Audio millivoltmeter

[clivel]

After a long hiatus I would like to get back into audio electronics. To this end I have been looking for test equipment to supplement my Tektronix TDS210 scope.

I have been looking on eBay for an analogue audio millivoltmeter. One meter that has caught my eye is the HP 3400a true RMS these seem to go for fairly reasonable prices. I have also seen a range of other HP analogue meters- HP 400, 403 etc.

Any recommendations would be appreciated.
Thanks,
Clive

 

[jackinnj]

After a long hiatus I would like to get back into audio electronics. To this end I have been looking for test equipment to supplement my Tektronix TDS210 scope.

I have been looking on eBay for an analogue audio millivoltmeter. One meter that has caught my eye is the HP 3400a true RMS these seem to go for fairly reasonable prices. I have also seen a range of other HP analogue meters- HP 400, 403 etc.

Any recommendations would be appreciated.
Thanks,
Clive

You might also want to look at the Fluke 8920A -- I found mine to be a tad more acurate than my 3403C.

These will give you a true, true RMS reading of a complex waveform.

You can DIY a really excellent true RMS meter with the Linear Technology LTC1968 -- I use one half of an LT6203 dual opamp as a driver for the LTC1968, the other half as a DC-Acurate filter.

If you need to measure micro-volts and tens of nano-volts the HP3581 wave analyzer is very good as well, but its bandwidth only extends to 50kHz or so.

 

[Conrad Hoffman]

I have some 3400A meters, and like them for some things, but they are showing their age and are a bit quirky. In fact, I have enough of them to reliably say their reliability isn't that great. They contain a few troublesome caps, and if you have a bad thermal converter, it's easier to just buy another meter. Unless you're a great troubleshooter, I wouldn't buy an "as-is" unit off eBay. I also have an HP400, and find that to be a much more trustworthy meter. It isn't true RMS, but the reality is that for audio use you'll usually be measuring sine waves anyway. IMO, it's important to have a needle-type meter around, as they're much better suited to peaking things up and seeing trends. For the budget minded, there was a very similar meter sold by Heathkit that isn't at all bad. For quantitative measurements, you should also have a good digital, be it hand held, or a bench meter. IMO, hand held meters are great, but they sacrifice update rate and bandwidth, for low power consumption. Once you use a good bench meter for a while, the hand helds seem a bit slow. As mentioned above, you can also build a very high performance meter.

 

[Nordic]

Rod seems to have a cheap DIY design too

And it is a generaly informative page on the topic.

http://sound.westhost.com/project16.htm

 

[jackinnj]

This is inexpensive and will work much better than the diode meter rectifier:

An externally hosted image should be here but it was not working when we last tested it.

and this is $10 worth of parts and is acurate to better than 0.1% up to a hundred kHz:

An externally hosted image should be here but it was not working when we last tested it.

 

[Nordic]

ok, you win

 

[clivel]

Thanks all for the advice.
I think I might give jackinnj's LTC1966 circuit a try, just need to hunt down a suitable meter movement.
Perhaps one of those super cheap HP 3400A's on eBay might just be the answer for a movement, attenuator and case, with the guts replaced by the LTC1966.

 

[jackinnj]

The LTC1968 has better bandwidth, the LTC1966 is more linear and more acurate. In any case the preamplification stage should get the input up to 100mV or so. I have a 1980's vintage synthesized signal generator liberated from one of the telcos via EBay post the internet bust -- still in cal -- so made this comparison:

An externally hosted image should be here but it was not working when we last tested it.

Both are MSSOP so the "toaster oven" approach to soldering the SMT device is best for DIYr's. I was able to do it even in my tiny lab -- even with my old fingers:

An externally hosted image should be here but it was not working when we last tested it.

 

[burbeck]

Jackinnj LTC1966

hi Jack,
the circuit you describe using the LTC 1966 is very interesting, i just got the parts and soldered the little devils onto dil converter boards
i thought i would buffer the output with an opamp to drive a meter movement

i intend to build an anttenuator and amplifier stage at the input hoping to measure down to 1mV FSD.

have you done any more work on it or have any recommendations.

regards

 

[jackinnj]

with the LT1966, 68 etc., or the AD536 you have to get the voltage up 100mV or 200mV to optimize the linearity. i think that a post-filter is necessary.

now we get to the interesting stuff -- for 1mV fs you probably want to be acurate down to at least 10uV -- you can go much lower with a low capacitance JFET front-end (not 2SK170 in other words).

when you get into the uV you have to protect the front end and use some bandwidth limiting -- at least to keep RFI and EMI out -- Boonton uses RFI protection, fuses etc., etc.

You could probably fit the RMS detector and preamplifier inside an HP 450D Attenuator --

 

[burbeck]

hi Jack,
thanks for that, i thought of using a jfet or mosfet input opamp for the front end, along with a home made attenuator network.

i realise this is ambitious and may not be easy, i was thinking of a 1 meg input impedance and a bandwith of 10 - 100kHz.

i have a boxfull of unused CA3130E's and CA3140's that i could try for starters, must be better chips out there these days?

i prefere to go the home made route (if possible)

i may well end up looking into your suggestion HP 450d

well i have to try these things
regards

 

[jackinnj]

You can purchase an HP 450D for less than $20 -- probably ships to the UK for $25 or so. It;s a 600 ohm device, probably handles 5W. It was also sold in a package with an HP True RMS Meter and a switching device.

http://s122635744.onlinehome.us/pictures/350d1.jpg

If you are going to use a JFET front end I suggest something like the BF244A which is much lower capacitance than the 2SK170 -- this is the device Joe Geller used in his resistor noise measurement article in Nuts n Volts.

 

[jackinnj]

I purchased an HP403B several years ago, just got around to playing around with it. The front was pretty sloppy but cleaned up with some isopropyl really nicely. Here's the unity gain front end schematic -- the first transistor is an HP Part Number, the second a 2N2189. I just substituted 2N4403. The input impedance is a high 1.6MegOhm -- attained via the positive feedback loop. Might get us away from using JFETs in a measurement amplifier:

An externally hosted image should be here but it was not working when we last tested it.