What are the practical uses of expensive, feature rich DMMs?

[MosherIV]

I don't know how to code or anything of the like.
Also when I look at pics of data acquisition programs the graphs always look to be aribrary without defined values displayed for each measurement. They also appear like they require set-up for each measurement. I just want to see a simple real-time continuous voltage graph on my screen without a load of complexity..

Oh 

I did a quick search and there is no simple program which you just run and connect to the HP3478A.

The closes thing I found is people talking about using National Instruments LabView

https://www.eevblog.com/forum/beginners/gpib-getting-started/

There is also this link with some info :
http://www.ko4bb.com/Test_Equipment/GPIB.php

When I get round to building my own GPIB interface, I will knock up a quick little utility to talk to my DMMs.
I will try to make it so that it can be configured for the HP3478A

Personally, I am not impressed with the Hantek thing, too proprietry. You would be locked in to whatever Hantek want to do.
It is also not a DMM - it is a data logger.

[alm]

No personal experience with the Hantek. 1% accuracy is pretty mediocre, even for a 3.5 digit handheld, but it may be sufficient for your application. Specifications are pretty brief, for example what is the maximum sampling rate?

Keep in mind that unlike something like a HP 3478A which has extensive documentation on its communication protocol, likely the only software that will ever support it is the one shipped by Hantek. Chinese manufacturers (even not so cheap ones like Rigol) do not have a good reputation for software. Make sure it has all the features that you need. I see no mention of it supporting multiple units connected to the same computer. If you want this, make sure to check if this is supported.

Note that they only list Windows versions up to Windows 8. No mention of Windows 10. That does not mean it will not work on Windows 10, but keep in mind that it will become a paperweight once the software stops working on your computer (upgrade to an incompatible Windows version, switch to Mac/Linux). The attached screenshot from the manual does not exactly fill me with confidence .

[coinmaster]

Yeah actually after doing a bit more digging it seems these hantek units have reliability issues too, so scratch that idea.
Damn I can't believe it is this difficult to find a data logging bench meter that doesn't cost an arm or require arcane knowledge.

[MosherIV]

I can't believe it is this difficult to find a data logging bench meter that doesn't cost an arm or require arcane knowledge.

Better beleive it.

The older GPIB based units pretty much standardized by LabView

I did use a HP34401A a few years back, I had to install the LXI drivers - real pain in the arse.
Once installed, the drivers also had a module which attached to MS Excel, which meant I could start Excel and connect to the DMM and log data directly into Excel. That was 

You could use a terminal program like TeraTerm to pass out the GPIB commands to the USB/Arduino/GPIB
TeraTerm will log all the communications (results) into a text file and the you can do a search and replace on the file later, turn it into a CSV file and then get Excel to turn it into a graph.

[alm]

The older GPIB based units pretty much standardized by LabView

There will usually be LabVIEW drivers. Because the communication is ASCII and well documented, it is also easy to develop your own software to interface with them. But as for off the shelf software, that is more limited. Sigrok has some support, but their support for bench meters is quite limited, and a front-end for DMMs is still incomplete.

I did use a HP34401A a few years back, I had to install the LXI drivers - real pain in the arse.

LXI drivers? Since when does the 34401A have Ethernet?

Once installed, the drivers also had a module which attached to MS Excel, which meant I could start Excel and connect to the DMM and log data directly into Excel. That was 

The HP/Agilent 34401A might be a decent choice. Sufficiently modern to be supported by BenchVue and the older (but completely free) IntuiLink for DMMs, but old enough to be available used for around $200 if you are patient. Has RS-232 in addition to GPIB, so you do not need GPIB if you only want to control a single meter.

You could use a terminal program like TeraTerm to pass out the GPIB commands to the USB/Arduino/GPIB
TeraTerm will log all the communications (results) into a text file and the you can do a search and replace on the file later, turn it into a CSV file and then get Excel to turn it into a graph.

Many meters can also be set to a talk only mode that will continuously output all readings to GPIB/RS-232. The only thing you need to do is capture it somehow.

[coinmaster]

Maybe I should forgo the idea of bench meters and use something like a Brymen meter?

[Fungus]

Maybe I should forgo the idea of bench meters and use something like a Brymen meter?

a) Make a list of what you actually need.

b) Make a separate list of what you would like.

c) Figure out how many hours a day you're going to be using it (and whether you need mains or battery power).

Cross off half the items on list (b).

Find something that meets those specs.

[coinmaster]

I've already listed what I need.
The issue is what I can afford.

