Producing a 10 millisecond 5 watt pulse using supercapacitors

[jwcornell]

Gang,

I have posted a few times in other areas on this project but an idea occurred to me.

I am working on a project to turn on and off a DC solenoid which requires a 10 millisecond pulse of five to ten volts at 5 watts. The polarity of the pulse determines the direction the solenoid is operated.

The trick here is that the location of the solenoid will be in remote outside areas where no power is available. Since the solenoid will only need to operate perhaps once every two the three days, very little power will be needed.

A reed switch attached to a vacuum gauge that moves very slowly (perhaps one or two psi per hour) will be used to trigger the pulse generator. The trigger point will be adjustable based upon the placement of the reed switch on the faceplate of the gauge.

My initial thoughts centered around the idea of using a 555 chip with a battery but I was having problems trying to figure out how to trigger the pulse only once since the reed switch was likely to be closed for quite some time (perhaps 10 - 12 hours) and vary based upon conditions. Since the device will be used outdoors, I started to think that maybe the use of a super capacitor coupled with a very small solar cell might make for a better solution. As the device will only need to create a pulse for ten milliseconds, it seems to me that this would eliminate the need for batteries and be considerably less problematic without the need to constantly check the battery status.

Ideally, the pulse would be generated only once for every time the reed switch closed regardless of how long the reed switch remained closed.

Has anyone worked with supercapacitors coupled with solar cells? Any ideas on how to design a circuit using this concept to achieve the above? Thanks in advance for your thoughts.

 

[Hero999]

If I've got the correct thread in mind, I believe I suggested using a capacitor.

In future, please stick to the same thread if you're talking about the same subject, even if it differs subtly from the original post.

 

[jwcornell]

Thanks Hero999. Will keep that in mind.

If I recall correctly, your suggestion on the use of a capacitor previously was to allow for the use of a smaller battery in order to supply the 10 milliseconds of higher power needed to trigger the operation of the solenoid. The idea I am exploring here is in eliminating the battery altogether and using a small solar cell coupled with a super capacitor to store the energy needed.

Your thoughts are much appreciated. Thanks.

 

[mneary]

Hardly needs a 'super' capacitor. 5 watts for 0.01 seconds is 0.05 Joule. E = 1/2 V^2 * C. Using E = 0.05 and V = 10, we solve for C. C = (2*E)/(V^2).

C = 0.1/100 = 1/1000 farad more commonly expressed as 1000 microfarads. Since the voltage will droop before all the energy is delivered, 2200 uF gives a safety margin.

As for solar power, an arbitrarily short charging time of one hour needs only (.1 Joule / 3600 seconds) = 30 microwatts. For simplicity I might use a series array of four of these 4V 1.6mA Solar Sensor-The Electronic Goldmine whose 4-5 mW would be serious overkill but the smallest I could find.

A standard 9V 250mAH battery would easily deliver a thousand of these pulses. You would have to check the battery every five years?

I assume that your pulse shaping circuitry is already solved in the other threads. I don't remember how much current those solutions needed.

 

[jwcornell]

Hi mneary. Great to hear from a fellow Californian!

I spoke with the manufacturer of the solenoid and you are correct. There is no need for a super capacitor given the amount of energy needed. He suggested using a 2700 mfd. He also thought that just having the capacitor dump it's charge into the coil of the solenoid would do the trick so long as at least 5 volts was produced but not to exceed 15 volts. He was however skeptical about being able to find the small solar cell which you have solved! Much appreciated. Will order some of these and try the concept out.

The problem with the battery is that the reed switch is likely to be closed for extended periods of time and I need to generate only a single pulse. Using a solar cell solution coupled with the storage capacitor gets around this problem and renders the device essentially carefree. Again, many thanks for your thoughts and will keep you apprised if this concept works.

 

[jwcornell]

Got a follow up question.

I've been reading some posts where there is some concern over the capacitor on a solar charged circuit potentially damaging the solar cells when discharging. There was some talk of putting diodes in the circuit to protect the solar cells.

When discharging such a small capacitor (2200 uF @ 10 volts) should I be concerned about the possibility of damaging the solar cells? If so, what diodes should I use and how should the circuit be designed? Many thanks.

 

[Oznog]

I don't see how a cap could "discharge" through a panel in a destructive way. It's only a leakage current, so it's not gonna blow through there. Without a blocking diode, a panel has dark leakage that will drain the cap but the energy's insignificant, except the cap will be drained. A battery generally shouldn't damage the panel but the panel could drain the battery badly overnight.

You need to be careful about 4 things. One, a panel will produce much more than its rated voltage when there's no load... or, it's just charging a cap. You could exceed the cap's voltage rating.
Two, the cap needs to have a low enough ESR that the I*R voltage drop doesn't interfere with its operation. In fact, many "supercaps" have pathetically high ESR , tens of ohms even (or more), and probably wouldn't be able to drive a coil without massive paralleling.
Three, the solenoid's min holding current. See, a cap might hold "X" amount of joules to begin with, but that's not the important calc. The calc is that 1 amp out of a 1F capacitor creates a voltage change of -1V/sec. Well, coil current is kinda difficult to calculate since it has a resistance and inductance, and the current needed to hold it is much less than the current needed to initially actuate it.
Four, be SURE you have a fast flyback diode on that coil, or else the voltage spike as it turns off will kill your transistor.

I doubt there is any problem with higher voltage on the coil, as long as it's pulsed. For example, I doubt even 40v would do anything harmful to it at all, as long as it's pulsed. The danger would be in the V^2/R heating of the coil when left on for sustained periods, due to the squared term it's much higher for extra voltage. However, given that the power source is solar, there's not even a lot of wattage possible. If the switch were to fail and get stuck on, the solar panel would be unable to charge the cap and the solenoid will only see a small voltage and current.

 

[jwcornell]

Thanks Oznog for your thoughtful response.

You mention four things to be concerned about.

1 I believe I can use a lower voltage solar cell to charge a higher rated voltage capacitor (please correct me if I'm wrong on this assumption). Since the capacitor is rated at 10 volts, I think I could use say a 2.5 volt solar cell which, as you point out, without a load might produce somewhat higher voltage but not presumably four times it's rated amount. Do you have any thoughts on what sort of multiples of rated output voltages solar cells are capable of producing?

2 Given my needs to trigger the solenoid, do you have some thoughts on what that maximum ESR should be?

3 Is there a way to calculate the holding current? This sounds like it might be one of those "just gotta try it" sort of things.

4 There is no transistor in this circuit although I am contemplating related circuits that may contain a transistor. Thanks for the heads up.

 

[jwcornell]

Oznog,

What about putting a voltage regulator on the solar cell charging circuit? Would that solve the problem?