Multichannel DMM setup for Automation (DCDC eff. testing)

[Frankupthehall]

Hi Folks,

I’m looking at expanding/upgrading my setup to be able to automate some DCDC efficiency measurements (Vin, Iin, Vout, Iout).

I currently own a Keysight 34450a that I would either include into the setup or use (sell) to upgrade my setup.

I’ve looked at two approaches and I’m interested in hearing feedback on the pros/cons to each avenue.
1.   Purchase 3x additional measures
2.   Upgrade to a scanning multimeter

Buying 3 additional meters was my first thought but I’m leaning away from it because:
-   3 more meters on my desk (it isn’t large)
-   I’d be purchasing lower end gear (SDM3055)
-   Will cost the same or more than if I were to upgrade

The upgrade to a scanning meter came when I saw that the Keithley 6500 had a scanning card. The current measurements appear to be limited to 1A. For higher currents, I thought a sense resistor setup could be a good viable solution around the current limit.

I’m drawn to the Keithley because it’ll still be a single meter on my bench (space-saving) and it’s also a higher-performing piece of equipment. The investment could potentially be about the same between the two options.

I’m wondering what other things I could be overlooking or should be considering when trying to decide between the two?

Particularly, what would be the major downside of using the 6500 with a scanner card over 3 additional meters?

[HKJ]

Why do you need meters? A good lab supply and a electronic load can do the measurement. It will, of course, depend on the range you want to test.

[Frankupthehall]

I want meters because of accuracy.

I have a Keithley 2231A-30-3 power supply, but I'm not sure I would rely on the readout for data collection that I'd pass onto a customer.

but more importantly, having measurements taken at a DUT's input and output connections is an important factor here.

[Hydron]

Just a heads up regarding the DMM6500 - there are a few DIY scanner card projects out there, so you can build your own with shunt resistors etc included as needed. I plan to build my own once my meter arrives next month.

But HKJ also has a good point - it may be much easier to do with accurate supplies/loads that have remote sensing (probably won't be cheaper than a DMM + DIY scanner though). Go and check out his software, might be of use too.

[Frankupthehall]

Good points.

Now I'm wondering what kind of E-Load and Power Supply I should be looking at if I went down that path?

I'm all about optimizing my investment so I appreciate the suggestions.

[HKJ]

Now I'm wondering what kind of E-Load and Power Supply I should be looking at if I went down that path?

I'm all about optimizing my investment so I appreciate the suggestions.

Go for a power supply and load with remote sense. If you want to use my software, that can do above testing automatic, you must find devices that uses a (virtual) serial port or network connection. I do support many of the cheaper loads (Itech B&K Maynuo and a few other). For power supplies my support there is not so many devices, but it is fairly easy to add a new SCPI power supply.
Something like the R&S HMC8041 power supply has 4-terminal connection on the front, I do support the HMC8043, i.e. it only has to be simplified to support HMC8041. The PS has about the same size as a bench multimeter and can be stacked above or below it, the same is the case for the electronic loads.

[wizard69]

Is a multimeter even the right solution here?   Also would you be able to wire up a standard DMM and dynamically switch between current and voltage readings thus reducing the meters required to two.   This probably depends upon your meter and test requirements.

A multichannel A to D card in a PC may be a smarter choice.  It depends of course on how much of this testing you will be doing.   If you get a card with 4 independent channels you can grab your data synchronously.

Scanner cards have their uses but can your test requirements handle the offset between the four readings?   Only you can answer this question.   With 4 multimeters you could trigger readings to happen at the same time.

By the way there is another option that is sorta half way between a bench DVM and an A/D card.   That is to use something like Keysights USB connected DVM module.   You should be able to put 4 or those in a small rack or use a scanner in the same rack.  For example: https://www.keysight.com/en/pdx-2905759-pn-U2741A/usb-modular-digital-multimeter-5-digit?cc=US&lc=eng.   You can also look into VXI but that ups your cost a bit.

Another option is a Raspberry PI with some A/D boards.   With the right libraries writing software here will be easier than trying to run a scanner or collection of DMM.   We are talking low end for the most part but the selection of cards is large, see this list: https://elinux.org/RPi_Expansion_Boards.   Actually you don't even need to be focused on PI here as other SBC platforms have similar capabilities.   The nice thing with this approach is that you can end up with a very nice test stand that easily sits on the network.

Finally don't dismiss using a Digilent Analog Discovery as a data collection device.

In the end there are so many options here that you can spend weeks trying to choose the best option.   Further the best option might be a combo of techniques.  Given that buying multimeters and scanners would be my last option

[Frankupthehall]

Wizard69,

I've been thinking about this almost all morning, and I believe HKJ's suggestion may actually be the one I like most.

While I'm trying to automate the measurement process, I don't expect to be doing high volumes of testing. I just like automating measurement processes (make my life and job easier).

What I like about HKJ's suggestion is that it'll require a single upgrade (Power Supply) and the purchase of my e-load (which was already planned, but not selected). It seems like the least disruption to my setup, while also giving me the capability that I need to run the tests.

As always though, I'm open to suggestions to any aspect of this.

[HKJ]

The solution I suggested is the one I have used to test many DC-DC converters, these test can be found here: https://lygte-info.dk/info/index%20UK.html

These test was done before TestController, but I included functions in TestController to do them automatic (A popup called Param sweeper). For even more automation a script can be used.
If testing at low mA current it may be a good idea to add a ammeter to the output of the converter, but after the remote sense connection, the electronic loads often as a error of a few mA. There is, of course, no problem including the meter in the TestController data collection.

[tautech]

SDM3055-SC for $ 728 is not an expensive solution however any of its scanner card or dual measurements are not instant as relays toggle through the measurement types selected.
HKJ does have a point about a sensing PSU and getting instantaneous V & I readings via SCPI from it much like HendriXML does in these threads:
https://www.eevblog.com/forum/testgear/this-graph-brings-tears-to-my-eye/
https://www.eevblog.com/forum/testgear/battery-charging-using-a-siglent-sds1104x-and-spd3303x/

Of course with the sensing and very low noise SPD1000X PSU's more accurate results could be obtained.

[Frankupthehall]

Tautech,

Looking at the SPD1000X PSU, I see that they're quite the overachievers.
Between an SPD1305X and a DL3021A, the whole package could come in at a reasonable amount.

I don't even think I would need to "swap" hardware which would be nice.

Certainly, a focus on the 4 wire supply can really help with simplifying the equipment needed for the testing.

HKJ,

Browsed your software and looks great! Though ultimately I expect I'll be writing my own Python App. the process is usually cathartic.

[coromonadalix]

hp 34970a ?? with added cards ?

[wizard69]

You can't dismiss the value in software either if that is what HKJ is offering.   There is no lack of options here so taking the avenue of least resistance seems sound.

Wizard69,

I've been thinking about this almost all morning, and I believe HKJ's suggestion may actually be the one I like most.

While I'm trying to automate the measurement process, I don't expect to be doing high volumes of testing. I just like automating measurement processes (make my life and job easier).

What I like about HKJ's suggestion is that it'll require a single upgrade (Power Supply) and the purchase of my e-load (which was already planned, but not selected). It seems like the least disruption to my setup, while also giving me the capability that I need to run the tests.

For some reason I thought you were already set for a power supply.

As always though, I'm open to suggestions to any aspect of this.

One thing my ramble about hardware didn't cover well is the time and effort to do the software.   If somebody already has a suitable software solution then that is the way you want to go.  If not or for some other reason you can't implement something that already exists, then you probably should consider what the overall software effort will be.