Multichannel DC Load for Small Currents

[Martin72]

Hi,

We need a dc load with multiple inputs, say 8 channels, with currents between 25mA and 2A.
Years ago, I´ve done a similar thing with 8 inputs, simply build with op-amps and darlington transistors (but it works...).
Unfortunately, all inputs there got the same ground level, now we need all inputs separated from each other.
Normally it wasn´t that big deal to change the old circuit into this, but I don´t have and wouldn´t get the time to do it.
Question:
Do anybody know a regular, buyable solution ?
I´ve searchging a bit, there a modular solutions from several brands but way too expensive and way too "overdressed" what current/power concerns.

Martin

[aristarchus]

If 150 Euro per channel is acceptable, then the KP184 could be a solution.

[doktor pyta]

regarding Kunkin KP184: not exactly suitable.
I'm working on upgrading this device and the lowest low frequency current ripple I can get is close to 50mA p-p.

[H.O]

25mA to 2A at what voltage?

There are some very cheap modules available from AliExpress, like this 25W unit and this slightly more expensive 30W unit.

The first one can apparently be controlled over serial but it's resolution seems to be 10mA while the resolution on the more expensive unit is said to be 1mA. I have no idea how well they actually meet those or any other specs. Out of curiosity I orderd a couple of the 25W units to take a look at (not expecting any marvels).

There are of course many other cheap units avaialble but those are generally just bare PCBs.

[aristarchus]

..the lowest low frequency current ripple I can get is close to 50mA p-p.

Dont know if you have the latest KP184 (V04), what I have here is hardware motherboard version 04 and showing "2001" at boot.
With this, the lowest mA measurement starts at 10mA and higher and has 1mA resolution, also regarding voltage starts measuring at 100mV and higher and again has 1mV resolution.


P.S. Not sure what exactly you mean, is it the "Dynamic" scenario settings that you refer?

[Martin72]

Hi,

25mA to 2A at what voltage?

6V.

There are some very cheap modules available from AliExpress, like this 25W unit and this slightly more expensive 30W unit.

Ordered both for testing...

[Martin72]

Hi,

The first one can apparently be controlled over serial but it's resolution seems to be 10mA while the resolution on the more expensive unit is said to be 1mA.

The "more" expensive unit arrived first, yesterday I could test it.

While it handles 1.2A with good accuracy and stability, setting it to 25mA was a disappointment.
Yout set the display to 25mA, real current was about 19mA.
So I set more, by a setting of 30mA in the display, current will be 24.5mA, set to 31mA, current will be appx 26.5mA...

[Martin72]

Today, the cheaper 25W models arrived.
They are more "crap" as the first one. 
So probably I must forget the idea of buying this ones and design my own load.

[H.O]

That's a bummer, an expected one but still...
Mine hasn't arrived yet.

[Martin72]

Will do some tests at the weekend  - Adjusted and displayed current vs. real flowing current in a table.

[Martin72]

The Loads...

Both seeming to have the same case, the "more" expensive one (WEL 3005) got a bad building quality as the frontplate foil wasn´t glued enough to the plate.

Saturday or sunday I´ll make several measurings.

[Martin72]

Inside the WEL3005....
µC is an 32F100 with 12bit DACs/ADCs.
On the bottomplate there is a lm358, lm321 and 2 dc/dc buck converters, probably for supplying the op-amp.
Current-shunt got a value of 50mOhm, maybe that is the reason, why the WEL3005 aren´t proper accurate on mA values. (1mA*50mOhm= 50µV....)

[H.O]

It's a bit of a stretch to claim 0-5A with 1mA resolution when using 12bit ADC/DAC, though they specify 1% +/- 1 digit so I guess on paper, in a perfect world it should be fine. Your 25mA test shows it's way out of spec in the lower range which, again, isn't really surprising.

Since your max current is 2A I suppose you could try doubling the shunt resistance making it into a 0-2.5A load. Obviously you need to scale the setting/reading accordingly but it might be worth a try.

According to tracking information mine arrived in Sweden 10 days ago, still haven't heard a thing from the Swedish post who handles the import charges etc.

[Martin72]

Hi,

That´s what I will do, increasing the shunt value - But first I´ll take the measurements, then drawing a schematic of it, then we will see if it makes any sense at all.

According to tracking information mine arrived in Sweden 10 days ago, still haven't heard a thing from the Swedish post who handles the import charges etc.

Maybe it´s still in the customs office like mine was.

[Martin72]

Measuring the WEL3005, the other one will follow up tomorrow.

The more current flows, the more the accuracy of adjusted, measured and displayed values fits together.
The measured values in the table are "one shot" ones, it drifts away after a while ( because it´s getting warm I guess) - Not that much, but some mAs.
Our selfbuild dc-load won´t drift that much.
Let´s see why in the next days.

Edit: Measures taken with 10Vdc loadvoltage.

[Martin72]

Here a snip from "our" DC-Load, the principle behind is easy as you can see it in the schematic.
My thoughts on the DC-Loads I´ve bought are, the reference voltage for the op-amp is given from the DAC instead of our one, by dividing the voltage through a trimpot.

[DaJMasta]

Chroma makes some load mainframes that can be pretty cheap used and can be pretty high performance, I've got a 6314 with four 63103s which have good low current specifications despite having a fair bit of capacity (the complete load is pretty heavy!), but I believe there are double channel modules that you could get 8 channels in one mainframe, and since it's RS232/GPIB controllable, it's not bad to connect up a second mainframe for extra channels with single channel modules.

If they are low voltage, you could also look into PXI options.  I think Pickering and NI both make battery simulators that can be configured as loads, and there may be manufacturers that build load modules, and it's certainly not too expensive to throw them in a chassis if you can get a hold of a bunch for a reasonable price.

[Martin72]

Thanks for your post.
Actually, we need a load with 16 isolated channels.
Meanwhile I was searching for a cheap solution, I´ve tune up our old 8-channel load into a all channel isolated one, actually make the the pcb design for it.
Nevertheless, I examine the WEL3005 electronic load, here a first pic of the load circuit.
As expected, the reference is coming from the DAC of the µC.