Switching Power supply

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
jackinnj said:


yes, the linear regulator burns watts. but DIYr's don't have to advertise the fact that their products are "green".

Sorry but probably I was not so clear:
I am not warried about the efficiency (better ... not in that case) but I think that I cannot obtain 10A@35V from this solution.

jackinnj said:

for the moment, Linear's site seems to be down, but I note that they do have a low noise gate driver, and that the 1A chip like the 3439 can be used to drive other devices. it is really worthwhile to take a look at the PDF or the Application Notes for these devices.

the LT3439 has come down in price -- $7 each -- pretty stemmy compared to a TL494 or SG3525 -- but you probably save in terms the magnetics and snubbers, if you want a low noise power supply.

one thing which both National and Linear take great effort to point out is that some of the current pathway's conduct a lot more than one would suspect (and for very brief periods of time) -- there are nanohenries everywhere!

Do you talk about that:

http://www.linear.com.cn/pdf/dn316f.pdf

If yes: this solution is not usable for my application (see first answer).

If not: plase tell me more, if you want, about your idea.
Thanks

Giuliano
 
At the end I had to choose between 2XIRFB20N50K e 2XIRGB20B60 so I estimated the power losses of the switches (I had some limitations: case not bigger than TO-220 and dissipator with Rth=1.7C/W).
The winner was: IRFB20N50K. What do you think about? Did I made some mistakes?
Thank a lot


I think the major differences between IGBTs and MOSFETs are slighty slower rise/fall times and higher conduction losses at low currents at the expense of lower losses at high currents

If you want to produce less EMI then rise and fall times have to be increased thus MOSFETs lose its advantage in faster rise/fall times

I'm working in a 2KW 160..250Vin to 450Vout PFC prototype [45Khz] and I've tried 2x IRFP460 from ST and 2x SKP10N60 from Infineon [essentially what I had at hand]. I'm using a magnetic snubber to keep dI/dt controlled during diode reverse recovery, since I'm using simple 2x MUR 860 diodes from OnSemi [this reducesdramatically both radiated EMI and switching losses]. I decided to discard IRFP460 in favour of SKP10N60 since rise and fall times were almost the same and IGBT conduction losses were slighty lower. Also, those IGBTs were much cheaper

Also I had a hell of >10Mhz ringing troubles that went away when I tried IGBTs

For high voltages I prefer IGBTs or bipolars since MOSFETs switch too fast [and show lots of 'gain' up to insane high frecuencies] and require an extra effort to slow them down. Also, with MOSFETs inductor winding techniques, PCB layout, etc.. are far more critical, everything seems to resonate at 25Mhz [or 40Mhz!!] when you connect it to a drain

Check out MJE13009, this bipolar shows around 70ns turn-off crossover times and 500mV Vce-sat for Ic=6A when properly driven [Ic/Ib1=5, Ib2=Ic/2], altough it's not usable to switch more than 350V. I like it for audio SMPS [no ringing at all and 1Kw full bridge output with a small heatsink and no forced cooling]
 
Eva said:

I think the major differences between IGBTs and MOSFETs are slighty slower rise/fall times and higher conduction losses at low currents at the expense of lower losses at high currents

If you want to produce less EMI then rise and fall times have to be increased thus MOSFETs lose its advantage in faster rise/fall times

I'm working in a 2KW 160..250Vin to 450Vout PFC prototype [45Khz] and I've tried 2x IRFP460 from ST and 2x SKP10N60 from Infineon [essentially what I had at hand]. I'm using a magnetic snubber to keep dI/dt controlled during diode reverse recovery, since I'm using simple 2x MUR 860 diodes from OnSemi [this reducesdramatically both radiated EMI and switching losses]. I decided to discard IRFP460 in favour of SKP10N60 since rise and fall times were almost the same and IGBT conduction losses were slighty lower. Also, those IGBTs were much cheaper

Also I had a hell of >10Mhz ringing troubles that went away when I tried IGBTs

For high voltages I prefer IGBTs or bipolars since MOSFETs switch too fast [and show lots of 'gain' up to insane high frecuencies] and require an extra effort to slow them down. Also, with MOSFETs inductor winding techniques, PCB layout, etc.. are far more critical, everything seems to resonate at 25Mhz [or 40Mhz!!] when you connect it to a drain

Check out MJE13009, this bipolar shows around 70ns turn-off crossover times and 500mV Vce-sat for Ic=6A when properly driven [Ic/Ib1=5, Ib2=Ic/2], altough it's not usable to switch more than 350V. I like it for audio SMPS [no ringing at all and 1Kw full bridge output with a small heatsink and no forced cooling]

Thanks a lot for all information.
I'l surely try this BJT but unfortunately, at the moment, only because I am curious.
Infact the (low cost) idea is to use an IC to drive the switches (in that case the new NCP1653 from onsemiconductor) so I think that its drive-current is not capable to drive two MJE19003.
I look forward to other SMPS discussion.
Thanks

Giuliano

jcarr said:
Giuliano:

I am using a switching PS (130kHz, 500W) to power a fully discrete amp as well as one based on LM3886s. If anything, I find the results to be superior to a normal 50~60Hz PS design. IME, it is essential to reduce the switching PS' harmonics (both transmitted and conducted) as much as possible.

