Home-brew 13.8v, 30A PSU regulation prob

ScottWang

Joined Aug 23, 2012
7,416
Including Scott: The first feedback resistor can be bypassed with a capacitor to minimize power supply ripple at the output. You can add a resistor in the range of 3.3M to 15M from the base of the output transistor back to the positive input of the difference amplifier. The increased Vbe under load will compensate for load droop. This method requires that you divide the reference voltage with resistors so there is an impedance to work with instead of trying to skew the reference voltage. Results of .01%, zero to full load can sometimes be achieved.

Been there, done that, almost broke my arm patting myself on the back.
#12, Thanks.
Which 723 circuit are you talking to?
 

bountyhunter

Joined Sep 7, 2009
2,512
I don't think loop gain is the problem because the OP design requires 3 ma drive current from the 723. (It is allegedly capable of 150 ma.) The usual method is to use a driver for a bunch of parallel output transistors, but this design uses Darlingtons which should accommodate the load nicely.

Still, we'll see when we get some measurements.
I don't think it's loop gain either, that's why I put that check last. I believe it's voltage drop issues somewhere along a ground line getting "gained up" by the control loop.
 

bountyhunter

Joined Sep 7, 2009
2,512
Bounty: The 723 lacks a sharp cutoff for current limiting, but it can be added with a 358, just as you did in your discrete design. The foldback feature of the 723 is very convenient in my opinion. In addition, the reference voltage stability can be enhanced by placing the chip in a zener regulated supply of about 12V to 15V, with or without a separate secondary winding. The op-amp inputs can be impedance matched to minimize thermal drift.
No question you can slap patches all over it.

My point was that back when the 723 was a viable product, there weren't cheap accurate discrete references available like there are now. For a 30A design, I would rather roll my own. For a 3A design, a 723 and a transistor will save a few nickels over a 3A IC regulator so that is a realistic choice. It can probably be made to work here, I just don't see it as the best choice if performance is the target.
 

#12

Joined Nov 30, 2010
18,224
#12, Thanks.
Which 723 circuit are you talking to?
I am not referring to any circuit on the datasheet. The whole point of, "tricks that aren't on the datasheet" is that they are not on the datasheet. You have to understand the LM723 in order to apply the modifications I described.
 

ScottWang

Joined Aug 23, 2012
7,416
I am not referring to any circuit on the datasheet. The whole point of, "tricks that aren't on the datasheet" is that they are not on the datasheet. You have to understand the LM723 in order to apply the modifications I described.
Ok, Thanks.
 

tvtech

Joined Mar 18, 2012
6
Sorry to digress...but I am learning here at last again and it feels wonderful. I too had 723 issues in the past which I did not know how to resolve.

Thanks Guys :)

tvtech
 
Last edited:

takao21203

Joined Apr 28, 2012
3,702
Normally these days one would design a linear regulator with discrete OpAmps.

You can for instance use a discrete 78l05 as voltage reference (with divider).

If you draw some Amps, then there will be some voltage dropped by the cables.

The resistance of the MOSFETs does not matter, as the feedback is taken after their outputs.

I have mounted TO220 MOSFETs at one time on VGA coolers for a linear supply. With 2 LM358 OpAmps, I archieved excellent stability (but I did not care for temp. compensation at all).

A 100mV deviation would be visible on the LED display but just on the margin.

In the end I built a TL494 PWM regulator, the cooling was too much effort. Using it under load for some months, everything required cooling- the toroid, the rectifier, and even the cables + capacitor turned hot.

I used a string of cables (different voltages), some metres long, and the voltage drop was considerable.
 

Thread Starter

Dave_UYZ

Joined Jan 16, 2014
29
Sincere thanks for the efforts so far.
I've ripped the wiring apart and will advise results a little later.

Notes:
All wiring is in thick conductors.
I would not say that the heat-sinks are of "bowling ball" size, but they are not small (say, about 5" cube?), and they are fan cooled. It was originally one long heat-sink with 8 transistors on it. I had to slice it into two.
You might ask "why this circuit"?
'Because I had the bits to do it that way' is the answer.
 

Attachments

ScottWang

Joined Aug 23, 2012
7,416
How is the isolation films for the power transistor? (mica films?)

