Old or New multimeter?

[DrStrangelove]

Originally Posted by pmmunro

Chauvin Aenaux even make a hybrid analogue/digital multimeter

https://catalog.chauvin-arnoux.co.uk...5011-3572.html

I rather wanted one of those years ago.

I want one again now.

But I'm not paying £500 for the privilege.

[russell_w_b]

Originally Posted by DrStrangelove

Chauvin Aenaux even make a hybrid analogue/digital multimeter

But I'm not paying £500 for the privilege.

I was given a Mastech M9704 multimeter, fully working, that has a lovely clear LCD display and a proper moving-coil meter combined. But I can't find a thing about it (one function I'm not sure of). Not as practical as a CA5011, though.

But cost me nowt.

I use both analogue and digital instruments. Broadly, but not exclusively, analogue for relative measurements and digital for absolute measurements, or where extreme sensitivity won't result in spurious readings.

[DrStrangelove]

Originally Posted by russell_w_b

I was given a Mastech M9704 multimeter, fully working, that has a lovely clear LCD display and a proper moving-coil meter combined. But I can't find a thing about it (one function I'm not sure of). Not as practical as a CA5011, though.

But cost me nowt.

Looks like a nice piece of kit.

No longer available apparently.

[newlite4]

I prefer to use a new analog multimeter most of the time such as the Maplin model that many of us have. Where close accuracy is required and parallax or pointer position is out of accurate range I reach for the Fluke 73. Top quality analog multimeters are still available such as the Simpson 260 but they are very expensive owing to the manufacturing processes to build the movements.
Neil

[The Philpott]

I would be interested to know how well a middle of the range modern DMM would compare with an Avo 8 MK III as regards frequency response from 5Hz to 15kHz. The manual for the latter claims only small variation in readings.

Dave

[Bazz4CQJ]

The Aneng 8008 DMM is budget rather than middle range but only claims 40Hz to 1kHz. I didn't know that the Avo 8 claims to go to 15kHz!

B

[G6Tanuki]

In what context is "frequency response" measured in a multimeter [analog or digital?] - I'm assuming that the AC ranges stick the signal into a rectifier which is then [hopefully] fed to the meter with some sort of separate scale to compensate for the non-linearity of the rectifier diodes.

An old AVO using selenium/copper-oxide-type rectifiers will behave a lot worse than a modern meter using Schottky diodes and op-amps in a 'zero offset' rectifier system.

[stickfly]

Thanks everyone,
This has been an interesting read.
When I posted this question I was thinking that DMM's didn't exist in the AVO's heyday so when an AVO was repaired or serviced the technician would be using another AVO to carry out the work.
I wasn't looking for the technicalities, more for your opinions really.
One thought was : When the DMM's did become available cheaply would they make the repairs to AVO's easier. (Apart from killing off the AVO which they invariably did).
I still think that using an AVO to repair another AVO is right (Old fashioned thought, maybe.)
I'm in the process of repairing a Model 7 & have been using a Model 7 Mk2, but I do have a DMM and have to admit that the DMM takes up much less space than an AVO
Regards ..... Gary

[Bazz4CQJ]

I think you could only do any serious repair/restoration work on any Avo product by referring to the Avo service manual and I think it very likely that they will call up a specific Avo instrument to do that.

The manual for the Avo VCM163 valve tester calls up the use of an Avo EA113 electronic multimeter, and I suspect that even when they were in production, they weren't very common.

A key issue here is that if the service manual calls up the use of a meter with x kOhms/volt input resistance, using any thing other than that (particularly a lower value) will give inaccurate readings.

B

[Radio Wrangler]

Originally Posted by Bazz4CQJ

The manual for the Avo VCM163 valve tester calls up the use of an Avo EA113 electronic multimeter, and I suspect that even when they were in production, they weren't very common.

A side effect of this is that the hugely inflated prices of the valve testers have infected the EA113 multimeter.

All test equipment manufacturers base their calibration/service procedures around the content of their own catalogues as much as possible. At HP we certainly kept some products going far beyond their natural production lives because of impact on support of other, but current, products.

Have a search for EA113 prices for a little amusement.

David

[lesmw0sec]

Originally Posted by stickfly

Thanks everyone,
This has been an interesting read.
When I posted this question I was thinking that DMM's didn't exist in the AVO's heyday so when an AVO was repaired or serviced the technician would be using another AVO to carry out the work.
I wasn't looking for the technicalities, more for your opinions really.
One thought was : When the DMM's did become available cheaply would they make the repairs to AVO's easier. (Apart from killing off the AVO which they invariably did).
I still think that using an AVO to repair another AVO is right (Old fashioned thought, maybe.)
I'm in the process of repairing a Model 7 & have been using a Model 7 Mk2, but I do have a DMM and have to admit that the DMM takes up much less space than an AVO
Regards ..... Gary

In my time with UKAEA in the 60's, we calibrated AVO's against a rather expensive potentiometer which was in turn set up using a Weston standard cell. About this time, the first (and accurate) digital voltmeters (not multi-meters) appeared, and we went over to using those. I seem to recall an accuracy of 0.01% plus or minus one digit at FSD.

