On Sun, 28 Jun 2020 21:56:33 -0400, Paul
wrote:
wrote:
On Sunday, 28 June 2020 19:13:47 UTC+1, Paul wrote:
T i m wrote:
On Sun, 28 Jun 2020 12:38:16 -0400, Paul
wrote:
It operates in its saturated mode (full on), and the losses
are a function of the channel resistance (I^2*R).
https://www.mouser.ca/datasheet/2/38...f55-956481.pdf
Rds-on = 0.018 ohms
At 10 amps, loss = 1.8W, while the load receives 12V*10A = 120W.
Then in the absence of a better alternative, that's part of the deal.
If you were obsessed, you could put ten MOSFETs in parallel :-)
And perhaps use a series resistor from the 8212, to limit the
peak current into the gates.
Paul
5w at 12v is 0.42A, not 10A. If a FET dissipates 1.8W doing that then you chose the wrong fet.
NT
I picked 10A because I don't know what the actual load is.
It was an illustration, using a number picked from the
air for the illustration.
Don't worry, our NT is very literal. ;-)
When the designer tested his creation with the light bulb,
that was for test purposes, and wasn't likely to be the
actual load.
;-)
Obviously if the load is insignificant, the series element
dissipation is likely to be insignificant as well.
Yup.
In the illustration, wasting 1.8W compared to 120W isn't too bad.
I have worse linear stuff running in the house right now,
stuff that isn't even close to being that good.
So, now we have ascertained that we will *have* to waste (valuable /
solar generated) energy to prevent damage to the battery from
over-discharge, the issue is how to do that.
Are the commercial charge controller solutions as 'good' when it comes
to any parasitic load then we might be able to come up with ourselves?
If we had something that could sleep for say 5 mins (depending on the
size of the load to the battery capacity (so rate of discharge), and
given no solar charge) and only wake long enough to check the battery
voltage and either toggle the output off or not before sleeping, as
long as the total energy consumed during that cycle was less than the
average of something more linear, then we may have a solution. ;-)
ESP32 / Deep sleep (10uA?)
https://lastminuteengineers.com/esp3...r-consumption/
Cheers, T i m