Short-circuiting solar panels

[mmmalmberg]

Greetings. I'm new here, not in the solar industry per se but I'm an artist who makes solar-powered robotic sculptures.

I'm charging a large (58F) 16.2V capacitor with a couple of small panels that have an output of something like an amp or two with an open circuit voltage of 22 or 23 volts. I want to use a crowbar circuit as a simple way of cutting power to the cap when it reaches 15V.

The Question: Can I damage a solar panel by short-circuiting it in full sun?

[billvon]

The Question: Can I damage a solar panel by short-circuiting it in full sun?

Short answer: no. (BTW you're doing that by trying to charge a 58F cap anyway.)

Are you really going to crowbar a 58F cap? From my calculations if your crowbar activates over the course of 1 millisecond you'll be dissipating 6 megawatts for a short time. I don't know of any crowbar that will survive that. Or are you going to crowbar the panel and have a diode to isolate the cap at that point? If that's the case it's probably easier to either limit the power via a big zener or open a circuit with a simple charge controller or linear regulator. The LM7815 will do this for you and is easy to use.

[inetdog]

Greetings. I'm new here, not in the solar industry per se but I'm an artist who makes solar-powered robotic sculptures.

I'm charging a large (58F) 16.2V capacitor with a couple of small panels that have an output of something like an amp or two with an open circuit voltage of 22 or 23 volts. I want to use a crowbar circuit as a simple way of cutting power to the cap when it reaches 15V.

The Question: Can I damage a solar panel by short-circuiting it in full sun?

No, you cannot damage a solar panel by short circuiting it. A given amount of current delivered either to a load or to a short circuit causes exactly the same amount of heating inside the panel. Since the cells are best modelled as current sources rather than voltage sources, the intuitions you have about their behavior based on experience with batteries and generators will mislead you.

Older design battery chargers using solar panels actually regulate their output voltage in just that way, by shorting out the panels.

The only thing you need to be careful about is that the panel must have some sort of protection against partial shading, in the form of bypass diodes. These are standard features in commercial and DIY solar panels, but confirm this with the panels you intend to use. (On the other hand, in a low voltage panel, it may not be a serious concern in the first place?)

[mmmalmberg]

Short answer: no. (BTW you're doing that by trying to charge a 58F cap anyway.)

Are you really going to crowbar a 58F cap? From my calculations if your crowbar activates over the course of 1 millisecond you'll be dissipating 6 megawatts for a short time. I don't know of any crowbar that will survive that. Or are you going to crowbar the panel and have a diode to isolate the cap at that point? If that's the case it's probably easier to either limit the power via a big zener or open a circuit with a simple charge controller or linear regulator. The LM7815 will do this for you and is easy to use.

Good points (why I'm asking the smart guys... Charging the 58F cap will pretty much never be from empty and being from solar panels will typically have a soft start in the a.m. But yeah it's pretty much like a short, I agree.

I'd really like to get a better sense of your suggestions.

With the zener, are you saying just stick a big fat zener across the panel? I've done this with a smaller version of the same thing, with a resistor in series with the zener.
Opening the circuit, I don't know where to get a charge controller for 15V.

Re. regulators, I already have an 80W 15V buck/boost regulator that I was planning to use, but I'm finding that bringing the cap up to 15V with a 15V supply is very slow and I want to be able to pump up the cap as quickly as possible, so I'm looking for simple alternatives to the 15V regulator. In fact I need something super simple as I need to ship the piece I'm working on in a couple of weeks and have MUCH to do besides sorting this out. Which is the appeal of the crowbar...

Speaking of which, 6 megawatts... nice. Maybe I do need a resistor to limit that... I'm very appreciative of your input - thanks!

[Sunking]

Re. regulators, I already have an 80W 15V buck/boost regulator that I was planning to use, but I'm finding that bringing the cap up to 15V with a 15V supply is very slow and I want to be able to pump up the cap as quickly as possible,

You really do not need anything, just connect the panel directly to the cap, and use a Zener Diode in series to limit the voltage if you want, but not really needed. It will still be painfully slow process, but as fast as you can get with a 80 watt panel. The bottle neck is the panel. When you connect th epanel to a dischargd cap the voltage will be ZERO, as the cap charges up the voltage will rise slowly. All you need to do is disconnect when you reach the desired voltage.

So if your panel Vmp is 18 volts you need a 3 volt Zener with a current capacity of at least 125% of the panels Isc rating. Bu tlike I said you do not need a regulator, just a means to disconnect when the cap reaches the desired voltage. You can do that with a simple voltage comparator and relay.

