The "proper" AT-to-split-supply mod!

[Pyr0Beast]

Re: The "proper" AT-to-split-supply mod!

Ordinary PSU would work just as well I guess. Ignore those 14V.

[kaboom]

Re: The "proper" AT-to-split-supply mod!

I wasn't sure about Infineon when they played 'me too' with MOSFETS. They should stick to RAM...

Gotta love it when things burn up.

It would really depend on the supplies that you have, how you'd modify them. The main issue is how to deal with regulation. You could keep the original 12V rectifiers and winding on the output inductor, their ampacity should be high enough. But you'll have to remove components/break traces to change from group regulation. Probably just a resistor/diode network to 'sum' all the outputs back to one error amp.

Don't even try it unless you have a variable voltage slightly higher than the 14V you want to regulate at. This can be a 15-20V laptop supply feeding a pot, to give you a full range of voltage, for example.

What I did with the Hipro AT that I made a monster out of was similar to this:
Find where 12V and 5V are summed and fed to the error amp.
Break that circuit/trace, but keep the 'lower' divider resistor.
Be careful here, the voltage set trimpot can be in parallel with the 'lower' divider resistor, and you can measure it wrong. Make sure it's in the center of its travel. With the value of the 'lower' resistor, we can now calculate the value of the upper resistor that will result in 2.5V at the junction of the two. This also goes to the error amp of the '494. For example:

Rl= 4.3K 2.5V/4300=581.4uA

If we want 14V out:

(14-2.5V=11.5V)/.0005814A=19.8K Rh=20K ohms
That's pretty much what I did. No compensation at this point.

While powering the control ckt from an external supply, feed the variable voltage to the top of the 'top' resistor.
Adjust the pot or variable supply up to the voltage you've designed for while watching the duty cycle at the driver transformer. It should fall to zero when, 14V in this case, is applied. It will pick up when you reduce the variable voltage.

I have no idea what's in the power supplies you've got, so the above may not even be applicable.

-Paul

[Scenic]

Re: The "proper" AT-to-split-supply mod!

the 14V 6.5A were from the original PSU's label.
the infamous Infineon 20N60C3 died. got hot enough to turn the pcb around it into coal (fell apart), slightly melted the PSU casing, left an outline in the aluminium heatsink and the tab of the TO220 package turned blue-ish.. ouch.

running it off the computer PSU is what i did for some time, but as it is (well.. was) my TV too, that wasn't too practical..
Plus the 430W Seasonic-Antec had to work quite hard, as the computer itself is loading it up quite a bit too (overclocked quadcore with a now-midrange graphics card (GTX260))

[kaboom]

Re: The "proper" AT-to-split-supply mod!

A-ha, little DC-DC converters instead of linear regulators to reduce the voltages. Everything now, as you well know, just uses the series-pass regs. The 5V output from the PS in a typical LCD feeds LDO regulators to make the 3.3V and 2.5/1.8V supplies.

14*6.5=91 watts for your LCD. At 13.8V, you need 6.6A. 12V in requires 7.6A. Currents at reduced voltage assume the supplies in the monitor stay just as efficient at lower voltage as compared to full voltage. I'd allow for 7.75A at 12V in.

For clarity, is that rating the listed output capacity of the power adapter or is it the from the nameplate on the monitor?

If you're confident of the supply in your computer, run the LCD off that while checking the voltage and current drawn. Calculate the wattage drawn. If it seems to be drawing too much power, have a look inside the LCD. Maybe that's what killed its power brick.

Just a thought...
-Paul

[Scenic]

Re: The "proper" AT-to-split-supply mod!

well.. this LCD is actually old enough to have seperate DC-DC converters in it.
one of them is for 5V. dunno how many amps (and it doesn't say on the label IIRC)

i've got a 13.8V SMPS that works (adjusted to 14.1V with the internal trimpot), but it's rated for 5A continuous, 6A peak. Voltage sags quite a bit with the monitor connected to it. I've piggybacked a temp-controlled fan on top to keep it from toasting itself, but that's not really optimal in any way..
i actually thought about beefing that PSU up and putting it into an old ATX PSU casing, but the layout and size prevent me from putting it in there.

totally cramped layout with traces everywhere, so i cant put extra mounting holes, and i cant mount it as it is cause nothing lines up.. (PCB is considerably smaller than that of an ATX PSU)

[kaboom]

Re: The "proper" AT-to-split-supply mod!

