regulate solar panel output

[justwantin]

My thanks to both you guys for this discussion. While the math and theory may say it can be done I think in this instance The 9 Volt panel is marginal. The 12V panel is (I believe) 10 watt but I can track that down. It was used to charge 12V batteries used for running radio telemetry and solenoids in an irrigation system.

Probably overkill but I'm interested in learning as well as project outcomes. Just directly hooking up a solar panel to a battery won't offer me any opportunities to beat my head against a wall.

 

[MikeMl]

In the discussion above, a crutial point is omitted. A 4Ahr battery will tolerate an overcharge at less than the 14 hour-rate and the energy will simply be converted to heat.
In our case the battery will take 60 hours to charge and this will take 12 days, so a regulator will never be needed.

I beg to differ. All of the SLA (VRLA or AGM) battery data sheets I have read are very emphatic that any chronic overcharge will degrade the battery, and void the warranty.

 

[Mr RB]

Maybe since the solar panel will only output full current (say at 1/14th C) for about 6 hours a day and there will always be some discharge over the other 18 hours then it won't really get to chronic overcharge overcharge situation?

 

[RCinFLA]

Solar cell model is an illumination based current source with a parallel inherent diode. At a given illumination level it will put out near same current at any voltage up to point where inherent diode starts to conduct. An unloaded panel has all its generated current go down the inherent diode yielding the N x 0.7 vdc open circuit voltage, therebouts, where N is the number of series cells in the panel. The inherent diode voltage is dependent on temp just like a regular diode. Maximum power point is just where the inherent diode starts to conduct. Usually point is selected so no more then 0.3% of available illumination current goes down the inherent diode. Roughly around 0.5 vdc per cell at room temp.

There is nothing wrong with putting a panel that has a Voc of 40 vdc across a 6 vdc battery. It will act as a current source. Problem is when battery is fully charged it will keep putting out the panel illumination based current.

Simple PWM solar controllers just duty cycles a series MOSFET switch to keep the battery voltage in proper range. You need a blocking diode to prevent backfeed at low illumination.

 

[colin55]

You said your panel was:

14x14cm

You wont get 10 watts out of a panel 14cm x 14cm. You might get 2 - 3 watts if you are lucky. That's why I said it could be connected all the time. But not all panels can be connected directly. It depends of the wattage of the panel and the ability of the battery to "absorb" any overcharge. A battery is normally charged at 1/10th of its AHr rating, for a period of 14 hours. This allows a nearly flat battery to be fully charged.
A battery looses a max of 3% of its charge per day.
You must stop charging the battery at the above rate after 14 hours.

If you want to charge a battery for more than 14 hours or keep charging it on a daily basis for an unknown number of hours, the charge-rate must be less than C/20. This means for each AHr you cannot deliver more than 50mA. In fact it is preferable to limit the current to 15-25mA to prevent the battery eventually drying out.
For a solar panel, you can only determine the charging current by measuring it when the battery is fully charged and when the panel is receiving full sunlight. This is the critical factor and even if this occurs for only 1-2 hours per day, it will create gassing and eventual drying-out of the battery.

 

[justwantin]

Sorry perhaps I was unclear. I don't have the firm specs for the 140X140mm 9V panel, only what I can best guess from a weatherd paper label on the back. My 12V panel is the equivalent Of a BP Solar Model SX10M PM 10W VMP 16.8 Pmax 9WVOC 21V.. It came out of a vineyard where it had supposedly gone bad. I popped off the cover on the leads and resoldered a corroded connection to get a good panel.

I'll use this panel for charging. Overkill for 6V SLA's but Maybe I'll get a 12V to play with or heat my workbench in winter/run a fan in the summer ;^)

 

[MikeMl]

Just a another idea. For small panels, a common way of regulating the charge to the battery is to hook the panel directly to the battery, and then install a SHUNT regulator. A shunt regulator does nothing until the battery voltage rises to the set point. None of the panel voltage headroom is wasted in the series drop across the regulator.

Google SHUNT REGULATOR, or look at a TL431 data sheet.

 

[justwantin]

Yes that's what I'm looking at using. The 12V panel I have came out of a system that used 2 panels in a series to charge 2 12V batteries in a series. I remember the diode to prevent drain back to panels at night and an off the shelf shunt regulator set up before the batteries. The system was originally spec'd for 12V. but charging was marginal with some RTU's not charging sufficiently to run solenoids for irrigation because of low voltage. Larger 20W panels would "get in the way" of vineyard operations so a 2nd 10W panel was installed below the first on endposts. They were mounted in such a way that they did not interfere with spraying and could be taken down easily for machine harvesting. The biggest drain was the radios that were periodically interrogated whether required or not. That's just the powesupply side of the mess and the shortcomings of this system whetted my appetite to know more.

