Connect with us

another powersupply question

Discussion in 'General Electronics Discussion' started by spankey666, Jan 23, 2012.

Scroll to continue with content
  1. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    Hi all,
    I am working on a project .i want to build a power supply using some 3.3kva transformers i have kicking around. these are single phase, 240 in 110v out. and i need adjustable voltage and constant current controls.
    i have posted this on other forums and got the usual "are you mad " and "you're gonna die" type responses so time to find somewhere that will actually give some helpful advice , hence me finding here..
    I have a basic understanding of electronics, but there are a few things i have never had dealings with , one is transformers, the other is mosfets and triacs. so some help regarding these would be appreciated. all of the linear circuits i have found basically just become heaters when high current is concerned.
    so can mosfets be used like this ??
    [​IMG]

    I've also been reading up on pwm controls, again and require some advice on them as well, but 1 question at a time.
    thanks
     
  2. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    If it were DC it could work, but not for AC, and for high current & voltage it would just become an unmanageable heat source as you say - since it's linear.
     
  3. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    thanks, yes its dc control, my idea was if this worked, possibly run several mosfets in parallel with cooling. This is where my lack of understanding comes in, i was under the assumption that when switched on, mosfets have very low resistance thus dont soak as much heat as bipolar transistors. and 60A mosfets are relativly cheap
     
  4. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    When switched hard on yes, but that's not what happens when you dim down that circuit. To cool 3kW's worth you'd have to resort to water cooling.
    PWM could be the way to go. It's a simple form of switchmode. The gate drive circuit would have to be quite different then.
    MOSFET's and Bipolar transistors will heat up just as much given the same (linear) working conditions - as will IGBT's.
    They simply have each their own area of "expertise" (voltage/ current/ speed - wise) where in a switchmode supply they will perform better or worse.
     
  5. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    excellent, thanks,
    can you recommend any links for ideas ? i have been googling this for days with little success with regards to cc-cv adjstable control on pwm. data sheets are a little confusing for us amateurs :) and most of the schematics ive found are for custom wound transformers and/or the dedicted ic's used are obsolete
     
  6. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    At that power level and voltage it's no small task, no matter what technology you use (except maybe thyristor control, that's easy).
    So is the 3.3kVA transformer delivering 110V AC, rectified 110V DC, or rectified and smoothed 155V DC?
    What is the intended use, and how much power do you want out of it (and at what voltages)?
     
  7. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    I have looked into triac control of the primary, again controlled via opto-diac with pwm. but again theres very little info out there. all references ive found with "dimmer" type circuits require the switching to follow the input sine wave, not from an external signal source
    As to power requirements, realistically 6-30v more most applications with as much current as possible, 10-110v on the odd occasion with lower currents required.
    Its basically a power supply for electroplating, my wife and my self are keen amateurs bordering semi professional when the need arises . most things are done in the lower voltage range but require the high current. my wife would like to experiment with titanium colouring which requires the higher voltages.
    Smoothing of the output can be as crude and unfiltered as you like.
    I didnt expect this to be easy, but didnt expect it to be this difficult either. :)
    so maybe a re-think. I have 6 of these transformers, so lets just concentrate on the 6-30v spectum. the higher voltage will be simple with just a variac type of control. cc/cv isnt required at that end of the scale so i would make more sense to make 2 separate supplies.
     
  8. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    Ok. Now, 3.3kVA at 110V is 30A. Smoothed to 153V you would get 21.5A max. Are those currents in the range you'll need?
    I had believed chrome plating to be a quite current density sensitive process and would not produce good results if fed by an unsmoothed DC source. Isn't that so?
    You could have gone linear if the transformers had a 24VAC (=32VDC) output, but with 153V DC you'll have to burn off kiloWatts of power.
    I'm not quite sure which regulating tecnology would be the easiest to apply though still being sufficiently good for the purpose. I'll think about it.
     
  9. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    well ive had constructive day :) as my previous post says, i'm leaving the high voltage alone for a while, as i need to get on with the lower voltage idea as this all seems to be more complicated than i expected. So i have taken apart a xmfr , which was no mean feat as full on potting compound , ready to rewind the secondary with 4awg . so now i should easily have 100a to work from at 24v . this should make life easier.
    I'm not into chrome, but anodic , nickel and electropolishing actually work better with a ripple. But even with 100a, electropolishing only allows for 72 square inches of surface area !!!! it requires minimum 200A per sq Foot of surface area :)
     
    Last edited: Jan 23, 2012
  10. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    Wow, that's big job too, but it'll be a lot easier to work with electrically - giving good results I'm sure.
    Make several identical and separate windings that you can connect in series or parallel as required.
     
  11. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    there wont be much room on the bobbin once its got 5mm wire wrapped round it LOL , but once ive got more of an idea of whats needed in secondary windings etc, i' do some more research and wait eagerly for your input.
     
