Running Pimeta v1 or v2 on single 9V cell - Page 3

200 mAH last about 5 hours

I have AD8620+ 2x HA-5002 + one led (10K RLED) + AD8022  + TLE2426 = 32.6 mA 

6x1.2V (2500 mAh each) last 70 hours.

you shouldn't have any problems with 9V battery. If the voltage gets too low sound starts to distort and dc offset raises. So must not leave amplifier on for long time. 

Those op amps work with about 8 volts. 

Quote:

Originally Posted by tangent 

I've updated the PIMETA v2 docs to distill some of what we've covered here.

Redrawn schematic, human-friendly

"O" that figures... lol. I should have looked at the PIMETA within the past few years.

I recall that being a 1M ohm resistor, not a 499 kilohm & 4.7pF. O.K v2 vs V1

Akin to the ground channel in the PIMETA v1.0, [C1 on Figure 4 in the BUF634 datasheet]. 10pF or 4.7pF that is the question. 220pF is way to big for an 8 Megohm buffer input impedance vs feedback loop impedance. Would the larger cap (10pF vs 4.7pF) make a bad difference in the audio band?

10pF it may be, as the boards are in, and that is what the silk screen calls for. If it gets installed at all, the provisions are there if need be. There are other ways, to defeat RFI.

P.S. the LMH6321 has a 250 kilohm input impedance, vs the BUF634's 8 megohm impedance.

well till later, I need some sleep. The OPA2134's and BUF634T's arrive today.

Quote:

Originally Posted by zarcondg 

I recall that being a 1M ohm resistor, not a 499 kilohm & 4.7pF. O.K v2 vs V1

Yes, I lowered those resistors because there's no reason for them to be 1M if the largest likely pot value is 50K.  If it makes you happy to use 1M instead, you can.  It will just raise the noise floor a bit.

Answered here in the docs.

Where are you getting 220pF?

I was exhausted, and I needed sleep. I don't want to hijack this thread, so I'll try to keep this short. (that didn't work)

Quote:

Originally Posted by tangent 

Yes, I lowered those resistors because there's no reason for them to be 1M if the largest likely pot value is 50K.  If it makes you happy to use 1M instead, you can.  It will just raise the noise floor a bit.

That makes sense, I saw that part of the schematic [PIMETA v1], scratched my head over the rather high value (thinking you were using the capacitive by-product of the resistor end-caps in the resistor package), and went back to the BUF634 data sheet Headphone amp [Figure 5, on page 9]. I'm using (mid-ref) 1k ohm (neg op-amp in) 10k or 22k ohm (BUF634 out), instead of there suggested feed-back loop (250ohm, 5kilohm). I'll eventually finish the page, and post the design (nothing special).

I wish I had seen that back in 2006 when I first drew up the op-amp and buffer schematic chicken scratch draft. design #17 of 23 (I think there was, excluding revisions to any particular topology). Lots of reading here, much appreciated.

BUF634 datasheet, Page 9, FIGURE 4.  "Boosting Op Amp Current.", [on chart in Figure 4] with OPA627, OPA132.

The OPA134 is not listed, scratches head. To bypass or not? Then I came across your (Tangentsoft) page on op-amps, there almost the same. Bypass it is.

What value? (Answer above link provided by Tangentsoft). It depends... lol

Back on toppic.
With PIMETA v1 battery life, the  BUF634 [bandwidth pin] resistor value may be of importance?

Would battery life testing with "Pink noise" or "White noise" into a predetermined ohm dummy load, rather then just a single sine wave be more accurate or useful at all? can it be done easliy without a full fledged lab?

[? not .]
 

Quote:

Originally Posted by zarcondg 

I wish I had seen that back in 2006 when I first drew up the op-amp and buffer schematic

That section is new to the PIMETA v2 docs, which didn't exist in 2006.

Realize that not everyone uses buffered op-amps for driving headphones.  Also, the BUF634 is a different beast.  As the PIMETA v2 docs say, you have to do tests to find out the best value for a given circuit configuration and load.

Very much so.  It's covered here, and in the datasheet section referenced from that.

Headphone amp designs of this sort are quite inefficient.  Most of the power that goes into the amp gets turned into heat, not into sound.  Testing the amp with a signal going through it won't usually give a huge difference in the measured value, and makes it difficult to compare results against quiescent current values, which are otherwise easy to estimate.  Signal testing has many other legitimate uses.

When you do signal testing, you should use signals that are as pure as possible.  If you don't have a pure source, it's often better to not bother with the test than try to make sense of the results you get from using a dirty source.