After more than a year I finally reassembled the old Microcare unit. Older forum members may remember I repaired it some time ago. It had blown the DC/DC converter stage used to drive the main power MOSFET on the high side. My solution was to rip out the discrete implementation and replace it with a small 12V-15V isolated converter.

Of course I did test it back in the day: It worked fine up to 140W and then it would reboot itself. I'm still hoping that problem has to do with the noisy power supply I was using to test it. I cannot find anything wrong on the board, so I'm going to assume it is fixed until I find that it isn't. If it isn't fixed, then the PIC-based control board will get chucked in favour of something based on the ATMega 328P. Time will tell.

Anyway, this post is to report on some current numbers. Now that it is fully assembled again, I measured the quiescent current (on a 12V battery setup), and I can report that it needs somewhere between 40mA and 50mA. My meter sadly doesn't do better than the nearest 10mA. Because I know that the very first item in circuit on the board is an SMPS, you could probably expect this number to drop to less than 20mA at 48V.

I noticed another interesting behavourial aspect. When I did a similar test with the Victron BlueSolar MPPT and a current limited supply to simulate the PV side, the behaviour was to run into the current limit and then back off cyclically, but the cycle was slow, maybe 1hz. My PSU has an LED that lights up when the current limiter is active. On the BlueSolar, it goes dim and bright cyclically at around 1Hz. On the MC this behaviour is different. It simply runs it into the current limit and then holds it there. So the algorithm is definitely a little different :-)

Thumbs up for trying again

But yes the Bluesolar mppt's I've seen the same, monitored power draw on the scope and its continually stepping up and down very slightly, does the math on 3 readings (i assume) then picks the best. that makes it slightly slow to respond to changes, it does 3 readings in a second but logically I think it's one adjustment per second.

 

And here she is, fully dressed again. Small video testing at 27V (ish), because that is as high as the small PSU goes that I'm using (buck converter on an old laptop brick). As predicted, the quiescent current drops quite a bit to only 32mA. It should drop even more as you go to 48V.

Also discovered, or rather rediscovered (have no idea what was going on in my brain when I wrote the original post) that the DMM does have a 200mA range that will give me more accurate numbers :-)

 

Okay. It seems the repair was a waste of time. Under full power, it simply resets the moment it has to start charging.

Good. That means the control circuitry is getting bumped for something AVR based. I'll harvest the SMPS from the old board, finding one with >65V input voltage is expensive enough.

I could even drive it directly from the PWM output of a Raspberry Pi if I wanted to. Next challenge... time...