What type of diode is best for protecting batteries in a weather station?

[sophiecentaur]

TL;DR

I need a diode to stop backup batteries (3.2V) from charging when there's an external supply.

It's a mod to a weather station which normally uses primary cells alone. There are four cells, connected as 2X2. I am providing dc power up there for another purpose so I thought it may as well supply the station as well. But I need to protect / isolate the batteries from the DC source with a diode.
I was hoping someone might have an idea about a diode that would supply say 0.5A (in short bursts) without too much V drop. What do people use these days? Germanium, Schottkey?
Or what about a fat capacitor to avoid a sag during the periods of high current? It's a very practical question and someone will know what't the best approach, I'm sure.

 

[hutchphd]

There are IC's that will do this essentially seamlessly. If you want to go old school and lightning proof there is always a double throw relay (NC) powered from the DC but it wouldn't be fast. That's the place I'd start, but I am old school by default.

 

[sophiecentaur]

There are IC's that will do this essentially seamlessly. If you want to go old school and lightning proof there is always a double throw relay (NC) powered from the DC but it wouldn't be fast. That's the place I'd start, but I am old school by default.

Relay! Really? I can see the appeal tho'. I guess the switching time wouldn't endanger the battery and the soldering wouldn't be a problem. I just don't get on with this 'ere newfangled leadless stuff. It won't even wet the tip of the iron whatever temperature I set it to and the surface mount things are a real problem for me. So I like to keep things chunky and simple; a relay might be the way forward - that is if you really don't think a diode would do it. It was always diodes for split charging of 12V lead acid batteries, until VSRs entered the world of boating. But 3V can lose a lot across a series diode. Hence my question.

 

[hutchphd]

Without knowing how much overhead you have on the 3.2 volts the diode might bite you. For simplicity you could test a Schottky diode and see if it works. I think there may be a way to hot wire a FET to reduce the "on" drop but I think the protection (internal) diodes defeat your purpose here. See if anyone actually knows a better way!

 

[sophiecentaur]

Without knowing how much overhead you have on the 3.2 volts the diode might bite you.

I was having some alarms, on a sporadic basis about the battery level and I measured the old batteries at 1.3V. I think the load is very suitable for primary cells as it seems to work in short bursts and they tend to recover in between.
But I really don't want to put much effort in because the mains supply doesn't fail very often (and the receiver would go down at the same time) I could just remove the batteries, I suppose.

Do you have a type number for a Schottky diode that would handle a few 100mA? I seem to remember they were mostly for small signals. (Obvs, the reverse voltage would be no trouble)

 

[DaveE]

Look at 1N5819 et.al., a very common ancient schottky. It will drop about 0.4V or so. If you think that's too much then you'll want to go with a MOSFET solution (or IC) for the right approach. If you insist on old school, then a relay; which I learned to hate in industry. Things that move in circuits are nearly always worse than things that can do the job without moving (if they don't get too hot). But I will confess to some hypersensitivity in this area.

 

[hutchphd]

Looks good to me. If you get issues you can always go with relay later.

 

[rbelli1]

Watch the reverse leakage on the diode. Some battery chemistries get angry if charged even at a small fraction of a mA.

BoB

 

[sophiecentaur]

Watch the reverse leakage on the diode. Some battery chemistries get angry if charged even at a small fraction of a mA.

BoB

Hah! Just when you thought it was safe to come out of the bunker.

D'you know what? I think I'll just do without the batteries and put an indicator LED on the station to let me know if the supply is on.
What I like about that solution is that involves minimal effort.

 

[rbelli1]

What kind of battery is this?

BoB

 

[sophiecentaur]

2 X1.5V (nominal) primary cells - bog standard.

 

[rbelli1]

Do you have room to upgrade to three cells? Then you can use sufficient standard diodes in series to drop the voltage to an acceptable level.

BoB

 

[Baluncore]

What do people use these days? Germanium, Schottkey?

We use a MOSFET, and control the gate.
The body diode conducts when the gate is off.

 

[sophiecentaur]

What kind of battery is this?

BoB

Do you have room to upgrade to three cells? Then you can use sufficient standard diodes in series to drop the voltage to an acceptable level.

