to a 100k for lots more noise.
the shorted input gives just the intrinsic noise of the amplifier. 0.1mVac is very low nV/rtHz
the 1k0 is equivalent to 4nV/rtHz.
You can use that to guess at what the amplifier is adding to the resistor noise.
100k is 40nV/rtHz, just in case the amp noise is obscured with the 1k0 measurement.
On the scope the peak to peak voltages look much bigger than the Vac reading that you get from your DMM. Firstly because of the noise added by the scope and secondly because Vpp is roughly 6times Vac.
A 100k hum + noise reading of 4mVac on your DMM will be very approximately 24mVpp on the scope, a 5mV/div setting will show 24mV = 48mm, up and down the scope screen, or 24mm above the line and 24mm below the line.
Last edited:
Oleg,
if your amp doesn't hum/buzz with shorted inputs (as you mentioned a few posts back), I believe there's no reason to disassemble it - my guess is that the problem is simply the bad (digital) volume control of your sources. You should try using a pot as volume control and running your sources at full volume to see if that is the problem - if you get no/substantially less audible noise that way, that would mean that your amp really is fine.
if your amp doesn't hum/buzz with shorted inputs (as you mentioned a few posts back), I believe there's no reason to disassemble it - my guess is that the problem is simply the bad (digital) volume control of your sources. You should try using a pot as volume control and running your sources at full volume to see if that is the problem - if you get no/substantially less audible noise that way, that would mean that your amp really is fine.
Hi AndrewT,
I repeated my earlier experiments just to confirm the result.
Amp is fully assembled. PGND of both channels connected to chassis.
I only have DMM with auto range.
Now the amp results:
Input open: tiny hum if I put my ear very near the twitter.
DMM measures 0.000 VAC (which means I can not measure the hum with my equipment).
1k0 resistor across the input: almost inaudible hum only when I literally put my ear into the twitter.
Input shorted: amp is literally inaudible.
Am I right to conclude that the amp itself has no ground loop issue?
To butizzle: I also blame my audio sources for the noise issue.
Oleg
I repeated my earlier experiments just to confirm the result.
Amp is fully assembled. PGND of both channels connected to chassis.
I only have DMM with auto range.
Now the amp results:
Input open: tiny hum if I put my ear very near the twitter.
DMM measures 0.000 VAC (which means I can not measure the hum with my equipment).
1k0 resistor across the input: almost inaudible hum only when I literally put my ear into the twitter.
Input shorted: amp is literally inaudible.
Am I right to conclude that the amp itself has no ground loop issue?
To butizzle: I also blame my audio sources for the noise issue.
Oleg
Oleg,
If you think that your noise is coming from your source or part of a system including your source, you may be interested in the white papers by Bill Whitlock. I'm guessing that you are already familiar with them, but just in case, here is a link.
JENSEN TRANSFORMERS, INC. - APPLICATION PAPERS AND SCHEMATICS
In particular, AN004 might be worth thinking about.
Whatever the source, I hope you find it. This kind of problem can drive a person crazy finding it.
Jac
If you think that your noise is coming from your source or part of a system including your source, you may be interested in the white papers by Bill Whitlock. I'm guessing that you are already familiar with them, but just in case, here is a link.
JENSEN TRANSFORMERS, INC. - APPLICATION PAPERS AND SCHEMATICS
In particular, AN004 might be worth thinking about.
Whatever the source, I hope you find it. This kind of problem can drive a person crazy finding it.
Jac
Rev 1.0 - June 2012 boards
I have a set of the "Revision 1.0 - June 2012" green boards that I'm finally getting around to doing something about.
Couple of questions, please:
I think there have been several group buys since I got mine - is this board still OK or have there been significant changes/improvements (to the boards design) over mine?
Assuming I'm going to use these Rev 1.0 boards, is this still the correct BOM (My_Ref FE - BOM) to use to buy parts?
https://drive.google.com/folderview?id=0B-qKRBuaKtzYYWVhZjE2Y2ItZjIyMC00NWJiLThlOGEtN2FjYjUxMzJlZjhk&usp=sharing
Thanks
I have a set of the "Revision 1.0 - June 2012" green boards that I'm finally getting around to doing something about.
Couple of questions, please:
I think there have been several group buys since I got mine - is this board still OK or have there been significant changes/improvements (to the boards design) over mine?
Assuming I'm going to use these Rev 1.0 boards, is this still the correct BOM (My_Ref FE - BOM) to use to buy parts?
https://drive.google.com/folderview?id=0B-qKRBuaKtzYYWVhZjE2Y2ItZjIyMC00NWJiLThlOGEtN2FjYjUxMzJlZjhk&usp=sharing
Thanks
Of course you should build amps with the Rev 1 PCB's. They are excellent amplifiers. The thread you have posted in is dedicated to the latest rev, the "FE", which incorporates a lot of surface mount parts and a few other features. The basic circuit design is the same. I have built several versions of both rev's and lived with each one for a year or more. I am not yet convinced that the newer FE version sounds better than the previous Rev 1.
There are several MyRef threads dedicated to Rev 1, containing extensive discussions about variations to the original basic and "ultimate" BOM. Search those threads if you want more information. This thread is currently occupied with discussing how to make these perfectly quiet amps quiet when they're attached to noisy sources.
Peace,
Tom E
There are several MyRef threads dedicated to Rev 1, containing extensive discussions about variations to the original basic and "ultimate" BOM. Search those threads if you want more information. This thread is currently occupied with discussing how to make these perfectly quiet amps quiet when they're attached to noisy sources.
Peace,
Tom E
They're RC or first bacth boards.
Minor changes, more recent PCBs have no more an alternate arrangement for C9/R10.
Absolutely. The BOM apply to all FE boards revisions.
Tom (Madisonears) probably have been confused by the green soldermask, most of FE boards are blue.