[mmmalmberg]

Just another point to make... I'm going to be crowbarring the cap while it's connected to solar panels as the voltage is inching up past 15V, and I've just realized that I need to use something like a mosfet rather than an scr, so that it will unlatch as soon as it gets back to 15V.

I think that will bring the 5 jillion watts down somewhat, make sense?

[mmmalmberg]

I really don't understand how that works, with a zener in series. I would think that would stop voltage from flowing up to a certain voltage and then open up. What am I missing? (Zeners are somewhat confusing to me I admit.. Also is Vmp the open-circuit voltage? Simple is what I'm looking for so I'm listening... Thanks!

You really do not need anything, just connect the panel directly to the cap, and use a Zener Diode in series to limit the voltage. It will still be painfully slow process, but as fast as you can get with a 80 watt panel. The bottle neck is the panel.

So if your panel Vmp is 18 volts you need a 3 volt Zener with a current capacity of at least 125% of the panels Isc rating.

[inetdog]

I really don't understand how that works, with a zener in series. I would think that would stop voltage from flowing up to a certain voltage and then open up. What am I missing? (Zeners are somewhat confusing to me I admit.. Also is Vmp the open-circuit voltage? Simple is what I'm looking for so I'm listening... Thanks!

If you put an ordinary diode in series between the panel and the cap, and put the crowbar circuit on the panel side of that diode, the crowbar will only pull current from the panels and will not be trying to discharge the capacitor. Much cleaner!

Or you can use a series Zener diode as suggested so that the capacitor never sees more then the panel voltage minus the Zener voltage. But that will leave the capacitor voltage to vary as the panel output voltage varies with temperature, etc.

You could also put a great hulking Zener at 15 volts in parallel with the panels, but would be a total waste, since it would have to be large enough to constantly dissipate that full output power of the panel.

Instead of a crowbar, which would not reset, you should use a shunt regulator which senses the current on the capacitor side of the rectifier diode and shorts out the panel output whenever the cap voltage is above your limit. That will "reset" as soon as the cap voltage drops below the limit. But you have to build some hysteresis into the system so that you do not end up with the shunt transistor partially turned on and heating up.

[Sunking]

A zener is very simple device. In the forward bias is just like any diode with a voltage drop of about .7 volts. In the reverse bias will hold a set voltage drop of whatever it is rated at up to its current capacity. So a 18 volt panel - 3 volt zener = 15 volts.

[mmmalmberg]

If you put an ordinary diode in series between the panel and the cap, and put the crowbar circuit on the panel side of that diode, the crowbar will only pull current from the panels and will not be trying to discharge the capacitor. Much cleaner!

I love that. And I should have diodes for each panel anyway to keep the panels from bleeding down the cap overnight with their leakage.

Or you can use a series Zener diode as suggested so that the capacitor never sees more then the panel voltage minus the Zener voltage. But that will leave the capacitor voltage to vary as the panel output voltage varies with temperature, etc.

Sorry I'm still not getting it. But in any case my objective here is to let the panel show 23V to the cap right up until it hits 15V, anticipating that it will hit 15V faster than if I somehow regulate the panel output to 15V. I guess that's intuition but with a lot of things, a greater differential give a more rapid transfer. Thermal conduction is like that for instance... I should say, if it isn't obvious, I'm not an electonics engineer I'm an artist who is self-teaching and learning from the likes of you all here, so some concepts are just going to smack me in the face with no comprehension until some moment of revelation maybe in the next day or two

You could also put a great hulking Zener at 15 volts in parallel with the panels, but would be a total waste, since it would have to be large enough to constantly dissipate that full output power of the panel.

Instead of a crowbar, which would not reset, you should use a shunt regulator which senses the current on the capacitor side of the rectifier diode and shorts out the panel output whenever the cap voltage is above your limit. That will "reset" as soon as the cap voltage drops below the limit. But you have to build some hysteresis into the system so that you do not end up with the shunt transistor partially turned on and heating up.

So I'm thinking about a circuit like this where the scr is replaced with a mosfet. Fair idea?

[inetdog]

I love that. And I should have diodes for each panel anyway to keep the panels from bleeding down the cap overnight with their leakage.

So I'm thinking about a circuit like this where the scr is replaced with a mosfet. Fair idea?

