Running Pimeta v1 or v2 on single 9V cell

I would expect right around 4 hours of useable battery life with a Pimeta 2 that is powered by a single 9 volt NiMH battery.
Right around 8 hours from a Pimeta with the same battery and a single buffer per channel.

9V = +/-4.5V rails, and this is when the battery is not near depletion. This is too low for AD8620. The recommended minimum supply voltage is +/-5V, and add the fact it is not a rail-to-rail opamp (it clips at 1V below each rail), results in a rather sub-optimal combination.

In case MisterX's hint is too subtle, what you have there is a battery, not a cell. A battery contains two or more cells. Chemistry dictates cell voltage. NiMHs are 1.2 V per cell, alkalines are 1.5 V per cell, lead/acid cells are 2 V, and lithium cells vary between 3.0 and 3.6 V depending on flavor. All of these are nominal voltages; do some Googling if you want all the hairy details.

A 9 V battery is composed of 3 to 8 of these cells, in series, depending on chemistry and the exact terminal voltage and capacity they're shooting for.

It's doable, with that chip. My tests found that it doesn't start clipping until about 6-7 V under headphone load. It's probably closer to 5 V with no load, as in a PIMETA v2, where the headphone load is handled by the buffers. You almost certainly lose out on some usable capacity, and the op-amp may not sound its very best, but I've done it successfully.

Try 2x9V in series, if you have an extra battery one hand. Realize, however, that having many cells in series brings its own problems. TANSTAAFL.

Also consider trying a rail-to-rail op-amp. The AD823 or AD8066, for instance.

Actually, the high current draw is one of the reasons not to string too many cells in series. There's a constant (small) amount of resistance per cell, so adding more is a tradeoff, not a clear win. Another problem with too many cells in series is an increased danger of cell reversal. There's a reason you don't find 18 V batteries on the commercial market.

(Before anyone goes Googling to find such a thing, I have no doubt one could be found. Before you post, please consider how many sigmas they are out from the peak of the sales volume curve. )

All true. The datasheet does say that, and the op-amp does clip as you describe.

The thing is, a 5 V rail-to-rail supply less a volt per rail leaves room for about 1 Vrms, enough for many headphones. The terminal voltage for a 9.6 V NiMH "9V" battery starts dropping fast below about 8 V, where we have room for nearly 2 Vrms to the headphones, which covers just about all dynamic headphones on the market.

1x9V may not be ideal, but it may be practical.

If it's important to run the amp on one 9v battery for some reason, the AD823 would be a reasonable choice.

The AD8620 loves to run from from a 16 to 24v supply, so a pair of good rechargeable 9v may be a better idea if you like the chip.

That's highly suspicious. To get 0.9 V of drop at 60 mA (your claimed current draw from the other thread) the internal impedance of that battery would have to be 15 Ω. Maha, unfortunately, does not offer datasheets for their 9V NiMH batteries. I've gathered what data I can on the web from competing manufacturers, and the highest number I saw is 6 Ω for a battery about two design generations back. Current ones are more like 2-3 Ω.

Either your battery is way out of spec -- dead -- or your PIMETA is drawing several times 60 mA.

I have a PowerEx 9V NiMH here, too, but it's dead at the moment. I'll have to wait until it charges up to test it.

I just did one test (maybe more later) on my Maha PowerEx 300 mAh battery. (Probably 7 cells, 8.4 V) After taking it off the charger and letting it sit for a few minutes to stabilize a bit, I put a 105 mA smarty load on it for a few minutes. I tested the voltage under load, removed the load, then gave the battery a minute or so to recover. I got about 5 Ω.

The datasheet specs are at 1 kHz, not DC like my test, so they're not giving it as much time to recover. My battery's probably still in spec at 1 kHz. I think my test is more fair for a near-constant load like a PIMETA, however.

New test, after leaving the same battery alone for 6 hours:

Starting voltage: 9.7637 V
After 60s of constant 105 mA load: 9.0930 V
After 60s of relaxation: 9.6408 V

Depending on how you want to look at it, it's 6.4 Ω or 5.2 Ω. Do you see behavior anything like that with your battery and amp, headfone?

IMO The better question still remains unanswered: what's the current draw when the amplifier is actually amplifying something?

If you don't mind I have a question regarding the C6 caps for the left and right channels.
Why are they not included on the schematic if they are considered "non-optional"?

Part Selection Guide

Because EAGLE's schematic editor exists primarily to support layout. There are a few places in the schematic where I had to compromise human readability for layout reasons. In the case of R6/C6, it's because all the other resistors use that dual footprint for two resistor options. Here, we're using it for two components.

While there are ways around these difficulties -- build a special R6/C6 component, in this case, to show the two separate parts on the schematic, even though it looks the same as the other resistors on the PCB -- better to just redraw the schematic in gschem, as I've done before. I'll get around to that one day.

Or a 2 second touch up with an image editor and nobody would be the wiser.

Well, you could build one, it's not hard. One IC, one power resistor, and a small collection of cabling and casework supplies. Throw in a heat sink, if you like.

If you really didn't want to do that, your PIMETA is probably suitable for this test. Just be sure to double-check current draw while doing the test to make sure it stays constant.

So the question is, is the Maha charger broken, or, more likely, is it right to refuse to charge those batteries?

I have a Maha charger, too -- 4x AA/AAA -- and the worst I can say about it is that it's oversensitive in determining that a cell is bad. If I get a red light and pull that cell out for a sec and put it back in, it generally charges right up. That charger's bad cell detection isn't worthless: if it rejects a cell several times, I know it's bad.

Not just a pretty picture:



I use one of these in it, though. And, I don't do hi-fi listening portably that much, so it doesn't get much exercise.

Tell you what, though, I'll swap that amp in for a desk amp I use here at home sometime over the weekend.

The 9V NiMH datasheets I have specify behavior at "2C" loads, or around 300 to 350 mA.

You should expect less than 4 hours of run time under such conditions. If that's enough, then your battery choice probably isn't a problem. If it is, take a look at that LiIon rechargeable I mentioned. You can't charge it inside the PIMETA, but I consider that a feature given how dangerous those things are.