[jboehle]

I know this is a dumb question but I don't know the answer. Measuring from the virtual ground (I use one of the empty ground plane holes for the Alps Blue pot - I used the Pana ECJ pot) to the OL gives me 11mV, and to OR gives <1mV. The docs on Tangent's site say to measure from vground to all of the output pins, so I figure that means to the OG too. Measuring to the OG gives me 102mV DC. Am I measuring from the correct points? Do I need to worry about the DC offset at the OG?

-Jason

 

[tangent]

Yes, you need to be worried about that offset. Since the signal is referred to virtual ground at the input, an offset through the ground channel will effectively give an offset to the signal. Don't plug your headphones into that unruly machine until you discipline it.

Is C6G installed?

 

[jboehle]

Yes, C6G is installed, a Kemet 10pF NP0 ceramic. I've tested all of the V+ and V- pins at every op-amp and buffer, and everything is fine there. I am using 2 fully-charged, brand new Plainview 9V NiMH batteries for the power source, wired in series using the pads on the PIMETA board. All 3 buffer positions are stacked 2 high. I am using 8620 for L/R and OPA627AP for G. The op-amps are not biased into class A. R8 is jumpered. I have BC 0.22uF caps on C1, and the C1 trace has been cut on the bottom of the board. I am using RLED and a red LED, which lights up as expected. All resistors are Vishay-Dale RN55 series, values are same as the PIMETA schematic. C4s are 6.8uF Wima.

I have an 8610 that I might try in the G channel. If that doesn't work, I will try just a single buffer in the G channel. There was really no reason for stacking the buffers in the ground channel, other than that I didn't read far enough in Tangent's docs to see that it doesn't noticeably improve anything before I did it. Other than that, any suggestions? I suppose I could've always burnt up a buffer when soldering the stacks together.

-Jason

 

[tangent]

Quote:

I suppose I could've always burnt up a buffer when soldering the stacks together.

Two easy ways to test that:

1. Swap buffer stacks among the channels and see if the problem moves with the stack.

2. Ground channel only: remove the buffers and jumper from pin 3 to pin 6. You can't do this simple test on the other channels because the multiloop configuration messes with it.

 

[jboehle]

Solved it! First I tried another OPA627AP, still had the *same* DC offset. Next tried a different, brand new single buffer in the ground channel, still the same offset. Replaced C6G with a different cap, still the same offset. Put in an 8610 and the offset dropped to a hair over 3mV, perfectly acceptable now!

Any idea why the OPA627AP doesn't work in the ground channel?

-Jason

 

[dokebi]

Because the 627 offset pin is connected to negative rail.

 

[jboehle]

Doh now I feel *REALLY* dumb.

-Jason

 

[tangent]

Quote:

Doh now I feel *REALLY* dumb.

You shouldn't. I should have remembered this, and I should have noted this in the docs. The docs (finally) mention this now.

EDIT: To make the OPA627 work on the ground channel, you just need to lift pin 5.

 

[m0nKeY bUsiNeSs]

Quote:

EDIT: To make the OPA627 work on the ground channel, you just need to lift pin 5.

By lift the pin do you just mean bend it so it isn't making contact with the pad or socket?

 

[dokebi]

Yes.

 

[jboehle]

Do you have to worry about the other 627 trim pin? Sorry my PIMETA is all cased up and working now, don't want to take it apart

-Jason

 

[tangent]

Quote:

Do you have to worry about the other 627 trim pin?

No. The other two nonstandard pins that are variously employed for offset nulling are 1 and 8, which aren't connected to anything on the PIMETA board.