Mosfet or Darlington for my hydrogen fuel cell?

Thread Starter

GammaRay

Joined Jun 11, 2012
17
I’m using a photovoltaic cell to charge up a supercap. When the supercap voltage reaches 2.2 volts, a voltage detector turns on a Darlington TIP120 and dumps the supercap’s charge into a hydrogen fuel cell to produce small amounts of hydrogen/oxygen. It works. However, in a quest for efficiency, I tried using a Mosfet (L1404Z) in place of the TIP120, but running into complexities, for example, different size loads seem to alter the turn on point that the Mosfet begins to conduct (as well as other problems). The TIP120 doesn’t have these problem. In terms of providing as much available v and current from the super cap to drive the fuel cell, my question for this forum… are there advantages of using a Mosfet instead of using the TIP120 to transfer the supercap charge to the fuel cell? Is the TIP120 a good choice? Is there a better choice?
 

Wendy

Joined Mar 24, 2008
21,848
Typically a MOSFET conducts much better and carries more current, though the low voltages you named are a problem. You will need a specialized MOSFET called a logic level MOSFET, and I'm not sure they go that low a voltage (but I think they do).

This is not HHO, is it?
 

wayneh

Joined Sep 9, 2010
16,102
As Bill said, MOSFETs are great switches but they need sufficient voltage on their gate to become saturated into switch mode. A BJT needs as little as 0.7V, with sufficient current, to turn on. A big advantage to the MOSFET is that it does NOT require a current, only a voltage.

I've forgotten the value, but you'll need a minimum voltage to get any electrolysis. You're trying to operate within an awfully narrow range if you run electrolysis at only 2.2V. Is there a reason you're not just using your PV panel directly? Why the cap and switch circuit? Your losses in all that may exceed the power devoted to electrolysis.

You might consider a voltage boost circuit. It would have a loss of course, but it might help overcome other losses such as your blocking diode.
 

Thread Starter

GammaRay

Joined Jun 11, 2012
17
Good points all and thank you. Reading up on Mosfets, specs say as a general rule power Mosfets require 10 volts greater than the source terminal to fully turn on, or 4 volts for logic level. I don't have that kind of voltage on this particular project, sooo, I suppose I must use a transistor.

With efficiency in mind, any suggestions for which type transistor to use to dump the supercap charge into the hydrogen fuel cell?
 

wayneh

Joined Sep 9, 2010
16,102
Generating H and O for by solar-powered electrolysis is a legitimate (if risky) activity. The OP hasn't divulged - perhaps intentionally - what he plans to do with the gases produced. It is a bit odd that he wants to lose so much power already stored as electricity in a supercap by wasting it making gas.
 

shortbus

Joined Sep 30, 2009
6,984
You have mosfets that are designed to switch full on. At "digital" voltage levels. Like the IRL510 as one example. You will need max 2 volt to turn it on. http://www.vishay.com/docs/91297/91297.pdf
According to the data sheet, at 2V that mosfet is barely on (from the graph you can't even tell the amperage it will be it is so low) If this is use it will heat-up the mosfet not act as a switch.

Threshold voltage is the turn-off voltage of a gate, not the turn-on voltage.:)
 

t06afre

Joined May 11, 2009
5,936
According to the data sheet, at 2V that mosfet is barely on (from the graph you can't even tell the amperage it will be it is so low) If this is use it will heat-up the mosfet not act as a switch.

Threshold voltage is the turn-off voltage of a gate, not the turn-on voltage.:)
:eek: Ouch. Was to quick. I used that device a couple of years ago. About 4 volt should turn it quite on. Then the resistance is about 0.76 Ohm at 5 volt about .54 Ohm.
 

Thread Starter

GammaRay

Joined Jun 11, 2012
17
I have a tiny table top fuel cell that starts producing at 1.5v. The point of this is to see if a photovoltiac cell can scavenge enough energy from ambient outdoor light (not direct sunlight) and accumulate enough energy over time in a cap to (intermittently) power a small fuel cell. Answer is yes it can, however, I suspect that using a darlington TIP120 to transfer the accumulated cap energy into the fuel cell may be a point of unnecessary loss. Although I'd prefer to use a Mosfet, looks like the low voltage is going to necessitate using a transistor (actually, I can go as high as 3v for the gate trigger). Question is, which transistor offers the best efficiency at =<3v ?
 
