AC reference voltage...options?

[Freq Band]

I'm trying to do some basic calibrations on various older test equipment.
Quite often an AC voltage standard is called for, and at different frequencies.

There are some items on ebay (Fluke 510a, HP 745a/b, etc), but those are around $200 +.

But would a sine wave ran through a power amp also give me a range of AC and freq. ?....in conjunction with an accurate analog AC meter, to check the voltage ?

=FB=

 

[Guest]

Yes.
Some modern PA amps may be used to power drill presses.

 

[Samuel Groner]

One cannot answer your question (unless you're calibrating drill presses...) without knowing:

* what precision
* what frequency range
* what amplitude range
* what source and load impedance

you need. 0.01% at 1 MHz and 75 Vrms is something different than 1% at 1 kHz and 1 Vrms.

Samuel

 

[JohnRoberts]

It depends on how often you need to do calibration, and resolution you need.

If you have a decent handheld VOM and access to calibrated reference, you could chart an offset table. Your handheld may not be very precise but it should be repeatable over a few days for calibration purposes.

Depending on the nature of measurements you want to make a simple oscillator and audio amplifier can be used but note, frequency response errors of the chain are additive so you can't assume an amplifiers 20dB of gain will be 20.0dB at 20 Hz and 20kHz.

Also be careful if running HF sinewaves or square waves through audio power amps they will have networks in the output stage that won't tolerate full voltage output at HF more than a few minutes at most.

FWIW I made a small piece of test equipment back in the '80s and it was reasonably easy to make the signal generator flat over a usable range. Meter response and linearity over a wide range of level and frequency was more difficult.

Have Fun

JR

 

[Guest]

FB, can you specify what you need exactly so we may help you seriously?

 

[PRR]

> But would a sine wave ran through a power amp also give me a range of AC and freq.?....in conjunction with an accurate analog AC meter, to check the voltage?

The basic scheme is:

a SINE generator with a presumed-flat output over frequency, and a low-freq AC meter to spot-check voltage

a SINE generator with an un-flat output over frequency and a wide-freq AC meter or 'scope (which can be checked for flatness against a square wave)

Most DVMs, even the $13 kind, are reasonably accurate above ~~20% of full scale and below 400Hz.

A genuine VTVM with 6AL5 signal diode will be flat from below 50Hz (you need to up-size some caps for 2Hz response) to well over 100KHz. They are non-linear below 1V but there is a special scale for that range. Such meters are a glut on eBay. The probe (always missing) is needed for exact DC readings but not for AC. If you "never saw a connector like that", open the case and solder to the guts.

The massive H-P 200AB will give as much Voltage as any sane calibration should need: 25V. However mine shows a 2dB droop 100Hz-200Hz on a DVM, and same 100-200, 1KC-2KC, and 10KC-20KC on a VTVM which I believe to be flat to far past 100KHz. The problem is getting the small-C end of the 200AB's tuning cap to balance, and I don't bother, just accept it as part of the game.

The Heath discrete transistor sig-gen, when set up, is pretty darn flat, and has its own meter with dubious absolute accuracy but apparently flat with frequency. So a 100Hz check against a DVM with 400Hz bandwidth sets it up for 10Hz to 111KHz. But it is only 10V output. But it does have a semi-precision decade scaler, so 10mV is easy. But the attenuator must be loaded in a specific impedance to give rated attenuation.

If you need small signals, you find or build attenuation networks.

If you need super large signals.... well, are you sure you do? Most instruments take a sane level at the internal input, with a passive attenuator in front. How far off can that passive attenuator get? Sure, 1,300V inputs will blow 600V compensation caps and give a horrible response error, and just moving a wire can cause 1pFd and 1% errors, so damage repair does need some thought and care. But if it does not seem wrong, I'd assume the factory or last-depot calibrated it better than I can, and better than I need.

To measure small signals, use a flat preamp to bring up to ~~1V and use your meter. (Or get an AC mV meter, but these are harder to find in good working order.)

If you really need BIG signals, audio loudspeaker amps are handy for several Amps and dozens of Volts up to 5KHz, but sustained output in the supersonic zone is risky.

If you never need more than a few mA signal, a 6V6 with 400V supply and 5K load resistor will give gain of ~~20, Vout near 100V peak, and Zout near 4K to up near a MHz. If you don't have the parts, bodge a $99 Valve Junior guitar amp: it has the right power supplies, the EL84 gives a bit more gain, you mostly disconnect the narrowband output transformer and use four 1K 5W film resistors and a 0.5uFd 400V coupling cap. Yank the 12AX7 and feed the power tube grid direct.

 

[CJ]

You might grab a good output transformer, like a Peerless or Acrosound, or just calibrate what you have, for the HV calibration.

Run it in reverse. You wont get much step up with a MacIntosh OT, they do not run at a very high ratio, depending on the model.

Your insertion loss should be minus 0.3db for a good piece of iron, which a good generator could drive.

And it should be flat from 20 20. Maybe load the sec with a little R, 100 K maybe?>

Good luck on the HP with the DC offset gyrator, or whtaever that spinning thing a ma jig is.

 

[Freq Band]

I'm not going to fool myself, or you gentleman.....much of this is over my head.
The tinkering I'm doing with older test equipment, is helping me swim a bit.
Reasons: 1) it's relatively inexpensive gear, and 2) I'm not going to learn what's really going on in a circuit, by punching keys on a computer keyboard.