If you using the traditional mica films, the mica films only has a good isolation function, but the function of the heat conduction is pretty bad, if you using the new material Artificial Graphite Sheet or something similar, then the effect of heat elimination that it will be better and better for the power transistors.
 

ian field

Joined Oct 27, 2012
6,536
I calculate 1.14 ma through the I Lim pot for a 5.7 amp load. If it is at 500 ohms, you are delivering .57 volts to the current limiting resistor in the 723 chip. That's enough to cause a 1% error. Turn your I Lim to zero ohms for accuracy tests.
Even with darlington pass transistors, you need pretty sturdy driver in a 30A PSU, A L200 regulator from ST (like a TO220 but more pins) might make life easier - for sheer brute force minimalism, a TL431 can handle 100mA. The On-semi datasheet includes an application example of a TL431 buck regulator - but I'm guessing the OP wants the low noise of a linear.
 

bountyhunter

Joined Sep 7, 2009
2,512
Even with darlington pass transistors, you need pretty sturdy driver in a 30A PSU, A L200 regulator from ST (like a TO220 but more pins) might make life easier -
For a 30A design, I would use a three transistor complimentary Darlington for the pass transistor. A PNP device collector driving into the base of the NPN-NPN Darlington. I usually assume a worst case min beta of 10 for the transistors so that gives a min current gain of 1000, so the driver device only has to handle about 30 mA..
 

#12

Joined Nov 30, 2010
18,224
Please read the datasheet. The MJ11016 has a typical gain of 10,000 at these currents and the transformer is delivering at least 10 volts of head room. That means 3 ma out of the LM723.
 

Thread Starter

Dave_UYZ

Joined Jan 16, 2014
29
Please read the datasheet. The MJ11016 has a typical gain of 10,000 at these currents and the transformer is delivering at least 10 volts of head room. That means 3 ma out of the LM723.
dc current gain is 1000 at 20A (and it falls to 200 at 30A), according to the ON data sheet for the MJ11016.
 

bountyhunter

Joined Sep 7, 2009
2,512
dc current gain is 1000 at 20A (and it falls to 200 at 30A), according to the ON data sheet for the MJ11016.
At 30A, each xsistor sees about 8A so the min spec of 1000 is probably valid. That's why I would use a PNP to drive the darlingtons, to keep the IC with the reference from any heating effect.The 723 can drive the transistors, but there will be some heat generated. Not mandatory, but very cheap to improve performance.
 
Last edited:

Thread Starter

Dave_UYZ

Joined Jan 16, 2014
29
Thanks, guys.

Following some problems in the garage and a minor disaster with the (separate) fan supply, I now have the cool transistors and, following the arrival of the loads (about 120W each, and usually used for heating the water for a hot drink in the car [see here] ).
I connected one and it drew about 12-15A for long enough to heat the water satisfactorily. Tomorrow I check the volts drops at half-scale current.

My thanks for all the help.
:)
 

Thread Starter

Dave_UYZ

Joined Jan 16, 2014
29
After a bit of re-wiring and even more checking, I have the voltage drop from 13.83 to 13.80v, under a 15A load. Thanks for all the help, chaps.

Now; it seems that my attempts to use a MOSFET as a switch result in dropping a fair voltage across it. As I had in mind to use this "Solid State Switch" as the final act in a crow-bar circuit, I'm now a bit stuck for the best way to create a high-speed switch-off in the event of a problem in the pass stage/s.

I really would rather not short the whole thing out and pop the fuse/s (I've tried to find a spare fuse at 3am on a Sunday morning, and it's a disappointing experience at best). A relay can take up to 300mSec to break the circuit so I reckon that's bit slow.

The usual 3423 chip is my ideal detector but what's the next bit please?

Could I use the SS Switch to break the dc line to the Pass stages and would that be of any benefit ?
 

ian field

Joined Oct 27, 2012
6,536
After a bit of re-wiring and even more checking, I have the voltage drop from 13.83 to 13.80v, under a 15A load. Thanks for all the help, chaps.

Now; it seems that my attempts to use a MOSFET as a switch result in dropping a fair voltage across it.
If you're using a MOSFET as a series pass source-follower, don't forget you need at least 6V Vgs to turn it fully on.

Some people use a charge-pump voltage doubling rectifier alongside the power one to develop a higher voltage to drive the gate.
 
Top