Easily measure diode capacitance and reverse recovery
Vishay » 1N4002, 1N4148, MUR880, UF4004
The other day Linda from Purchasing came to me with a problem: Lou from Operations needed to source a replacement for a shorted diode on a switching power supply. The darned thing was marked with a strange part number that no amount of Googling could decipher.
There was a recognizable logo marking, but that manufacturer could not provide a data sheet. The part number was from a previously acquired company and was unique to a specific customer. We were on our own.
Fortunately there was a second identical unit in for repair, and Lou was able to provide me with a good diode of the same type. Now all I had to do was figure out what it was. A standard rectifier diode? A zener? Schottky? Reverse voltage breakdown rating? Junction capacitance? Recovery time?
From the DO-41 package size it was easy to deduce that the rating was a watt. It was also easy to inject various currents and measure the forward voltage drop to determine that it was not a Schottky diode. Hooking up a few power supplies in series and gradually increasing the reverse voltage (with adequate series current limiting resistance in case a zener threshold was reached) proved it was not a zener diode – at least not below 200 volts.
The required PIV rating could be resolved by initially substituting a test diode with a high voltage rating, and scoping later.
This left only the unknown junction capacitance, Cj, and reverse recovery time, Trr. This is the time that a diode remains conducting when suddenly switched from forward to reverse biased. I had to figure out a way to measure these parameters. Not with exotic equipment; just enough to get in the ballpark – in other words, a function generator with a falltime of 40 ns, and a 100 MHz scope, which was all I had to work with.
The test setup was easy: Drive the diode-under-test (DUT) with a 5 V pulse – DC offset set mostly negative to bias the diode on only during positive peaks. Scope both sides of the DUT and trigger on the turn-off edge. Varying the DC voltage offset controls the DUT forward voltage and conduction current. Measure the DUT conduction current as the voltage dropped across its series 50 Ω resistor.
The first thing to do was to characterize the test setup. How does a Mickey Mouse® test like this relate to real diodes? That was determined by first measuring some known DUTs for comparison and determining the validity of the setup. I tested the following DUTs and found the results quite interesting:
The time scale of 100 ns/div is constant for all images for ease of comparison.
This shows how a standard rectifier diode is fine for use at 50 or 60 Hz, where the gradual smooth change into reverse bias takes much longer than the diode's TRR. But you can see that in abrupt switch mode operation, the diode becomes a virtual short circuit for a large percentage of the cycle period. Not good.
Now let's compare the above to a 1N4148.
Next let's compare a MUR880.
The small TRR makes this diode suitable for high current switchmode use, but even so, one of the limits to the operating frequency is how fast the diodes come out of conduction. Synchronously driven FETs instead of diodes can overcome this limitation.
Finally we look at the unknown diode.
Based on the above measurements, the UF4004 or UF4007 would be a good choice to substitute for the unknown fast-recovery DUT. I told Linda to get both and have Lou try the 1000 V UF4007 first, measure the peak VR, and, if low enough, use the 400 V UF4004 whose forward VI curve more closely matched the unknown.
Materials on the topic
Inside a huge PCB factory: https://www.youtube.com/watch?v=_XCznQFV-Mw
News on theme: