Multimeter madness

[patricks]

Hi everyone!

Does it take a while for digital multimeters to "settle" before giving an accurate low voltage reading?
The reason I ask is that I've just finished building a phantom power box so that I can use my condenser mics with a very nice preamp that doesn't supply phantom power. I built a regulated 48V power supply, and applied it to the XLR connectors using this schematic.

With the unit plugged in, but the phantom switched off, if I measure between XLR pins 2 and 3 (the ones that carry the audio signal), my (cheap) digital multimeter reads about 0.5V when first connected, then settles down to less than 0.1V over the course of 30 seconds or so. Switching phantom power on doesn't change the reading, but if I measure the voltage between XLR pin 1 and 2 or 1 and 3 (10V because of the zeners) and then go back to pins 2 and 3, the reading jumps to nearly 1V before dropping gradually and settling back at 0.1V or less over the course of a minute.

Is it normal for meters to take a little while to settle on a reading like this? The preamp uses a THAT 1512 chip and is direct-coupled, so I don't want to fry the chip by applying a DC voltage to the inputs.

Cheers
Pat

[GibsonGM]

Well, it IS phantom power, right? It's what you wanted! 

Most DMMs, mine included, (but maybe a $500 Fluke EXcluded) take a moment to sample a voltage.  More or less, depends on the quality (price) of the meter.  Mine samples 3x per second, moderate rate.  RS crap meters might only sample 1x per sec, or even longer!   

Anyway, there is often a bit of voltage that remains in a circuit after you open the switch. The caps in there, C7, and R3, etc...if there is a path to ground in any way possible, they will discharge over some time and you'll see a slowly decreasing voltage (altho R10 is low, so it won't take as long as it could).  RC time constant at work.

AND, sometimes there is a weird phenomenon where caps sort of "recharge" (for lack of a better way to express it) after you've drained them, and with no power in the circuit!  No worries, they won't charge up to enough to be more than a worry about static for sensitive items, but no danger to YOU.  There is a term in physics for this which I can't recall at the moment...like static electricity...

The way around the worry re. your chip may be to attach a bleeder somewhere - I know nothing about how sensitive it may be - thinking it's not an issue?  How much voltage is present when operating - gotta be more than this residual, right?  What if your caps ever got leaky, as they do?   The protection zeners are 10V...that may tell us something about where you need to keep your DC?  Gotta data sheet?

[patricks]

Hi Gibson, thanks for the response! Yup, it's what I wanted, I'm probably just being over-cautious
What you've said about multimeter sampling and capacitor charging/discharging makes perfect sense. Another check this morning (after the circuit had time to discharge fully overnight) showed that even switching on the power supply to the phantom box (with the phantom supply switches off) gives a reading of up to 1.3V across XLR pins 2 and 3 initially. It settles down quite quickly, but that'll be the capacitors doing their morning stretches.

The chip runs off a +/- 15V supply (datasheet here), and from what I can tell, the limitation to the voltage across the input pins is mostly about current. Any voltage difference between the pins will probably be amplified as noise, though, so I think I still want to make the differential as small as possible. I'm guessing the 6.8k resistors aren't as well matched as they could be, so I'll improve that.
The preamp I have keeps the circuitry nice and simple, and doesn't have a lot of the components recommended for use with phantom power. I think I'll use my dynamic mic for today's recording session, then match the 6.8k resistors and add the protective components to the preamp later. Better safe than sorry

Cheers
Pat

[GibsonGM]

Sometimes we worry too much.
Sometimes we don't worry when we should.
The problem is to find out when we NEED to worry! 

[patricks]

'zactly
In this case, I know I haven't taken all the recommended steps to protect the preamp, so I'll revise and reassess. Thanks again for the help!
Cheers
Pat

[PRR]

Seconds isn't the meter. (I had a meter with a long settling time; but it was special.)

The circuit really has that half-volt. Most likely there are caps without ample bleeder resistors. If it is half-built, try adding the stuff you know you should have.

It is not impossible that a Phantom power mike input with a non-Phantom microphone will show DC voltage audio-common "forever". The important thing is no voltage one audio to the other.

A circuit of the actual situation might bring a more specific explanation. (Drawing-out the circuit, you may see it yourself.)

[GibsonGM]

Ya, I probably wasn't entirely clear...cheap meters tend to 'walk', esp. if measuring something that's changing.

In the OP's case, the voltage is actually there, not bled off...

[patricks]

Thanks guys, I'll post a schematic hopefully later today.
Yeah, there's a measurable voltage from audio to common, which doesn't worry me, it's the half volt from audio to audio that's the concern.
Cheers
Pat

[patricks]

OK, here's the schematic in chunks. First, the power supply:

Then, the schematic of how it's delivered to the mic:

And the input to the preamp:

The schematic to the preamp isn't mine, so I've left component values off and only posted the input to the THAT 1512 chip. It's just to show that the input is direct coupled and it's close enough to the datasheet design guide that I shouldn't be stepping on any toes.

Given that I can measure a voltage between XLR pins 2 and 3 with the unit plugged in but the "phantom" switch off, I'm thinking maybe the 10uF capacitor between +48V and ground is the culprit?
After some further though, I'm not sure that mismatched resistors is the issue. I measured them out at 6.75k before installing them, so they'll be within 10 ohms of each other, which is what I've read is recommended. If this all looks above board, I'll install the protective diodes in the preamp (as recommended in the datasheet) and see how it fares.

[PRR]

I assume the LED is wired the other way.

Now put say 47K across the 10uFd. In a half-second it will bleed the 10u from 48V to 18V,  then to 7V 2.5V and 1V in 2 seconds. It takes infinite time to reach dead-nutz-ZERO, but it will be unreadable in less than a minute.

The THAT circuit is probably 99% one of the plans in the data-sheet. There are not that many ways to skin this cat.

[patricks]

I assume the LED is wired the other way.

Whoops! Yup, typo on my part.
I'll try adding the 47k resistor across the 10uF cap and see how it goes. As it is, the voltage settles to 0.1V or less in around a minute anyway, but anything that helps it along is good. Thanks again!
Cheers
Pat