  12. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    I was thinking of dividing it into four 6V sections, thus you could get 6V 550A, 12V 275A, 18V 140A, or 24V 137A. Wow, winding a single strand (?) 5mm wire..
     
  13. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    Hahaha.... i thought i was getting carried away :D

    by the time you get to 380 amps , you need roughly 12mm wire LOL,
    i was thinking thin stuff for any fly back control , powering low current control circuits.
     
  14. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    even at 30v wouldnt a linear control still be a massive heat dump ?
    ive found this on the web, if it is true constant current adjustable and not adjustable current limit it would work, would be even better if a way to control scr rectification instead. :)
    [​IMG]

    [​IMG]
     
    Last edited: Jan 24, 2012
  15. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    Oh yes indeed, with 5V and 100A out it would dissipate 2500W.. A tracking preregulator would definitely come in handy.
    Splitting up the secondary is a (manual) way of achieving that btw..
    Here is the service manual for a HP range of high current lab supplies (0-10V 0-100A / 0-40V 0-50A). It's 183 pages and 10MB so beware before downloading..
    On the last page is a not-so-good-quality schematic diagram, but it serves to show the complexity such "bell & whistles" beasts can achieve.
    Here and here on the other hand are examples of simple lab-supply shematics that surely can be expanded as needed.
    A variable current limit is usually x-x'A and not so accurate or stable, whereas an adjustable constant current is 0-xA and quite accurate & stable.
    Your first diagram might work, and could be expanded, but it's drawn in a way that makes me sceptical about the competence of the originator.
    I'm sceptical to lab-supply diagrams with an LM317 in them too btw..
    Also, just as an example of high-current power supplies that can be bought, here is a 13.8-18V 60A 1kW switchmode PSU available for $130.
    Come to think of it, an EV motor controller might be just what you need. They take like 48-192V DC in and can deliver 100's of Amps.
    I don't know much about them though.
     
    Last edited: Jan 25, 2012
  16. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    ive loked at dc/dc converters, etc. havent found anything that can easily adapted. so back to linear circuits.but how about instead of running everything through a transistor bank that gets hot, is it not possible with the right circuit and ic to instead of switching on a transistor, but switch an ac switching circuit.straight from the transformer, kind of like this ?
    [​IMG]
    ive separated the ic and transistor supply for obvious reasons :)
    or even the same idea, but instead of a scr switching rectifier on the secondary, control via opto-coupler a simple triac "dimmer" on the primary ? this would be easier and cut down component costs, but i dont know what would happen with current output if the primary is dimmed like this [​IMG]
    i really appreciate your help on this matter
     
  17. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    I don't think you can control thyristors in the analog way you suggest in the first diagram. They need phase related/timed trigger pulses.
    The second diagram is just an on/off control and needs phase timing too in order to work. The rectified DC output will be very "crude" though.
    Using triac dimming on the primary is probably the cheapest and easiest way to control the output voltage, but you'll find it hard to add a current limiter circuit.
    Here is some triac control theory (application note) btw..
     
    Last edited: Feb 1, 2012
  18. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    i found that sheet yesterday, page 6 it shows a dc motor control. which is kind of what i was thinking about. does an opto-coupler diac not work in a similar way to a standard diac ? ie break down voltage when triggered with light ? i was under the impression that it would still be phase controlling as it would be in sync with the ac signal, but only break down when triggered by the led. i have only found reference to this with ac motor speed controllers, but would this not work on the lv side of a transformer providing the voltage is above the diac breakdown of 35v aprox ? and then be triggered by a suitable pwm control.

    I'm just bouncing ideas around LOL
     
  19. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    That's an optocoupler triac, not a diac, and it doesn't behave anything like a diac. It triggers the triac as soon as the LED receives a sufficient current
    A diac relies on the phase shift of an R-C filter to achieve phase related trigging.
    Here it's the rising voltage (33V) that breaks down & triggers the diac (and then the triac).
    Notice that the motor speed controller has two R-C filters in series. This is to improve trigging on inductive loads. Use the circuit w/o rectifying.
    Since a triac has an almost current-independent voltage drop of 1-2V it would be impractical to put it on the secondary side if you can put it on the primary.
    You'd have to use a low-voltage diac to make it work on a 24V circuit. I don't know if such are made.
    I believe electroplating works best with low ripple, so I think some electronics on the secondary is desirable.
    If you rectify 6V AC you get 6.5V DC. If you need 5V out and 100A you'll only dissipate 150W in an analog regulator.
    PWM is unrelated to diacs and triac trigging btw..
     
  20. spankey666

    spankey666

    19
    0
    Jan 22, 2012
    this has zero volt cross over detection and switches at 20v but only when the led is on. http://kubuntu.free.fr/wiki/data/fp/moc3041.pdf

    so switching by square wave with variable mark space ratio would work to create an "average" output... No ?
     
    Last edited: Feb 5, 2012
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-