Please take a look at the attached diagram. What I want to do is take a small transformer and get around +24, +12, -12 and -24 volts DC to run into four 3 terminal regulators (two 7800 series and two 7900 series, of course).
SOMETHING on the edge of my mind is telling me that there's a hidden "gotcha" in this design. I'm not seeing it, but it's gnawing at me, so before I buy parts just to let the smoke out of them, please see if there's a problem here.
(GEEZ I feel so stupid.... an EE for decades and a simple power supply is bothering me).
Polarity of capacitors for the negative supplies is reversed..
Labels for -12 and -24 seem swapped..
And: You are using 1/2 wave rectification. You MIGHT be able to get away with this IF the capacitors are large enough AND the current needed is not too high. It would be better if all the capacitors were connected to ground on one side..
What are the currents you need?
What is the transformer rated at for current?
Can you get TWO of those transformers? If so you can use a fullwafe center-tapped configuration and get + 12.7 - 12.7 and +25.4 and - 25.4 (1.414 [AKA Square root of 2] times the RMS AC). Reality check: You will have about .7V drop in silicon restifiers.
Polarity of capacitors for the negative supplies is reversed..
Labels for -12 and -24 seem swapped..
And: You are using 1/2 wave rectification. You MIGHT be able to get away with this IF the capacitors are large enough AND the current needed is not too high. It would be better if all the capacitors were connected to ground on one side..
What are the currents you need?
What is the transformer rated at for current?
Can you get TWO of those transformers? If so you can use a fullwafe center-tapped configuration and get + 12.7 - 12.7 and +25.4 and - 25.4 (1.414 [AKA Square root of 2] times the RMS AC). Reality check: You will have about .7V drop in silicon restifiers.
You are correct. I drew the negative rail caps backwards AND the negative side voltage labels are backwards too.. My bad. (told ya I was feeling stupid!)
As far as capacitor voltage or polarity or diode current or the "square root of two" thing, that's not relevant to my question.
What I want to know is: Is there a hidden "criss-cross" diode short across the transformer that I'm not seeing?
terryking228:
Can you get TWO of those transformers?
Sure I can. I could also get a single transformer with +9 and +18 taps around a center-tap.
The reason I was asking this at all was because for a dual-tap transformer I would have to order one from Digi-Key whereas a single center-tap transformer I can get at Radio Shack.
What I am building is a low power (maybe 1 or two amp) bi-polar power supply to test op-amps and such.
I've got no bipolar power supplies here! To test something, I need to tie two wall warts together. That's ridiculous!
That's why I want to build that supply (+/- 12 and +/- 5).
Hi, Diodes connect only to capacitors. When the waveform forward-biases a diode it will conduct and charge a capacitor. It's unusual that the "24V" capacitors are riding on the "12V" capacitors, but they will crage that way. The issue is what DIScharge you will have..
If you have two of those transformers and connect the (correct--phasing) ends together, you have in effect a center-tapped transformer with 9-0-9 Vac and 18-0-18VAC. If you use four sets of 2 diodes to make 4 outputs you will have about + and -12.7 and + and - 25 VDC. And there will be full-wave rectification. So you have high enough voltages to regulate down to + and -5 and + and -12 At 2 amps you will need regulators in TO-3 or TO-220 case and decent heat sinks.
The attached photo is of the sketches I made to see if the idea would work. I drew the first one on the left with 4 individual diodes, then thought "all I did was make a bridge rectifier" and re-drew it as a bridge - then saw THAT wouldn't work, so I drew more and realized it's really not a bridge and when I was done I confused myself so much I locked up and couldn't think.
But still I feel there's SOMETHING wrong with the idea........
terryking228:
Hi, Diodes connect only to capacitors. When the waveform forward-biases a diode it will conduct and charge a capacitor. It's unusual that the "24V" capacitors are riding on the "12V" capacitors, but they will crage that way. The issue is what DIScharge you will have..
If you have two of those transformers and connect the (correct--phasing) ends together, you have in effect a center-tapped transformer with 9-0-9 Vac and 18-0-18VAC. If you use four sets of 2 diodes to make 4 outputs you will have about + and -12.7 and + and - 25 VDC. And there will be full-wave rectification. So you have high enough voltages to regulate down to + and -5 and + and -12 At 2 amps you will need regulators in TO-3 or TO-220 case and decent heat sinks.
The more I think about it, the more I think I should just get a nice toroidal transformer from Digi-Key (a center tapped transformer with 9 volt and 18 volt windings) like this:
arduinoaleman:
You will never get what you want from your transformer whithout using electronic voltage transformers.
Turning 18 Volts into 24 Volts is like a miracle. I know, there are circuits which can do this.
However, you will need electronic switching circuits and the output in mA ore Amperes might be limited.
You don't understand ac circuits. 18V transformer secondary is 18V rms, ie 25.5V peak. Until you
learn about such things please refrain from confusing the answers here.
The +/-24V capacitors should connect to ground, not boosted off the +/-12V lines. If your current consumption on the 24V lines exceeds the 12V current consumption then when it switches off and draws the capacitors down, it will reverse polarize the "24V" caps. They will explode and spew corrosive electrolyte across the inside of your box.
For light loads (opamps) then it's pretty easy to make the caps big enough to get the ripple down to an acceptable level. If you're running motors off this then you'll notice they will perform poorly and make a buzzing noise, unless the caps are gigantic.
Krupski:
Sure it can. And it is. Drawn as CIRCUIT COMMON (a triangle) which is not the same thing as GROUND.
The frame of the transformer will be GROUNDED. Anything at 0 volts DC will be connected to CIRCUIT COMMON.
And the two are completely separate.
Unfortunately, circuit common is also frequently called ground, even in circuits that are not actually grounded. It is an absurdly stupid convention that unfortunately persists in EE through force of habit.
In order to be disambiguated from a circuit "ground", actual ground is often called "earth".
After lunch I can whip this up in a simulator and let you know what it does.
I think the "gotcha" is that you would need to use the center tap directly as a common (GND). In that case, I think a circuit like this (image from the web, but use 12 0 12 transformer) would work OK. Would only need to add the 7805 and 7905 and required caps to get quad output.
OK I lied, it wasn't after lunch, it was much later than that.
I whipped it up in LTSpice and didn't see any issues. The biggest issue is what someone already pointed out, each output if only half-wave rectified so you'll need larger caps on the outputs to prevent the voltage from sagging too much.
dlloyd:
I think the "gotcha" is that you would need to use the center tap directly as a common (GND). In that case, I think a circuit like this (image from the web, but use 12 0 12 transformer) would work OK. Would only need to add the 7805 and 7905 and required caps to get quad output.
That is only necessary for full-wave rectification. With larger caps, half-wave is perfectly acceptable, given the OP's requirement for multiple rails at slightly odd voltages.
@Jiggy: does your simulation actually have dummy loads on the power supply? Can you simulate transients like switching on and off?
And it is. Drawn as CIRCUIT COMMON (a triangle) which is not the same thing as GROUND.