Suitable fet for solar charging project

CoNative

Hello,

New guy here. I'm electronics newbie and I'm wondering if you all could help with a solar battery charger project I'm working on. I picked up a couple of these panels: http://www.parallax.com/Store/Components/Optoelectronics/tabid/152/txtSearch/solar/List/0/SortField/4/ProductID/619/Default.aspxWith the two panels in parallel and running through a schottkey diode, I was able to charge 4x AA 2000mah nimh batteries from about 4.3V to about 5.35V in about 10 hours.

So then I tried integrating my own battery charging circuit into a project I'm working on based on a Picaxe. When the batteries get low, <4.8V, it will turn on the solar panel, and shut off when they're charged (above 5.2V). The current draw of my project is only about 1.6mA when idle (which it will be most of the time) and occasionally will drive a motor (<300mA) for very short periods of time. Since it will mostly be idle, I just want the solar panels to keep the batteries topped up since this will be on 24/7.

I have the Picaxe turn on the TIP32C (through the n-channel mosfet) and when the battery voltage is low and turn it off when the batteries are charged. I tested this using a partially depleted battery pack and fully charged battery pack to see that it was switching the solar panel on and off. The problem I've run into is that when I actually put this out in the sun for a day with a low battery pack, the charging was very slow, much slower than when I tested it it by itself. The battery voltage after 9 hours in the sun only rose from 4.78V to 4.85V. I haven't had a chance to see what the current from the panel was after it went through the PNP... but I'm guessing that is where the problem is. It nearly looses what was charged overnight, so this isn't going to work well. I need help picking components to do this sort of thing. Anyone have some ideas?

Thanks



BTW, D2 is there just incase I happen to put in Alkaline batteries on accident or the panel actually does put out 6V it won't damage the Pic.

Beau Schwabe

I have a very similar setup but the scale is a bit larger (600Watts) and I'm monitoring the voltages of the battery and Solar.

Looking at your circuit I would lower the 10k resistor to the base of the transistor to something between 1k and 470 Ohms... Also it looks like you have the Emitter and Collector reversed on that transistor. Since you are using a transistor, you can also potentially remove D1.

Tracy Allen

It looks like the PNP transistor is in backwards. The emitter should go to the solar panel and the collector toward the battery. The gain of the TIP32 is about x30, so to get a charging current of 200 mA, it needs a base current of about 7mA, so the resistor R1 should be more like 750 Ω.

An alternative I think would be to have only one NPN or Nmosfet, to break the ground side of the solar panel instead of the high side. The FQP30N06 could do that on its own.

Ariba

Many Battery charge regulator for solar panles just short the Solar panel to swith it off. The FET that makes the short needs to withstand the short circuit current off the solar module, but the demand is very low in this state, because the voltage is near to zero:

Andy

Mark_T

Firstly the PNP is in backwards, swap emitter/collector as people have pointed out. Secondly R1 is way too large a value. The TIP32C has a low gain, not sure what at 167mA, but perhaps 30 to 100 or so. Assume 30 in worst case, you need about 10 to 15mA from the base to guarantee saturation, and R1 should be about 330 ohms to provide that.

Since the current is quite low a power transistor is unnecessary, a 0.5A switching transistor in TO192 case should be sufficient - my transistor collection throws up 2N2907, ZTX550 but there are many other high-current switchers out there. Or a p-channel logic-level FET (though that will require a shottky diode as well I think to prevent discharge through solar panel when no sun). In fact that might be needed anyway as Vebo for TIP32C is only -5V which is sailing close to the wind.

Also you need to detect when the solar panel is able to provide enough power, otherwise you'll just be wasting 15mA into the base of that transistor at night!!

CoNative

Thanks for the replies.

Wouldn't shorting the solar panels damage them?

If the two panels together are capable of 334mA, I don't think the Picaxe would have enough current (20mA max on output pin) to fully turn on the TIP32C, right?

I tried using just the FQP30N06L to switch the low side of the solar panel, but for some reason wasn't able to get the solar panel to turn off completely...i think... I'm going to give it another go today, I was probably doing something stupid. How do you determine what value resistor for the pulldown on the gate of the FQP30N06L? Should I use a lower value resistor for the pulldown than the 10K?

That's a good point about wasting power into the base of the mosfet at night when the batteries are low. I would need to leave D1 in the circuit in order to do the following:


If battery voltage is low and the solar panels need to be switched on, I can switch Q1's gate high and read the voltage across LED1 (which is behind D1) on Pin 11 of the Pic. If the solar panel isn't providing enough power, switch Q1's gate low to save power. Repeat every x minutes until the sun comes up and solar panel is able to provide enough power and then leave Q1 high until batteries are charge or the sun goes down again

Thanks again.

