Multimeter selection for production testing

[tszaboo]

I'm setting up a production line testing tool. It an in-circuit functional testing JIG, that measures a bunch of things, and makes sure that the board we ship is within specifications. There is one test that's still missing, measuring the current consumption of the DUT. The original design called for a Keysight 34465A for this. And well, it's not within the budget to buy a new one.
So I'm searching for an instrument that can do current measurements, connect to the test system with Ethernet (USB might be acceptable). I though to ask the collective brainpower of the EEVBlog forum, because you know more instruments than I do. So here are specs:
- Desktop multimeter
- Measures ~100uA with 10uA accuracy
- Measures ~2A with 1% accuracy
- Can switch between these ranges without disconnecting the DUT
- Has Ethernet and SCPI, maybe USB
- Doesn't look like a toy
- Not an absolute nightmare of software to automate

My baseline is now the Siglent SDM3045X which I think would fulfill the role. I'm basing this on no actual evidence, just a guess, having never used Siglent multimeters. Honorable mention are all the Owon multimeters that look like a toy, and the Hantek 365* that looks like the piece but doesn't have uA range. And the Keysight U2741A that's somehow even more expensive than their 6.5 digit desktop meters.

[Antonio90]

I apologize if this doesn't add anything but, maybe it is cheaper to source a DMM with good voltage specs and a good shunt?

EDIT: It probably doesn't add much. I can test later today current measurements for the SDM3045, but your requirements are well within spec, and, at least with the "EasyDMM" software, logging is really not a problem via LAN. Haven't tested it otherwise.

[tszaboo]

I apologize if this doesn't add anything but, maybe it is easier to source a DMM with better voltage specs and a good shunt? As the measurement is automated converting from V to A shouldn't be an issue.

It needs to measure 100uA and ~2A on the same contacts. If I would make the system from scratch, I could do a range switching with a relay ie, but the rest of the system is already put together. I also don't like to measure ~100uV because thermal effects.

[tautech]

My baseline is now the Siglent SDM3045X which I think would fulfill the role.

Have you considered the SC models for their 16 channels of acquisition ?
They start from SDM3055-SC so a bit dearer than the SDM3045X but all can do dual measurements, regardless of the model.

Unfortunately they need be a -SC model from the factory as the scanner card cannot be retrofitted.

[2N3055]

I'm setting up a production line testing tool. It an in-circuit functional testing JIG, that measures a bunch of things, and makes sure that the board we ship is within specifications. There is one test that's still missing, measuring the current consumption of the DUT. The original design called for a Keysight 34465A for this. And well, it's not within the budget to buy a new one.
So I'm searching for an instrument that can do current measurements, connect to the test system with Ethernet (USB might be acceptable). I though to ask the collective brainpower of the EEVBlog forum, because you know more instruments than I do. So here are specs:
- Desktop multimeter
- Measures ~100uA with 10uA accuracy
- Measures ~2A with 1% accuracy
- Can switch between these ranges without disconnecting the DUT
- Has Ethernet and SCPI, maybe USB
- Doesn't look like a toy
- Not an absolute nightmare of software to automate

My baseline is now the Siglent SDM3045X which I think would fulfill the role. I'm basing this on no actual evidence, just a guess, having never used Siglent multimeters. Honorable mention are all the Owon multimeters that look like a toy, and the Hantek 365* that looks like the piece but doesn't have uA range. And the Keysight U2741A that's somehow even more expensive than their 6.5 digit desktop meters.

I have a question about this:

"- Can switch between these ranges without disconnecting the DUT"

Does that mean :

- I don't want to change cables from LO/HIGH range inputs on meter.

or

- DUT cannot ever be disconnected while we change ranges in meter. Like in an device is measured in standby and then in active mode but without reset because meter disconnected one shunt and then connected the other. It must preserve connection and only change shunt value.

[tszaboo]

My baseline is now the Siglent SDM3045X which I think would fulfill the role.

