Milivoltmeter for audio work

[BeBuLamar]

I was over the tapeheads website where they discussed milivoltmeter for measuring output from tape recorders. Most would say something like the Leader 185/186 etc.. is best but those are not in production any more. Besides I don't think those need to be analog with a galvanometer. A DMM doesn't seem to have good accuracy for the bandwidth from 20-20kHz. I have the Fluke 189 and 289 and I think they are OK but not quite good for those types of measurement.
What would you guys think is a good choice besides buying from the old analog meter from ebay.

[rsjsouza]

I don't know the Leader meters, but I don't quite understand how a galvanometer could have better "accuracy" than a 50000 count meter, regardless of the bandwidth. The Fluke 189 has 100kHz of bandwidth and can show voltages with a minute fraction of a 100-division galvanometer.

One aspect that I think may give this impression is the natural dampening of a galvanometer when shown variable voltages on its inputs - the typical 3 readings per second DMM tends to show a cacophony of numbers under these conditions. However, the bargraph can help show trends and variance.

So, answering your question: I would experiment using the 189 under the same measurement conditions of an analog VOM to see if the information you get makes sense or is useable.

(I haven't done measurements in cassete tapes in decades, so I might be missing something)

[TimFox]

What bandwidth do you need for the audio voltage measurement?
I'm not familiar with current production, but older analog and digital-readout units such as the
-hp- 400GL (average responding, analog meter, 0.1 mV min full-scale, 4 MHz),
-hp 3400A (true rms, analog meter, 1 mV min full-scale. 10 MHz bandwidth),
-hp 3403C (true rms, digital display, 10 mV min full-scale, 100 MHz bandwidth),
and Marconi 2610 (true rms, digital display, 2 mV min full-scale, 25 MHz bandwidth) are nice.  I own one each of all but the 400GL.
I don't know how old the units used by your tapeheads posters are, but the traditional AC millivoltmeters were the tube-based -hp- 400 series, not as wide a bandwidth as the later solid-state units such as the 400GL .

[BeBuLamar]

The meters you suggested are all out of production. As I stated in my post I want to find something current. There must be some made for those types of work today.

[ci11]

You may want to look at some "psophometers" Rohde & Schwarz and Siemens made them in Germany, and Boonton, UP and Ballantne made them in the US.

These are very high precision millivolt meters that can read down to -100dbV. Their response can be dead flat between 15Hz to 100kHz or weighted through various filters, even custom external ones. This is great to dialing in tape head bias. They were used by analog telephone networks to measure noise floor levels to ensure standards compliance and of course, telco's had all the money to afford the very best. Stunning pieces of analog engineering that are no longer made.

The ones I use are Rohde & Schwarz UPGRs. There is one on German eBay right now: https://www.ebay.de/itm/304428989805?hash=item46e161bd6d:g:4o8AAOSwPURiOfUF

BTW - I use a 189 also and there is no way I would compare its bargraph - or any bargraph on any DMM - to how you can characterize AF signals.

[nctnico]

The meters you suggested are all out of production. As I stated in my post I want to find something current. There must be some made for those types of work today.

The Vici VC8145 bench DMM has a good accuracy over the audio frequency spectrum but you have to stay withing certain levels because of the true-rms converter it uses. ACV measurements are specified for 0.8% up to 20kHz while staying within 75% of the measurement range.

As you already found out a lot of DMMs are optimised for 50Hz only. You need to check the specs carefully but there are DMMs out there that are also suitable for audio work.

[Fungus]

One aspect that I think may give this impression is the natural dampening of a galvanometer when shown variable voltages on its inputs - the typical 3 readings per second DMM tends to show a cacophony of numbers under these conditions. However, the bargraph can help show trends and variance.

The 187's lowest range is 50mV so the 51 segment bargraph will struggle to show 1mV wiggles.

One possibility is to use a $2 opamp to amplify the signal so the bargraph shows it better.

