No load voltage at the filter output is 43.19VDC
Loaded to 3.65 Amps voltage drops to 39.71VDC.
I presume 14.2 Amps load will drop it to 35VDC approximately.

Don't presume, do an actual test.

With the size of your transformer, I think you should do an actual AC or DC loading test until the transformer secondary AC terminal voltage drops about 5%. That current would then likely to be the normal current rating from your transformer. Higher loading your transformer will heat up.

Record the DC loading and voltage at this load. Your power supply maximum voltage output will then be ~5V or more lower than this to give headroom for the regulator transistors to work properly. Of course someone else would go as low as 2V headroom but the output voltage regulation will suffer.

 

Yes....
I made a test on my assumption and it did drop to 35VDC at 14 Amps.
As I have seen by practice just now minimizing the differential voltage tends to worsen load regulation.

So you want me to do an AC load test till it drops to 5%.
OK. getting on it now. be right back.
I'll post the measurements this time

 

I am getting a 5% difference at the former terminal with 12.2Amps
From 31.5VAC to 29.9VAC at 12.2 Amps.

 

I am getting a 5% difference at the former terminal with 12.2Amps
From 31.5VAC to 29.9VAC at 12.2 Amps.

OK. The 5% voltage drops occurs at 12.2A AC loading.

Now depending on who you believe and which formula you use, the current max. for you power supply output can range anywhere from 7.6A to 11A. You have also done a test to confirm the smoothing capacitor voltage maintains at 35V DC at 14A DC loading.

So I would say your DC power supply can output a maximum of 30V and somewhere between 7.6A to 11A.

You can probably get away with 11A output if the duty of the power supply is not 100% full load for more than a few hours. For a few minutes, it can even go higher than 11A.

 

Tough luck, all the work for nothing.
Oh well. I think I will maintain at 10 Amps.
I will increase the filter to see the voltage sag at 10 amps.
From there I will go on.

By the way, why the so much drop after the bridge, do you think I need a low Vf bridge or diodes

PS. I believe what I see and feel and advice from experienced ones ( but I will experiment on the advice )

 

Hi,

Without having seen your schematic, and assuming that you're using a capacitive input filter (and no filter inductor) for smoothing, the DC current value that charges the filter cap, is not a smooth, flat DC current that follows the sinusoidal RMS output of the transformer when under load. Due to the filter capacitor, the phase angle for charging the cap is normally quite low, causing the actual bridge current in real time (not the rms value) to be very much larger than the average output current of the power supply. And the higher the cap value, the worse this phenomenon gets. Now, since the actual current out of the transformer secondary and thru the bridge is high compared to the average DC output current, you must look at the bridge's datasheet and you'll see that the Vf of the bridge (forward voltage drop) is much higher at higher currents and typically the current will be 5 to 10 times higher than the average output current, albeit for only a short time during each half cycle, but you get a much higher voltage drop from the bridge no matter what in this case and the only thing that might help is to lower the filter cap value. There are of course, equations for computing the best cap value for your specifications. If you have a scope, I encourage you to look at the output bridge current or the cap charging current coming from the bridge. You can do this with a small series resistor of perhaps 100 milliohms at 2 watts or 1 ohm at 10-20 watts. If you're going to do this, put the series resistor in the ground side between the bridge minus and the cap minus and put the scope ground lead on the bridge minus side and the scope center lead on the cap minus side and you'll see what I'm talking about. You don't need a fast scope to see this. The scaling of course will be 1amp/volt if you use 1 ohm, and 10amps/volt if you use a 100 milliohm resistor. If you hookup your scope this way, you can also see the cap voltage at the same time using the other channel (assuming a dual-trace scope). The ground strap for the other channel stays at the bridge minus and the center lead would go to the cap positive. If you connect it any other way, you're likely to get a fire or an explosion, so be careful.

