how can you tell if the switching psu transformer is iso-transformer?
i read little about isolation tranformer that imply somekind of shielding and grounding manner. but, i don't know how to determine if a transformer is an isolation one.

Posted on 2006-07-26 01:35:12 by dion
dion, like I said before, I don't remember the theory behind all this and am too lazy to think too hard about it but, unless you are into very creative design techniques, I wouldn't do this at all.  What's to say the lower voltage output PSU could handle the voltage and current of the higher voltage PSU?  The output circuitry of both would be different and the potential difference between the two would cause feedback from one into the other and I can just see some minor explosion in the works here. 
Posted on 2006-07-26 06:51:44 by drhowarddrfine
dion,


how can you tell if the switching psu transformer is iso-transformer?

好bsp; 好bsp; 好bsp;If you got your PS from a computer, you can assume with a high degree of certainly that it is not isolated.好bsp; The computer manufactures don't spend any more money than necessary, and power transformers are not cheap.好bsp; Have you researched this subject like this link?好bsp; http://computer.howstuffworks.com/power-supply2.htm好bsp; Unless you know what you are doing and have the circuit schematic, it is best you don't follow through on your idea.好bsp; Ratch
Posted on 2006-07-26 08:05:38 by Ratch
imho, connecting two PSU in parallel or serial is absolutely possible, and can be done with no additional parts. You only have to look-out for grounding in the serial connection - as pdixon mentioned. In the serial connection, you just have to cut the 'ground' wire of the PSU that supplies the positive charge of the resulting PS. Btw, removing the "ground" from high-voltage devices isn't as serious as people in the US (who are used to 3-wire power) would think. In my country, the standard is 2 wires - hot/neutral. We only do grounding for boilers and large heaters. (my PCs, TVs, AC, ovens, etc. are not grounded).

I see all these well-backedup posts here, but maybe you forgot the final part of the PSU schematic: large capacitors, diodes and regulators. The diodes necessary for parallel connection of 2 PSU are already present in the PSU. The capacitors, filtering-out spikes, are already there. Precise voltage-control is already there. High-current detection + power-down are integrated (I've personally short-circuited several different PSU lots of times, those same PSU never failed in the next several years).
Btw, knowing exactly your PSU's schematic is rarely possible - for the simple fact that usually the manufacturer... erases the driver-IC's label on purpose.

dion: welcome back :) . In your schematic that uses 2 PSU, the only thing you'll have to watch out for is the interface between ICs that use the 3.3V supply, and those ICs that use 5V supply. There's a CMOS/HCMOS IC that serves exactly this purpose (4050, HC4050). You could skip using this IC in places, but you'll have to refer to the output and input bipolar/mos transistors and take note they'll just share ground. Units (parts of the schematic) that are in schematic-type "common collector/emiter/source/drain" are immune to trouble with different supplies, iirc from experience.
Posted on 2006-07-26 12:58:48 by Ultrano
i guess i had it enough right now. thanks you guys...
Posted on 2006-07-26 23:27:16 by dion
  What's the matter dion? Having a mental overload protection shut down?  :lol: Just a quick statement on isolation Xformer. What you do is use an ohm/continuity meter and make sure there is no connection from any primary wires to any secondary wires. In other words an isolation Xformer lets say with no taps so that your power in is AC house current and some specific voltage out like 12.6V AC then what you have internally is a large number of turns of wire on the primary with the two leads sticking out. These two leads have continuity because they are obviosly one single wire. In an isolation Xformer the secondary is simply another coil of wire wrapped over top of the primary ciol winding. Now these wires are laminated which is similar to a plastic coating so there is no physical contact between the two coils or the primary & secondary windings. They are inductively coupled. The electro-magnetic field from the primary winding creates an electro-magnetic field in the secondary. In a conventional Xformer they use lamenated plates for the core and in the case of switch-mode being it runs at a higher frequency than 50 0r 60 hertz, smaller Xformers can be made with the core material usually being ferrite or some high permeability powdered iron core.
Posted on 2006-07-28 04:38:53 by mrgone
Power supplies come in many many shapes and sizes and functions. When joining several together, you're mostly concerned if they have input to output isolation. Most linear supplies get this for free with the step-down transformer, a switch mode may or may not have such isolation. You can check this with an ohmmeter, unplug it from the wall and see if you get continuity from any AC power pin to any output pin. Personally, I would not buy a supply that did not have input to output isolation for safety reasons.

If the output are well and truly isolated, then it is safe to stack multiple power supplies in series for a higher voltage, combine them at ground for several different voltages, and such.

If not isolated, you can still connect as long as the connection back to the input lines is connected together on the outputs.

I've occasionally run two power supplies in parallel. They were linear, isolated units with the same output voltage. What happens is one unit drives the output (the one with the slightly higher output setting) while the other coasts in shut down (as it "sees' the higher voltage, and it throttles back trying to reduce the output voltage). With a linear current limit (ie, no fold-back), once the max current output of the higher supply is reached, it reduces its output voltage to the point where the other suply snaps into regulation, and now that supply feeds the increacing current. So one runs at max, the other runs at the difference.

This works best with a fixed load, if the load changes a lot the output can show a lot of ripple as one supply cuts in and out.

(Just passin thru, tossin in my 2 cents)
Posted on 2006-08-07 11:38:59 by Ernie
Sorry, I don't want to hijack the thread or bring it offtopic - but it's nice seeing that you're still around Ernie.
Posted on 2006-08-07 12:47:21 by JimmyClif