Are you considering a solar trailer for your ebike? Do not underestimate the impact of even small amounts of shading on your solar array, especially if you're using SunPower cells.
As a 10 year veteran of the residential solar industry with extensive experience modeling PV energy production on shade-challenged rooftops, I thought I was pretty knowledgeable about the impact of partial shading on solar modules. However, I’ve recently learned something about SunPower cells which surprised me.
Try to guess what the power drop will be due to the following shading situation. About 25% of the cells are shaded so you might imagine that it would be around 25%, right?
Actually, it's closer to 90%.
Huh? ... but why? These are two 94W panels made by Solbian with genuine SunPower cells. At over US $7/watt, they are definitely a premium product and not your garden variety $2/watt Chinese semi-flex panels made with SunPower's assembly line reject cells. I've bought enough of the latter to appreciate the difference.
And yes, these modules have bypass diodes installed. If you’re unfamiliar with bypass diodes as used in photovoltaic panels, I recommend you read this first
. I won't cover the basics here.
If you had asked me to guess what the drop would be in the image above, I probably would have guessed "a little over 50%". My reasoning would have been that since at least 1-2 cells are shaded in 3 out of 6 substrings then 3 out of 6 of the bypass diodes should be activated so that the current is shunted around them and that since each active bypass diode dissipates several watts of energy, that adds up to a little over 50% of the total. If this was a traditional mono or poly silicon module made with non-SunPower cells, I might have been right but as I've recently learned SunPower cells are different.
Here's the unshaded array with all 60 cells connected in series to one Genasun MPPT boost controller. 148 watts from a nominal 188 watt array is typical given the slightly less than optimal tilt angle and cell temperature around 50-60°C. The green lines indicate which cells are part of the same substring group, each with its own a bypass diode.
Now look what happens when I shade just three cells (2 on one substring and 1 on another). The power drops from 148W to 19W.
And here's what happens with 20 cells shaded. Surely the bypass diodes should activate at this point and we should be getting full power from 4 out of 6 substrings, no? For some reason, the bypass diodes are not activated and the shaded cells are opposing the current flow.
Now look what happens when I completely block the same 20 cells. A "hard shadow" if you will. (Please disregard the paper "cells" on my shading implement. I wasn't planning to publish these results.)
Interesting. We're now getting 88W which is consistent with 2 bypass diodes activating (148W total - 2/6*148 for the bypassed cells - a few watts for the heat dissipated by the 2 active diodes).
And finally, here's the same situation but this time using 2 MPPT controllers. We're only getting 70W instead of 88W with the hard shadow so it looks like the bypass diodes are not activated here and the power from the shaded panel is close to zero while the unshaded panel remains unaffected because it is now connected to a separate MPPT charge controller.
For the two of you who are still reading after all of that, here's the payoff which got me to post this. I direct your attention to this SunPower white paper
. From pages 20-24 we learn:
a typical conventional cell has a breakdown voltage of approximately -15V to -20V, whereas the SunPower cell’s breakdown voltage is only about -5.5V for its second generation Maxeon cells and -2.5V for its third generation Maxeon cells.
SunPower does include diodes in its J-boxes, but the diodes do not turn on when only one cell is shaded. The voltage drop
across a single reverse-biased cell is not sufficient to drive significant current through the diode. SunPower includes diodes only to increase the production of the system in the case that several cells in the same substring go into reverse bias. In this case, the diodes limit the total amount of power that can be dissipated by reverse-biased cells.
Except when they fail to activate due to the lower breakdown voltage and the power loss is significantly greater than the contribution of the shaded cells.
My final solar trailer build will consist of 4 PV modules with 4x6 cells each. The two modules closest to the bike will each have their own MPPT tracker and the two furthest from the bike will be connected in series to a third MPPT tracker. I'm also planning to use "smart bypass diode" devices (SM74611 made by TI
) which have a forward voltage of only 26mV (take that, Schottky diodes!) and power dissipation of around 300mW which will help with the tiny junction boxes I'm integrating into my panels. I'm not sure they'll fare better than the results above but I think they're worth a shot.
And yes, I know, a solar array mounted over my head would virtually eliminate this problem but I'm still pretty sold on this trailer idea.
PS: Apologies to all the electrical engineers out there for any terms I may have butchered above. I dropped out of engineering school and majored with a BA in French.
EDIT May 19, 2020: Fixed link rot.