Even though trying to calibrate an AC voltmeter is not audio related, I'm sure there are methods here that will strike an audio tune.

One item I'd like to try to calibrate (or at least check for close-to-factory specs), is an:
HP 3410a AC Microvoltmeter.
http://bama.edebris.com/manuals/hp/3410a/
Page 11 here: (also see note below*)
http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1967-05.pdf

...which recommends an HP 734br Voltmeter Calibrator
Range: 3mV to 3v
Accuracy: +0.2% at 400 Hz

I have (most are not with recent cal):
several oscillators, one up to 20v PP with sweep.
HP 600 ohm attenuator.
Hp 400e voltmeter
3581a wave analyzer
a cal'd HP 3457a digital AC/DC voltmeter
100kHz Tek analog O'scope
Gertsch RT-61 ratio transformer (400Hz max)
http://www.tucker.com/images/images_spec/00000412.pdf

....and other stuff...
http://i5.photobucket.com/albums/y177/Midiot/DSCN2695.jpg

I've read other posts that recommend not to cal stuff yourself, but even if I fail, I don't mind trying.

*see page 16. I may have an opportunity to get one of these.....but as you can guess, not recently cal'd. :wink:

=FB=

 

[Samuel Groner]

The HP 3457A is surely precise enough to calibrate a HP 3410A. Either use one of your oscillators as source or just get the sine wave from an ordinary audio DA converter.

I see no need for a power amp if the maximum level is just 2 V..?

Samuel

 

[Freq Band]

Alright. I needn't make it complicated if it doesn't need to be.
Other topics on the subject warned against using an IC digital meter for AC at higher frequencies, or low levels. According to the 3457a manual, it seems ok.

http://www.gellerlabs.com/34401A%20AC%20zero.htm
http://cp.literature.agilent.com/litweb/pdf/5988-5513EN.pdf

excerpt:
""...Low-Level AC Measurement
Errors — When measuring ac
voltages less than 100 mV, be
aware that these measurements
are especially susceptible to
errors introduced by extraneous
noise sources. An exposed test
lead will act as an antenna and a
properly functioning multimeter
will measure the signals received.
The entire measurement path,
including the power line, acts as a
loop antenna. Circulating currents
in the loop will create error
voltages across any impedances in
series with the multimeter’s input.
For this reason, apply low-level ac
voltages to the multimeter through
shielded cables, and connect the
shield to the input LO terminal.
Make sure the multimeter and the
ac source are connected to the
same electrical outlet whenever
possible, and also minimize the area
of any ground loops that cannot
be avoided. A high-impedance
source is more susceptible to
noise pickup than a low-impedance
source.
..... ""

=FB=

 

[Guest]

I use Wavetek functional generator, from 0.01 Hz to 10 MHz, and HP 400 vacuum tube voltmeter. More than plenty for audio filters, EQs, etc...

Speaking of drill presses, I bought one Chinese press with weak motor, so powering it from a PA amp and functional generator helps to make it usable and regulate speed without swapping a belt.

 

[Freq Band]

OK, so I think I have understood the basics about generating and measuring "x" AC volts at "x" freq.

But another problem arises when trying to generate and measure low millivolts, say 10mV...because in that region, I'm probably measuring noise as well, correct?
And I assume proper impedance matching/correction is important here, too.
(When adjusting the oscillator's output attenuator to near 10mV, all 4 meters I have gave different readings, and I could not duplicate any reading, even with shielded BNC cables. The HP 3457a would at times not give consistent readings either.)

I do have this Gertsch "ratio transformer".
I've read these are used to calibrate/check the division settings on AC voltmeters.
As I understand it, less noise is produced by dividing a larger voltage....instead of dividing 1V to get 10mV, it may be better to divide 10 or 100V using the proper ratio. (??)

Max input voltage on this RT-61 unit:
2.1volts x f (cps), or 350v... whichever is less.

I'm trying to figure out what my best/stable AC source would be, to get exactly 10v or 100v, @400Hz. This probably is what was discussed above and I'll experiment.

------------
I apologize if this subject is a bit off-topic to this forum.


=FB=

 

[Guest]

Do you mean 10 microvolt?

 

[Freq Band]

I might have been tired when I made the above post.

I have four AC volt meters. Three are analog, one, digital.
It turns out that two of the analogs are not working correctly, or are temperamental.

I re-measured everything.

So I am able to get 0.01 volts (10 millivolts) from my oscillators,
And I don't think I was correct about the noise statement.
Noise should not be a problem at 10 mV, for the situation I'm using it in, correct ?

Regardless of that, I am still curious if dividing a 10v signal, would give less noise than dividing a 1V signal.....when dividing it down to the micro or nano volt region. I may try, and see if I can find an answer. I could be wrong about that as well.

Thanks Wavebourn, for staying with me. I'm a bit embarrassed, this is basic stuff, but I have a Masters in Art, not EE. I used to live two blocks from Cal Tech.....none of it rubbed off onto me I'm afraid.

=FB=

 

[Guest]

Most probaly your milivoltmeters are not calibrated, or frequencies are out of their ranges. I've found using a Wavetek signal generator (that does not care of a frequency giving the same output level) that HP 400 vacuum tube voltmeter is quite accurate on audio frequencies.