So then you also recommend a boost circuit? It does seem like for these applications these are the way to go, and Be80be and Atomsoft know their stuff, so I'll see what a MC34063 can do.
So using the site Atomsoft posted (
**broken link removed**)
If I split the strings into two sets of 10 each, which it seems I could easily do, the output voltage would only be 35, I could use 6 of the 1.3 volt batteries that fit in the tube in series for 7.8V for that the calculator suggests these values:
Ct=321 pF
Ipk=405 mA
Rsc=0.741 Ohm
Lmin=135 uH
Co=29 uF
R=180 Ohm
R1=1k R2=27k (35V)
I'm not sure how using the closest real values will effect this though, converting to nearest real values it would be
Ct=330 pF
Rsc=1 Ohm
Lmin=150 uH (will one of these weird little ones work?:
https://www.taydaelectronics.com/servlet/the-468/150uH-Inductor/Detail)
Co=47 uF (not very close to 29, should I combine to get closer?)
R=220 Ohm
R1=1k R2=33k
While were on components, am I right in thinking that Ct is ceramic and Co should be electrolytic? And will a 1N4001 work for the diode shown at right?
So, what if I use one converter circuit to get a 35 volt output, use a ULN2803 after that (rated at 50V 500mA per gate), with each gate controlling a string of 10 LEDs in series, put the 7.8 volts from 6 AA size batteries through a 5V regulator to run the microcontroller, and connect the 8 control pins of the ULN2803 to 8 outputs on the microcontroller. I'm going to keep rewriting my plan till it's good.