Tracy Allen

Oh, right! The reverse diode in the mosfet allows current to flow even when the main channel is off. The flow of current in that circuit when charging is counter-clockwise, so the (-) side of the solar panel will sit at about 1 diode drop above ground.

For simplicity the short circuit across the panel as suggested by Andy can't be beat. No, the short circuit across that small panel will not damage it or the wiring. The diode D1 prevents back-current.

CoNative

Tracy/Andy, wouldn't it be wasting current keeping the solar panel switched off in that design?

Beau Schwabe

Shorting the Solar panels? ... yes, I know everybody...well "many" anyway seem to do it. Honestly that doesn't help the cause towards solar panel efficiency. Even the minimalist in me would look for an alternative solution that makes better sense.

Tracy Allen

The purpose of shorting the solar panel is to avoid overcharging the battery per the original question. Yes, it is dumping current into a short circuit when the battery reaches the point of being topped off, so if you have some other place to use it, well, fine. Divert it to charge another battery, or heat your hot chocolate, or break down water into hydrogen and oxygen. It doesn't matter whether you dump the current or shut off the current, potential energy is not being used. But can you call it wasted? What if the solar panel hadn't been in the way at that particular spot on earth?

Ariba

I was also surprised when I saw a circuit that shorts the solarpanel. But these solarcells are different from batteries. In fact the inefficiency is the same if you make a short or switch it off. See the attached picture. In one case you have a lot of voltage and no current, in the other case you have a lot of current but no voltage.
As Tracy says: of course you waste energy that you get for free from the sun, but as long as you can't use it for anything else you have to waste it.

Andy

CoNative

Sorry, i didn't mean to start a debate about wasting the solar panels energy. It's interesting that shorting doesn't damage them. What i meant to ask was about wasting the batteries current sourcing the mosfet's gate high so the solar panel is essentially switched off. Switching the low side seems like it would be more efficient on the batteries. With the pic sending current to Q1's gate, I measured the current tonite (no sun on the panels though) and it was only 0.5mA... does that sound right?

I'm up for trying Mark's suggestion too with using more suitable parts. Instead of using the n-channel, could i use an NPN to switch the low side?

Tracy Allen

What point in the circuit did you measure 0.5mA? Oh, and which circuit is it, the one from post #1?

One nice thing about mosfets is that their gate draws zero DC current. So it doesn't matter if it is left high or low. The pulldown resistor from gate to source does draw current, but that resistor can be a high value, 100k or 1M, so it only draws microamps.

Tracy Allen

Here is a picture that helps to clarify what goes on when the mosfet is in the low side of the solar panel. The circuit would work fine if you put a mechanical relay there or if the mosfet did not have the reverse diode. But with the reverse diode, and when the solar panel SP has a voltage greater than battery B, even if the mosfet is off (gate connected to source), there is a current path around the circuit through the two forward biased diodes. So charging continues even when the mosfet is off.

A bipolar transistor would be better, but when you look at it closely, there is still a "sneak" path through the base-collector junction and through the PIC's subtrate diod.

CoNative

I measured the load between the battery and entire project using my original circuit in post #1 (btw, I had the PNP installed correctly I just screwed up the orientation in the schematic). My code checks the battery voltage 5 seconds after power up. I watched the load stabilize at 1.53mA, then rise to 2.03mA after the check with a battery pack that needs to be charged. With a set of fully charged batteries, the current would stay at 1.53mA after the check. So I assumed the .5mA increase was the current needed to keep the mosfet switched on for the battery pack that was low.

Last night, I built the circuit I posted in #6 and the current only increased from 1.53mA to 1.71mA after the voltage check (on the low batteries). So less than .2mA rise with the mosfet switched on in that configuration. That makes sense about your schematic showing how the charging still occurs if the panel voltage is higher than the battery voltage.... I saw that behavior. Just shining a flashlight into the panel, with the batteries disconnected, LED1 would turn on, although very dim. As soon as I would connect the batteries, the LED would turn off until after the Pic ran the voltage check and switched on the mosfet, then the LED was much brighter (as bright as just lighting the LED with the solar panel outside of my circuit). But in the real case scenario, the sunlight is going to win every time and possibly overcharge the batteries.

Wouldn't my circuit in post #1 work if I used more appropriate parts (like n-channel without reverse diode or NPN and one of the PNPs Mark suggested)? It seemed to work like I intended other than not charging the batteries very fast. I have a BC547 NPN transistor I was thinking I could use to switch the ZTX550 PNP Mark suggested, shouldn't that work?

Or is using a relay or shorting the panel my only options?

Thanks again for the help.