Nope. Not right at all for the purpose. The SCR is used in this circuit specifically because it will "trigger" at a certain input. Below that point, it will be drawing zero current and therefore not heating up. Above that point and it will trigger, pulling enough current to blow a fuse or breaker in the malfunctioning power supply but not dropping much voltage or having to conduct the current for very long and so not dissipating too much heat.

If you stick a linear device like a MOSFET in there, you will have a linear regulator, in which the FET will draw just enough current to keep the panel output at 15 volts and so the FET will be getting almost the full output power of the panel to dissipate as heat. That takes a big FET on a big heat sink.

I am not feeling interested and helpful enough to find you a circuit which will do exactly what you want (sorry, maybe someone else will ), but the key word you need to search for is "hysteresis". Or look at the circuitry for a "shunt-regulated solar charge controller", which is designed to do almost exactly what you want.

[mmmalmberg]

Nope. Not right at all for the purpose. The SCR is used in this circuit specifically because it will "trigger" at a certain input. Below that point, it will be drawing zero current and therefore not heating up. Above that point and it will trigger, pulling enough current to blow a fuse or breaker in the malfunctioning power supply but not dropping much voltage or having to conduct the current for very long and so not dissipating too much heat.

If you stick a linear device like a MOSFET in there, you will have a linear regulator, in which the FET will draw just enough current to keep the panel output at 15 volts and so the FET will be getting almost the full output power of the panel to dissipate as heat. That takes a big FET on a big heat sink.

I am not feeling interested and helpful enough to find you a circuit which will do exactly what you want (sorry, maybe someone else will ), but the key word you need to search for is "hysteresis". Or look at the circuitry for a "shunt-regulated solar charge controller", which is designed to do almost exactly what you want.

Thanks inetdog, makes sense and very helpful. Maybe I can do something like use the scr output to saturate a mosfet? Learning as I go... I guess I was thinking the zener in that circuit would provide the hard on/off but perhaps they are not so hard-edged when they break down...

[billvon]

With the zener, are you saying just stick a big fat zener across the panel? I've done this with a smaller version of the same thing, with a resistor in series with the zener.

Sure that works as well, provided the zener is chosen carefully and its power rating is not exceeded. Resistors can help but most of the dissipation will still be in the zener.

Opening the circuit, I don't know where to get a charge controller for 15V.

http://www.ti.com/lit/ds/symlink/ua7805.pdf

Use the 7815. See page 8 of that data sheet for connection info. You need a 7815, a .33 to 1uF cap on the input and a .1uF cap on the output. Parts available from Radio Shack and/or Digikey. Add a diode on the output (an MBR350 would work fine) if you don't want the small amount of discharge from the 7815 to slowly draw down the caps.

Re. regulators, I already have an 80W 15V buck/boost regulator that I was planning to use, but I'm finding that bringing the cap up to 15V with a 15V supply is very slow and I want to be able to pump up the cap as quickly as possible

It is likely that you are collapsing the input with the buck/boost. It will always try to draw as much power as possible to regulate its output voltage, and that means it may be sucking the panel down to very low voltages (5 volts or so.) In that case the panel's output will be greatly decreased.

so I'm looking for simple alternatives to the 15V regulator. In fact I need something super simple as I need to ship the piece I'm working on in a couple of weeks and have MUCH to do besides sorting this out. Which is the appeal of the crowbar...

The crowbar may not do what you want. If it's an SCR, for example, when will it ever release the panel to charge again? You'd have to wait for night for the panel to get below the SCR's holding voltage.

[Sunking]

FWIW you do not even need expensive Zener diodes. All you need is a series string of plain ole diodes to give you the voltage drop you require. Each Diode has a voltage drop of .6 to .7 volts. So for 3 volts 3/.7= 4 to 5 diodes of the proper current capacity which should be roughly 125% of the panel Isc. That is as simple as it gets and CHEAP.

[inetdog]

Thanks inetdog, makes sense and very helpful. Maybe I can do something like use the scr output to saturate a mosfet? Learning as I go... I guess I was thinking the zener in that circuit would provide the hard on/off but perhaps they are not so hard-edged when they break down...

A Zener diode does not "break down" in the same way that an SCR does. If you look at the voltage versus current curve you will see that they conduct a little bit (leakage) at below the threshold but that in the knee of the curve the current will vary with voltage. What you need is something which exhibits "negative resistance" in that the current goes directly to a very high value at the threshold and does not stop until the voltage goes back well below the threshold.

If you look at a small enough section of the curve, a Zener still looks linear.