Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Help designing 1.6KW Isolated AC/DC with Constant Current Output

Not open for further replies.


Newbie level 6
Jul 9, 2017
Reaction score
Trophy points
Activity points
I need to design a power supply that will regulate 35A CC. The proposed input power is 3PH 70VAC 2800ARMS off the secondary off a 480V transformer. I need a little guidance to get on the correct path to base a design. Application is industrial equipment.

I have 2 input power options: Use the transformer secondary or use mains input.
If I use mains input the range is 200VAC-575VAC
If I use the secondary winding the range is 60VAC-140VAC (85VDC-265VDC) 70VAC nominal but I need to account for user miswiring in the primary, AKA feeding 480V into the 220V tap. The system has an OVLO but that would not protect this supply.
Required Output is ~35VDC @ 35A with a duty cycle of 10%
I am leaning towards a Push Pull design but not sure the best way to drive it and regulate the output. Should I go with a self resonant design with a closed loop chopper on the output or a IC based design such as the LM5030.
I have seen some interesting designs in the Audio world that use self resonant push pull driving 2 parallel ETD39 transformers at 1000W output, but not sure if that is the correct path to take. The main requirement is that the supply is isolated from the mains so I don't have any sneek paths for the 2800A to wreak havok on.
Any experienced feedback is appreciated. I know I have a lot to learn.

if you want something quick, and size and weight are not an issue, then an SCR solution would work just fine.
--- Updated ---

what do you mean exactly by duty cycle ?

Also, if you want all HF SMPS design for smallness, then what about 3 Phase rectifier, followed by Phase shift full bridge.
(though i am not sure what is the rise time of the 35A pulses)
If its quick then add some Bucks to the output of the PSFB

Yes, per the above, a filtered rectifier, followed by even a standard hard switched H bridge - would get you there quickly & safely, with good longevity for industrial application.

For 200VAC-575VAC, SiC 1200V a good choice, ditto for rectifying diodes ...

I see, yes that duty is 10% so even though its 1225W, its only 122W on Average.
It would be interesting to see what is the highest f(sw) you could use for that Full Bridge....specially when its input is from the 575VAC
(Though presumably 575VAC is referring to the line to line voltage of a three phase system....and its not each phase which is 575VAC?)
And Ref Half Bridge, we can onvestigate the realms of sandwich wound transformer or not?
In a Full Bridge, the leakage inductance, AYK, acts as a nice turn on snubber, so why sandwhich wind when some leakage L is useful?
I used to think Full Bridges need sandwich winding to keep the ringing easier to snub on the secondary diodes...but since your FB will be quite high step down, and NS/NP quite small, then the secondary referred leakage is going to be small even if you dont sandwich wind.

Ive often wondered about a sandwich wound FB, with an added primary side "external leakage inductor"...and diodes to the rails to quell the ringing from this added inductor.
Last edited:

Also, if you want all HF SMPS design for smallness, then what about 3 Phase rectifier, followed by Phase shift full bridge.
(though i am not sure what is the rise time of the 35A pulses)
If its quick then add some Bucks to the output of the PSFB
I would prefer a standard full bridge over a PSFB as this is my first higher powered design and I want to keep it as simple as possible. I do need the output to have a quick response, but I am not sure what that speed is yet. I would most likely prototype the supply and add the bucks if needed after. I will be hand winding the first transformer so I also need to design around what cores and bobbins are available and being fairly new to magnetics, this is a tough universe to learn to navigate. We do have some planar transformers that we use in other switching applications that I might be able to use for this.
Can anyone recommend a full bridge controller IC that they have used and liked. TI has a ton of them that I will start browsing. I also do want to use a higher switching speed to keep the magnetics down. I am limited to a size of about L 250mm x W 120mm x H 100mm for the PCB. Caps and chokes can be external if needed.

The full bridge and half bridge folders of this course, show some controllers for full bridge

SMPS course_little folders
[moderator action: removed link to external file server]

Also, you can use any current mode controller as a full bridge controller...just use the flipflop cct to give you the full bridge outputs.
Last edited by a moderator:

Constant current implies rapidly increasing power, for example, charging a cap at 20A, when its terminal volts are 10V = 200W, trying to push 20A into the same cap at 200V = 4kW

Better to design for constant power converter.

@Warpspeed because, from the OP 1st post: " The proposed input power is 3PH 70VAC 2800ARMS off the secondary off a 480V transformer "

well, you get much better power factor if you rectify 3 phases as opposed to just 2 ( single phase )

Your high Ampere spec makes it feasible to use chokes to reduce AC voltage. Each phase can use 12 mH (and a large capacitor for power factor correction). Thus each phase contributes a share of overall power, staggered so supply to the load is reasonably smooth.

Values are customized for unchanging load. More effort is required to include current regulation.

3 phase 70VAC 3 chokes 3 PFC caps 6 diodes load 1 ohm gets 39 vDC.png

putting 680uF across an AC supply may not be the wisest course of action, at 35V, 60Hz, this is 9 amps ac in the cap - not really sure why they are even there ... ?

Further - the above arrangement could only be constructed if the user had access to all 6 wires from the transformer - this is often not the case.

Just further again - while the middle source is full wave rectified ( FWR ) are the outer ones ?

Also the left had side short would be a problem for mismatched phase voltages ....

Originally I discovered the chokes causing severe power factor error, giving the idea power is wasted. Hence I added the large corrective capacitors in my post #13. It brings down Ampere flow in the system.

The simulation below is simplified. 3 phase delta in the obvious triangular arrangement. Three wires allow access to three joins. Chokes are relocated. (For some reason their Henry value can be reduced.)
Six diodes rectify AC. Load is in the center region.
Generators are oriented properly as seen by location of the white dots.
Green trace is voltage. Yellow trace is current.

I omit the large capacitors. This causes increased power factor error which goes hand-in-hand with greater current in the wires (twice as much as in post #13). However according to articles explaining power factor, this doesn't mean the generators must produce greater current half of which is wasted. It means there is greater current 'sloshing' around the loops. The generators only need to produce the same amount of energy as before with PFC caps installed.

It's up to the builder's decision whether it's worthwhile to correct power factor error.

3-phase 70vAC 3 chokes (large PF error) 6 diodes load 1 ohm gets 35VDC.png

it's now a classic 6 diode, 3 phase rectifier running off a delta connected source

Not open for further replies.

Part and Inventory Search

Welcome to