Register to reply

Diodes in parallel

by John1397
Tags: diodes, parallel
Share this thread:
John1397
#1
Dec9-12, 06:09 PM
P: 65
Can one take 4 - 1 amp diodes check them with diode tester and if they match hook them in parallel and get a 3.6 amp load to pass thru them?

John

Phys.Org News Partner Engineering news on Phys.org
Research project on accident-avoiding vehicle concluded
Smaller artificial magnetic conductors allow for more compact antenna hardware
Autonomy expert led team in developing "self-flying planes"
AlephZero
#2
Dec9-12, 07:57 PM
Engineering
Sci Advisor
HW Helper
Thanks
P: 7,275
No. The diodes won't "match" perfectly. Whichever one takes most current will get hotter than the others. This will decrease its internal resustance so it will take even more current ... until something bad happens.

You might be able to get away with putting a small resistor in series with each diode to limit the problem, but that would waste power.

Just buy one diode with a high enough current rating. A 5A diode only costs cents/pennies, not dollars.
yungman
#3
Dec9-12, 08:29 PM
P: 3,904
NO!!!

Enthalpy
#4
Dec10-12, 03:07 PM
P: 661
Diodes in parallel

- Buy one proper diode
- They could at high current and under some conditions... Improbable at 1A.

It needs that the direct voltage increases with temperature or at least that the current share increases less quickly with temperature than dissipation improves.

One possibility is that the series resistance drops much voltage (hence the strong current). This drop increases with temperature and can compensate the threshold decrease.

Not very common, and better stay away of this.

By the way, bipolar transistors have the same worry with temperature stability. Their emitter injects a denser current at hotter spots, which would lead to a local runaway, but manufacturers add series resistances everywhere in the emitter to spread the current evenly.
John1397
#5
Dec11-12, 07:30 AM
P: 65
I knew that it does not work if the are not exactly matched I once put two 1 amp diodes in parallel in a circuit but the current passing thru is only 1 amp so this would probably be ok. I used a 3 amp diode in a circuit that passes 2 amps is this big enough or should you have a 5 amp diode?

John
AlephZero
#6
Dec11-12, 08:15 AM
Engineering
Sci Advisor
HW Helper
Thanks
P: 7,275
Quote Quote by John1397 View Post
I used a 3 amp diode in a circuit that passes 2 amps is this big enough or should you have a 5 amp diode?
It depends on the circuit. A circuit that takes 2 amps "on average" may have higher peak currents. Most diodes will stand short term overloads, but using them like that doesn't improve their reliability.

Allowing a "safety factor" above the theoretical currents and voltages is usually a good idea, but you can't make a general rule about what the safety factor should be.
Windadct
#7
Dec13-12, 07:30 PM
P: 569
Just to be annoying to yungman --- YES you can.... but only for about 4 seconds....... Look up Negative Temperature coefficient --- that is the property that makes this NOT work.
Just buy a bigger diode. Even in very large power systems ( > 1000A ) it is difficult that is why this exists - it is a single piece of SI -->> http://www.semikron.com/products/dat...0_02890170.pdf
analognoise
#8
Dec14-12, 05:32 AM
P: 1
Everyone says no, but if we're talking about hypothetical, identical, exactly matched perfect diodes, there's actually no difference between four diodes in parallel and one especially fat diode.

The semiconductor junction is defined by the width and length. Two identical diodes in parallel are exactly the same as one "double wide" diode.

In other words: Don't try it in real life, but if you answered "No" on a semiconductor device physics test, I think you'd be wrong; can someone verify this? I'd have to go hunting for the answer. I'll do the hunting if nobody else chimes in. :)
Windadct
#9
Dec14-12, 07:50 AM
P: 569
For cases in high current this is done, the diodes are usually screened and matched by the manufacturer for Vf - and the mechanical design needs to ensure the diodes ( usually stud or capsule types) are equally cooled. But the impact to reliability and serviceability is severe so this practice is avoided when possible. In the OPs specific case - it is way too easy to go with the larger diode.
skeptic2
#10
Dec14-12, 02:01 PM
P: 1,822
Once for a high reliability circuit I paralleled two 5 volt regulators through diodes. Each one was capable of supplying enough current for the whole circuit by itself. I don't remember the current requirements for the circuit but let's say it was 0.5 amps. The positive output of each regulator went through a 1 ohm resistor to the anode of the diode. The cathodes were tied together.

To do this you need to consider that the diode with the lower voltage drop will conduct more current and heat up which further lowers its voltage drop. The other diode will conduct less current so its voltage drop will increase. As long as the difference in voltage drops across the series resistors is greater than the difference in voltage drops across the diodes, you're okay. You must be sure to calculate for worst case. This circuit survived accelerated life testing between maximum and minimum specified temperatures.
Windadct
#11
Dec14-12, 02:15 PM
P: 569
Actually on a test - my guess would be they expect the answer to be no - because the NTC on the Vf is so pronounced, that is the reason the question would even be on the test. For example - you would not see a question asking if you can parallel resistors.

And even absolutely equivalent diodes may not "share nice" depending on the circuit thermals - mechanical design.

As for Skeptic's case - this is not really parallel as there are separate sources, and referred to as auctionering diodes, quite common in hiRel. A similar misunderstood case is on VFDs - that share a DC bus, typically you need to add a diode between each drive and the shared DC bus - to ensure one drives rectifier does not supply the DC to all of the other drives sharing the bus - because technically all of their input rectifiers are in parallel.


Register to reply

Related Discussions
Infinitely Long Parallel Rectangular Parallel Strips (H field) Introductory Physics Homework 4
Diodes in series and parallel Introductory Physics Homework 1
Can you add voltages in parallel if you use diodes? Electrical Engineering 2
Why are diodes nonlinear? Electrical Engineering 5
Question regarding diodes Electrical Engineering 2