arrgg - ES1J failed

Aarrgg - my ES1J failed. It was in a 20V to 350V step-up, and delivering only 35mA.

--
 Thanks, 
    - Win
Reply to
Winfield Hill
Loading thread data ...

It doesn't seem to be avalanche rated.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

The diode smoked and melted its solder joints.

I realize 20:350 = 18x isn't a very attractive step-up ratio. Operating at 10W out, it means a 35mA output current is matched with a nearly 1.6A peak input current, making a 2-volt peak drop on the ES1J diode. But 3W * 5% = 200mW. I observed 2.7W loss indicating 82% efficiency. But the diode appeared to be the only part that got hot. I think accounting for the diode's reverse-recovery time mostly causes heating in the MOSFET switch, which didn't get hot.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Oh, maybe it fried from reverse recovery.

It's a "fast" diode, whatever that means.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Really? You've got that thing operating in hard switching / CCM, at /that/ high of a duty cycle?

That's as impressive as scary...

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

You could use a step-up transformer, and/or a multiplier to moderate the duty cycle.

formatting link

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Yes, it's really not wise. See my answer to John's next post. I just thought that before replacing the inductor with a transformer on the PCB, allowing a lower, more sensible duty cycle, I'd see what the limits were. I was surprised to experience the diode failure.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

ES1J are SMA diodes rated for 1A and 600V. Later today I'd like to experiment with three different manufacturer's parts, plus ES2J types, plus US1J parts. Both of these have higher current ratings. But they also have higher capacitance and longer reverse-recovery times.

The ES1J curve says 5pF at 100V, whereas the others are twice that. My inductor, using an RM8 core, has 8pF of capacitance. The MOSFET, an IPP65R225C7, 0.225 ohms, has 20pF at 100V.

After these experiments, to satisfy curiosity, I'll set them aside and replace the inductor with a proper inductor/transformer, which I've already wound.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Yes, my PCB is laid out for a proper inductor/ transformer, and I've already wound two candidates. The design is a small bipolar 200V 20W converter, with 1:2:2 step-up, which will employ a 33% duty, much more sensible, and I expect no problems. But then, this diode failure was a surprise.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Even if you provide lots of copper a semiconductor in the little SMA package won't survive that for long.

That depends on who wins the wrestling match. Which in turn depends on the Rdson in the FET and if that is low enough the diode loses big.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Yes, at 1/4 ohm, Rds is quite low. So you're saying that in CCM the diode is still conducting, and the FET switches on and the diode refuses to switch off right away... Ouch. Both diode pads ended up detached from the PCB. I'm not sure what to do about it except to use lower flyback step-up ratios, run at lower power levels and avoid CCM.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

What's wrong with DCM? Too much inductor loss?

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

The specs for that diode don't indicate, but another diode I was looking at indicates that the total reverse recovery charge goes up with the rate of change of current, which makes sense. That suggests that increasing the switching time (which will increase dissipation in the FET) will reduce dissipation in the diode. There may be a happy medium there.

Sylvia.

Reply to
Sylvia Else

Larger inductor needed (saturation current)? Higher component peak currents? More EMI?

There are a number of reasons CCM is preferred (though not always possible, obviously).

Reply to
krw

At the higher power levels I was experimenting with, 10 to 15W and above, CCM is preferred.

But my ES1J, cathode at +350V, still conducting substantial current -- its anode suddenly being forced to ground, despite its reverse-recovery- time refusal to go along, WHAMMO! Smoke city.

OK, I really don't understand how to analyze the battle between the MOSFET and the diode, all I know is after a few minutes it melted.

But if the diode was on a transformer secondary winding, where it should be, instead of on the main inductor (i.e. transformer primary), then there'd be some leakage inductance to give it the 25-75ns it needs to open up and be a diode.

Note, I had a 22-ohm gate resistor on the MOSFET and observed a 50ns fall time for the 350V, but evidently that wasn't good enough.

Next my already-wound RM8 transformer goes into place, the duty cycle gets into a reasonable region, and the peak diode current drops. It'd better, because my goal is a safe 25W.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Hmm, before you plop in the new transformer, try a...

*scratches furiously* ~10uH inductor in series with the diode (it needs to be good for a few amps). Put a resistor of about... *scratches a little more* 100 ohms? in parallel with it.

The resistor will get hot, and the inductor will increase overshoot considerably (your supply inductor is probably about the same... which, again, might be a hint about things..), and you'll probably need to:

...Do this, but more. At least this way, turn-on can be sharp (use a R || D

  • R in place of a single gate resistor), but turn-off needs to be just soft enough not to overshoot and blow the FET. Which is, well, again, warning signs... :)

Dropping frequency until it pops into DCM is the best option though.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

No, DCM will work at those power levels with RM8-style inductors like the one I made. If I reduce the inductance and increase the PWM OFF time, it should be possible to stay in DCM.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Or try one of these but they are big:

formatting link

It's a cool company name :-)

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

OK, I'll play with it a little more.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

That is most likely because the FET won the arm-wrestling match. The diode got overpowered at the end of every conduction cycle, tolerated that for a while and then ... PHUT.

Or if you must use CCM use a SiC diode? The larger Vf shouldn't matter in this application. Only thing is that AFAIK only large versions are available and in a production run it costs around 50c more.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.