Have you considered the SC models for their 16 channels of acquisition ?
They start from SDM3055-SC so a bit dearer than the SDM3045X but all can do dual measurements, regardless of the model.

Unfortunately they need be a -SC model from the factory as the scanner card cannot be retrofitted.

I didn't. While I can see how this could be useful to do the test setup in the first place, for this setup, only current measurement is needed. And this thing will be dedicated for the production line for a few years, so minimum product...
I think if I design a new test jig, than that scanner card can be very useful to measure more things.

I'm setting up a production line testing tool. It an in-circuit functional testing JIG, that measures a bunch of things, and makes sure that the board we ship is within specifications. There is one test that's still missing, measuring the current consumption of the DUT. The original design called for a Keysight 34465A for this. And well, it's not within the budget to buy a new one.
So I'm searching for an instrument that can do current measurements, connect to the test system with Ethernet (USB might be acceptable). I though to ask the collective brainpower of the EEVBlog forum, because you know more instruments than I do. So here are specs:
- Desktop multimeter
- Measures ~100uA with 10uA accuracy
- Measures ~2A with 1% accuracy
- Can switch between these ranges without disconnecting the DUT
- Has Ethernet and SCPI, maybe USB
- Doesn't look like a toy
- Not an absolute nightmare of software to automate

My baseline is now the Siglent SDM3045X which I think would fulfill the role. I'm basing this on no actual evidence, just a guess, having never used Siglent multimeters. Honorable mention are all the Owon multimeters that look like a toy, and the Hantek 365* that looks like the piece but doesn't have uA range. And the Keysight U2741A that's somehow even more expensive than their 6.5 digit desktop meters.

I have a question about this:

"- Can switch between these ranges without disconnecting the DUT"

Does that mean :

- I don't want to change cables from LO/HIGH range inputs on meter.

or

- DUT cannot ever be disconnected while we change ranges in meter. Like in an device is measured in standby and then in active mode but without reset because meter disconnected one shunt and then connected the other. It must preserve connection and only change shunt value.

Standby and active without reset. The measurement happens on the same cable. Imagine a DMM doing break before make, autoranging and resetting the DUT every time.

[2N3055]

Standby and active without reset. The measurement happens on the same cable. Imagine a DMM doing break before make, autoranging and resetting the DUT every time.

That is exactly why I asked.
I don't have meter in question, so I can't test, but I would verify that this is the case.
Maybe Tautech or somebody else with the meter can check this for you.

Otherwise, I agree that meter can easily do the job you specified.
It already has enough accuracy specification reserve.
SCPI part should be easy enough for you.

[Caliaxy]

[….] and the Hantek 365* that looks like the piece but doesn't have uA range.

FWIW Hantek HDM3065 does have a 100.000 uA range (like the Keysight 34465A it tries to imitate).

[joeqsmith]

For production, I would have thought calibration and service was a big deal but never made the list. 

**
Also, I would think it would need to be something you could replicate (build more fixtures assuming expansion).

[tszaboo]

[….] and the Hantek 365* that looks like the piece but doesn't have uA range.

FWIW Hantek HDM3065 does have a 100.000 uA range (like the Keysight 34465A it tries to imitate).

The HDM3065 does, but it's a different meter than the one I talked about:

For production, I would have thought calibration and service was a big deal but never made the list. 

**
Also, I would think it would need to be something you could replicate (build more fixtures assuming expansion).

Maybe it should make it to the list. Although for equipment like this, I think I wouldn't do accredited calibration. I also don't calibrate externally the 12 bit ADC that is doing the other tests in the jig, ie, measuring the 3.3V if it's 3.3V. ISO9001 leaves a lot of freedom on how to do things. Simple performance verification of the current ranges, with another meter, once a year, would 100% fulfill the requirements.

[mawyatt]

If one is measuring on 100uA scale (standby) with a DMM and the DUT then demands 2A (becomes active at turn on), wouldn't the DMM burden voltage max out until the DMM could auto-range or follows command range change to the higher current range?

