[Syzygies]

I've always bought the party line on power cap voltages, but now I find myself wanting to fit 12 NiMH AAA cells into my Hammond J16 PIMETA, to give me a bit more voltage before the AD843's start clipping as the battery runs out. However, I used 16V caps, and 12 NiMH cells briefly reach 18V fully charged.

I suddenly don't see the point of pulling them now. They cost, what, $3? If my virtual ground collapses, I've got worse troubles, I can replace them with 25V caps during that repair.

Or could cap failure with a ground collapse wreck my $200 Ety 4s phones, something I could avoid by replacing the caps now?

 

[drewd]

There are two ways to look at the situation - as long as nothing goes wrong with your amplifier, no worries.

If the virtual ground collapses and the 16V caps are sitting across 18V, they're not going to instantly blow up. The dielectric will have to heat up until the cap fails. It won't happen quickly and you'll notice an obvious problem with the amplifier well before that happens.

As a practical matter, you're probably safe. From an engineering standpoint, it's bad design because you've allowed negative safety margin for a critical part. If the cap fails due to overheating (admittedly, an unlikely event in your situation), it's hard to say if it will fail open or short. The failure mechanism is boiling off of the electrolytic material which has the potential of allowing the two plates to contact each other.

But if you're going to bust open the case and make changes to the batteries, why not take a few extra minutes and replace the four caps? Then you have no worries.

-Drew

 

[Syzygies]

I'd agree. Unfortunately, I don't have 25V caps that fit until my next order, and I can't imagine life without this amp, so I'm going to wing it until the caps come. I'm committed to this hobby, I'm willing to always have on hand the parts for another amp, various friends have expressed interest in amps, so I'll "overstock" from here on out. (Funny how it takes a long time to work out how to make the first one, but making another is then easy.)

Most of the time, the battery will be below 16V. However, planes tend to crash on takeoff and landing, I'd expect the ground to go just after charging, Murphy's law and all that...

 

[mono]

By the time I pull batteries off a charger, put in a device and it sits for a moment, the batteries are already down to ~ 1.45V. This is considering almost every battery I've ever had, (though I admit I can't remember measuring every single one) so I'd be surprised if you ever see the full 18V.

I don't think you will have any problems, being that the caps must be slightly over-spec'd, the amp won't be operating at ~18V for long if ever, and the complete collapse of virtual ground is more significant than 4 caps going out, which they are even less likely to do being so underused, the ESR is heating them up very very little relative to their temp rating, it's entirely conceivable that they could run at 20V+ and any heat would disperse long before a critical range. But, suppose they did eventually pop, at that point you'd notice and could replace them. From such a mild condition they shouldn't explode violently.

How low do you want to run the batteries though?
I've a Pimeta that runs from ~ 7.2V, 8.8V max. It can run OPA2107 at that voltage. It has a few tantalums stuffed onto the board in odd locations but I'm not even sure it needed them, was just playing around and had a few extra tants. Point being, NiMH cells shouldn't be ran much lower than 1.0V anyway, you might be better off using some other strategy than getting every last mA possible out of them. See if your Pimeta works properly per the application @ 6V... If it does you'd be looking at draining each cell (on avg.) below 0.5V which is too low, IMO, especially after the cells age and maybe don't hold same capacity anymore, making it more likely to reverse one if it's the first to be emptied since they're in series.

 

[tangent]

Quote:

Or could cap failure with a ground collapse wreck my $200 Ety 4s phones

Other way 'round, actually. Vground collapse probably kills your phones, no matter what the rail caps do. Think about it: IG and OG are the same (if only indirectly, in the case of the PIMETA) and the headphone jack's ground lug is tied to OG. So, both channels now see DC-only voltages if music is going through the amp.

EDIT: I should say that the reason I promulgate this "party line" is for problems during building the amp. No one wants to replace their caps several times during the build while they work out problems in other parts of the circuit. Vground should not ever collapse once the amp is working. But then, airplanes should never fall out of the sky, either.

 

[Syzygies]

Quote:

Originally Posted by tangent EDIT: I should say that the reason I promulgate this "party line" is for problems during building the amp.

Interesting, this suggests using one power supply while building and testing, and allowing a higher voltage supply after the amp checks out.

My 12 AAA cells charge to about 17.3 volts, and are a shade above 16V at first afterwards, quickly descending below 16V. I think I'll leave these caps in; my charging jack disconnects the amp, so it only sees the 16V.

Edit: After a seriously full charge, my battery pack starts out at 17.2V in initial use.