[Fungus]

I've already listed what I need.
The issue is what I can afford.

But I do need at least 4 channels and the ability to measure at least up to 300v.

You can easily build an external voltage divider that divides by 10. That would open up a lot of possibilities, eg. that Siglent with the multiplexer option.

[tautech]

I've already listed what I need.
The issue is what I can afford.

It always is but you also need consider future needs. Take a step back and have a good think about this.

Even though something might be a little beyond budget now, consider; first cost = last cost.

[coinmaster]

To reiterate my needs:
1)Minimum of 4 channels
2)1 Gohm input impedance @ 200v (I often measure circuits with impedances between 10Mohm to 100Mohm)
3)300V+ range
4) (simple!) Data logging
5) I only need voltage readings, current and resistance measurements aren't required.

My wants:
Wall powered
Stackable

You can easily build an external voltage divider that divides by 10. That would open up a lot of possibilities.

Maybe, what would you recommend?

It always is but you also need consider future needs. Take a step back and have a good think about this.

Even though something might be a little beyond budget now, consider; first cost = last cost.

I agree but I already need about $5,000 for other equipment on top of this. I gotta strike a balance.

[coinmaster]

Maybe one of these https://www.picotech.com/data-logger/adc-20-adc-24/precision-data-acquisition with a 1 gig resistor in series with each input?

[coinmaster]

Wait nevermind it's got a 39 mv minimum input meaning a 39v minimum reading. That won't do.
Looks like I may go back to the SC1016. The manual says "insulation resistance- 1 Gohm minimum" is that referring to the input impedance?
I'm not sure I like the idea of relay switches though, what's the point of logging start-up conditions if you log in 1 second intervals.

[tautech]

Looks like I may go back to the SC1016. The manual says "insulation resistance- 1 Gohm minimum" is that referring to the input impedance?

No, inter channel isolation. (leakage)

I'm not sure I like the idea of relay switches though, what's the point of logging start-up conditions if you log in 1 second intervals.

AFAIK 1s is typical for all these logging cards.
There's better tools for checking startup conditions.

On the SC1016 the relays are miniature types and the contact life is only stated for current measuring operation where voltage readings will be much kinder on their service life.
They're 6.5 x 10.6mm SMD Japanese NEC UD2-4. 5NU
They've been laid out in such a way that it's made replacement if needed very possible.

[Fungus]

To reiterate my needs:
1)Minimum of 4 channels
2)1 Gohm input impedance @ 200v (I often measure circuits with impedances between 10Mohm to 100Mohm)
3)300V+ range

You can easily build an external voltage divider that divides by 10. That would open up a lot of possibilities.

Maybe, what would you recommend?

Some packets of these:

http://www.ebay.com/itm/231661826547

http://www.ebay.com/itm/192130735925

Put them in series to get your 1GOhm impedance, divide your input voltage by 10 so you'll see 0-30V at your multimeter.

(putting 3x300M in series gives you some insurance on case one of them fails short).

Trimming them for accuracy is left as an exercise for the reader. It might be as simple as mixing/matching by hand to make matched sets of resistors which are close enough or you might need a trimmer of some sort. I don't know what accuracy you need.

The point is: You'll pay through the nose for that input capability on a general purpose multimeter when it's not difficult to build it yourself.

Edit: Of course that doesn't work because of the comparatively low resistance of the meter in parallel with the bottom resistor in the divider chain. Still, the principle stands. Resistor dividers are your friend when you want to measure high voltages.

[alm]

Pretty sure you will not find a DMM with those specifications. If they have a > 1 GOhm input impedance, it is usually up to 20 V max. Off-the-shelf solutions that come to mind:

• Use a high voltage probe to divide the input voltage and feed it either into four separate meters or a scanner card.
• Use an electrometer. Electrometer have a very high input impedance (often in the TOhm). Most will only go up to 200 V, however. If you think bench DMMs are expensive, do not look at list prices for electrometers. You can sometimes find a Keithley 619 in dual channel configuration on eBay for $200-$300.

Put them in series to get your 1GOhm impedance, divide your input voltage by 10 so you'll see 0-30V at your multimeter.

(putting 3x300M in series gives you some insurance on case one of them fails short).

I do not like this design. What is the voltage on the output of the 'divider' if the connection to the meter becomes open? Could be a loose wire, bad contact, a relay switching (if using a multiplexer), the meter resetting/switching modes or even a fusible resistor that opens. I would put say a 1.11 MOhm resistor in parallel with the output to the meter, as is the case for many commercial high voltage probes for the exact same reason. This will increase your attenuation ratio by a factor of ten, obviously.