If so this would starve the DUT of voltage and as mentioned likely incur a reset!! Maybe a low drop SBD across the DMM could prevent the reset by allowing the initial higher "turn on" current to flow thru the diode while the DMM auto-ranges or followed commanded range change, while keeping the diode current insignificant during DMM measurements as the normal DMM burden voltage is low enough not to allow much diode current.

Also consider an "Active Clamp Diode" if you have a control line available. Here a simple Power MOS (P or N depending on where it's located) device shunts DMM terminals and is active when transitioning from standby (100ua range) to active turn on (2 amp range).

Best,

[bdunham7]

Unless I've missed something, none of these meters mentioned--not even the unaffordable Keysight--can do what you are asking for.  You need to measure 2A on the 10A range but even a long-scale meter isn't going to measure to 10µA accuracy on that range--that would be 1ppm. Everything talked about so far has a separate circuit for the 10A range, so how is the meter going to 'switch' between those ranges?  There are older meters that either have single current ranges with 0.1R shunts instead of 0.01R, such as the Fluke 8842A or multiple shunts with a single input such as the HP 34401A.  I just tried both of them. 

The Fluke 8842A will do what you want except it is limited to 1.99999A and relies on amplification to create a 200mA scale that amazingly enough can measure 100µA as accurately as you need (in my test) but is not specified to be that good.  It is also GPIB-only.

The 34401A works, has the single input with multiple ranges/shunts (3A/0.1R and 10mA/5R) and is specified to easily meet your needs.  However, it does not have make-before-break when switching ranges.  It's quick (a double-pole relay switches over) and you could probably deal with it by using a bypass capacitor and not using autorange but rather ranging up (via software control) before applying the 2A current.  The 34401A has RS232 that can be adapted with a USB cable and the command set is sufficient to do the simple task you're requesting.  The only remaining issue is burden voltage which is about 900mV @ 2A. 

Edit:  I just looked and I was wrong about the current inputs on the Siglents.  And the 34465A lower range current input is actually 3A.  So the Siglent 3045X works just fine except you'll have to check the make/break operation between ranges.

[2N3055]

In automated test where you control both instrument and test equipment it is simple to create test code for DUT that will work in synch with test flow and switch meter in proper range before drawing large current.
That would be sufficient to verify operating modes and hardware.

[tszaboo]

If one is measuring on 100uA scale (standby) with a DMM and the DUT then demands 2A (becomes active at turn on), wouldn't the DMM burden voltage max out until the DMM could auto-range or follows command range change to the higher current range?

Oh believe me I know this issue. Constant problem for firmware validation, the current changing 3-4 magnitudes. We have a Keithley 2280S that will do this with reasonable accuracy. But you do it exactly like this:

In automated test where you control both instrument and test equipment it is simple to create test code for DUT that will work in synch with test flow and switch meter in proper range before drawing large current.
That would be sufficient to verify operating modes and hardware.

In the end you only need 2-3 values to verify if a device works properly. Idle current at different power supplies enabled, and peak currents during RF.

[mawyatt]

The KS34465A and SDM3065X have no problem handing these current ranges for a single set of terminals. KS has a 3A max terminal that also supports 1ua range (as well as a 10A range on another terminal) and the SDM has 10A terminal that supports a 200ua range.

Also the DMM6500 terminal that surrorts 3A and 10ua range.

We just checked all of these to be sure

Best,

[2N3055]

The KS34465A and SDM3065X have no problem handing these current ranges for a single set of terminals. KS has a 3A max terminal that also supports 1ua range (as well as a 10A range on another terminal) and the SDM has 10A terminal that supports a 200ua range.

Also the DMM6500 terminal that surrorts 3A and 10ua range.

We just checked all of these to be sure

Best,

That I have no doubt, and that is what Tszaboo said was specified in original testing proc (KS34465A  was).

But if slight adjustment to test scripts can enable him to use SDM3045X that is 4-5x price reduction, and he will still be well over accuracy specs needed.

