Linear Voltage Regulator Design

[cyw1984]

If I want to output 0-30V

Vin =

[Staigen]

Hi cyw1984

If I want to output 0-30V
Vin =

[audioguru]

An LM358 is very slow, about the same speed as the output transistor. A modern faster opamp would have better transient response. It wouldn't be difficult to compensate it with a little capacitor across the 40k resistor.

40k isn't a standard value so instead I would use 33k in series with a 15k or 20k trimpot, to adjust the gain of the circuit to exactly 3.000000. (Lots of bits in the DAC) ;D

The poor little PNP transistor #3 doesn't have a resistor to limit its base and collector currents. An emitter resistor for it would accomplish both.

The darlington connected output transistors #1 and #2 don't have resistors to turn them off. If they get warm then they will have thermal runaway without resistors. Their reaction time will be very slow without resistors. :(

[Guest Alun]

An LM358 is very slow

13V/us isn't that slow.

This thread has an interesting discussion on linear regulators, you don't have you build the voltage doubler if you don't want to, you could use a p-channel MOSFET as you're not bothered about the dropout voltage and on resistance.

[audioguru]

Hi Alun,
An old LM358 is half an old LM324. They are about the 1st low-power opamps ever made so sure they are darn slow.
They slew 2.5V in 5us which is only 0.5V/us. Their output is slew-rate limited above only 5kHz. :'(

Your P-channel Mosfet in a low-dropout regulator is a good idea to keep the supply voltage of the opamp down to its max rating. ;D

[Guest Alun]

Sorry, my mistake I was thinking of the LF351 for some reason.

I agree a faster op-amp would be better and as the load is connected to the drain of the MOSFET which is faster than most op-amps it will help speed things up even more. ;D

[cyw1984]

Hi Alun,
An old LM358 is half an old LM324. They are about the 1st low-power opamps ever made so sure they are darn slow.
They slew 2.5V in 5us which is only 0.5V/us. Their output is slew-rate limited above only 5kHz. :'(

Your P-channel Mosfet in a low-dropout regulator is a good idea to keep the supply voltage of the opamp down to its max rating. ;D

Then....Which op-amp should I use?????

[Guest Alun]

Try the CA3140E and have a go with my idea if have any p-chanel MOSFETs available.

[cyw1984]

Try the CA3140E and have a go with my idea if have any p-chanel MOSFETs available.

i SEE....It 's 4.5M high speed Op- amp
....Is your design need remove 555 timer ??
Also, which P-channel MOSFETs and Diode do you used??

[cyw1984]

It is my new design for my circuit

If LM358 is slow op-amp...which opamp should I use???


Also, how to add on somthing to improve this circuit . Such as capicator to reduce noise and zener diode make the voltage stable???

Thanks a lot

[ante]

What’s the DAC look like? How many bits?  ???

[audioguru]

Hi Cyw,
You didn't finish your circuit with the important stuff I mentioned, so I added/changed a few things.
About the only thing that remains is adding current limiting or regulation. If you draw too much current or short the output, it will blow-up! ;D

[Guest Alun]

Just one question, why use so many transistors?
The op-amp already has a very high gain, adding more transistors won't increase the performance much, infact such a high open loop gain is likely to lead to instabiliy and oscilation. I think you'll find that one p-channel MOSFET will out perform all those transistors. ;D

[Staigen]

Hey boys, and old man Audioguru
Lets stop a while and think, and dont rush away
There is so much we dont know about this regulator, but there is at least some data given. We know:
The input voltage is 40 Volts
The output voltage: 0 to 30 Volts, depending on Vref
Vref: 0 to 10 Volt from a DAC

Something we dont know, and i belive we must know:
Output current
In some instances the output resistance of the DAC

Something that would be nice to know:
Data about the DAC(not necessarily needed, except above)

The 3140 IC have a max voltage rating of only 36 Volt, so why stress it above that, use an another IC

---

Audioguru, i have seen a lot of commercially designed circuits with 2N3055, both switched and linear, and the resistor between base and emitter has been in the range of 100 to 220 ohms, dont you think that 10 K is a little bit high? And also the resistor between the 2N3053:s base and the output, its usally betwwen 1 K and 1,5 K? What is the purpose of the 1 K resistor between the base of the 2N3053 and the collector of the 2N3906? And no offset adjust?

