[Filburt]

Quote:

Originally Posted by tangent Why are you testing first with headphones? I'll bet you find that DC offset is unacceptably high. Yes, that's likely. Can you try some low value resistors (10-100) in place of the L1s? If that fixes it, the ferrites aren't the right sort for the problem you're having. One problem with the ferrite solution is that it requires careful tuning: the ferrite has to match the problem, it's not a guaranteed fix. I'm having second thoughts about recommending it as a default.

I tested it first with headphones because, for some reason, it didn't occur to me that it'd have problems. Yes, it was a bit of naiveté on my part...oh well.

Should I order different ferrites? The ones I got are the 50 ohm ones you have in your parts list from Digikey. I don't have any low value SM resistors handy, so I can't test it with those.

I checked the DC offset...it's _gigantic_, assuming I'm testing it right. I used MisterX's instructions, and measured ~1,2v on the right and ~3,9v on the left, which is obviously atrocious levels of offset. This was with the switch just off of the on position (e.g. around the lowest volume position).

While I don't have the spare resistors, I do have the C3 caps. Should I try those?

Edit: Okay, tried C3. Now I get 2,8 on the right and 0,5 on the left...

 

[tangent]

Quote:

Originally Posted by Filburt Should I order different ferrites?

If you don't have a scope, you don't have the information you need to select the proper ferrite. This is one reason I'm beginning to backpedal on ferrites being the default choice. No question, ferrites are the best way to go if you can get away with it, but it's not a foolproof option. (No aspersions being cast here, really.)

Quote:

I don't have any low value SM resistors handy

I guess you didn't plan on adding R1? The low values you'd use there would be suitable.

You could solder-tack some 1/8W resistors to the pads instead. Even 1/4W...you'd just need to bend the leads back under the part.

Quote:

Okay, tried C3. Now I get 2,8 on the right and 0,5 on the left...

What units? Volts? millivolts? Megavolts?

 

[Filburt]

Oh, duh, I totally forgot about R1. Wouldn't that be such a high value that it would seriously damage the performance?

The units on the numbers were volts, not millivolts. Sorry about the commas, if that's being confusing. Right is 2.7v and left is 0.5v, whereas before C3 right was 1.2v and left was 3.9v. This is, admittedly, enourmous DC offset. I was surprised to see such a huge shift post-C3 installation.

What I'm wondering is how others got ferrites to work sufficiently, then. Maybe someone could tell me which value worked for them?

 

[Filburt]

Okay, I replaced all L1s with 75 ohm resistors. Now I'm getting 4,8v on the right and 5,7v on the left. Still a ton of offset.

 

[tangent]

Quote:

Originally Posted by Filburt Okay, I replaced all L1s with 75 ohm resistors. Now I'm getting 4,8v on the right and 5,7v on the left. Still a ton of offset.

Then I wonder if we're chasing the right thing. Are the chips still getting hot? What's the amp's current draw?

 

[Filburt]

Quote:

Originally Posted by tangent Then I wonder if we're chasing the right thing. Are the chips still getting hot? What's the amp's current draw?

How do I measure the current draw?

The LM6172 is getting warmish, while the AD8397 is getting very hot.

 

[Teerawit]

Quote:

Originally Posted by Filburt How do I measure the current draw?

Break the connection between the power supply and your amp. Insert your ammeter (in series) to reconnect the power supply to your amp. What does your ammeter read when the amp is powered?

 

[Filburt]

Quote:

Originally Posted by Teerawit Break the connection between the power supply and your amp. Insert your ammeter (in series) to reconnect the power supply to your amp. What does your ammeter read when the amp is powered?

Er, so I pull out one of the power supply wires, like the positive lead, then what? Stick it to the positive probe on my multimeter and then poke that into the positive power-in?

 

[shimage]

Quote:

Originally Posted by Filburt Er, so I pull out one of the power supply wires, like the positive lead, then what? Stick it to the positive probe on my multimeter and then poke that into the positive power-in?

I think you have it, although I'm little confused by what you're saying. MisterX posted a great picture of it the first time you asked how to measure current draw ...

 

[Filburt]

Quote:

Originally Posted by shimage I think you have it, although I'm little confused by what you're saying. MisterX posted a great picture of it the first time you asked how to measure current draw ...

Oh yeah...forgot about that! Thanks MisterX!

 

[Filburt]

Okay, measured current draw. It's 44mA idle. This is with the AD8397/LM6172 combo. This seems like it might be a touch high, but maybe it's okay.

 

[Filburt]

Okay, now it's reading 5,2v on the right and 6,7v on the left...I haven't changed anything yet...

 

[Filburt]

Teerawit had me measure IG to OG, and that came out at basically 0v (it was fluctuating between 0,001 and 0,000v on my multimeter...mostly at 0,000).

 

[Filburt]

Okay, tried jumpering L1 for the hell of it after removing the resistors from L1 on the Ad8397 (still had it on the LM6172).

Got 5,4v on the right and 6,9v on the left, which is the exact same dc offset as when 75 ohm resistors were in (1,5v) but pushed up 0,2v.

 

[Filburt]

Okay, forgot to put C2 in (doh). Now I get 5,55 on the right and 5,35 on the left. So now we have a reversal of offset. This is with L1s jumpered.

Now here's something odd. I have tried putting ferrites back into L1, as well as resistors, and all they do is *increase* my dc offset by a lot. When I put ferrties in, for instance, my offset shot up to 1,2v between the two, as opposed to the 200mv jumpered. This is really strange.