I still recommend checking if SDM3045X switches between needed ranges seamlessly or is it "brake-before-make" when switching shunts.

[mawyatt]

@2N3055,

Agree, the SDM3065X would easily fit the requirements, so suspect the SDM3045X would suffice.

Also as you mention, if the OP has control over sequencing, then seems that the DMM could be set to 2A range before invoking the higher current Active Turn On DUT mode.

Only issue there is if selected DMM does the "make before break" current range transition which is easily verified, but can't imagine that this would ever be "break before make" as this would introduce a dangerous condition.

Best,

[bdunham7]

I still recommend checking if SDM3045X switches between needed ranges seamlessly or is it "brake-before-make" when switching shunts.

Specifically with the SDM3045X there is also the issue of the 'ranges within ranges' where it switches internally at 2V even in the 6V range.  I wonder if that also happens at 2A in the 6A range--that might be problematic or at least super-annoying since this is right at the specfied test point.  Locking it in the 10A range (with less resolution) or getting an SDM3055 instead would be solutions to that issue,if I'm not imagining it.

[2N3055]

@2N3055,

Agree, the SDM3065X would easily fit the requirements, so suspect the SDM3045X would suffice.

Also as you mention, if the OP has control over sequencing, then seems that the DMM could be set to 2A range before invoking the higher current Active Turn On DUT mode.

Only issue there is if selected DMM does the "make before break" current range transition which is easily verified, but can't imagine that this would ever be "break before make" as this would introduce a dangerous condition.

Best,

Well,
I'm a "better safe than sorry" type of guy.
I would rather him test.
Just in case...

And to make sure people understood me correctly, problem is that if meter goes temporarily open circuit when switching between ranges. It should not but better check.
Although that can be worked around too with bypass relay that you enable relay, switch, disable relay, measure.

[2N3055]

I still recommend checking if SDM3045X switches between needed ranges seamlessly or is it "brake-before-make" when switching shunts.

Specifically with the SDM3045X there is also the issue of the 'ranges within ranges' where it switches internally at 2V even in the 6V range.  I wonder if that also happens at 2A in the 6A range--that might be problematic or at least super-annoying since this is right at the specfied test point.  Locking it in the 10A range (with less resolution) or getting an SDM3055 instead would be solutions to that issue,if I'm not imagining it.

I personally would recommend upgrade SDM3055 to better future proof it for not much more money.
But even locking SDM3045X in 10A range would be good enough for 1% needed.
But I agree I would go with SDM3055 and have no compromise option for less than 500€

[bdunham7]

And to make sure people understood me correctly, problem is that if meter goes temporarily open circuit when switching between ranges. It should not but better check.
Although that can be worked around too with bypass relay that you enable relay, switch, disable relay, measure.

You would think it would be simple enough to design them so they don't, but I've seen quite a few meters that have an interruption in the current shunts when switching ranges, including the HP 34401A I just tested. It's typically a few milliseconds as relays change position--in the case of the 34401A it is a set of DPST contacts switching from one to the other.  In the OP's case, a simple bypass capacitor along with making sure to do the switching at the 100µA level should be good.

[2N3055]

And to make sure people understood me correctly, problem is that if meter goes temporarily open circuit when switching between ranges. It should not but better check.
Although that can be worked around too with bypass relay that you enable relay, switch, disable relay, measure.

You would think it would be simple enough to design them so they don't, but I've seen quite a few meters that have an interruption in the current shunts when switching ranges, including the HP 34401A I just tested. It's typically a few milliseconds as relays change position--in the case of the 34401A it is a set of DPST contacts switching from one to the other.  In the OP's case, a simple bypass capacitor along with making sure to do the switching at the 100µA level should be good.

Yes, that is why I pointed out that this needs to be tested, because I also saw it on several occasions.
Not to mention when you use scanner...

As I said, it is not a deal breaker, it is easy enough to put in a bypass element (a capacitor, a scripted relay, whatnot) but you must know how it behaves.
You can even design your own shunt+switch and calibrate, measure voltage and calculate.