parallel diodes

[evandude]

I am doing a simple battery-backup system for the computer in my car. in other words, a schottky diode feeding the input, with a small lead-acid battery powering it. It only has to survive the few seconds of engine cranking.

Anyway, since the computer draws up to 2 amps, I can't run just one shottky. Also, the forward voltage increases with current, and I know that the computer needs as high a voltage input as possible, so less forward voltage is better. So I was planning to run several of them in parallel, to minimize current and thus forward voltage.
I know that people usually advise not running diodes in parallel, in case one passes too much current and blows up thus straining the rest.

Do you think I can get away with it? and how many diodes should I use? I have 10, I was thinking of using 5 in parallel.

I know i can buy schottkys rated at much higher currents from digikey/etc... but I don't really want to spend $10+ for shipping/handling on one or two diodes. i got these from glitchbuster for under 5 bucks.

 

[Styx]

get a higher rated current diode - only way.
Since you said it was only a cost thing that is stopping it, as opposed to tech, just spend the extra

 

[evandude]

I was really hoping that wasn't the answer.

Anyway, I guess i'll cross my fingers and hope that i can find a suitable schottky somewhere in the stockpiles of one of my professors here so I don't have to pay $10+ for one.

 

[Styx]

Soz man, parallel diodes are just very dodgy unless you can ensure all lead imedances (of everything) are the same (resistance, inductance and also closeness together for capacitance)


parallel diodes are used - I have an IGBT module at woth with 8 in parallel to get to 800A. BUT this was assembled by DYNEX at die level with the bonding very well done.

you could try it if you want, but I would personally always go for 1

 

[Nigel Goodwin]

If paralleling diodes, just as with transistors, you should include a resistor in series with each one, in order to ensure the current is equally shared between them.

However, I have seen quite a few commercial uses of two rectifiers in parallel! - but it's not good practice.

 

[k7elp60]

I had a very similar problem. I had a vehicle that I wanted to run some amateur radio equipment in, but didn't want to run power cables to the battery thru the firewall.
I had a 70amp/hour gel cell. I was able to parallel 5ea schottky diodes and use power from the cigarette lighter socket. The key is to match the forward voltage drop. Having a power supply and a load for testing powersupplies made it easy. I just adjusted the load for a 2.5 amp current and measured the forward voltage drop until I had 5 with the same forward drop, plus or minus a few millivolts. As an added precaution I put a 15 amp circuit breaker in series. I was able to run the 100 transmitter on the battery, and the battery was being charged when the radio was not it use or on receive mode. The schottky diodes gave enough of a voltage drop that the gel cell battery never got overcharged, yet was charged fully.

On high current rectifers(excess of 100 amps) a lot of the times manufacturers will specify the exact forward voltage drop for a specific rectifier, and the diodes are matched for forward voltage.

 

[Styx]

yup k7elp60 matching is the best that you can do (still not the best IF you can get a single on its own)

That is what DYNEX did for my 800A device (8 die in //el)


A mate at work worked on traction drives at 6kV and he had to series up a load of IGBT's to get the blocking voltage. He had to match the switching characteristics - tuning where needed the Gate resistor so all were switching at the same time.


So in this case not only will the one-state drop be significant BUT the switching is going to be the killer. IF one switches on faster than the others it will take the full current. Equally if one switches off slower than others it will see the full current


The switching characteristic is the real killer - V and I together

 

[Roff]

Schottky diodes have virtually zero switching time. Even for ordinary diodes, recovery times will not be a problem for manual switching, unless maximum instantaneous current is exceeded, because the excessive power peak will be so short that it will not have time to appreciably overheat the junction. This may not be true at 800 amps, but we're talking about a few amps here.
I agree with Nigel about the use of current-balancing resistors. A bigger diode, as mentioned, is a simpler solution.

 

[evandude]

However I am using schottkys specifically for the lowest possible forward voltage. unless i was using very low-value resistors (even an ohm would be doubling my voltage drop at a half an amp per diode) it wouldn't work. The idea is that when the car is on, the voltage will be high enough to charge the secondary battery even with the diode drop. with the car off (but computer left on for a little while) the secondary battery will slowly drop until it is one diode drop below the main battery voltage. (then the diodes would switch on and allow current to come from the main battery, powering the computer) if the diode drop is large enough, the computer will start to screw up and/or reset before that point, and thus the system won't be much good.

Also, when the car and computer are off, there is still a small current drawn by the computer, so the secondary battery would, as before, drop to a diode drop below the main battery voltage. If that voltage is 0.7v or even 1v, then the battery will be forced to recharge that 1v or so when the car is turned on, which seems like extra wear and tear on the battery, and also would make it so i had to have the car running for a certain amount of time before the backup battery was charged enough to run the computer when the car was turned off.

What I was thinking was that I could put a power resistor in parallel with the diode(s)... say, 50 or 100 ohms. That wouldn't make much difference when the computer was actually running, but in the case where the car and computer were both off, where current draw is only a couple dozen milliamps, it would keep the secondary battery charged so at least help with one of the problems. And even when the starter motor was cranking, and main battery voltage dipped down to 6 or 7 volts (the entire reason for using this backup system) it shouldn't allow significant enough current to flow back and reduce the secondary battery voltage...

My other option is to drop $85 on a commercial DC-DC converter specifically intended for this purpose, that can handle engine cranking without losing output... but being a poor college student makes that a little unreasonable until a few months from now.

Any thoughts?

 

[williB]

evandude , this may seem like a silly answer but , why not disconnect the computer's backup battery from the car battery during starting ..
i was actually thinking of putting a pc in my car also..wish i had a laptop ,but i do have a couple of mother boards..
btw what are ya gonna do with the auto PC..?
i was gonna collect data with mine..

 

[evandude]

I have it set up to play music and do GPS navigation, including voice instructions and even voice recognition for giving it voice commands.

I don't have a proper website made up for it yet, but here's my repository of pictures if you're interested.

**broken link removed**

Yeah, I guess I could manually switch off the power from the main battery when starting the car. Considering I have a couple of large capacitors and a big inductor on the power wire to the computer, i could probably even have the power for the computer fed through a relay, which was actuated by a simple voltage comparator circuit, so that whenever the main battery voltage fell below, say, 11.5 or 11.75 volts, it would switch off the relay. the small delay before the relay actually released should be taken care of by the caps, inductor, and backup battery, I hope. I would probably have to bypass the relay with a resistor also, otherwise the computer would totally drain the backup battery when the computer was off with the small standby current... but as long as I kept the resistor to 50 or 100 ohms, no more than 120 mA or so would be fed backwards when the relay was open during engine cranking.

 

[Nigel Goodwin]

evandude , this may seem like a silly answer but , why not disconnect the computer's backup battery from the car battery during starting ..

You could do it automatically with a relay wired to the starter motor feed!, so that as soon as power is applied to the starter motor the relay opens and disconnects the computer battery.

 

[fredsm]

Old PeeCee PSU's generally have Schottkey's on
the 5V output. Look for something like an SC10C3M
(good for 10A, commonly paralleled on the board).
Should be just the ticket.
<als>

 

[evandude]

Whoa, awesome. PC power supplies aren't very hard for me to find, either... I used to work at an electronics recycling place, they have pallets full of them

by 'old' PSU's I assume you mean AT, rather than the modern ATX? If all else fails, I could tear apart the one I use as a bench power supply and see if there are any in there... I wonder how it would work without them? :lol: