Re: FET leakage question


John Griessen
 

On 03/18/2016 12:40 PM, edbreya [TekScopes] wrote:
I have to ask again, what is the maximum PV voltage available? Is it 2-2.2V?
Sorry, yes -- PV panel is 4 cell amorphous Si giving 2.0 to 2.2 V at peak power, higher
without load.
Is that per PV element, and how many are there? Your
schematic shows 2 PVs, paralleled through MOSFET switches.
I'll add as many as a system needs power, with all the panels the same since they have setup charges,
minimum orders and are laborious to make the interconnect for.

It can be more efficient to put them in series and run at higher
voltage. That's why I asked previously if you have full control of various aspects, meaning that you can connect up the elements
as needed.
The goal is to have high reliability, and run on just one, even when more are connected.
Run for several unattended years is the selling point I want.


Depending on the PV voltage available, you can buck or linear regulate down, or boost up. If the voltage is only 2.2V, then the
booster would work, as I described previously, without need for disconnecting the PV
The disconnection is for fail-over reliability, and round robin pumping up to higher volts. It's a required feature.

- it cannot charge or overcharge the battery
since it can't put out enough voltage without the booster in operation. Also, if the PVs are to be paralleled, and are illuminated
uniformly, there's no need for isolation, so Q3 and Q4 aren't needed. If they are illuminated uniformly, then you also have the
option to run at higher voltage with a buck or linear regulator.
Long series strings of PV cells fail that much faster. Not going that route. Planning on selling replacement
PV panels 30x20mm small with flat flex cable attached so installation is a little finger assembly with
a Phillips screwdriver the only tool needed.


I presume Q2 is the synchronous rectifier.
Yes.

If this is a very small scale PV/converter system like a few watts or less

15 to 100 mW charge ability, system consumes 2 mW average unless some extra sensors or actuators need more.

think if you're that worried about leakage, just use a smaller device with higher Rds and less leakage, or (if using boost) just
skip the synchronous rectifier and go with regular. Or, just measure the actual leakage - you may be pleasantly surprised that
it's much less than you expected. The other thing is that unless the system has to operate at quite a high temperature, the
leakage should be quite low.
Ambient with "enclosure in the sun" greenhouse effect is the worst case. I'm going to study that Siliconix paper
to figure out how to estimate leakage at temperature of 65 deg C when leakage at 25 deg C is specified.

Thanks for reviewing this Ed.

On 03/18/2016 12:55 PM, David Wise [TekScopes] wrote:
This is one tasty problem. There are so many avenues of trade-off. If I were doing this for fun, I could go for weeks.
:-) Well I'll be sure to post some scope photos later!

On 03/18/2016 01:08 PM, Jim Cotton [TekScopes] wrote:

Another consideration is you want shottky diodes on each string that you tie in
parallel to the power point tracker in case a string is shadowed.
I've read and understood about that phenomenon, and am sidestepping it differently -- any panel that is
under output level voltage is going to be isolated within a few tenths of a second. Not by diode action,
but by measuring with ADCs and deciding by a program to isolate it. There won't be much potential to
damage PV cells by forward breakdown while shaded because only 4 are in series anywhere in my system design,
and those four are always physically adjacent.

Generally in front of each PPT there are a number of series strings which are then
paralleled with blocking diodes. You want a high enough input voltage for the PPT to be efficient.
My panels are 30x20mm small. I'll have an over supply by 300 to 500% of power, so I don't need
to get every bit out of them. The goal is reliability over time more than high power flow to weight
as in a sun racer. Because of the low voltages and only 4 cells in series, I decided to use FET switches
and programmed timers to throw them instead of diode switching by how they are wired.

John

Join TekScopes@groups.io to automatically receive all group messages.