I know it's less than optimal, but you could try a sealed beam headlight as a load on the 5V output and put ~12V into that LCD. The logic and signal stuff will be powered from regulators that reduce the nominal 14V input anyway. Just the backlight inverters and audio amp, if any, use the DC input voltage directly.

Put an ammeter in series in case you're worried about things getting 'interesting.'

-Paul

[Scenic]

Re: The "proper" AT-to-split-supply mod!

if i had the knowledge, i could convert one of those tons of low wattage (200-250W) ATX PSUs i've got for my 21" LCD of which the power brick died.. :/

needs 14V 6.5A.. oddball value.. so no luck on ebay.. except cheapo replacement PSUs from hongkong.. no thanks..

[kaboom]

Re: The "proper" AT-to-split-supply mod!

Go get 'em!

I almost built the big supply from scratch, but realized ~80% of the old Hipro could be reused. Driver TX, PCB, quad opamp, '494 and enclosure. Rather than redundantly duplicate what I already had, the thing got 'recycled.' Saves time when you don't have to do the PCB layout, etch, drill, etc.

Eww... one of those cheap AT/ATX crossover bastards. I've seen what was essentially an AT PS with a small PCB for the standby/aux mounted across the two heatsinks! Lousy design too, the primary snubber resistor for the aux ran real HOT. Probably dissipated as much heat as all the standby ckts on the old mobos put together. Did the 'JNC' label have red stylized letters on a white background?

The fact that you got such high voltages out of it by manually poking the feedback is interesting. This would indicate that unusually low duty cycles were used to 'regulate it down' to the 'puter voltages. I'll guarantee that thing was discontinuous, even till a few tens of watts out! Makes for nice high peak ripple current through the coupled inductor and caps. High output ripple voltage, too- what with no PI filters.

They left the cap over the RC combo out 'cause they didn't want to AC comp it. Perhaps because of the stupidly low duty cycle?? The output inductor and capacitance of the rectifiers ring every time the 'spikes' from discontinuous operation excite it. The AC comp cap would feed these parasites into the error amp and upset it, so they left it out. They're counting on the extra phase margin...

The Deers/Allieds/Apevias are pretty bad, but those crossover things were just junk. 'Crown' comes to mind immediately when I think of the AT/ATX trash. Just bastardize an existing design instead of starting over properly.

I wouldn't have thought one of those things would be around so long after ATX became standard.

The later it gets, the more transistors like to explode, so don't go too far in each day. Just sayin'...

-Paul

[Th3_uN1Qu3]

I have said myself that building a PSU from scratch is better than modding an AT/ATX supply. Well, inspired by kaboom's work, i'm ready to prove myself wrong.

I have a JNC "200W" AT supply. If you think you've seen gutless, you ain't seen this. There is zero input filtering - not even the room for it. The primary diodes are 2A. The heatsinks are ridiculously tiny, they don't even have fins. The caps are Capxon, SD and E.V.A.Top. No PI filters (but at least the board is silkscreened for them). It does have TO-220 output rectifiers tho...

Funny thing, the board is printed for both AT AND ATX versions! Talk about low budget. I've been using it as an adjustable power supply for powering the control circuit of the power supply i am building from scratch. I wired a pot from ground to + input of error amp and that was it. Btw i managed to blow all Capxons on the output when i turned the PSU up to 28v - if there is an overvoltage protection circuit, i don't see it working. The SD cap on -12v escaped because it's rated for 50v.

Well, i've been getting accustomed to compensating the TL494 (which btw is a lot easier than the textbook examples, because it uses a non-inverting error amp which gives you an extra 180 deg of phase headroom, and it seems not being able to have negative gain doesn't give trouble at all). The typical compensation for the TL494 is type 2, but the 2nd capacitor, that is to be put over the R-C network is usually missing. Either way, the 494 is easy to work with.

The current state of mods: I have removed the 5v rail from the feedback loop. Next up are the -5 and -12v rails. I have temporarily installed 470uF 25v Rubycon caps for both 5v and 12v. The fan is now running from what used to be the 5v rail, this will be changed. It will be running from a temperature controlled circuit.

Goal: Transformation to +/-45v with both rails regulated. Step by step, and with lots of pics.