Regards,
Kamran Kazem

 

Circuit corrected. Post Removed

 

This came to me via PM:

Hey
I donno why
But my circuit is producing a low freq oscillation when I load it to 8 amps or more. And the I have 500mV difference at load regulation.
That is too much.
Can I see ur LM723 just the regulator part.
Thanks

I already told you there is nothing special about the circuit.

See below:

 

Thanks mate.
I just wanted to know the bypass caps and how to rearrange them to eliminate the audible oscillation I am getting. The oscillation is the reason for the terrible load regulation. I can measure AC voltage at the base of the drivers .
I think my layout and poor choice of caps is giving me all this trouble. I am on it now. The LCD is on my desk top as a second monitor for now. Be right back

 

All right gentleman.
I have made changes to my own schematic to improve the load regulation.
at a loading of 10 amps at any given voltage up to 30VDC ( output is limited to at 30V due to the huge drop at the caps from 40V to 35V at max loading, so to increase the load regulation the differential voltage is kept at 5V minimum)
Previously my circuit produced a low frequency oscillation when loaded and load regulation was at 2.2% . and now it's down to 0.16% .

The changes were not made in accordance with the LChung's diagram but I will be comparing it with his to further improve the regulation after this post.

Diagram removed for improvement

 

The standard "rule of thumb" for capacitor voltage rating is 2x the peak voltage they will have across their terminals.

Therefore, C4, C5, C8, and C9 must be rated for 84v or more.

Have you calculated ripple voltage on your two 15,000uF caps with a 10A load and a 50Hz input?

It would be a good idea to provide some start-up current limiting between your transformer and the caps, or you will likely fry your rectifier bridge. Ask me how I know this.

 

OK. I'm asking now, Do tell.
Do you think a 35A bridge will get fried. I mean I do discharge them every time now that I once got big bang out of it when I was fiddling with the caps without discharging 'em .. almost threw m out of the chair.
The bridge is still doing fine yet and I have been powering it up a couple of times when experimenting.
I did not measure the ripple but wouldn't the caps heat up if there is too much ripple.
I can measure the ripple at max loading just to be sure.
As for going higher voltage for the caps what would be the advantages in the case of a linear supply.

 

OK. I'm asking now, Do tell.
Do you think a 35A bridge will get fried.

I have a 16v 10A unregulated supply with a 35A 400V bridge sitting next to me.
One of the rectifiers in the bridge is burned out due to excessive current during power-up.

I mean I do discharge them every time now that I once got big bang out of it when I was fiddling with the caps without discharging 'em .. almost threw m out of the chair.
The bridge is still doing fine yet and I have been powering it up a couple of times when experimenting.

It's up to you. If you don't provide any current limiting during start-up, be prepared to replace the bridge occasionally.

I did not measure the ripple but wouldn't the caps heat up if there is too much ripple.

The caps will heat up if there is too much ripple, and/or if the caps do not have a high enough voltage rating. 8v is not enough of a margin for voltage rating. You should have at least double the supply voltage.

I can measure the ripple at max loading just to be sure.
As for going higher voltage for the caps what would be the advantages in the case of a linear supply.

The closer you get to the voltage rating of the cap, the higher the leakage current through the cap. This causes power dissipation in the cap. If the power dissipation in the cap is high enough, the electrolyte will boil, and the cap will rupture, getting electrolyte and aluminum confetti all over everything.

You will have a mess.

 

Oki Doki..sarge..I salute u once again.
Taking you word for it.

but.....
Surge current limiting, I will need to switch out the current limiter after some time...no!!!
Like use a relay to short the resistor after the DC bus is stable.

 

Yes, you'd need to eliminate the current limiting once the caps were pretty well charged up.

You could use a relay powered by the filter caps. A high-current automotive-type relay would be cheap and readily available. Use a resistor in series with the relay coil to fine-tune the relay coil's trip level. After the filter is up to perhaps 2/3 to 3/4 normal voltage, you won't need the inrush current limiter.

 

Diagram removed for improvements

 

Rifaa,
If you use a DPDT switch for power on, you can use one side to apply power, and the other side connected to some power resistors to drain the filter caps rapidly.

