Potential problems with series capacitors?

[MarkE]

I could hone in on more specific advice if you would tell me how much current you hope to draw when the circuit is discharging, and how fast you want to charge them.  Or you can work it out yourself.

Ignoring any leakage within the caps themselves the big issue will be when you have a low impedance across them.  In that case the charge will redistribute and can potentially harm the smaller capacitance devices with over voltage.  With your JT circuit that should only be during charging.  So you can apply the voltage clamps only during charging.  In that case, I would be inclined to use a MOSFET buffer with a TLV431, or LM4040.  The TLV431 would be a bit easier as it is a 1.24V device.  An external voltage divider sets the voltage.

[d3x0r]

Well; my this crank generator generates up to about .35A  (350mA); I intend the circuit to run in less than 100mA and preferably less than 20mA... But; I found a workaround to just use less than 2.7V to run it instead of 9V... (increase the secondary:primary winding ratio)  http://www.overunity.com/14591/lasersaber-strikes-again-a-joule-thief-king/msg409615/#msg409615   
results in lower voltage requirement, same current draw
(as the voltage decreases the current draw does also... so from 10V down to 4V goes from 20mA to less than 10mA... where 2.0 down to 0.6 goes from 40mA to 20mA)  overall I'm sure it's the same sort of power.
I'm still not getting micro-amps... well I did get microamps but then the LEDs are like 10% or less brightness... and really not very useful.  LS is getting 20 minutes on 1F at 9V sort of range... or about 60:1 ratio of crank time to output (15 seconds:15minutes)... maybe it's a difference of crank charger....
But; maybe I don't need to worry about the series caps to get higher voltages... even at 1.4V I have pretty nice output at the moment; and that's 20LEDs... I'm pretty sure I can balance that out...

[MarkE]

1200S @ 1F ~833uA/V so the discharge is pretty low and you will not want a balancer sucking off current on the order of 100uA.  If you do go with a balancing circuit, insert a switch so that you can cut it off during discharge.  You can use a FET for the switch, but that will increase both the minimum resistance of the balancer and the leakage versus a decent mechanical switch or relay.  If you can do without a balancer, that is all the better.  Simple is good.