Designing Adjustable 0-30V 100A CV/CC Power Supply Using 3kVA 35V Transformer

Hi, I am now looking to make a power supply adjustable cv 0-30v and would really like it to be adjustable cc as well. I have found various circuits for Cv, but very little on CC hence my plea for help
. I already have a 3kva 35v transformer (no secondary tappings) which i would like to use.
I have bought various purpose built switched mode Power supplies, but these dont tend to last more than 6 months so its getting rather expensive and tedious
power output doesnt need to be that smooth as it will be used for electroplating.
My limited knowledge needs me to keep the design as simple as possible
any help and pointing in the right direction will be greatly appreciated
Thanks in advance.

Hi, it is possible to design a relatively simple switched mode to perform your requirements, of course a linear supply is not viable due to the heat dissipation at these power levels so SMPSU is the only viable option. Making a CC unit should not be too difficult but for the plating application I suggest maybe some current hysteresis will be needed to stop rapid PWM changes as these may result in device heating and reduction in component life unless correct filtering is used.

Linear Technology, Texas instruments and Analog Devices are the obvious 3 PWM controller suppliers for the controller device, the other parts are available from RS or Farnell, Mouser etc. The transformer you have will provide mains isolation but I am not sure what rules may apply to the equipment itself if this is a commercial application, of course it would be necessary to obtain CE LVD on the equipment minimum I guess since it is connected to the mains supply and SMPSU's can emit high EMI over a wide bandwidth from harminic content so these considerations must be part of the plan. Have you looked at Vicor their web site is at www.vicorpower.com they have modular solutions that can be added in to create pretty much most requirements.

I am not a power supply designer but I can investigate further for you and maybe put together a design I can model it forst and run some design simulations also but if this is for a commercial setup then please consider the CE implications here, getting CE is very expensive and takes time which add much to the cost and above all one must consider personal and public safety at the end of the day. Feel free to contact me further if you would like further assistance in this design. Regards, Andy. Email andy(_at_)physics.org

thank you for replying,
i am not a commercial concern, just a very enthusiastic hobbyist with limited resources.
the old skool way of doing this was simply a large transformer with full wave rectification and a large dimmer switch type control on the primary and just watch the needle. but later thinking and research has deemed CC more accurate with controlled deposits especially when anodic coatings are required ie on aluminium..

I've been searching google all night , and very little comes up especially for the high current i require. so maybe , like the old skool, any regulation could be done on the primary side where current is lower ??? I dont know.

email addresses dont show on this forum

Thanks again

Hi again, I think for CC control, isolation and EMI it would be better and easier dealing with the low voltage side (secondary) here. I will take a look at the range of PWM controllers around and get back to you later if that helps, maybe there is a power supply expert who provides a solution that is safe and workable for you in the meantime but I will try and sort a design out for you.

Get back to you ASAP. Andy.

i have done some more searching and photoshopping, and have the pic below..
i have paralleled the transistors to be able to use easily available ones at 30A each. not sure if this will work though . also there is a 0.2ohm resistor , is this required ??
also the op-amp requires a -6v supply, would there be any issues running a transformer off the secondary of the main transformer ? like ive drawn.
if the concepts crap please tell me, i can take it LOL

Hi Sarah, ok looked over your schematic and it is not good, a linear solution will have to dissipate 3kW so it will become more of a heater than a regulator, also the base control should be on the output side of the regulator not connected to the 35V line as this will just bias the transistors on always with no control, the control current has to be driven into the base of the transistors to vary the drive.

The only practical solution is a switched mode one, PWM is then the simplest to make using a simple PWM controller with output filtering and some input filtering and this must be on the Secondary or Low Voltage side of the mains transformer not on the mains side, any mains control will only result in high levels of electrocal noise EMI and make it difficult to set up a CC system. The solution is a simple PWM controller with power MOSFET switchers on the output followed by filtering and a CC sensing output. I will sketch out a schematic over the weekend for you, I have several meetings today as it is Friday. I will get back to you on this design. Andy.

I have done more searching and reading, and all i can find are linear circuits. nearly all references to pwm are for switching LEDs. could you possibly recommend an ic controller that isnt obsolete ? every time i find something it is using out of manufacture components.
I have come up with the pic below. it is constant current only, and to be honest, thats really all i need. i nearly always have the volts on max and just set the current on my existing supply. i am still struggling to understand the terminology. current limit and constant current seem to amount to the same thing in real terms dont they ?
would somebody more knowledgeable than myself be so kind as to look over this circuit and advise on is suitability for my needs. or advise on something that is.

thank you in advance
Sarah

Sarah, the schematic you show is suitable for a load that is placed between the positive rail and the current limiting element. The circuit you have is a low power regulator and not worth scaling up to 35 volts at 100 amps.

