Shorting a solar panel

[inetdog]

Thank you:)
I make solar-powered robotic sculptures; the project I'm working on is for ISEA2012 in Albuquerque.

I currently have a buck-boost regulator between the panel and the capacitor, but I'm finding it slow to charge the cap all the way up to 15V with a 15V supply - the rate of charge drops dramatically as the voltage in the cap rises. I'm hoping that a direct connect from the panels, which have an open circuit voltage of about 23VDC, will top the cap off more quickly.

I'm looking at using a crowbar-type circuit modified by using a mosfet instead of an scr so that it doesn't latch. I'm not an engineer so I kind of hack my way through this stuff 'til it starts working:)

I think a fundamental problem is that you are starting with a crowbar as a model. A crowbar is not meant to regulate, it is meant to protect. In essence it shuts down a defective power source by jamming a crowbar into the works! It is not meant to keep the voltage at a particular level, it is meant to crash the system totally if the voltage goes higher.

The only reason that a crowbar approach is relevant to your problem is that short-circuiting the panel will not damage the panels or blow any fuses or breakers. But by the same token it means that you need additional circuitry beyond what is in the basic crowbar design to make it do what you want. I understand that you would like to use as few components as possible, but there is a minimum that will be required.

[fix_it_guy]

Hi mmmalmberg, I like your work. Looks like you and I would get along nicely, my shop is filled with pieces and parts and and all kinds of stuff you would like to use. If I can't fix something I take all the parts out of it so I can fix the next one.

[bill von novak]

That's hugely interesting and relevant. Is the same true with battery storage?

Yes, that equation always holds. However it is usually not an issue because batteries operate over a narrow voltage range. About the only time you see it is when you have a battery bank that's getting low and you get less and less total power out of it at the same current i.e. if your bank is at 12V at 100 amps you're getting 1200 watts delivered to your inverter, but if it's at 10.5 volts at 100 amps you're only getting 1050 watts delivered.

[inetdog]

Yes, that equation always holds. However it is usually not an issue because batteries operate over a narrow voltage range. About the only time you see it is when you have a battery bank that's getting low and you get less and less total power out of it at the same current i.e. if your bank is at 12V at 100 amps you're getting 1200 watts delivered to your inverter, but if it's at 10.5 volts at 100 amps you're only getting 1050 watts delivered.

The underlying situation is not quite that simple. For both battery and capacitor the current flowing in for a given fixed voltage drops toward zero but never gets there. If you have a fixed amount of power (rate or energy transfer) then you will be able to deliver all of it to the battery while it is in a low State Of Charge (SOC). And then as you get closer and closer to the Float voltage, the current becomes constant at a rate which just offsets the self-discharge rate. But in a capacitor, if you were able to deliver a fixed amount of power, the initial current would have to be infinite. Not a practical scenario.

But a solar panel actually delivers a nearly constant current as a function of voltage as long as the panel Voc is enough greater than the load voltage. So if that were the only factor, the voltage on the capacitor would increase linearly as charge it. But since the voltage is very low in the early stages, less power is going in and at the end, the power is at its greatest.

But the real problem that you are facing is that you (theoretically, of course) can never charge a capacitor to 15 volts using a 15 volt power source. The internal resistance will cause the current to be proportional to the difference between the supply voltage and the capacitor voltage, and that will go toward zero but never actually reach it!

So to charge the capacitor to exactly 15 volts in a reasonable time, you have to start with a voltage source of more than 15 volts and NOT reduce the current coming into the capacitor until you actually reach 15 volts. That is what your circuit needs to do. You do not need a buck/boost regulator at all if the panel voltage is reasonably matched to your target voltage. You just want to send the full current from the panel to the capacitor until it reaches your target, then stop. For that purpose, a series switching element to cut off the current is much more reasonable than trying to short out the panel.

[BB.]

Also note that the equation E=1/2 * CV² is in "Watt Seconds"... I.e., a 14.5 volt to 10.5 volt 58 Farad capacitor stores, in battery terms:

• E (watt*hours) = 1/2 * 58 F (14.5v² - 10.5v²) * 1 Hour / 3,600 Seconds = 0.81 Watt*Hours

A similarly rated 12 volt battery would be rated:

• 0.81 WH / 12.5 volts = 0.07 Amp*Hour @ 12 volts

A single AA battery also stores around 1 Watt*Hour of power... (I think I got all the units right--but is is possible I messed something up--been a long time).

Big/super caps are not really all that they are cracked up to be a battery replacement in most applications.

If you can work with low voltages--A pair of NiCad batteries in series (2-3 volts) might be a good answer. They can be run down to zero volts, and two batteries in series is low enough voltage that they cannot damage each other if one is of lesser capacity.

For example any multi-cell rechargeable battery--say 6-24 cells for 12-48 volt battery bank, if one cell is "weaker" than the rest, the once cell can go to zero volts and actually be "reverse charged" by the other cells).

