My series blocking needs to pass 8 A and I am wondering if I can use 4 diodes (rated 2 A each) instead of 8 A.
We have a cost constraint, hence we are looking for an alternative to replace a high spec'd diode (8 A).
Below is the idea that I have in mind.
I appreciate any suggestions/lessons regarding this. Thanks!
Single diodes in parallel will not share current perfectly.
They also won't share perfectly imperfectly, so we can conclude that the total current rating is somewhere between 2A and 8A.
How much inbetween, depends on manufacturing variance, temperature matching, tempco, and V(I) curve.
As a rough feel (but short of doing actual engineering; I have no datasheet to reference, anyway), I would probably expect more like 6-8 in parallel for the required rating here.
A single say DPAK diode is looking pretty attractive for that case.
Single loads with single diodes, are absolutely no problem; the currents are set independent of each diode.
If the blocking voltage is low enough to allow you to use Schottky diodes, use them. If you need higher blocking voltage, i.e. silicon diodes, you need to add resistors in each branch to avoid thermal runaway due to negative tempco on the forward voltage on Si diodes. Check the datasheet for exact values, but it will be in the range of 200 mV at max current.
EDIT: I misremembered. Loose Schottkys may also experience thermal runaway. Pre-packaged ones, for example two with common cathode in TO-220 are normally thermally bonded strong enough to not requiring series resistors. Sorry about that. Again, check the datasheet for max difference in Vf over temperature and apply series resistance accordingly if you absolutely can’t use a single one.
Best solution is to just buy a 8+ A capable diode in the first place. Plenty of them available in the market.
Diode current sharing can be bad. This can be mitigated by overframing. In other words, by using more than four 2 ampere diodes for your 8 ampere need.
Both normal Si diodes and Schottky diodes have a negative temperature coefficient which is not good for current sharing. A common heatsink with good thermal coupling will help sharing.
Real diodes have a semiconductor junction with its well documented voltage equations like Erbers/Moll in series with bulk resistance.
The voltage across the bulk resistance is more significant in power devices driven at or near thier maximum current rating. For example, the voltage drop across a rectifier diode in a car alternator will be around 1.1 volts at the alternator's nameplate rating, so it can be seen that extra power wasting series resistors may not be needed.
Fortunately this series bulk resistance has a positive temperature coefficient which could balance out the undesirable negative temperature cofficient. Look at the curves carefully and you may get away with this.
From experience, some of the new SiC Schottky diodes have a stable overall positive temperature coefficient at reasonable currents and shared well on the bench.
Alternative to a single high spec'd diode
It appears the goal of the circuit is to protect against reverse connection of a DC power supply. This makes the solution simpler than rectifying AC current would be.
A good alternative is the classic P-MOSFET circuit (schematic from this question):
For a more detailed explanation of the options, see Onsemi application note AND90146.
