0
\$\begingroup\$

I am fooling around with a solar panel and I am trying to make a circuit, which would monitor its voltage and eventually switch the load on and off.
The thing is, I don't have a good "reference" voltage to compare with. The solar panel can produce anywhere from 0 to ~20V. If it is high enough I can use zener or even voltage regulator, but when it is (almost) dark both of these options will drop in voltage.
So, I was wondering if it is a good practice to use a battery (like a coin cell or something) just to feed one input of a comparator. It doesn't need to be precise, and I think there will be virtually no drain from the battery, so it will last very long.
Another option is to stick with the zener and abuse the gate threshold voltage of the NMOS (whos gate is connected to the comparator output). I mean, if the comparator outputs either ~0V or the zener voltage and that zener voltage drops low enough, the MOSFET won't turn on anyway.
I will probably ask a separate question with more details of what I am trying to do and whether it makes sense at all, but right now I am just asking whether using a battery as a reference voltage for a comparator (and nothing else) is something common, does it make any sense, etc.

\$\endgroup\$
  • 1
    \$\begingroup\$ This sounds infinitely much like fixing a problem that doesn't need fixing. When your solar panels output less than 1V, what, exactly, are you turning on/off with the ouput of the comperator? The answer is either: Something powered by something else -> Supply for your reference. Or: Nothing, because there is no power -> Why the hell do you need the comparator then in the first place? \$\endgroup\$ – Asmyldof Apr 30 '16 at 17:43
  • 1
    \$\begingroup\$ There are a lot of pitfalls associated with switching a load connected to a solar panel based on voltage. For example, when you disconnect the load, the panel voltage will typically go up by a lot. This may cause your simple comparator to reconnect the load (after some odd microseconds). But reconnecting the load will cause the voltage to drop back down, so your comparator will disconnect again, etc. This is an oscillation, and can be bad or undesireable for lots of reasons. There are various ways to deal with it, but they can't be explained in a comment. \$\endgroup\$ – mkeith Apr 30 '16 at 17:55
  • \$\begingroup\$ @mkeith: Actually, this is exactly the effect I am hoping to abuse - by rapidly switching the load on and off I could keep the panel near its "maximum power voltage" region, hopefully squeezing a bit more power in less-than-optimal conditions. Most probably I would be just wasting power, but it is still fun to try and experiment. I could even learn something in the process. But I agree with you - this is another topic, which I intend to post as soon as I've made up my mind enough to ask a meaningful question :) \$\endgroup\$ – Mishony Apr 30 '16 at 18:21
  • \$\begingroup\$ In order for this to work well, you will probably want to use a capacitor on the solar panel side, and an inductor, diode and capacitor on the load side. Basically you are going to be designing a hystertic buck converter that maintains Vin at Vpp. The only practical application I can think of for this is a solar battery charger. An interesting idea that can probably be made to work. You will need to work out a way to avoid over-charging the battery. \$\endgroup\$ – mkeith Apr 30 '16 at 18:32
  • \$\begingroup\$ Why not just combine the two - a small re-chargable battery and a precision voltage reference run from it to get you over those long winter nights. \$\endgroup\$ – JIm Dearden Apr 30 '16 at 18:56
4
\$\begingroup\$

Don't use a battery as a voltage reference. To quote my firmer boss and mentor: "Battery voltage is a nonlinear function of everything. [That includes temperature.]"

Zener voltage reference, or a bandgap voltage reference would work for you. They are available in a variety of voltages.

Keep in mind also that most A/D converters inside microcontrollers have a built-in voltage reference, which can be enabled from code.

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$

Actually, this is exactly the effect I am hoping to abuse - by rapidly switching the load on and off I could keep the panel near its "maximum power voltage" region, hopefully squeezing a bit more power in less-than-optimal conditions.

What you are after is called MPPT or Maximum Power Point Tracking:

Solar cells have a complex relationship between temperature and total resistance that produces a non-linear output efficiency which can be analyzed based on the I-V curve. It is the purpose of the MPPT system to sample the output of the PV cells and apply the proper resistance (load) to obtain maximum power for any given environmental conditions.

Of the three types listed, "perturb and observe" most closely matches what you are attempting.

You certainly could use a battery as a reference, but batteries are inherently non-linear, the panel output will be non-linear, and the load will likely be also. That's a whole lot of variables which will likely make the system highly unstable and unpredictable. I'd suggest researching MPPTs and see what others have done for more ideas.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ A specific solar cell will have a specific maximum power voltage that doesn't vary too much with light level. So it is half-way decent to regulate the panel voltage. There are some MPPT units that do this. But this only makes sense when the load can absorb the maximum power for long periods of time. The load power will vary with illumination level. I think it is a decent idea for a battery charger. \$\endgroup\$ – mkeith Apr 30 '16 at 18:42
  • \$\begingroup\$ Being far from expert, I think what I am trying to do is a poor man's "Maximum Power Point Control", not MPPT ... the difference (in my head) being that I am not trying to actively track the MPP but instead just assume it is somewhere around a known voltage (17.5 according to the panel specification and more like 12.5 according to my measurements) and try to keep the panel voltage close to it. I realize that even if it works it would be pretty bad compared to off-the-shelf solar charge controllers, but my goal is to experiment and learn, I don't really plan to use it for anything specific. \$\endgroup\$ – Mishony Apr 30 '16 at 18:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.