1

Hi I have a chicken coop door opener that is battery operated. The batteries tend to run down every other month even though the thing is basically just a digital alarm clock with a stepper motor attached.

Anyway I have some Ni-MH batteries and a 12V car solar panel that would be good to recycle into a solution for powering this door opener.

In full sun the panel seems to put out 22v in shade its about 14v.

The clock is on 24hrs a day. So ideally the batteries would recharge fully during the day while supplying power to the clock and have enough juice in them to get through the night.

I'm really quite a novice when it comes to all this so I was hoping someone could point me to a circuit to charge these batteries with a solar panel or some other advice would be welcome.

I'm pretty sure the charger is this one (I've had the thing for over a decade so can't be too sure, no other markings on it) http://www.jaycar.com.au/Ecotech/Solar-Power/Battery-Chargers/Smart-Solar-Battery-Charger/p/MB3501

parts to recycle

The batteries are setup to provide 6 volts (roughly) to the chook door clock.

EDIT

I've added a few more images showing the input values and output values.

Peter
  • 113
  • 7
  • 1
    Any other charger specs. Current on short circuit and/or (even better) into 4 cells in full sun would be useful. Cells can be in almost any charge state for that test. – Russell McMahon Mar 23 '15 at 03:58
  • noob question...how do I tell you the current on short circuit? Do you want me to stick the 4 cells into a battery pack and connect the + & - ends to the panel somehow? – Peter Mar 23 '15 at 04:28
  • I think the charger is this one http://www.jaycar.com.au/Ecotech/Solar-Power/Battery-Chargers/Smart-Solar-Battery-Charger/p/MB3501 – Peter Mar 23 '15 at 04:30
  • Isc - shorting panel into current range of meter with enough rating will not hurt panel or meter. If that's the correct panel my visual guesstimate of Wattage / Isc was spot on :-). | Current into battery will be close to Isc in full sun as you are loading Vmp (max power) down to about 5V. This will not hurt panel BUT charging MUST be stopped when battery is full as those cells will cook at 120 mA when charged. My answer so far applies as is. – Russell McMahon Mar 23 '15 at 05:40
  • And yes. Panel+ - meter on amps - 4 cells - Panel-ve. A contortionist OR an engineer with enough skill luck time and patience can do it on the ground with no battery holder (ask me how I know :-) ). Two people helps :-). BUT my answer seems about right. Where are you located? – Russell McMahon Mar 23 '15 at 05:42
  • i'm in canberra – Peter Mar 23 '15 at 09:09
  • This should be useful. 1st line of table 4 is sunshine hours equivalent per day ~~. eg 6.85 hours average/day in December http://www.gaisma.com/en/location/canberra.html – Russell McMahon Mar 23 '15 at 10:05
  • Circuit diagram and notes added. – Russell McMahon Mar 23 '15 at 10:54
  • http://www.vishay.com/docs/68633/sup53p06.pdf datasheet for SUP53P06 – Peter Mar 25 '15 at 06:29
  • Is the voltage shown with or without battery. WITH battery you get Vbattery = < 5.5 V until charged. WITH battery you should get about 5.5V - maybe connect a 1K load where battery goes to get some load current - and even better a capacitor - say 100 uF or higher across the 1k. – Russell McMahon Mar 25 '15 at 10:12
  • There is a modest chance that the cct is not stable as is and is oscillating - with battery load I'd expect it to be OK. - as above - is this with or without battery? – Russell McMahon Mar 25 '15 at 10:13
  • without the battery – Peter Mar 25 '15 at 10:18
  • BD650 from Jaycar will work better as turns on at about 1.5V or less on base. BUT I'd expect the FET to work. – Russell McMahon Mar 25 '15 at 12:09
  • I just tried to make sense of you photos - a few words can be worth 1000 pictures sometimes :-). IF you are applying the PV panel with 4.7V OC then getting 3.9V is no surp[rise. When you have 20V OC the connected output should be correct but it is not obvious from the photos that you did that. – Russell McMahon Mar 25 '15 at 12:12
  • Your photo angles + detail level + sharpness combine to make it wholly uncertain where a number of important connections go. The LM336 connections appear wrong but I cannot be sure. Note that the datasheet pinout is shown from the bottom view. – Russell McMahon Mar 25 '15 at 12:22

1 Answers1

3

Under charge currents of << C/1 a Nimh cell is fully charged at about 1.45V. Using 1.4V/cell gives you slight lee way at the loss of a small amount of capacity. Slightly lower again is even safer.

If you don't mind wasting solar energy (and that should not be a problem here), feed the batteries through a diode (if there is not one in the panel already, and clamp the battery voltage at 5.6V (1.4V/cell). [ Or 5.4V at 1.35V/cell for good safety].
I assume battery temperatures are in the 20 -30C range usually - best voltage will vary somewhat with temperature but that should work well enough.

That panel is perhaps 2 Watts (you may have a spec there or can measure short circuit current in full sun). If so then Imax is about atts/Vmax_power = say 2W / 15V =~~ 130 mA. Actual could be 50 200 MA - neither extreme being too likely.

100 - 200 mA is too much for a cheap TL431 clamp regulator by itself.
A TL431 driving a TO220 P Channel MOSFET (plus a few resistors) or an N Channel MOSFET plus any small PNP transistor will give you a clamp regulator suitable for the task.

TL431 divider string top resistor can be maybe 100k so drain on battery when there is no sun is around 50 uA = not a problem in this application.

Or you could use a standard series regulator such as an LM317 plus 2 resistors set to 5.6V would work but backfeed via the regulator and resistors adds slight complexity.

There are other ways but the TL431 + MOSFET clamp should work well enough.

TL431 "turns on" when gate voltage >= 2.5V.
Z1 on pulls Q1 gate low turns Q1 on which dissipates excess energy in Q1 + Rload.
Rload is optional if MOSFET can dissipate all energy OK - but usually using a resistor avoids needing a heatsink.
Rload = V/I <= (Vbattery_max - V_FET_on) / I_panel_max
V_FET_ON is the voltage drop across the fully on MOSFET
= Rdson x I_panel_max.
With a MOSFET with Rdson = say 0.1 Ohm then on voltage at say 150 mA = = V = IR = 0.15 x 0.1 = 15 millivolts, so a half decent FET needs minimal allowance for on voltage.

Say V_FET_on = 0.1V, Imax = 150 mA, Vbat max = 5.5V.
Rload = V/I = (5.5 - 0.1) / 0.150 = <= 36 Ohms. 33 Ohms OK.
Lower OK but FET will then make up some of load and dissipation may be higher.

schematic

simulate this circuit – Schematic created using CircuitLab

LM336 datasheet - All LM336 datasheets I found were poor (Faichild, TI, LT).NONE gave adj pin current.
If Vclamp is not correct R3 & R2 may need to be lowered while maintaining 1.2:1 ratio.
eg 33k : 27K

Russell McMahon
  • 147,325
  • 18
  • 210
  • 386
  • Yes please. if its no trouble. – Peter Mar 23 '15 at 04:33
  • I think i get the gist of what is going on there but I'm not sure how to go about choosing the bits and pieces to make it happen. – Peter Mar 23 '15 at 22:16
  • I have tried to edit it on a breadboard cos its the way I think about this stuff http://123d.circuits.io/circuits/670335-the-unnamed-circuit/ does it look right? – Peter Mar 23 '15 at 22:52
  • @Peter See amended circuit. Jaycar part numbers added. Z1 can be LM336. BD650 is PNP darlington - should work well. Other parts are std - show cct to Jaycar if having problems with parts – Russell McMahon Mar 24 '15 at 00:03
  • @Peter - Your breadboard is wrong in several places. As it has no labelling it is somewhat hard to read. I could do but no need as you just need to follow circuit correctly and it should work OK. || Your FET has two pins shorted. You seem to have an extra diode. You have 2 x Rs in series for no obvious reason. ... . Just follow circuit given and ... . – Russell McMahon Mar 24 '15 at 00:15
  • Thanks for editing your image...it will help me find the parts better. – Peter Mar 24 '15 at 00:32
  • I think I've made the circuit correctly, but like you description says there's some wiggle room in the resistance. R0 is that purple bit of wire, the others are as you specified. Does the voltage seem right to you? I thought we were chasing 5.5v out. It looks like we've fallen short. advice welcome. – Peter Mar 25 '15 at 06:27
  • Russell - I'm coming to this question late in its lifecycle but - wouldn't it be better if the voltage clamp is before the series diode? I know that the temperature coefficient of the diode now becomes a problem but at least the quiescent current of the circuit is zero when the solar panel is in darkness. – Dwayne Reid Apr 05 '15 at 04:52
  • @DwayneReid Only sort-of maybe. The clamp proper can be on the PV side but the sense resistors really need to be on the battery side so the TL431 is controlled by actual battery voltage. Iq when battery is "Just right" is ~ 3V/68k =~ 45 uA or about 1 mAh / day. In this context that is entirely acceptable. | Having the divider on the battery side allows accurate clamp voltage of charged battery. Placing the divider on the PV side means Vdiode is part of the sense voltage. Whether a Silicon (as shown) or Schottky diode is used the voltage drop between heavy charge and final whisper ... – Russell McMahon Apr 05 '15 at 07:33
  • ... of current is maybe 0.2 or 0.3V. This is significant in battery charged terms and if it cannot be stably allowed for may cause under or overcharge. Over charge is especially bad as above a certain V charging never stops. | Note that when the clamp is not working ONLY R3 & R4 draw current - the rest is vvvv close to 0 (under 1 uA). Any time Q1 or Z pass current they are intended to – Russell McMahon Apr 05 '15 at 08:12