I assume you would like a power supply that will supply a constant current to a load that is placed between common and the constant current source.

With the transformer you have you can obtain a 50 volt rail. Then a buck regulator can be used in constant current mode. The transformer you have is probably rated at 3.5 kVA into a resistive load. Into a simple bridge rectifier it will be good for perhaps half that power. This is due to the high charging current pulses into the rectifier/filter causing high RMS current and transformer winding heating. Now factor in a power supply efficiency (after the transformer losses) of 90% and you can build a 1600 watt CC power supply using your transformer. If this is acceptable we can go from there.

Thank you for replying,
As i will be using this mostly in the 12-21v range 1600w will be more than enough for my immediate needs. so if i can realistically have a max 70-80 amps available within the 8-21v range if would be fantastic.

Let's pin down the specs a bit more. What is the nature of the load, how much current (or voltage) ripple can it tolerate, and what output impedance is required?

An ideal CC source has an infinite impedance but that is probably not needed by your load. So a number is needed to design to, such as 1 ohm.

If a low output impedance (a voltage source) will work for your load that would be the easiest to build. To drop the 50 volt rail to 15 volts and lower a Buck regulator is a good way to go. But, it is inherently a low impedance source. To make it a current source requires loop gain and that can lead to problems. Being that you have (I think) never built a high current switching power supply it would be best to build an open loop circuit where the voltage is adjusted. Some output impedance can be provided with a series resistor. Even simpler than a Buck regulator is a Variac (variable autotransformer) at the AC input.

What is the AC input voltage?

The load is as stated above, this is for electroplating. mainly small items of jewelry made from aluminium.
with the anodic process, as the layer builds it becomes insulating, so the voltage has to increase to keep the current at the base metal constant. hence the need for CC control. large fluctuations in current need to be kept at a minimum as with small components it can cause issues , obviously as the load size (area) increases, small fluctuations are less problematic. but pulse (ripple ?) is actually beneficial to the process i am playing with other plating practices including cobalt ,chromium and gold as well as electropolishing which are better suited/ require a cv control in small setups like mine hence my original post. so if both ways are possible then thats the way i'de really like to go..
The input voltage is 240v 50hz
I dont know what you mean by output impedance , but load resistance is dictated by the surface area of the component.
i have read up on variac control, and switching from the low voltage side via opto isolators etc. but again theres very little i can find on the net to implement this method.

the circuit posted above is for hydrogen production by passing a cc through an electrolyte, same sort of thing that i am doing, heres a link http://www.hhocarfuelcell.com/pwm-circuit/constant-current-pwm.php

It so happens i was speaking to a fiend about this yesterday, and after some rummaging around in his shed he came back with a little metal box that contains the exact circuit above that was for one of his old projects that he never got around to finishing. so over the weekend ii will have a play with a small transformer and see what happens.
I am wondering whether or not a simple 2 diode + scr rectifier set up on the transformer o/p controlled by pwm? instead of the mosfet. and bridge rectifier pic (below).
i envisage current detection is via a shunt resistor in series with the load which should be of the right resistance to allow a meter for readings to be put in parallel with.
you are correct, i have never made a high power supply , is it really that obvious ?

Sarah, you said your plating process likes DC with a bit of AC ripple riding on top. Will the process tolerate high current DC pulses of a low duty cycle. For example, 50 amps at 20% duty giving an average current of 10 amps?

Sarah,

I took a look at the products at American Plating Power. They list several SCR power supplies that are DC with just 8% ripple. So, it looks like pulsed high current DC is not used.

That means that for your SCR controlled power supply a ripple filter is needed. That would consist of following the rectifier with an inductor and a capacitor. The trick here is to make it a closed loop system that does not oscillate with different loads.

Have you got an oscilloscope and time to get it working?

i dont own scope, but i should be able to borrow one. as for time,i could do with sorting something out asap as my digimess has packed up again !! the current tracking is all over the place

please forgive me for asking so many stupid questions, but by asking, i can build up a bigger picture in my mind as to what can and cant be done in real terms. this not only relates to this topic, but to my general understanding of electronic theory. so i thank you all for your patience and input.
so my next dumb question. a triac based "dimmer" on the primary of a transformer wont work, this much ive established (except on some electric welders ) would the same approach work on the ac from the secondary pre rectifier ? as then the load will be resistive ? i havent found any reference to this being implemented, so theres obviously a reason why not, any suggestions ?
Edited and added
Also all of the triac circuits ive found work from 120v. is there any reason they cant be used at lower voltages ? or is this because they aare usually used in line main situations ?
ive since found this http://www.littelfuse.com/data/en/Application_Notes/AN1003.pdf page 6.
by my thinking (dangerous i know) if said idea works at 35v, the diac could be replaced with an opto-coupler (opto-diac) and then a simple voltage and current sensing circuit on the output could trigger the triac ????