Author Topic: 36V power supply - Design considerations  (Read 7788 times)

0 Members and 1 Guest are viewing this topic.

Offline made2hackTopic starter

  • Frequent Contributor
  • **
  • Posts: 424
  • Country: 00
    • Made2Hack
36V power supply - Design considerations
« on: April 05, 2014, 11:28:02 am »
Hello All,

So, excuse my perhaps rather open ended question but I'm a bit lost.

Here is what I need:



I want to make a power supply unit from 230VAC mains to 36VDC with 1A of current, or 36 - 40 Watts (to have some head room).

The problem is that I am at a complete loss as to what kind of transformer to choose. I can't wrap my head around it. Do I need toroidal? Isolation? What VA rating?

Should I build a linear PSU or switch-mode?

I will be powering one LED @ 36V / 1A (or around 36W), so i estimate that circuit should be made to handle 40W (for some extra headroom). And of course we assume that the PSU circuit is not 100% efficient, then what rating of transformer do I go for? 40VA? 48VA?

So, do I get a transformer with 2x18V secondary voltages? 1x40V and then buck it later using some voltage regulator? It seems that either the VA ratings are small, or they are large, somehow I can't find a middle ground. Or for that matter a 36V secondary voltage transformer?  :-//

Any ideas as to how I can design this relatively efficiently or without too much complexity? Is it even possible?

Thanks,

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18022
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: 36V power supply - Design considerations
« Reply #1 on: April 05, 2014, 11:36:47 am »
I'd go for 50VA because most power input circuits involve a rectifier and tank capacitor, because the rectified power from the transformer is 100Hz pulsing then while that voltage is below the capacitor charge it uses no power but as soon as it goes over there will be a current inrush so the transformer needs to be able to deal with it.

Any type transformer is fine for your power. just go with what you can get.
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 6939
  • Country: nl
Re: 36V power supply - Design considerations
« Reply #2 on: April 05, 2014, 11:50:59 am »
A One Hung Low PSU from China for this voltage/wattage is like 4$ and a Mean Well will run you 25$ ... you could get the One Hung Low version simply to get an idea how (not) to design one :)

PS.  they are generally switchers going from rectified mains (with passive PFC) to 36V, probably single switch flyback.
« Last Edit: April 05, 2014, 11:54:42 am by Marco »
 

Offline made2hackTopic starter

  • Frequent Contributor
  • **
  • Posts: 424
  • Country: 00
    • Made2Hack
Re: 36V power supply - Design considerations
« Reply #3 on: April 05, 2014, 12:11:07 pm »
I would definitely consider a Meanwell since they are @ the moment the duck's guts when it comes to LED PSU's.

However, I also want to learn to build it myself since that is most fun, except the transformer part is what I don't understand at the moment.

I just realized, if my bridge rectifier drops 1.2V, that means that 36V output transformer is no good, correct? I need a minimum 36V + 1.2V transformer to get my desired 36VDC out?

So how do I achieve this? 38V transformer? Do they exist? 40V transformer? or 48V? Do I loose efficiency when dropping that many volts?

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18022
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: 36V power supply - Design considerations
« Reply #4 on: April 05, 2014, 12:13:32 pm »
if you want something efficient then you need a SMPS. Bear in mind that a 36V transformer will peak at 36*1.414 so you will end up with 36V anyway and this is one reason to oversize your transformer.
 

Offline made2hackTopic starter

  • Frequent Contributor
  • **
  • Posts: 424
  • Country: 00
    • Made2Hack
Re: 36V power supply - Design considerations
« Reply #5 on: April 05, 2014, 12:22:25 pm »
@Simon,

So a 36V transformer, or a 2x18V transformer should be ok? Because the peaks will always be higher? Irrespective if I'm dropping 1.2V over the bridge rect?

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18022
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: 36V power supply - Design considerations
« Reply #6 on: April 05, 2014, 12:42:44 pm »
you will still loose 1.2V off anything coming out of the transformer so peak outout of the 36V transformer is (36*1.414)-1.2 = 49.7V
 

Offline madires

  • Super Contributor
  • ***
  • Posts: 8134
  • Country: de
  • A qualified hobbyist ;)
Re: 36V power supply - Design considerations
« Reply #7 on: April 05, 2014, 12:49:55 pm »
2x15V should be fine also: 2 * 15V * sqrt(2) = 42.4V. And a 50VA transformer fits perfectly.
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18022
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: 36V power supply - Design considerations
« Reply #8 on: April 05, 2014, 12:52:37 pm »
Well this all really depends on the LED
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5135
  • Country: ro
  • .
Re: 36V power supply - Design considerations
« Reply #9 on: April 05, 2014, 03:13:44 pm »
Transformers are specified in  AC rms , which means that when you rectify the output voltage to DC the voltage increases and current decreases.  There's also some loss in the bridge rectifier diodes.

