[Open_electroporator] Determining resistance before pulsing

[John Griessen]

On 9/11/18 11:24 AM, Nathan McCorkle wrote:
> What should be our method for checking the resistance of the sample?

1.
That means to create a recipe to get some low volts that are safe to put across the sample,
then measure with the ADC.  It should be steady state, not decaying exponential,
and if it has ripple/spikiness it should not be changing or decaying ripple either.

The 30k or 40k total load resistors soldered on the output need to be taken off first,
and use a sample that has all the volts.
Using R-cuvette higher than 30k will make recipes completely different -- needing some experimenting.
Using lower than 30k is outside the design spec and cannot do.
The volts will go higher with less of a load.

So that's the spec for measuring R-cuvette.  (above)




2.
Now, this makes me think of how we get a spiky steady state with the machine and
you've asked before what to do analog-wise and I said, "Not much, it's HV."
But some small HV filtering could be done and could even be left in all the time.

If left in all the time it will affect measuring R's that are high as its Resistance is,
but that can be OK.  Do we care about high R cuvette values?  Not really -- just low ones
that need more power to zap than we can deliver.  The top trace in the scope photo is
20 us time per horiz div. and 500V per vert div.  If we put an RC filter with time constant 20 usecs the
ripple would be much less, but the parts would need to be 3kV rated.  They can be made up of chins of lower rated parts
and they make the pcb bigger, but could be good.  Any in line R's will decrease volts, so we don't want too much
compared to the 30k ohm design minimum for R-cuvette.  Choose 5% drop, and R is close to 1.6k, so use standard values
of 6 or 4 0805 size R's to get it 400 ohms or 267 ohms each  390 x4  and 270 x 6 are standard values to use

the capacitance to put with 1600 ohms to get the 20 usecs time constant is .00002/1600 = 13 nF.

13/4 is 3.25, and standard values of 3.3 are usually available, so we'll look for 3.3nF 1kv rated caps...
Hey, they're available for 8 to 11 cents.

This is something to try out soon.  I'll ask for parts.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: culture_shock_pulser_debug-2018-09-10-1.jpg
Type: image/jpeg
Size: 241666 bytes
Desc: not available
URL: <http://cibolo.us/pipermail/open_electroporator/attachments/20180911/7b7456ed/attachment-0001.jpg>