hi power smps

[Eva]

Originally posted by N-Channel
Intgreal Body diodes,

Also, if you do consider PFC'ing the front-end, you will need to replace the normal standard-recovery rectifiers with ultra-fast rectifiers rated at at least 800V and 15A. The reason for this is now, there are high-frequency pulses being drawn from the AC line: XXkHz (whatever frequency the PFC is running at), modulated at the AC Line frequency of 60Hz.

Steve

That's not really required. The common practice is to add a small capacitor, say 100nF to 470nF depending on power level, across the + and - terminals of a standard diode bridge in order to prevent most of the HF current ripple from reaching the diodes. A considerably higher capacitance is obviously required at the AC side of the diode bridge in order to finish the filtering job.

Also, when multi-kilowatt outputs are required, sometimes it's easier to go modular and stack multiple smaller SMPS units with their outputs connected in series or parallel (depending on voltage requirements) and proper active current/voltage sharing.

[jamesrnz]

Steve and Eva,

I am thinking about the diode bridge and with standard or uf recovery diodes. I suppose the UF diodes would be good if I were depleting the capacity of the doubler but if i were not a small cap would work I will have to do some overhead experiments to see how those little electrons and coulombs are getting along..


I was refering to diodes in series with the mosfets (i was unclear)
in Offline full-bridge SMPS....need help

posts 289 and 290....

i was thinking that the instonaneuos reverse currents might be reduced.......saving mosfets..


I totally agree with safety as the most important aspect, after all if you kill yourself you cant keep playing.

When I was wil the Voice of America (VOA), we had some people die due to their unattentiveness, after all a megawatt transmitter operating at 30kv and 50 amps is not a good thing to commune with.

thank you both...

jimbo

[N-Channel]

Jimbo:

Diodes in series with MOSFETs: I have heard of using Schottky rectifiers of sufficient Ampacity. Since it is in series with and not across the +HV line, there is no need to use 600V Ultrafasts. This would address the concerns Instantaneous reverse current.

Diodes in the Main Rectifier bridge: I seem to remember somewhere that the main rectifiers should now be fast-recovery or ultrafast-recovery, because of the high-frequency pulses modulated at 50Hz or 60Hz are being drawn through them. I think it was Brown's chapter on PFC, or maybe it was an old Motorola (ONSemi) Application Note. Something like "Now the bulk input cap is reduced to 1-2mF, and the high-ripple pulse DC is fed to the PFC Converter circuit. This 1-2mF cap removes the high-frequency component in the RF range, but that's all. The xxkHz PFC component is still present.

However, if EVA says it's ok, then I would accept that as Gospel, and go with a conventional bridge.

[EDIT]: After going back and re-reading the PFC Section, I realize that you could have either MURs OR the 1-2mF cap after the std-recoveries. Having both would be redundant. Doh!

Regarding her comments on the really high power stuff, I have to agree that modularizing it, and putting the secondaired in series is easier. I can't believe I forgot this. I recently tried to fix a neighbor's 2500W Inverter and saw that each section (2- 1250W sections in parallel) used four smaller transformers, with their 40V secondaries in series. This ensures proper current sharing, and parallelling the filtered/rectified high-voltage DCs together was fed to the DC-AC 60Hz Bridge Output section.

Although this is for DC-AC step-up, the same principle applies for an off-line SMPS of multi-kW capacity.

[Eva]

Those diodes bypassing the own MOSFET body diodes are only required in certain situations in which the body diodes would be subject to hard switching otherwise. That's the case for class D amplifiers, motor drivers, and those phase-shifted bridges required to operate at very light loads, but *not* for conventional half and full bridges driving a transformer.

Body diodes have been always a huge pitfall in MOSFET transistors. They suffer from strong charge storage phenomena resulting in a very long and "peaky" reverse recovery process. Furthermore, if a certain dIrec/dt slope is exceeded during reverse recovery or a certain dV/dt slope is exceeded when Vds rises just after recovery, the MOSFET will LATCH itself into an ON state (routinely resulting in the four devices of a full bridge exploding).

High voltage MOSFETs are particularly prone to that failure mode in circuits involving any body-diode switching. That's because the performance of their body diodes is particularly poor and their Rds-on is not low enough in order to shunt the diode most of the time in order to prevent charge storage as it happens in low-voltage MOSFETs. That's a good reason for using IGBTs instead in troublesome circuits, altough there are new generations of high voltage MOSFET transistors with improved body diodes at the expense of higher Rds-on and prices. Check "C3" and "CFD" series from Infineon for instance.

Note that using the own MOSFET to shunt the body diode after it has been conducting may also lead to failure modes as it happens in phase shifted bridges at light loads. That's because charge may remain stored in the diode, thus keeping the recovery process unfinished until an attempt is made to turn off the MOSFET (and if the other side turns on too quickly and tries to "push" the diode above its weak limits... BOOM!!).

You may find some interesting reports googling for "mosfet body diode failure modes"

It's quite disappointing to see MOSFET devices flattered everywhere while their failure modes and weaknesses are just ignored. Most people know that an IGBT can latch itself into an ON state, but nobody seems to know that MOSFETs suffer from similar pitfalls. Also, nobody seems to know that high voltage MOSFETs also suffer from small current tails, like SMPS IGBTs, due to internal gate spread resistance (not all MOS cells are turned off at the same time).

[jamesrnz]

Ok,

I believe that steve had posted a pfc jpg and i have searched for it with no luck.

i downloded i and looked over it and have a question....

there is a 1n5407 across the boost inductor. why? it seems as this would limit the lower voltage that the pfc could go to by creating a half wave rec bypassing the pfc.

i did not see that in the appnotes from ti or onsemi........


jimbo

[jamesrnz]

steve,
it looks like my switching freq is 83.3k from pin 4 of the 33025 so i gues the transformerr is 41.6k.

i need to kick it up abit . this could cause heating also.?

jimbo

[N-Channel]

Jimbo,

Here is that Active PFC. And here is the link to the original posting: PFC Switched-Mode Power Supply

I started that thread, and it sparked alot of good exchanges and idea sharing. Look at post #14 & beyond............

Steve

[N-Channel]

Jimbo,

Forgot to answer your other question: 83.3kHz (Osc) 41.6kHz for the transformer is OK, but you ar egood upto a couple of 100kHz before core losses become significant.

I finally dug up the equation for the frequerncy:

F(osc) = 1/[C(t) (0.7R(t) + 3R(d)], where R(d) is the deadtime resistor between pins 5&7. Mostly this is zero, but can also be 100 [FONT=symbol]W/[FONT=symbol].

So say, for R(t)=2.21K, R(d)=0, and C(t)= 6800pF, your clock frequency will be:

F(osc) = 1/{0.01E-6 x [(0.7 x 2000) + (3 x 0)]}
= ~95kHz, or 47.5kHz on the transformer.

[jamesrnz]

steve,

i found the pfc schematic i was looking for. it was posted be cocoholic.

i sent him an email with no answer back yet.

i have attached his scchematic.

and was wondering why the 1n5407 across the boost inductor?



jimbo

[Eva]

The diode across the boost inductor is only employed during startup. It prevents excess energy from being stored in the inductor and then dumped back to the output capacitors, as too much energy is likely to be transferred, thus causing an overvoltage condition.