There's a few simple formulas you can use to approximate everything when doing full bridge rectification (using a bridge rectifier)

V dc peak =  Vac x 1.414  (edit: then substract drop on rectifier, see paragraph below)
I dc = Iac x 0.68

So if you get a 24v AC 50VA transformer  ( 50/24 = 2.08A)  the rectification will give you 34v at about 1.4A . Then, depending on the bridge rectifier you choose, you need to substract 2 times the voltage drop of an individual diode in the bridge rectifier because at any time, there are two diode rectifying the AC wave. So, with 0.8-1.1v drop per diode, you'll actually have a peak dc voltage  of about 32v.

This is peak DC voltage so you need a capacitor to store some energy to keep the voltage up when the voltage goes down a certain voltage.  There's another simple formula to approximate the voltage ...  Capacitance = Current / ( 2 x mains ac frequency x Vdrop)  where Vrop is how much you're willing to let the voltage go down from the peak voltage.

So, if you want for example the voltage to stay between 30v and 32v at 1.4A, then you should aim for C = 1.4/2x60x2 = 0.0058  or at least 5800uF.

So with 24v ac transformer you can see you're pretty much left with 30v dc... and then you have to keep in mind linear regulators will have a voltage drop of about 2-2.5v.It's obvious it wont work if you go with linear regulator route, but it could still work with a led driver chip that boosts the 30v to 36v .
You could also go with a larger transformer, but then if you plan to use linear regulators you have be careful of their maximum voltage - for example a 36v ac transformer may go over 50v dc which can be too much for linear regulators.
 
« Last Edit: April 05, 2014, 03:37:27 pm by mariush »
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18022
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: 36V power supply - Design considerations
« Reply #10 on: April 05, 2014, 03:29:27 pm »
Vdc peak = (Vout RMS*1.414) - Bridge drop
 

Offline made2hackTopic starter

  • Frequent Contributor
  • **
  • Posts: 424
  • Country: 00
    • Made2Hack
Re: 36V power supply - Design considerations
« Reply #11 on: April 06, 2014, 05:49:03 pm »
Thanks guys for the input. I will read up and come back with some ideas.

What if, for example I used the 24v transformer as mariush stated, giving me around 30-32, should I then "re-transform it" to 36vdc? is that possible? Or use some kind of boost ic with inductors and what not? Will they be large?

Also, any idea (I know it's vague) of how large the psu will be? I want it to kind of be "wall wart" size or at the very most maybe laptop psu size. It is for a lamp and some aesthetic consideration needs to go into it.

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18022
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: 36V power supply - Design considerations
« Reply #12 on: April 06, 2014, 05:55:49 pm »
maybe if you told us about the LED(s) you are driving we will have a better idea.

I'd alsomst be inclined to rectify skip the capacitor then have a switch mode converter with a nice big filter cap and use it in constant current mode by taking your feedback off a small value series resistor.

that's off the top of my head but i think it would be less painful on the transformer and more efficient.
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5135
  • Country: ro
  • .
Re: 36V power supply - Design considerations
« Reply #13 on: April 06, 2014, 06:06:13 pm »
I don't see you using enough LEDs in a lamp to need 36 v at 1A ... 36 watts of led lightning require some cooling and they'll make very bright light, both not really compatible with a lamp.  Well, I guess it depends what you mean by lamp, if it's nightstand lamp or some reflector type lamp.

I told you how you can do it yourself - if you want to, you can just get a universal 12-19v laptop adapter and then make a small led driver you can hide in the lamp behind the heatsink that's going to keep the leds cool.

Go to http://uk.farnell.com/led-drivers  or http://www.digikey.com/product-search/en/integrated-circuits-ics/pmic-led-drivers/2556628  and pick one or several ICs that can boost that 12-19v to the voltage and current needed by your leds. 
If each led needs less than the dc voltage you have using such an adapter, you can get a "buck" led driver, or you can put several in series and use a "boost" led driver.  Or you can use one driver for each led.

Just browse those and pick one that seems easy to use, most of those have example circuits in the datasheet.
 

Offline made2hackTopic starter

  • Frequent Contributor
  • **
  • Posts: 424
  • Country: 00
    • Made2Hack
Re: 36V power supply - Design considerations
« Reply #14 on: April 07, 2014, 11:05:33 am »
I wasn't really expecting the LED Driver IC's to be able to boost from 12 - 36V. Last time I asked in the forums about boosting from 12 - 40, the considerations were the size of the inductor would be too large, thus it would be easier to buck.

That solution proved ok for another LED setup and it works. However, that relied on a one hung low dc/dc boost driver from china.

This time, I want to go from AC rms directly to my LED. Thus the questions regarding building the PSU. But I don't mind boosting in smaller steps, say from 24v to 36v. I will look at the led drivers to see if they can do boost.

As for the LED, it will be a Nichia NSBWL121A COB module. It is a 40W power LED. It can be used @ 350ma / 31V for 10W or 1512lm, up to as high as 36V / 1100mA or 40W, for 4032lm. I will integrate a PWM switch to dim the unit. Datasheet if anyone is interested: Nichia NSBWL121A



Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf