Interesting. They also oscillate with moderate sized capacitive loads, like things with decoupling capacitors, or solar cells. I suspect that the oscillation damaged an expensive device that I was testing once. The newer ones are somewhat less bad about stability into capacitive loads, but I think nowhere near as good as the SMUs in IC testers.
Awhile back I got an HP 4145B for $275 on eBay, plus about the same for a complete test breakout box. It has four very nice SMUs that work from below 100 fA to 100 mA, 1 mV to 100V, but you do have to deal with floppy disks and maybe GPIB.
Cheers
Phil Hobbs
We buy a Keysight box for all our test stands. It's great for automated test, but the front-panel interface is so bad it's unusable. And the box is so long it won't fit on a regular test bench.
That's another cool product someone could do, a nice affordable several-channel SMU. I'd buy some. Single-channel SMUs go for about $5K.
They should do AC too. Can't think why they don't.
Chris Jones wrote: ================
** Not when you live in the real world instead of just theorising like posters here do ALL the time.Having that parallel diode gives a huge *warning* if you mistakenly connect a battery in reverse.
Sparks fly, amp meters hammer over.
The diode will fail short if the abuse is continued, saving the PSU from damage.
All my bench DC supplies have such diode and I have never blown one yet.
.... Phil
Yes, the evaluation board for the AD5522 seems like a nice starting point. I bought one and tried plotting the IV curves of some LEDs, which worked fine. It only comes with software to demonstrate that the chip works, but if someone would write proper software for it, even the eval board might be useful as a quad SMU, and it is fairly inexpensive. They do tie force and sense together on the PCB, but a scalpel and some enamelled wire will fix that. Of course if someone would design a proper PCB for it, with 4mm sockets or BNC connectors on a front panel and isolated power supply and isolated data lines, maybe one channel with a higher current range, that would be better. Maybe one channel could be adapted for lower currents too, by using the external sense resistor option and adding a couple of op-amps to buffer the low-leakage signals.
Here is another guy who has been trying to use the AD5522 eval board as a quad SMU:
Yes, that AD5522 part can be updated quite fast, if you added a FPGA to drive it then at least audio frequencies ought to be feasible.
FPGA for sure, to close the various loops. Maybe use some delta-sigma isolated ADCs, like DUM7703.
It's an interesting problem. All channels should be fully isolated, very quiet, and have a big voltage range and an absurdly wide current range. I suppose stored output capacitance could be bad news, or necessary, or something, but complicates the low (uA or pA) current ranges.
I'd like to see the schematic of the output stage of a good SMU.
I'm envisioning lots of relays. And some really cool curve tracer apps.
ptechnology.com:
upply.
an
re
as a
tion.
the
about
y when
.
s,
at
The
ds,
or
om
t-2
.
Cool. Lots of parts. Lots of relays.
We recently did a nice benchtop delay generator. It's a sweet platform for other instruments, and a 4-channel SMU is a possibility. May as well study what's out there.
We'll probably do a pulse generator first, since we're halfway there.
IIRC one of Jim Williams's books has a chapter on that sort of stuff. The guy uses a cascade of lead/lag networks to make a '1/2 pole' rolloff, so regardless of the capacitance you hang on it you always have a 45-degree phase margin. Of course 45 degrees corresponds to a fair amount of overshoot, but at least it won't oscillate.
The HP 4145B service manual has schematics. Pretty old-school, of course. Artekmanuals.com sells nicely remastered ones for pretty cheap, or you can get the A version from Keysight or at various places online.
Relays are good.
Cheers
Phil Hobbs
Really?!?!? $275?? Was it functional? I was looking for one about 6 months ago and prices then are about what they are now...~$1800-$4000. That is doing better than the current bitcoin takeoff... J
A series diode sounds like a good idea when charging a battery. A parallel diode (reverse) on the PS output is good for inductive loads... and maybe other things I don't know about. But weren't you first talking about a diode across the pass element in the power supply? Just trying to keep my diodes straight. :^) George h.
Works great. It did come without a boot disk, but I already had one so it was no big deal. Of course it was an actual auction and not a dealer. The four triax cables were nearly as much as the box. ;)
Cheers
Phil Hobbs
Have something to add? Share your thoughts — no account required.
Ask the community — no account required