SMUs (source measure units) explained

[nexus]

Hi all,

Being a relatively young engineer, I did not have any knowledge of what a SMU (source measure unit) was until an older friend and EE mentor mentioned that such a (great) thing exists about 2 years ago.

Since then, my curiosity was peaked and I have come to use and understand SMUs as both very accurate and versatile power analysis and measurement instruments on the bench. As with all the content on my site, my intention is to spread the knowledge that has been passed to me and spark someone else's interest in the same way that mine was. 

Enjoy and let me know what you think: https://neonkev.com/2020/09/21/precision-swiss-army-knife-the-smu-source-measure-unit/

Thanks!

[exe]

I absolutely love idea of SMUs, but they are expensive as hell. Even old models such as 2400 are sold for many hundreds of $$.

It would be great to have an open-hardware smu unit. Unfortuntely, several attempts failed from the past. The closest implementation that I'm aware of is https://www.eevblog.com/forum/projects/eez-h25005-a-possible-successor-of-eez-h24005-programmable-power-supply/ . It's not exactly a match to commercial SMUs in terms of precision and dynamic range, but seems to be the so far the best. I ordered one, hope to have a chance to play with it soon.

I'm slowly designing my own devices in spare time that will have some characteristics of an SMU (ability to sink current, multiple current ranges, accurate readback), but it take quite some time to finish: https://www.eevblog.com/forum/projects/how-to-design-fast-bench-supply-with-cc-and-cv/

[jogri]

I absolutely love idea of SMUs, but they are expensive as hell. Even old models such as 2400 are sold for many hundreds of $$.

They are not that expensive if you're willing to spend the time repairing one… Broken Keithley 236s sell for ~300-500 USD, don't consist of proprietary, unavailable Voodoo-magic parts and can be upgraded to the 1.1 kV 237 model for roughly 150 USDs. I got mine for under 300 bucks, replaced a single resistor and it worked again.

(Spoiler alarm: There are currently no broken 236s on eBay, seems like companies don't get rid of old test equipment at the moment (i got my unit a few months ago). Maybe in a few months when the insolvency auctions start...)

[exe]

Keithley 236

It's too big for my bench . My dream SMU would be Keithley 2461. Hope to find one with a broken resistor for $300-500 one day .

[2N3055]

I absolutely love idea of SMUs, but they are expensive as hell. Even old models such as 2400 are sold for many hundreds of $$.

It would be great to have an open-hardware smu unit. Unfortuntely, several attempts failed from the past. The closest implementation that I'm aware of is https://www.eevblog.com/forum/projects/eez-h25005-a-possible-successor-of-eez-h24005-programmable-power-supply/ . It's not exactly a match to commercial SMUs in terms of precision and dynamic range, but seems to be the so far the best. I ordered one, hope to have a chance to play with it soon.

I'm slowly designing my own devices in spare time that will have some characteristics of an SMU (ability to sink current, multiple current ranges, accurate readback), but it take quite some time to finish: https://www.eevblog.com/forum/projects/how-to-design-fast-bench-supply-with-cc-and-cv/

h25005 and h24005 are just power supplies, like any other.

SMU is by definition precision meter AND precision 4-quadrant source with precision current limiter and/or dedicated current source mode with programmable compliance voltage.

There is nothing in open source even close to that. I know Marco Reps was working on OSMU but that went quiet some time ago...

[exe]

For me two-quadrant operation is enough, so I'm working on this.

[TheUnnamedNewbie]

If you don't need full four-quandrant, and can live with two-quadrant (only positive voltage, source/sink current), and need to buy NEW, the R&S NGM/NGL line of supplies can fill a hole between full SMUs costing 10+k USD/EUR and basic lab supplies.

[nexus]

I came across this hobby site where a brave individual is working on a purely open source SMU instrument. It appears to be in prototype stage with some promising results! https://poormanssmu.wordpress.com/2020/06/05/first-prototype/#more-716

Analog devices also appears to be working on a low cost, low power SMU device that can be cascaded to control up to 8 channels, each with +/- 12 volt, 40ma capability. Not quite the dynamic range for high power devices, but for the cost it is an outstanding four quadrant power module with sub millivolt and microamp resolution.
https://www.analog.com/en/design-center/evaluation-hardware-and-software/evaluation-boards-kits/dc2591a.html#eb-overview

[bson]

Four-quadrant is most useful in that it can go wherever your circuit has a voltage or current source.  Especially current sources can be a pain to actually build and easily become projects in themselves.  This is why you often see a stack of SMUs on a bench - because many times you need all sorts of bias voltages and current loads.  Obviously bandwidth and impedance becomes paramount.  (Low impedance for voltage sources, high impedance for current sources.)