Good luck, jonathan carr

Thanks a lot for sharing your experience.
If you want could you say what device you used for the application?
Thanks

Giuliano
 
Note that optimum base drive for bipolar transistors requires a instantaneous base current proportional to the instantaneous collector current, and this involves current transformers. There is no place for exotic driving ICs

An externally hosted image should be here but it was not working when we last tested it.


Note : The center winding on the upper side of T2 should have the reference point in the right terminal instead of the left one

This classic schematic illustrates proportional drive for bipolar switching transistors [2x 2SC4242, half bridge, 35Khz]. The control IC [IC1] only has the capability to turn off one or both transistors by shorting the current transformer [T2] in one direction or in both directions, but turn-on process happens entirely due to positive current feedback [very smooth process]. The control IC is only capable of prohibiting turn-on but can't force it or speed it up

Also, the transistors doesn't instantaneously turn off when the current transformer is shorted, instead, shorting only starts the base charge removal process that lasts for 1..2uS [then the transistor turns off]. The turn-off dV/dt and dI/dt are entirely dependent on switching transistor properties [very smooth for bipolars]. The crossover time depends on Ib2 [the amount of negative base current used for charge removal] and Vbe2 [the amount of reverse biasing caused during crossover period due to current transformer leakage inductance - hard to understand until you try it]

This cheap and ancient aproach produces lower EMI than any similar circuit using MOSFETs. It's still being used in nearly all PC power supplies, they pass EMI tests with just a small common mode filter [most of the EMI is produced in the secondary rectifier section anyway]
 
Eva said:
Note that optimum base drive for bipolar transistors requires a instantaneous base current proportional to the instantaneous collector current, and this involves current transformers. There is no place for exotic driving ICs

Gosh :headbash: it's true!
Sorry, sometimes it's better if I take a while to think before answer :) .
Thank you for schematic and explanations.
I'll study the schematic as soon as possible.

Giuliano
 
if you have a ham radio club nearby (don't know what the situation is in Port, NL, ESP, IT etc.) the most recent issue of QEX has a really thorough article on ATX rebuild -- the supply is redesigned for amateur service so that the output is constant from about 100 milliamps to 20 amps, here are some snippets from the schematic:
An externally hosted image should be here but it was not working when we last tested it.
[p]
An externally hosted image should be here but it was not working when we last tested it.
 
jackinnj said:
if you have a ham radio club nearby (don't know what the situation is in Port, NL, ESP, IT etc.) the most recent issue of QEX has a really thorough article on ATX rebuild -- the supply is redesigned for amateur service so that the output is constant from about 100 milliamps to 20 amps, here are some snippets from the schematic:

Thanks a lot
I'll study it too
I hope to have enough time :bawling:
bye

Giuliano
 
allanon77 said:


Thanks a lot
I'll study it too
I hope to have enough time
bye

Giuliano

don't get upset -- the whole thing can be built on and around an existing AT printed circuit board -- you just cut out a lot of the extraneous circuitry -- you arent' running multiple disc drives etc.

you'll have to do some serious thinking about the transformer, control loop etc. since your needs are different...

the interesting thing about the QEX mod is that it is sufficiently quiet to be used with a ham radio receiver where a signal can be some 10's of microvolts into a 50 ohm front end (a very adequate signal, called 599, is 100uV into 50 ohms) -- take a close look at the multiple LC filters on the output.
 
Hi,
Not that I'm an expert in the field but.....

As for the input stage I would use some sort of PFC topology,
with FETs of course, to create the DC link voltage. The FET in the
PFC stage is circuit common referenced and need only a "low side"
driver= no esoteric pulse transformer design. Also the benefits of
using PFC (boost) for the input are wide input voltage and low
THD on the mains side.

After the PFC stage there must be the DC/DC stage which uses a
transformer to get the isolation barrier. There are a variety of topologies
to consider here. You just want to bring the DC link voltage from
the PFC down to what you want as output. A half or full bridge will
need some level shifting of the gate commands for the upper FETs.
Here you need to consider either a driver with level shift (high side driver)
or use gate pulse X-formers.
Also a flyback design could do the job.

Personally I would just stick with the old iron core and big caps unless
mass production is called for.
But if you really want SMPS it is perfectly possible but keep the safety
goggles on and use at least a variable ac source. Make sure all
low voltage circuits are running OK before any HV is applied or
things can get expensive.

/ Mattias
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.