 

[justwantin]

Just a another idea. For small panels, a common way of regulating the charge to the battery is to hook the panel directly to the battery, and then install a SHUNT regulator. A shunt regulator does nothing until the battery voltage rises to the set point. None of the panel voltage headroom is wasted in the series drop across the regulator.

Google SHUNT REGULATOR, or look at a TL431 data sheet.

I've picked up a couple TL431's and have looked at the datasheet. I had a go (the only way I learn) with the shunt circuit from the data sheet on a breadboard with no joy. I used a variable bench top power supply and a couple combos of resistors for R1 and 2 using the equation but Vout remained pretty much the same as Vin. I'm quite sure that I have cathode anode a ref wired correctly. Some circuits show a resisitor between Vin and R1. Is this resistor necessary and what value should it be. I may have smoked one of the 431's already but resistance between pins remains the same as the untried 431. I hesitate to try the unused one until I'm sure I've got the basic circuit right.

 

[MikeMl]

Yes, a shunt regulator must work against a series resistance. When you use it with a solar panel, the series resistance is "built-in" to the solar panel, because by it's very nature, the panel is current-limited. When testing a shunt regulator, you either must use a current-limited power supply, or if your supply is not current-limited, then put a resistor between the supply and the regulator.

Say you are building a 7V shunt regulator, and your supply is 10V. The maximum power that a LM431 should dissipate is a 1/2W, so the max current that it should shunt to ground is I=P/E=0.5/7 = 71mA. Since the series resistor will drop (10-7)=3V @ 71mA, R=E/I= 3/0.071= 42Ω

The resistor power rating should be P=I*E= 0.071*3= 0.21W, so a 1/2W resistor is appropriate.

If you need to increase the power dissipation of the LM431 shunt regulator beyond its intrinsic capability when using in on a 12V battery, it is simple to add an external transistor which increase the current capability...

 

[colin55]

Just put the shunt regulator directly across the solar panel. You can't use a series resistor. What is the series resistor going to do?

 

[MikeMl]

Just put the shunt regulator directly across the solar panel. You can't use a series resistor. What is the series resistor going to do?

Do you ever read the preceeding posts? If you had, you might have learned that justwantin hooked a shunt regulator directly across his power supply and likely blew it up!

 

[justwantin]

Thanks for that I'll see what I can do given that info. I had a feeling the current out of my supply might be the problem but I didn't know enough to know why it would be though, or even what to google for.

 

[justwantin]

The maximum power that a LM431 should dissipate is a 1/2W

I haven't thrown a resister at this yet. I've been digesting your reply. I've just about understand how you have come to your answer for resistor sizing but why max at .5W?

There's probably a little variation between manufacturers but dissapation seems to be spec'd at better than .7W on the sheets I've looked at. In my case I have a National in a TO-92 package and max is .78W. Have you factored in a safety margin? and if so is there a general rule of thumb for such margins?

 

[justwantin]

Thanks Mike,

I believe I have a handle on this now, at least off my benchtop supply, and have learned allot too. One last thing. would I place a diode between the tl431 and the solar panel or battery?

 

[MikeMl]

...I've just about understand how you have come to your answer for resistor sizing but why max at .5W?

There's probably a little variation between manufacturers but dissapation seems to be spec'd at better than .7W on the sheets I've looked at. In my case I have a National in a TO-92 package and max is .78W. Have you factored in a safety margin? and if so is there a general rule of thumb for such margins?

Yes, I put in a safety factor. At 0.78W, the TO-92 will be stinking hot.

With a shunt regulator, the regulator itself dissipates most when the load is the lightest. OTOH, the series resistor dissipation doesn't change as a function of load; only as a function of input voltage...

 

[MikeMl]

If you want to prevent the battery discharging either into the resistors around the 431, or into the panel itself, then put a shottky diode between the reg and the battery, pointing at the positive pole on the battery. Set the regulator voltage ~0.5V higher than the battery chemistry dictates.

 

[justwantin]

Are you talking about something like a 1n5817? The system I was involved with had 1n4004 or 1n4001 diodes. I take it there is a difference and it matters.

 

[MikeMl]

Shottkys have a lower forward drop and have a sharper knee; Silicon diodes will work as long as you offset the voltage higher.

 

[justwantin]

Thanks again I have somewhere around .4V forward drop using a 1n4004 but will look at a schottky next time I'm in a shop as my stock of resistors leaves me with using a 1k and 1.8k to get around 7.2v before the diode. The breadboard circuit is charging off the 9v panel as I write. Almost time to move on to something else.

The tl431 is on the hot side of warm but everything seems fine. I am not using a resisitor to regulate Vin as was the case off my benchtop PS. I have noted in other discussion that people refer to a resisitor on the solar panel output. Is that standard for panels? Neither my 9V or 12V have anything but wires off the backside of the panels.