-Bill

[bill von novak]

But a solar panel actually delivers a nearly constant current as a function of voltage as long as the panel Voc is enough greater than the load voltage. So if that were the only factor, the voltage on the capacitor would increase linearly as charge it. But since the voltage is very low in the early stages, less power is going in and at the end, the power is at its greatest.

Agreed, and that's true of battery systems as well with PWM controllers. (Which is one reason MPPT controllers can do a lot for you; they can match a cold panel to deeply discharged batteries and not waste as much power.)

That is what your circuit needs to do. You do not need a buck/boost regulator at all if the panel voltage is reasonably matched to your target voltage. You just want to send the full current from the panel to the capacitor until it reaches your target, then stop. For that purpose, a series switching element to cut off the current is much more reasonable than trying to short out the panel.

Also agreed. At best a buck/boost will improve your charge times, but it needs at least a simple MPPT function to accomplish that. AFAIK no such device is available off the shelf.

[mmmalmberg]

But the real problem that you are facing is that you (theoretically, of course) can never charge a capacitor to 15 volts using a 15 volt power source. The internal resistance will cause the current to be proportional to the difference between the supply voltage and the capacitor voltage, and that will go toward zero but never actually reach it!

THIS is exactly what I'm trying to fix. I've got theoretically 22 or 23 V open circuit from the panel, so it seems to me that if I feed that directly to the panel, that would be best and I just need a way to stop feeding the cap when it gets to 15V.

So to charge the capacitor to exactly 15 volts in a reasonable time, you have to start with a voltage source of more than 15 volts and NOT reduce the current coming into the capacitor until you actually reach 15 volts. That is what your circuit needs to do. You do not need a buck/boost regulator at all if the panel voltage is reasonably matched to your target voltage. You just want to send the full current from the panel to the capacitor until it reaches your target, then stop. For that purpose, a series switching element to cut off the current is much more reasonable than trying to short out the panel.

I'm right with you on that. I just don't have a really simple way to do it (the shunt circuit seems like it can be quite simple) and I also like that the shunt circuit does not put anything between the panel and the cap until the target voltage is reached.

Two questions:
Do you know of a super simple way to open the circuit rather than short it?
What is the real problem that a number of people seem to have with shorting the panels? For me it's heat, anything else? Re. heat I was thinking I could use the shunt to divert power to a small cooling fan. But I'd love to have a good, clean and super efficient way to open the circuit.

[mmmalmberg]

Big/super caps are not really all that they are cracked up to be a battery replacement in most applications.

If you can work with low voltages--A pair of NiCad batteries in series (2-3 volts) might be a good answer. They can be run down to zero volts, and two batteries in series is low enough voltage that they cannot damage each other if one is of lesser capacity.

-Bill

I actually have a 1.2 AH SLA that will be charged by the panel at the same time through a current limiter at 150 mA and a small morningstar charge controller. The cap is intended to take the heavy I/O demands of the motors in my system because in the last iteration of this piece, which had the SLA only, I was killing the SLA's in six months from having to charge and discharge them way beyond their intended limits. This is why I need the cap to run at 15V, so that I can do the heavy lifting with the caps, and then in the evening as the light fades, the caps go down and the battery allows some extended time in the evenings to keep doing things, run some lights etc.

[inetdog]

THIS is exactly what I'm trying to fix. I've got theoretically 22 or 23 V open circuit from the panel, so it seems to me that if I feed that directly to the panel, that would be best and I just need a way to stop feeding the cap when it gets to 15V.

I'm right with you on that. I just don't have a really simple way to do it (the shunt circuit seems like it can be quite simple) and I also like that the shunt circuit does not put anything between the panel and the cap until the target voltage is reached.

Two questions:
Do you know of a super simple way to open the circuit rather than short it?
What is the real problem that a number of people seem to have with shorting the panels? For me it's heat, anything else? Re. heat I was thinking I could use the shunt to divert power to a small cooling fan. But I'd love to have a good, clean and super efficient way to open the circuit.

Well, the problem most people have with shorting the panels is just that you are wasting energy. There is no point in diverting that energy to a fan or to heating unless you NEED that fan or heating. The panel will not get any hotter delivering current into a short circuit than it does just sitting in sun with an open circuit.

Shorting the panels as a way of regulating charge to battery causes big current spikes that generate RF noise, can damage a partially shaded string of panels without bypass diodes (or can overstress underrated bypass diodes used in some cheap panels), and is just inelegant. The shorting device in the controller has to be able to handle the Isc current continuously.

A simple series FET or bipolar transistor with the right drive circuitry will provide a good cutoff. Or you could go "solid-state" and use a relay.

[vtmaps]

The panel will not get any hotter delivering current into a short circuit than it does just sitting in sun with an open circuit.

That is certainly true. In fact, the shorted panel will run cooler than the unloaded panel. --vtMaps