[pmmunro]

As far as I know, there never was an official service manual for the Avometer Model 8 Mk III but the manual for the Mk I & II simply states that a "suitable Precision AC and DC Voltmeter" and Ammeter" should be available to check the accuracy of a repaired meter. (1st attachment)

The nearest thing to a Service Manual supplement for a Model 9 Mk II, which was contemporary with the Model 8 Mk III, is the "Recommended Spares List" which gives the required accuracy figures. In fact it should usually be possible to better these.

The general rule is that an instrument should ideally be calibrated against a standard which is one order of magnitude or better than the instrument under test.

Avo made a "Precision" Avometer (2nd attachment) which doesn't quite meet the 10 times rule but, from the description, was intended for such requirements as checking ordinary Avometers.

The third attachment is from a 1970 short-form catalogue and shows the factory test set-up. The panel meter looks like it might possibly be a Weston (Sangamo).

Concerning the frequency range, the claimed response was to 15 kHz for the the 1970 Models 8 & 9. These used germanium diodes for the rectifier.

PMM

[The Philpott]

Avo 8 MkIII:-

A rather flat line!

[russell_w_b]

Originally Posted by The Philpott

I would be interested to know how well a middle of the range modern DMM would compare with an Avo 8 MK III as regards frequency response from 5Hz to 15kHz.

I accept that a Fluke 87V is a little more than a 'middle-of-the-range' DMM, but the error at 40kHZ is typically:-

* -2.23% at less than 1.3V.
* -1.21% between 1.3V and 5V.
* -4.89% between 6V and 14V.
* -3.18% between 15 and 25V.

All measurements WRT a R&S oscilloscope reading directly in Volts RMS.

[mhennessy]

Originally Posted by Bazz4CQJ

The Aneng 8008 DMM is budget rather than middle range but only claims 40Hz to 1kHz. I didn't know that the Avo 8 claims to go to 15kHz!

B

These cheap multimeters using the DTM0660 (and similar) chipsets are absolutely incredible - in terms of functionality, it's amazing what you get for the money. Just don't use them in a high energy situation - bench use only.

But the AC bandwidth is their biggest problem.

They are 3dB down by 3kHz, and drop like a stone thereafter.

Whereas a more traditional meter might well be inaccurate at higher frequencies, but it will at least tell you that there is something there. Certainly, within the usual audio range of frequencies...

Why do they do this? Because they implement the true-RMS conversion in the digital domain.

The RMS converter chip in a traditional "True-RMS" multimeter is likely to be the single most expensive part on the PCB. They are essentially a mini analogue computer. In single quantities they might cost anywhere between £10 and £20, but it's possible to spend quite a bit more if you want.

Whereas doing this in the digital domain is easy today - no faffing about with expensive processes to make high quality analogue circuitry. But you have to sample the signal before you can calculate the RMS value, and that's great up to a point, but after that, it's Nyquist's Revenge!

For someone not aware of this, it's a real trap. I mention it on my reviews, but after the passage of time - about 5 years - I do think that I could have placed a stronger emphasis on this. Perhaps I'll update them with that in mind...

[Wendymott]

In my quest for "lazy itus".. I have made 3 Lab type bench meters as per photo.. When trying to develop my VHF Transceiver I decided that my AVO 8 Mk3 was too inconvenient as a bench meter, having to switch ranges etc. The HP Dmm 3478 was too "digital".. and in the wrong location.. on the top of the scope stack.
So analog meters it had to be. I was fortunate to buy a 10uA FSD meter "ex Radio shack".. multimeter, and I had the other two in stock. A bit of plastic work soon had 3 meters operational, using the HP to calibrate my home generated scales.
At one time, I, as maybe many of you, thought digital meters the "B all and end all" of the modern age. But not in my workshop anymore. Digital has its place, but Analog has a greater place in my world.

[Radio Wrangler]

That opens up the minefield of AC waveform assumptions in AC range calibration.

Bog standard meters use a rectifier and maybe a filter capacitor, or just rely on the moment of inertia/inductance of the movement. They actually read close to the peak voltage of the applied waveform, but the manufacturers have set them so the scale is marked for the RMS voltage, assuming the waveform is sinusoidal. Any other waveform and you're on your own! Avometers, most analogue meters and most DVMs fit into this category.

Other meters sport "True RMS" AC ranges. Sounds good, dunnit? An extra bullet point on the marketing brag-sheet. It is good. As Mark has said, the bandwidth may be a bit restricted, but you now get some idea what your meter means when waveforms are not perfect. In the late 70'a Analog Devices inc. brought out some analogue-computing RMS to DC converters. They'd recruited Barrie Gilbert to their staff and he's the king of non-linear analogue processing. All of a sudden voltmeter makers high enough up the market to afford them put true RMS AC detection in their posher models.

