Question regarding electronic load

[brumbarchris]

Hello,
I occasionally develop various 5V, 3V3 power supplies for the projects I work on. Sometimes, these tend to be switched, rather than linear and some are a bit more complex than the most basic ones that you get, and I normally like to check their output variation, noise and load step response using an electronic load.

Consequently, I have occasionally looked at getting electronic loads. However, I have had difficulties in getting a programmable electronic load in the 2W...10W range. I have looked both at reputable manufacturers (Keysight, Chroma etc) as well as low-value brands (B&K Precision, RND, PeakTech) but most of what's available seems to start at 250W, 300W, 400W. Why is that the case? Is there not a market case for say, 10W rated electronic loads?

I know that there are some schematics for such things floating around, but why did they not make it commercially?

Regards,
Cristian

[tautech]

Max load and available resolution are very different things. 
Have a look at a SDL1020X-E and the latest posts in the SDL1000X thread. 

[brumbarchris]

Hello,
I have, but it is not really related to my original question is it? Or am I missing something obvious?

I am aware of the differences between resolution and max load.

But why do not 10W max load, much cheaper,  electronic loads exist?

Cristian

[Hydron]

Most of the cost in the load will be UI, enclosure, control circuitry, none of which will change whether it is 10W or 300W. Not a lot of point sacrificing the utility of having a higher power device to save the cost of say 5 FETs, a bigger heatsink and some jellybean opamps/sense resistors (take a look at a teardown of a load - in the one I have each additional power MOSFET uses 2 op-amps, some 0603/0805 passives and a low precision shunt resistor) - it won't decrease the selling cost enough to matter.

If you really want a low powered load and are _also_ in the market for a power supply, GW Instek have their GPP range power supplies, which have a 50W load mode for each of the 2 main channels (100W in the case of the single channel version) as a bit of a bonus feature. Not the most sophisticated load (only CV, CC and CR modes and some limitations at that, e.g. 1.5 V minimum in CV mode, 1R minimum in CR) but a useful addition to what is already a well featured unit (I have the 4 channel version which definitely seems the best value, is not perfect but the only alternative for similar features is spending $$$ for a Keysight or R&S or something).

[2N3055]

10W range can be easily served with mosfets switching in power resistors..

Also , all you need is mosfet, heatsink, power resistor (or few if you want ranges) and two opamps to get controlled load.
Connect to signal generator for  programming.. Measure current over main shunt, voltage on output..
Power with batteries for fully floating solution.

Or lookup those Jim Williams loads that can achieve insane speeds, that you can't buy commercially..

Also, those mini loads on E-bay will valiantly serve that power range. For 30-40 USD

So plenty solutions..

[nctnico]

In this power range I'd get an NPN darlington transistor and a 10 turn potmeter (and a resistor in series) to set the base current (perhaps a modulation input can be added as well). The biggest advantage is that the collector of an NPN transistor is a true current sink without needing a feedback loop (except for changing the potmeter when the transistor gets hot). A true current sink has a really high impedance so if a power supply is unstable it will show. And only 4 parts are needed: wall wart for power, series resistor, 10 turn potmeter (2k Ohm or so), a beefy NPN darlington and a heatsink.

[MegaVolt]

Consequently, I have occasionally looked at getting electronic loads. However, I have had difficulties in getting a programmable electronic load in the 2W...10W range. I have looked both at reputable manufacturers (Keysight, Chroma etc) as well as low-value brands (B&K Precision, RND, PeakTech) but most of what's available seems to start at 250W, 300W, 400W. Why is that the case? Is there not a market case for say, 10W rated electronic loads?

Check out Keithley's 24xx SMUs, 26xx are 4-quadrant power supplies that can be loaded electronically. Their power is just in the region of 20-30W. It is expensive.

If you need cheap, then for this there is Aliexpress which just offers similar toys for example EBD-USB +

[brumbarchris]

Thank you very much for your replies, they really covered a broad range of potential reasons.

Regards,
Cristian

[Kean]

Not exactly what OP was asking about, but Richtek have a nice little tester for checking pulse loads on small DC/DC converters.
It is based around a pre-programmed mcu generating a range of frequency/duty cycle signals, but saves having to roll your own.
Available for only US$30 or so from Digikey or Element14.  The schematic (very simple circuit) and usage docs are provided.
https://www.richtek.com/Home/Design%20Support/Reference%20Design/RD0004

[brumbarchris]

This Richtek thing is a very good suggestion, thanks! It is a reasonable compromise to an electronic load where you can directly set the load current in mA.

Cristian.

[David Hess]

In this power range I'd get an NPN darlington transistor and a 10 turn potmeter (and a resistor in series) to set the base current (perhaps a modulation input can be added as well). The biggest advantage is that the collector of an NPN transistor is a true current sink without needing a feedback loop (except for changing the potmeter when the transistor gets hot). A true current sink has a really high impedance so if a power supply is unstable it will show. And only 4 parts are needed: wall wart for power, series resistor, 10 turn potmeter (2k Ohm or so), a beefy NPN darlington and a heatsink.

That works much better if you add a low value resistance in series with the emitter to remove the temperature dependence.

In this application, 10 watts at 3 volts is 3.3 amps.  The voltage drop at maximum current should be at least 1.2 volts (two Vbe drops for a Darlington) but given the low voltage, 0.6 volts (one Vbe) is acceptable.  So 0.6 volts at 3.3 amps is 0.18 ohms.  Adding this resistor will considerably improve performance.

In this application the supply being tested can be used as the source for the base current; there is no need for a separate power supply.  I have done it this way all the time when I needed an instant power resistor for load testing.

[exe]

There are many Chinese electronic loads. I bought one myself, don't remember its rating, but it seems to be handling 100W okayish (the load reported temperature 65C). Don't remember the price, I think it was anywhere in $15-35 range. It's not the most advanced one, but serves me well.

For fun I pumped 10A into it, it worked, I think smelled something, so turned it off.

PS min current is 10mA on my unit.