BoB

There are 2 alkaline in series and that pair seems to be in parallel with another similar pair. We may think DODGY but it was still running on the original set, after probably three years, continuously.
The four cells are in a moulded compartment and the whole unit seems to be quite intelligently weatherproofed. Inside shows no sign of wet but a spider sometimes gets into the 'two bucket' style rain gauge and there's a hint of dust. And it's 'out in all weathers' as they say.

But honestly, all this aggro, when I could just be replacing the batteries with a DC supply. The supply would also supply the fan (thermometer) which, at present is only operating when the PV cell gets sunshine. Of course, that's the only occasion when the unit is likely to be far above ambient temperature.
The contributions to the thread are much appreciated, chaps. It's nice to get down and dirty with electronics sometimes.

 

[Jay57]

I would be very surprised if a Schottky would not work. Very commonly used in many products just as your are describing. First the electronics are usually able to operate at much lower voltages than supplied. 2.5V for your weather station would not surprise me. We develop RF integrated circuits and our proto boards are powered from 5V wall bricks. We use Schottky there to protect against accidental polarity reversals. Looking now at one of our boards Schematics I see we use a MBRS240LT3G from On Semi. Its 2A device so at 0.5A its foreword drop is as others suggested 0.2 at 25C and only increases to 0.25 at 85C. The reverse leakage is about 10uA at 2V reverse which is not going to hurt your battery's at all. Just keep the charged as a benefit. Good luck.

 

[Baluncore]

We use Schottky there to protect against accidental polarity reversals.

Welcome to PF.
I would not use a series diode. I use an N-chan MOSFET. Connect source to the load input, drain to the supply negative, and the gate through a 10k resistor to the load positive. The body diode conducts when first connected, then the MOSFET turns on. Get the supply polarity wrong and the body diode blocks current, and the MOSFET is turned hard off.

 

[DaveE]

Welcome to PF.
I would not use a series diode. I use an N-chan MOSFET. Connect source to the load input, drain to the supply negative, and the gate through a 10k resistor to the load positive. The body diode conducts when first connected, then the MOSFET turns on. Get the supply polarity wrong and the body diode blocks current, and the MOSFET is turned hard off.

Yes. Polarity reversal protection is a different (easier) problem. The series MOSFET is the standard approach. I most often see the P-Channel version in the positive lead, but it's exactly the same idea. It's essentially the same as (some) synchronous rectifiers in SMPS outputs.
https://components101.com/articles/design-guide-pmos-mosfet-for-reverse-voltage-polarity-protection

 

[DaveE]

Welcome to PF.
I would not use a series diode. I use an N-chan MOSFET. Connect source to the load input, drain to the supply negative, and the gate through a 10k resistor to the load positive. The body diode conducts when first connected, then the MOSFET turns on. Get the supply polarity wrong and the body diode blocks current, and the MOSFET is turned hard off.

This was my first thought at a good solution. Series MOSFETS for isolation, perhaps like a solid state relay, but simpler. The drain/source connections are obvious, but in <5 min of sketching, I didn't come up with a simple way to control the gates. The problem is the voltage difference between the sources may be +, -, or 0. Plus they share the same voltage at the load.

I think this is a subset of the hot-swap modular redundant power supply problem, which IIRC, always have a control IC if they don't want diodes (schottkys + remote sensing). It's solvable, but not a trivial problem.

edit: This is the part where someone shows me a really simple elegant solution for the MOSFET control that I'm not creative enough to think of. Happens more often that I care to admit, LOL.

 

[rbelli1]

https://www.digikey.com/en/products/detail/stmicroelectronics/BAT41/603518

This is a very low leakage schottky diode I found.

BoB

 

[Tom.G]

For low forward voltage diodes (around 300mV) try this search:
https://www.google.com/search?&q=low+vf+schottky

Cheers,
Tom

 

[Helena Wells]

If you want low forward voltage diodes you can try Ge diodes as well.

 

[sophiecentaur]

Having taken off the inner cover of the unit, I find there are two series diodes, isolating the two batteries from each other.
I can't read the type numbers easily but I plan to do away with a battery alternative and just use an external supply from inside the shed.
Thanks for all the contributions though.

 

[hutchphd]

Then you can leave batteries in or not without worry...just supply the external voltage at the diode junction...no real backcurrent...or as you say don't worry about it. Either will do fine.