RMS calibration means that whatever the waveform (within limits!) the meter would indicate the DC voltage with equivalent heating power.

True RMS helps with waveforms you can't correct for by mental arithmetic. But if you have a weird waveform and the voltmeter tells you the RMS value, you'd not have been able to point to it on a scope picture of the waveform very easily at all.

Where true RMS shines is if you need to measure the level of noise. Nothing else can touch it. Everything else has certain types of noise which drives it doo-lally. The old HP8970 noise figure meter used a diode detector, and as it was comparing two different levels of noise, you might think RMS/peak errors would come out in the wash, but no, the noise levels came from different mechanisms and their proportions changed to give the difference in level. So you were metering the ratio between apples and oranges. It really needed proper RMS detection. The bandwidth had to be flat over 4MHz, so far beyond ADI's converters, and non-linearity of readings would be a big problem. So I A to D'd the signal with a fast (for the time) flash converter with plenty of bits, and made a DSP engine on a large programmable logic device which did further down-conversion, quadrature split, digital filtering, RMS detection and finally results integration. So the digital true RMS can be done, and 25 years ago 4MHz was achieved with something like 0.02dB deviation from accurate logarithmic scale over 25dB range of RMS level.

Where matters of waveform assumptions really mess with denizens of this forum is that classic valve characteristics meters run the valve with AC from the mains. The valves rectify their own anode and screen voltages (and currents) as half-wave rectifiers. No meters have scales for this, and what does it mean when you turn the knob to set Va=250v? It's getting a sinewave, and IA is averaged from the half-wave rectified current.

So for checking/calibration VCMs there are waveform average assumptions made right left and centre. And you do need to use a meter with known scale/waveform relationship because you are using it off-piste.

So if you use an AC voltmeter on the mains, and the waveform is not perfect, you get an error on a classic diode rectified meter. How much depends on how grotty your waveform is.

Playing the above logic backwards, the accuracy of a VCM is affected by the mains voltage it gets, which is why there is the "Set ~" procedure, but beyond this, the waveform has a contribution to the error. People's experience show it's not too bad unless some reason gives you something especially bad. Running from a generator isn't a good idea. Don't think about those 'Constant voltage transformers' once used to fix iffy mains voltage, their output tends to be more like a square wave.

So there is no one multimeter does everything, and they all have little gotchas that you become wary of with experience. The gotchas turn up in everyday work, they aren't exclusive to specialist labs.

It's a learning process. With time you grow experience of what you can trust and under what circumstances, and when to go looking for some help.

The biggest thing to be aware of is the risks of measuring on systems with very high power availability. You need properly rated probes, leads and meter. It can cost you your life or your eyesight.

David

[stuarth]

Most analogue meters have AC ranges which are are average responding, but are calibrated for the equivalent RMS values assuming a sine wave. Most are DC coupled, so will show the presence of DC voltage, although some only have half wave rectification, so don't show the presence of negative DC voltages.

I used to find a Heathkit Valve Voltmeter (V7-AU) useful for fixing transistor audio amps for 2 reasons,
1/ it was much higher impedance than passive meters on the lower voltage ranges needed on transistor pre-amps with high impedance bias circuits, and
2/ on AC ranges, the Heathkit, and many similar VTVMs, use a peak to peak responding rectifier, and have P-P scales which read the correct P-P value of the waveform, and corresponding RMS scales assuming a sine wave. If you hooked it onto the output of a transistor power anp in good working order, and drove it hard eg with music, the P-P reading should approach the supply voltage, a useful and quick check that all is well. The AC ranges are also AC coupled with a high voltage blocking capacitor, so you could easily measure P-P ripple on power supplies.

With Analogue meters, the circuits are relatively simple, so you can see what sort of rectifier response it has, the equivalent circuit for Ohms ranges so you can see what voltage and current you are delivering to the device under test etc.

The same is not true for digital meters.

I was suprised to find my digital meter, which claims to show the true RMS voltage, i.e. the DC voltage which would give the same heating efffect, says the true RMS voltage of an AA cell is zero. So it shows the equivalent DC voltage except the DC part! I can see wanting to measure the AC component of the voltage (or current) on a node carrying DC, eg the ripple on a power rail, (but P-P would be better for that), but what if you really wanted the true RMS with AC and DC present, eg half wave rectified mains?. I've no idea what a so-called true RMS meter would read for that.

Are all true RMS meters like that, or can some measure actual true RMS of any voltage (or current)?

Stuart

[mhennessy]

Yes, many multimeters are AC coupled on AC, so only give you the AC component.

Some multimeters have an AC+DC mode, which gives you both together. For example, see the Fluke 187/9 or 287/289.

Alternatively, you can measure both and sum them: http://powersupply.blogs.keysight.co...c-voltage.html

[stickfly]

When I started this thread I never expected things to go this far.
Regards ..... Gary