If you want to protect the supply (and yourself) in case of power outages, you might use push buttons to latch/unlatch a line powered relay. If the power fails or drops momentarily, the relay drops out, and you have to push the ON button to start it again.

 

I did thought of that.
But the thing is I am planing to use a pic to control the current and short circuit and temperature too. So I can use soft touch for power.
This is going to be a great supply. Since I am taking great care in the first stages.
And also for the great help from u folks I am getting to make it too.
The first stage is rock solid regulation. I have am trying to beat L Chung you know . and his supply is like a rock at 15 amps . I have never seen that.

He is 1388. I will be ... ( i still figuring if you should know )

Thanks though.

 

The standard "rule of thumb" for capacitor voltage rating is 2x the peak voltage they will have across their terminals.

Therefore, C4, C5, C8, and C9 must be rated for 84v or more.

Have you calculated ripple voltage on your two 15,000uF caps with a 10A load and a 50Hz input?

This is not what the capacitor manufacturers say; the one I talked to says they have no such rule of thumb. The Cornell-Dubilier factory engineer says that you're ok operating at the rated voltage; the AVX factory engineer says the same thing. That means don't go over the rated voltage on ripple peaks, but there's no need to derate 50%. The capacitor will last at least for the rated life with a voltage equal to the rated voltage applied for the rated lifetime.

There's sometimes a surge voltage rating slightly higher than continuous rated voltage. You can go up to that surge rating for a short time.

If you derate 50%, you will get a somewhat longer life (see the curves in the data sheet), but if you don't derate, you will get the rated life, which will be thousands of hours with rated voltage continuously applied.

And, that's at rated temperature. If the cap temperature is less than 85° C, which is typical for ordinary electrolytics, the life will be much longer. Keeping the cap well below rated maximum temperature will have a much greater effect on life than will 50% derating.

It would be a good idea to provide some start-up current limiting between your transformer and the caps, or you will likely fry your rectifier bridge. Ask me how I know this.

In post #63, R!f@@ tells us that "I am getting a 5% difference at the former terminal with 12.2Amps. From 31.5VAC to 29.9VAC at 12.2 Amps."

From this we can calculate that the series resistance of his transformer is (31.5-29.9)/12.2 = .13115Ω This means that the peak short circuit current (and maximum possible surge to the bridge) would be 1.414*31.5/.13115 = 340 amps.

The typical surge rating of a 35 amp bridge (http://www.vishay.com/docs/88612/gbpc12.pdf; R!f@@ should look up the specs for his particular unit) is 400 amps. And, that's if the surge occurs when the diodes are already at their maximum allowed temperature. If the diodes are cold when the surge occurs, the allowable current is much more.

The resistance of his transformer alone (to say nothing of the ESR of the caps, the resistance of the diodes, the resistance of the house wiring to the wall socket) is sufficient to limit the surge to less than a bridge rating of 400 amps (at maximum temperature).

He could measure the surge with a scope to allay any possible fears that he might be exceeding the alllowable surge current.

I have a 16v 10A unregulated supply with a 35A 400V bridge sitting next to me.
One of the rectifiers in the bridge is burned out due to excessive current during power-up.

Must have been a situation where the series resistance of your transformer was low enough to allow a surge that exceeded the rating of your bridge.

That doesn't seem to be that case for R!f@@'s transformer, if he uses a good quality 35 amp bridge with 400 amp surge rating.

 

From this we can calculate that the series resistance of his transformer is (31.5-29.9)/12.2 = .13115Ω This means that the peak short circuit current (and maximum possible surge to the bridge) would be 1.414*31.5/.13115 = 340 amps.

This a new trick for me.

All ur knowledge is flowing via my vains .
Beware people, when I rise I will conquer you all, and you all bow down to me. Ha ha ha ha ha!!!
" heard that in a cartoon"

Thanks, Elector.. any comment is always welcome.
Better to be safe than sorry.
Besides Sgt's idea is already made, so it's cool.