[bdunham7]

I was over the tapeheads website where they discussed milivoltmeter for measuring output from tape recorders. Most would say something like the Leader 185/186 etc.. is best but those are not in production any more. Besides I don't think those need to be analog with a galvanometer. A DMM doesn't seem to have good accuracy for the bandwidth from 20-20kHz. I have the Fluke 189 and 289 and I think they are OK but not quite good for those types of measurement.
What would you guys think is a good choice besides buying from the old analog meter from ebay.

The issue isn't accuracy or bandwidth--the F189 and F289 have that covered quite well--it is sensitivity.  A 6.5-digit DMM with a 100mVAC range will have a resolution of 100nV, but will be too noisy and inaccurate below 1mV.  A meter like the old HP403B or the Leaders will have a 1mV full scale range much better suited for such readings.

As far as modern equivalents to the old analog AC voltmeters, AFAIK it is slim pickings.  There's the Tonghui TH1912, a rather strange 38,000 count meter that I've never personally seen, but it does have a 3.8mVAC range.  I don't know of any better-class vendors with an equivalent product, although there are some Aliexpress no-name versions out there.  Another alternative would be a low-noise low-bandwidth preamp, but really your best bet is probably to do what most other people do and get one of those old HP or Leader AC voltmeters.  Other than not being current--meaning you have to restore and maintain one yourself--there isn't much advantage to more modern designs if you could even find them.

[rsjsouza]

The meters you suggested are all out of production. As I stated in my post I want to find something current. There must be some made for those types of work today.

The closest I know is the Sanwa EM7000, which has FET inputs and claims to measure from 40 to 1MHz (in low voltage). Its biggest drawback is the absence of a mV scale, though.

I was over the tapeheads website where they discussed milivoltmeter for measuring output from tape recorders. Most would say something like the Leader 185/186 etc.. is best but those are not in production any more. Besides I don't think those need to be analog with a galvanometer. A DMM doesn't seem to have good accuracy for the bandwidth from 20-20kHz. I have the Fluke 189 and 289 and I think they are OK but not quite good for those types of measurement.
What would you guys think is a good choice besides buying from the old analog meter from ebay.

The issue isn't accuracy or bandwidth--the F189 and F289 have that covered quite well--it is sensitivity.  A 6.5-digit DMM with a 100mVAC range will have a resolution of 100nV, but will be too noisy and inaccurate below 1mV.  A meter like the old HP403B or the Leaders will have a 1mV full scale range much better suited for such readings.

Thanks for the clarification, although I still struggle to see how the range from 100nV through 1mV (a whole 10000 counts within the full 1M counts of the 6-1/2 DMM) would be any noisier or less accurate than a galvanometer at 1mV FS with 10µV of resolution. At those levels, noise due to probing or other sources would be a problem with both instruments anyways. Oh well, perhaps I need to crack open one of my remaining cassete players and probe away to see for myself the inadequacy of my DMMs... 

[bdunham7]

Thanks for the clarification, although I still struggle to see how the range from 100nV through 1mV (a whole 10000 counts within the full 1M counts of the 6-1/2 DMM) would be any noisier or less accurate than a galvanometer at 1mV FS with 10µV of resolution. At those levels, noise due to probing or other sources would be a problem with both instruments anyways. Oh well, perhaps I need to crack open one of my remaining cassete players and probe away to see for myself the inadequacy of my DMMs... 

Well, a good 6.5 digit DMM might give a decent attempt at it, but they typically are not specified for at least the bottom 1% of the range, if not more.  So below 1mV you are on your own and unless you know your model, you don't know if the low voltages will be lost in the noise of residual counts or zeroed out by the firmware.  Keysight's TrueVolt meters might be better at low-voltage AC, but even they are not specified below 1mV.  The old analog models seem to be quite accurate when I test them using divided voltages--typically within a needles-width of being spot on or perhaps 1-2%. 

If you are measuring noise, you may want a TRMS model like the HP3400A.   I had one but I let it get away and now they go for big bucks.

[ehobby]

I have several HP meters that would work well (mostly analog 3400s, 410C, 400FL, 3456A, etc.) but I splurged a few years ago and bought a GW Instek GVT-427B. It is a bit of a specialty meter as it is a two needle (two channel) meter. It covers 10Hz to 1Mhz and has scales from 300uV (actually 350uv) to 100V (-90dBm to +43dBm) full scale. I can easily read down to 5uV.