Last edited:

wayneh

Joined Sep 9, 2010
16,102
The point of this is to see if a photovoltiac cell can scavenge enough energy from ambient outdoor light (not direct sunlight) and accumulate enough energy over time in a cap to (intermittently) power a small fuel cell.
This still makes no sense to me. The voltage across the supercap is limited to the voltage of the PV cell (unless/until you add a boost circuit). The current used for electrolysis is limited to what the panel can make at the voltage where electrolysis happens, you have said 1.5V. So what I'm saying is, everything is limited by the PV panel power output, which I would think you would want to maximize. Putting a store-and-dump circuit in the way just adds a source of loss. Supercaps leak, circuit resistances burn off power, and they add nothing. Get rid of all that if you want to make more gas.
 

Thread Starter

GammaRay

Joined Jun 11, 2012
17
The solar cell is small, it puts out enough v (when unloaded), but does not put out enough current to drive the fuel cell directly. Allowing the energy to accumulate in a large electrolytic or supercap and then periodically dumping it into the fuel cell does provide enough current (for periodic cyclical operation). Doing it now using a TIP120.

My question for this forum is what switch is most efficient to use, with 3v available for the gate, for minimum transfer loss (I upped it to 3v). t06afre offered up the IRL3502 @ 0.009 Ohm @ 3v.

0.009 Ohm would be acceptable. Anybody see a reason the IRL3502 would not be a good choice in this application?
 

THE_RB

Joined Feb 11, 2008
5,430
Supercaps have a high internal resistance, I have one for a human powered flashlight and the internal resistance of the 2.3v 3.3F cap is in the tens of ohms. They are not a great storage device for "dumping" a high current pulse as much wil be lost in the caps internal resistance. A bank of standard 6.3v electros or other caps might be a better option.

Re the transistor, something like a BC337 will have avery low on voltage maybe 0.1v and turns on at 0.6v and has a working beta of maybe 200 so will switch 200mA with only 1mA base. They can be a good choice for low voltage high efficiency circuits.

A step up would be one of the Zetex brand high performance transistors, I have used them also and they offer advantages at higher currents >400mA but at lower currents <300mA a hand picked BC337 can be better.
 

wayneh

Joined Sep 9, 2010
16,102
The solar cell is small, it puts out enough v (when unloaded), but does not put out enough current to drive the fuel cell directly.
Rubbish. Electrolysis occurs at the minimum voltage, period. Low current is slow, high current is fast. You gain nothing by accumulating current (charge) and then applying it later at a higher rate. Now, if your electrolysis cell somehow overloads the PV panel, bringing it's voltage down below where electrolysis can occur, that's where you should focus. Are you claiming there can be a current drain into the fuel cell in the absence of gas production?
 
Last edited:

Thread Starter

GammaRay

Joined Jun 11, 2012
17
My lil table top fuel cell requires 1.5v and 30ma to begin hydrogen production. Less than 30ma, nothing happens, at least nothing perceptible. Admittedly, the fuel cell is years old, it may be time to buy a new one.

Will try both the BC337 and IRL3502 and compare it against the TIP120.

I appreciate the helpful advise.
 

wayneh

Joined Sep 9, 2010
16,102
Less than 30ma, nothing happens, at least nothing perceptible.
There's no reason I can think of for this. Gas production depends on voltage, and gas production is simply proportional to current. I suspect it's just very slow. Your panel may not be supplying 30mA at the threshold voltage.

It would be an easy and interesting experiment to just let it run for a while and compare results to the pulsing approach. Have you tried that?

I could understand there being a bit of a voltage hurdle that must be overcome to initiate conduction, after which the voltage can be lower. And again, maybe your panel cannot overcome this hurdle by itself. Maybe what you need is a way to "kick-start" the cell into conduction whenever the panel voltage should be high enough but current is not flowing.

If your problem is that the panel cannot overcome the threshold voltage, you should consider a boost circuit. Just pull the circuit out of a solar garden light. It'll give you roughly double your panel voltage at a high frequency square wave. I believe a boost circuit is far more efficient than a store-and-dump circuit.

Here's a good reference, BTW.
 

Thread Starter

GammaRay

Joined Jun 11, 2012
17
Wayneh, great suggestion about using the boost circuit from a solar light (joule thief?) Will definitely give it a try. I've got a couple of those laying around. Good direction, thanks.
 
Last edited:

wayneh

Joined Sep 9, 2010
16,102
Be sure to come back and let us know how it went. One thing you need to know is that, because a boost obviously doesn't create power, the amperage will go down in proportion to the voltage going up. The power transfer might be over 80%, so a 2X voltage boost will give maybe 40% of the original current.
 
Top