Most of space-based hardware operates on solar panels composed of lots of photovoltaic cells connected in short series that produce around 28V; lots of these are connected in parallel to produce desired amperage.

Meanwhile, many devices - in particular, ion propulsion - requires very high voltages to operate. They use massive Power Processing Units to provide the desired voltages, producing a lot of heat and losing power in the process.

Would it be advantageous to use solar panels that connect lots and lots of solar cells in series, generating high voltage directly? (and similarly, connect thermocouples in RTG in series)? Stepping the voltage down for various electronics use should be at the very least similarly hard as stepping it up, and the devices that require high voltage could be supplied through a simple regulator, without need to step up the voltage by a factor of a couple thousands.

  • $\begingroup$ Having input in paralel means you only loose some power if one set of cells fails. Having all (or many) in series means a greater loss. This might be one of the reasons. $\endgroup$
    – Hennes
    Jun 13, 2016 at 16:09
  • $\begingroup$ @Hennes: They could still be bypassed. Imagine the cell connections going back and forth across the width of the panel, in shape of UUUUUUU with a line of tiny circuits along the edge, sensing the voltage across of each of the U segments, and detecting a break just short-circuiting it, connecting the neighbor segments. $\endgroup$
    – SF.
    Jun 13, 2016 at 17:04
  • $\begingroup$ @SF. That's done simply with a diode that is biased in the nominal current flow direction. Under nominal operation, the diode is reverse-biased, but if a cell or group of cells need to be bypassed due to shadowing, the diode becomes forward-biased and begins conducting. $\endgroup$
    – Tristan
    Jul 26, 2016 at 15:25
  • $\begingroup$ There a simple voltage regulators for voltages like 3; 5; 6; 12; and 15 V, but there are no simple one part voltage regulators for some kilovolts when voltages are to high for a single diode or transistor. Stepping up voltages by a factor of some hundreds to some thousands is nowadays easier than to build a DC in, DC out voltage regulator for kilovolts. $\endgroup$
    – Uwe
    Nov 27, 2017 at 17:21

2 Answers 2


No, it doesn't make sense, for two key reasons.

First of all, it is a loss of redundancy. If one of the cells fails, then the entire string fails. You really don't want that to happen... If you build in a shunting circuit of some kind to manage a failed cell, then the voltage could fluctuate, causing additional issues.

Secondly, solar cells placed in series need to match each other. It's harder to match the amount required for a really high voltage than is required to match the 28V power. Badly matched cells results in lost power inside the cells, and a really bad match could damage the cells. They are matched according to their IV curves. Basically, you want a set of cells to have the same amount of current running through them given the same voltage. Otherwise, some of the energy must be lost.

  • $\begingroup$ For redundancy, the string can be subdivided into bypassable segments. What is this cell matching? $\endgroup$
    – SF.
    Jun 13, 2016 at 17:08
  • $\begingroup$ That could be done, true, but it involves more work. Will add more details about matching. $\endgroup$
    – PearsonArtPhoto
    Jun 13, 2016 at 17:53
  • 1
    $\begingroup$ @SF. I think a bypass is impossible without loosing some power in the non-bypass mode, at least for sensing; And in normal condition, the current goes through all bypass circuits, adding up. $\endgroup$ Jun 13, 2016 at 23:26
  • 1
    $\begingroup$ It might not make sense, but that's exactly how it's done. The ISS solar arrays stack up several hundred cells in series to get to the working voltage. $\endgroup$
    – Tristan
    Jul 21, 2016 at 21:16
  • 1
    $\begingroup$ @SF This is exactly what is done. In fact, modern solar cells for space applications have bypass diodes integrated directly into the unit. See spectrolab.com/cic.htm for a datasheet. $\endgroup$
    – Tristan
    Jul 26, 2016 at 15:24

When ion propulsion is used to travel huge distances while accelerating slowly to very high speed, a flexible power distribution is needed.

The amount of available electrical power from solar cells decreases with the square of the distance to sun. The power of RTGs decreases with time due to the decay of nuclear fuel and the degradation of thermocouples. If much power is available, ion thrusters should work with full thrust. But after many years much less power is available and should be reservated for control and communication of the space craft as well as scientific payload. If some power is left, ion thrusters may work at reduced thrust.

But if some solar cells or thermocuples are reserved for generation of high voltage only, the may not be used later exclusively for control and communication and no longer for thrust. Solar cell arrays may generate some kilowatt at a distance of 1 AU, but at a distance of 5 AU only 1/25 of initial power is available, only some hundred watts.

Using a single voltage of about 28 V allows a flexible power distribution to high and low priority current consumers.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.