# additional 24v circuit to a computer smps

Discussion in 'General Electronics Chat' started by sunshine, Jun 18, 2008.

1. ### sunshine Thread Starter Member

Jun 18, 2008
13
0
hi

i have a spare computer smps and i wanted to add an additional circuit to provide 24v at a high current, and the perfect candidate was this http://www.reuk.co.uk/OtherImages/power-transistor-voltage-regulator.jpg.

i want to take it's input from the main transformer output, the one that is connected directly to the schottky diode. i wonder if that's gonna do the job or is it going to be a mess. also i wanted to know whether i have to add a capacitor at the input of the new circuit to smooth things up or that will affect the schottky diode operation.

i don't want my adjustment to be destructive, as i will need the default voltages that the unit provides.

thanx

2. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
Well, the SMPS is reasonably efficient - but to add a linear regulator to the mix would cause quite a bit of wasted power.

What are the output voltages vs amps the supply currently has, and what are your power requirements for each voltage?

How much current do you expect to get out of the 24v supply?

Your best bet is likely a DC-DC boost converter running from the 12v supply. (corrected from 24v supply)
Whatever current you get from the 24v output will reduce the other current outputs porportionately, less the loss in the boost converter. For example, if you were to use a DC-DC boost converter that was 92% efficient, you decided to power it using the +12v supply that originally had an 8A output independent of the other voltages, of which you needed to retain 2 amps out, you would then have:
(8A-2A)x .92 x 12/24 = 2.76A output @ 24v.

Last edited: Jun 19, 2008
3. ### sunshine Thread Starter Member

Jun 18, 2008
13
0

actually you have a good point, DC-DC boost converter is a nice idea. i chose a linear converter in the first place because it was my quick and dirty solution, but the datasheet never mentioned the power consumption for the module and i guess that the power transistor doesn't eat much too !? or am i wrong?

i don't have exact figures about current/voltage outputs of the supply, let's just assume that it gives 10- A @12v, and 20- A @5v and any lower volts are derived from the 5 branch.

for my 24v branch any thing higher than 2A would be great, but for DC-DC converter i guess that i won't be able to get toroids or beads to create an inductive element and i have to adapt a solution that uses capacitive elements almost solely. i think that might work.

4. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
http://www.maxim-ic.com/appnotes.cfm/an_pk/910
12V to 24v DC-DC boost supply by Maxim-Dallas.
3A output, 93% efficiency.
In order to get the same level of efficiency, you'll need to use the exact components (or equivalent replacements) they specified for L1, D1 and N1.

5. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
What you have shown is not a linear converter, it is a linear regulatorwith a pass transistor.

The pass transistor is operated in the linear region; ie: a variable resistor. It will get quite toasty warm. The heat is wasted power.

There should be a label on the side of the supply with at LEAST a manufacturer's name and part number. From there you can search for a datasheet, or the manufacturer's specifications.

DC-DC converters using capacitive switching are usually for low power applications, perhaps up to a couple hundred mA.

The exact inductor called out on the schematic I linked to in the last post is carried by Digi-Key.
http://parts.digikey.com/1/parts/667689-inductor-xfrmr-9-9uh-1-60a-smd-vp5-0155-r.html

Note that you will need to enclose your boost supply in a grounded metal case to keep RFI emissions under control. Otherwise, it may cause considerable interference with AM radio reception.

6. ### sunshine Thread Starter Member

Jun 18, 2008
13
0
this is a very neat design, i really like it. but considering the complexity to function ratio (if we can calculate that), then the only result is that this circuit is almost a complete unit by itself. i mean it only needs a step down transformer (which is much affordable to anybody) and a couple of capacitors and a bridge and it's totally functioning on it's own.

what i need is a less than 8 component circuit, that i can put on a max 2.5x2.5 cm card that can add an additional terminal to supply. i don't know what size would be a card carrying all these components but i guess it'll be bigger. and also the components are nice, but i cannot get such components let's say for now and some time to come.

well thank you anyway you posted a good idea.