It is great for audio work especially stereo audio. You can put a single audio tone into both channels of an amplifier and read the resulting signal levels at a similar point on both channels further into the amplifier to see if the signals are tracking/amplifying at the same levels. As you change the input frequency or amplitude of the input signal you can see how each channel tracks by watching the two needles move. In theory they should track each other relatively closely. If there is a frequency/filter or bias problem you will see the needles track differently. Though you can get actual measurements it is often more handy to just watch the needles move in relation to each other. You can see the two signals on two separate voltmeters but it may be very difficult or really slow you down to try to keep track of two different needles on two separate meters.

Another good use is to be able to put one channel on the input of a circuit and the other on the output, then adjust the voltage scales to get the needles in a good position near each other (or exactly on top of each other if the amp has a gain of, for example, 10 or 100). As you change the input voltage and frequency the needles should track the same up and down the scale. If not, the amp stage is non-linear or is doing some frequency filtering. This is great for both troubleshooting or working on a new design.

There are a lot of interesting things you can do with a two channel dual needle voltmeter you probably would not try or easily do with a single channel voltmeter, especially a DMM. It does work well as a regular single channel voltmeter, or you could look at it like it as two single channel voltmeters (the channels are independent). Another consideration is if you are looking at a signal with moderate movement it is easy to track on the analog meter vs. spinning digits on a DMM. You can look at two signal points on a scope but if the changes are at a lower level vs. the size of the signal it can be very hard to read the change between traces, whereas the needle(s) movement (convergence or divergence) is easy to see.

I suggest you look into one of these meters as your primary audio related AC voltmeter. My primary meters are Fluke DMMs for most of the things I do. However for audio and analog device work, an analog meter is usually more useful. For stereo work the GTV-427B is what I use.

[david77]

Any half decent DMM will do for most of the things you might want to do on audio stuff. You mentioned Tapehead.net, so I assume you want to align a cassette or R2R deck. This is one of the areas where I still prefer the analog meter. It's absolutely possible to set up a recorder using a DMM but it somehow feels more natural to watch a needle while trying to align the heads or eq.
I would actually try to get hold of one of the old used meters that have been mentioned. I usually use a HP 400F or a nice old tube voltmeter that I restored. There also where many clones of the HP400 series available until relatively recently.

[hpw]

If you may get an old HP 400F as lowest -80dB range as 0.1mV   

IMHO in today's measurements, a spectrum analyzer using a GOOD sound card & attenuation if required, gets you more about noise (hum, hiss, ..) and signals analyzing.. just carefully on tube gear 

[BeBuLamar]

One aspect that I think may give this impression is the natural dampening of a galvanometer when shown variable voltages on its inputs - the typical 3 readings per second DMM tends to show a cacophony of numbers under these conditions. However, the bargraph can help show trends and variance.

The 187's lowest range is 50mV so the 51 segment bargraph will struggle to show 1mV wiggles.

One possibility is to use a $2 opamp to amplify the signal so the bargraph shows it better.

There is no need to measure that low. The reading is generally higher than 100mV.

[Fungus]

There is no need to measure that low. The reading is generally higher than 100mV.

What exactly is the problem with the DMM? They have the bandwidth, the resolution, a bar graph... what's wrong with what they show?

If the signal is 100mV then maybe an oscilloscope is the best tool. A cheap DSO will show far more than any meter.

[thared33]

I've heard many audio engineers say they use the Fluke 8060A for this very purpose, but it's long out of production. It's what I use and it's very accurate.

[rob77]

just for the sake of curiosity what is exactly the measurement in question ?

[rsjsouza]

There is no need to measure that low. The reading is generally higher than 100mV.

What exactly is the problem with the DMM? They have the bandwidth, the resolution, a bar graph... what's wrong with what they show?

If the signal is 100mV then maybe an oscilloscope is the best tool. A cheap DSO will show far more than any meter.

Indeed. If that is the case, any decent modern equipment will do. Heck, evem the aforementioned 8060A has enough resolution and sensitivity.

[BravoV]

What exactly is the problem with the DMM? They have the bandwidth, the resolution, a bar graph... what's wrong with what they show?

Yep, but looking at the Op's response, looks like the Op prefers the old style analog meter with wiggly needle.

To Op, if you still insist on analog meter, GW-Instek still produce them, its still listed in their web site.

Something looks like this below, but newer version called GVT-417B (with "B" suffix), or try search used one like below older model.

Click to enlarge ...

[bdunham7]

There is no need to measure that low. The reading is generally higher than 100mV.

Then your F289 will handle that just fine.  It may seem 'twitchier' than an analog meter, especially measuring noise--but it's not wrong about that.  The analog meters have a deceptive calmness that some prefer.

[ehobby]

Another option is the Heathkit IM-25. I have restored two of these, better than the original accuracy, and have one on the bench. It has a big meter so a bit easier to read, covers the 10hz to 100khz range and can read down to 1mv. AC accuracy (~4% including reading the needle position and panel meter accuracy) is not as good as for some of the HP gear but for audio, it really does not matter that much when working with audio. Really.

When I think about fixing audio devices some of the most important things to check for are linearity, tracking between channels, distortion and hum/noise. Hum and noise are easy to see on a scope. Gross distortion can be seen on a scope. Small amounts of distortion can be measures on a THD or IMD analyzer (or even a PC sound based solution). Linearity and tracking are easy to see on an analog dual needle ACVM, a little harder on a scope or single ACVM. A DMM will also work but best if the signal is fairly static. I used a Fluke 8060A through the 80s for audio level work. It is a really great DMM, holds its accuracy well and fairly cheap on epay. My bench Fluke 45 is what I use in place of the 8060A for some measurements where I need an actual accurate TRMS measurement for.

If you understand what you are measuring, it is easy to pick the right tool for the job. Just make sure you have the right tools. If you are only going to do one or two audio device repairs, it is probably ok to just use a good DMM and lots of fairly static measurements, along with your scope. If you are really interested in audio, audio repairs and perhaps audio design, get some real analog audio service/test equipment and make things easy on yourself right from the start.

[Fungus]

Linearity and tracking are easy to see on an analog dual needle ACVM, a little harder on a scope

On a modern DSO? You can display the difference between two signals using the math channel. It can probably display phase difference between them numerically, too.

[SmallCog]

What exactly is the problem with the DMM? They have the bandwidth, the resolution, a bar graph... what's wrong with what they show?

Yep, but looking at the Op's response, looks like the Op prefers the old style analog meter with wiggly needle.

To Op, if you still insist on analog meter, GW-Instek still produce them, its still listed in their web site.

Something looks like this below, but newer version called GVT-417B (with "B" suffix), or try search used one like below older model.

Click to enlarge ...

At work we had a draconian rule come down on us from above that we could have no electrically powered equipment >x years old.

Health and safety rifa madness! Anyway a very nice HP mV meter went in the e-waste bin and was replaced with one of these GW Instek’s

I’ve never used it but I know it took months to arrive, we honestly thought they’d scammed us.

Entirely possible nobody has actually used it… but Health and Safety paid to replace all our old stuff.

[J-R]

Experimented with some 0-20kHz signals and various DMMs and generally speaking I was able to get reasonable results down to about 0.2mVAC.  Any DMM with a reasonably fast analog bar graph seemed to be an option if you were wanting to monitor an audio signal.

But I am curious what the specific process will be with the tape head measurements the OP is referring to?

[jsantoro]

As David77 said, an old HP 400F or Leader 181/186 would be my choice over a new Instek or that Tronson brand dual needle HP/Leader knockoff. I use 2 Leaders daily and would not give them up for anything. I check cal yearly on all my stuff and these 2 have not budged in years. When you need to null/peak a signal, count DBs etc. analog is the way to go and easier on the eyes. If you are working on analog tape and will be doing it a lot  this is  what I would recommend. Not You couldn't use a DMM to do it but it should have readout in DB like the 8060 and read AC flat up to whatever frequencies you are working with .