Anyone have a suggestion for the best offline flyback design tool? I see ones from Power Integrations, Infineon, TI, ON Semiconductor, and others.
Has anyone designed a power supply using one of these tools? I'm mainly interested in the transformer design since I will determine which devices to use for the switching circuit.
For example, the BVdss rating for the Power Integration devices is 700 volts. So if I use their tool to design the transformer, in some cases I would have to use an external device such as an NCP2117 driving an external Mosfet.
The designs I would be creating would be 100 watts or less with a maximum output voltage of 35 volts. Input VAC range would be 85-135.
Has anyone designed a power supply using one of these tools? I'm mainly interested in the transformer design since I will determine which devices to use for the switching circuit.
For example, the BVdss rating for the Power Integration devices is 700 volts. So if I use their tool to design the transformer, in some cases I would have to use an external device such as an NCP2117 driving an external Mosfet.
The designs I would be creating would be 100 watts or less with a maximum output voltage of 35 volts. Input VAC range would be 85-135.
Voltwide, very interesting. I played with your spread sheet. The calculations are very close to what the Power Intergrations tool calculates.
I just finished a 45 volt, 2 amp supply based on PI tool calculator output values. I used a EFD30, N87 core gapped for 787 uH.
Output voltage varies from 46.7 volts at 93 ma load to 46.9 volts at 2.35 amp load. The only annomally I see is with a 1.2 amp load the output jumps to 48.1 volts.
I just finished a 45 volt, 2 amp supply based on PI tool calculator output values. I used a EFD30, N87 core gapped for 787 uH.
Output voltage varies from 46.7 volts at 93 ma load to 46.9 volts at 2.35 amp load. The only annomally I see is with a 1.2 amp load the output jumps to 48.1 volts.
Nice to see that PI have adopted my calculations
Irregularities of output voltage often indicate a change of operating mode, from discontinuous to continuous or vice versa.
My calculations are strictly based on discontinuous "critical conduction mode" aka "boundary conduction mode", aka "quasi-resonant zero voltage switching" (QR-ZVS), aka "ringing choke" converter. There is no fixed frequency, operating cycle frequency drops with load, i.e. lowest frequency at max load.
Over the years I designed a series of such off-line converters. My favourite controllers are 8pin PFC-controllers like L6565 from ST, an off-the-shelf product that simply does the job. Others like FAN7527 may work as well.
Concerning the ferrite, you may have a look at ER28 cores, often seen in far east smps.
Center pole is round, and the bobbin is very wide enabling good coupling essential for flyback.
Irregularities of output voltage often indicate a change of operating mode, from discontinuous to continuous or vice versa.
My calculations are strictly based on discontinuous "critical conduction mode" aka "boundary conduction mode", aka "quasi-resonant zero voltage switching" (QR-ZVS), aka "ringing choke" converter. There is no fixed frequency, operating cycle frequency drops with load, i.e. lowest frequency at max load.
Over the years I designed a series of such off-line converters. My favourite controllers are 8pin PFC-controllers like L6565 from ST, an off-the-shelf product that simply does the job. Others like FAN7527 may work as well.
Concerning the ferrite, you may have a look at ER28 cores, often seen in far east smps.
Center pole is round, and the bobbin is very wide enabling good coupling essential for flyback.
Last edited:
I have a short question about smps, I want to build a inverter this winter who is for welding, I need a constant voltage source.
Am I right when I say that I need a type with voltage feedback and a current limit protection (only for as it stick).
and a current source have no voltage feedback and only a adjustable current limiter, (voltage version a fixed current limit).
This is for a co2 welder it needs constant voltage.
It is the only question to be shure.
thanks
Am I right when I say that I need a type with voltage feedback and a current limit protection (only for as it stick).
and a current source have no voltage feedback and only a adjustable current limiter, (voltage version a fixed current limit).
This is for a co2 welder it needs constant voltage.
It is the only question to be shure.
thanks
I have found this.
maybe it is of use for designing, looks good.
Design of switch power supplies
regards
kees
maybe it is of use for designing, looks good.
Design of switch power supplies
regards
kees
Schematic and transformer data docs are attached.
Note: Supply is designed to operate between 85-135 VAC. So it can't be used on European lines without re-designing the transformer. Also, even though the transformer document states 45V @ 1 amp, I have loaded it to 2.35 amps and let it run for 15 minutes with no problem.
Note: Supply is designed to operate between 85-135 VAC. So it can't be used on European lines without re-designing the transformer. Also, even though the transformer document states 45V @ 1 amp, I have loaded it to 2.35 amps and let it run for 15 minutes with no problem.
Theoretical yes, practical no!
If you choose 50um copper foil isolated by 50um tape, and add the interwinding gaps, you achieve a quite poor copper filling factor.
Litz wire is much better!
A look at the schematics reveals a flyback converter operating at 65kHz fixed freqency. MOSFET Source shunt is monitored by controller indicating this unit operates in current mode. If you increase supply voltage, primary current rises faster and as the controller terminates on-time at a given current level, the on-time gets shorter with increasing supply. This is the way all these 85~270V converters work. From that point of view you do not have to redesign xformer to comply with EU mains.
Usually the transformer is designed for about 130V flyback voltage. This is per se independent of supply voltage.
By the way everything looks familiar, no strange details for me in the schematics.
I think it would be very helpful for you to monitor the primary current with a scope with the probe across the PowerMOSWFET source shunt. Best trigger signal is at the gate of PowerMOSFET. It is very informative to watch primary current at different load levels, see on-time increase with load, and possibly the boundary between discontinuous and continuous mode.
Needless to say that you take care for safety, i.e. smps must be supplied via an isolating transformer, otherwise bumm!
Last edited:
Theoretical yes, practical no!
If you choose 50um copper foil isolated by 50um tape, and add the interwinding gaps, you achieve a quite poor copper filling factor.
Litz wire is much better!
Thanks for your reaction, I talk about big stuff a full bridge 150 amp weld inverter,, there she use litze for the primairy and bandcopper for the secondary like 1.5 mm thick and 38 mm wide, this because litze that thick is a challence.
Just al these smps are a challence, you can do terroristic attack with it.
regards nice weekend.
kees
I put your transformer data into my spreadsheet. Being written for boundary mode this makes only sense at the the specific load where discontinuous mode ends. In that case I played with the parameters until 65kHz frequency was calculated.
Looking at the data I found the boundary at about 24W input power or close to 0.5amp output current. Under full load you are definitely in continuous mode.
I wonder that you are happy with that xformer: Flyback voltage is quite low with 70V, ferrite core enters saturation, all in all I assume you need about twice the primary turns for 65kHz operating frequency. Furthermore the transformer stays much cooler with litz wire than with single stranded enamlled copper wire. If you cannot acces litz wire, a tri- or quad-filar winding of thinner wire helps a lot.
You will find your application in column #5 in the updated spreadsheet:
Looking at the data I found the boundary at about 24W input power or close to 0.5amp output current. Under full load you are definitely in continuous mode.
I wonder that you are happy with that xformer: Flyback voltage is quite low with 70V, ferrite core enters saturation, all in all I assume you need about twice the primary turns for 65kHz operating frequency. Furthermore the transformer stays much cooler with litz wire than with single stranded enamlled copper wire. If you cannot acces litz wire, a tri- or quad-filar winding of thinner wire helps a lot.
You will find your application in column #5 in the updated spreadsheet:
To be honest, I never dealed with such power. Specially flyback is imho best below 150W. Upto 400W I had quite good results with LLC converters.
There's a demo version of POWER 456 available on the Ridley Engineering website -- Ridley Engineering | - Software
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.