[Fungus]

I do not like this design. What is the voltage on the output of the 'divider' if the connection to the meter becomes open? Could be a loose wire, bad contact, a relay switching (if using a multiplexer), the meter resetting/switching modes or even a fusible resistor that opens. I would put say a 1.11 MOhm resistor in parallel with the output to the meter, as is the case for many commercial high voltage probes for the exact same reason. This will increase your attenuation ratio by a factor of ten, obviously.

Which wire, exactly? Going through all possible failure modes is good.

The basic design is what it is though. Any high voltage probe or multimeter will eventually boil down to a resistor divider chain which reduces the voltage to a manageable level. The rest is just semantics and component choices.

[alm]

The wires connecting the resistors, possibly through a multiplexer, to a DMM or datalogger. Since these are outside the box that contains the external resistors, I would consider them more likely to become disconnected than an internal solder connection.

An additional advantage is that the input impedance has less influence on the accuracy.

[Fungus]

The wires connecting the resistors, possibly through a multiplexer, to a DMM or datalogger. Since these are outside the box that contains the external resistors, I would consider them more likely to become disconnected than an internal solder connection.

I fail to see how it's any different than measuring a smaller voltage without the resistor divider.

If it's all floating, nothing much happens. The reading will be wrong.

It it isn't floating then disconnecting a single wire doesn't do anything except make the reading go to zero (red wire) or maybe slightly less accurate (black wire).

If we're building something for a definite, known setup we can plan accordingly.

An additional advantage is that the input impedance has less influence on the accuracy.

The OP has stated he wants 1 Gigaohm impedance. If it's noisy, add capacitance.

[3db]

I was going to contribute but don't really get what the OP expects.
How accurate do you expect your measurement to be.
Millivolts at 300 volts ?

[alm]

I fail to see how it's any different than measuring a smaller voltage without the resistor divider.

See attached schematic. The top one is how I interpreted your description, and the bottom one is my suggested modification. JP1 signifies a connector, a relay, a fusible resistor or any other reason a connection may open. If JP1 opens, than there will be 300 V across it (with an output impedance of about 1 GOhm). Any wiring capacitance will also be charged to this voltage. That may be a problem if the downstream components can not handle 300 V. Think a DMM with a scanner card that is only rated for 100 V. Same if the meters input is AC coupled. The parallel resistor's only downside is that it increases the attenuation ratio to 100:1.

An additional advantage is that the input impedance has less influence on the accuracy.

The OP has stated he wants 1 Gigaohm impedance. If it's noisy, add capacitance.

What if the input impedance changes slightly due to range switches or AC/DC modes? That effect is decreased by a factor of ten by the parallel resistor.

[Fungus]

What if the input impedance changes slightly due to range switches or AC/DC modes? That effect is decreased by a factor of ten by the parallel resistor.

Yep.

The first step would be to pick/buy a meter and find all that stuff out. After that you can design a circuit that matches the meter and required accuracy. Maybe it's better to divide the input voltage by 100 (or even 1000) so the impedance of the meter is relatively high compared to the sense resistor. Maybe it's better with the meter in series.

The OP seems to have a very definite input voltage characteristic in mind so it shouldn't be difficult.

[nctnico]

Creating a divider with a 1G Ohm input resistance is not going to be trivial to say the least. In one of my designs I have a 100M Ohm input impedance and as I expected it doesn't work very well due to parasitic capacitances (measuring voltages which aren't there). As a general rule of thumb: using resistors over 1M Ohm means trouble.

[Fungus]

Creating a divider with a 1G Ohm input resistance is not going to be trivial to say the least. In one of my designs I have a 100M Ohm input impedance and as I expected it doesn't work very well due to parasitic capacitances

Edit: Yep. The the more I think about it, the more the 1GOhm requirement throws a spanner into the works.

I guess the real solution is to divide the input voltage by 500-1000x to give a final sense resistor of 1-2MOhms. Use a suitable instrumentation op-amp to buffer the voltage across that resistor and send it to the multimeter.

[David Hess]

Simple / portable meters usually have a 10 M input impedance on all ranges.

But commonly available *cheap* handheld meters usually have an input resistance which varies with range and an input resistance down to 1 megohm is common.  This causes a massive problem with some accessories like high voltage probes but is irrelevant for electrical applications as opposed to electronics applications.  Good handheld meters have a precision 10 megohm input resistance on DC *and* AC volt ranges.