---

Alun, your circuit is a good one, exept that the 3140 IC are not guaranteed for 40 Volts and there is no voltage limit on the MOSFET:s gate! And no offset adjust.

---

Whats so wrong with the original circuit, exept that it have the wrong IC and no resistors from the base to the emitter on some transistors? It also lack protection for the base on one transistor and it have no offset adjustment! But, it have output short protection, so it is much safer! The shematic is here under:

//Staigen

[cyw1984]

What’s the DAC look like? How many bits?

[cyw1984]

Hi Cyw,
You didn't finish your circuit with the important stuff I mentioned, so I added/changed a few things.
About the only thing that remains is adding current limiting or regulation. If you draw too much current or short the output, it will blow-up! ;D

Thanks a lot...But I have few question about this....

1. DC sypply voltage only 36V, Is ir really suitable???

2. How to caculate the resistor values (10K) between PNP Base and Emitter and NPN base to output(10K and 100K)???

3. What is the values of capacitor should I use??

4.I know it haven't current limiting....I just for testing . did you added current limiting on it??

Thanks a lot...Sorry for my trouble and poor english

[cyw1984]

Hey boys, and old man Audioguru
Lets stop a while and think, and dont rush away
There is so much we dont know about this regulator, but there is at least some data given. We know:
The input voltage is 40 Volts
The output voltage: 0 to 30 Volts, depending on Vref
Vref: 0 to 10 Volt from a DAC

Something we dont know, and i belive we must know:
Output current
In some instances the output resistance of the DAC

Something that would be nice to know:
Data about the DAC(not necessarily needed, except above)

The 3140 IC have a max voltage rating of only 36 Volt, so why stress it above that, use an another IC

---

Audioguru, i have seen a lot of commercially designed circuits with 2N3055, both switched and linear, and the resistor between base and emitter has been in the range of 100 to 220 ohms, dont you think that 10 K is a little bit high? And also the resistor between the 2N3053:s base and the output, its usally betwwen 1 K and 1,5 K? What is the purpose of the 1 K resistor between the base of the 2N3053 and the collector of the 2N3906? And no offset adjust?

---

Alun, your circuit is a good one, exept that the 3140 IC are not guaranteed for 40 Volts and there is no voltage limit on the MOSFET:s gate! And no offset adjust.

---

Whats so wrong with the original circuit, exept that it have the wrong IC and no resistors from the base to the emitter on some transistors? It also lack protection for the base on one transistor and it have no offset adjustment! But, it have output short protection, so it is much safer! The shematic is here under:

//Staigen

It is 12BIT DAC 0-10V
also........It about 0-2V
but....current limit I will design later

[Guest Alun]

Alun, your circuit is a good one, exept that the 3140 IC are not guaranteed for 40 Volts

Use a higher voltage IC or a separate power supply for the IC.

and there is no voltage limit on the MOSFET:s gate!

Just add a zenner, see attachment.

And no offset adjust.

Why do you need to offfset adjust? Any error in the output can be taken care of with a small trimmer resistor.

Whats so wrong with the original circuit, exept that it have the wrong IC and no resistors from the base to the emitter on some transistors?

You could also power the IC from a zenner diode shunt regulator to get round the voltage problem but this won't work with my MOSFET circuit, however I can solve this problem too.

I chose a MOSFET because it's simpler, will work out cheaper, has a lower drop out voltage (not important here though), will have a lower quesant current and a far better transient response.

current limit I will design later

Even though you haven't considered limiting the current at this stage it's still very important to know the magnitude, this will give you an idea of the componant ratings the design techniques you should use. By looking at the componats you've used and your origional design I assume you want the output current to be around 10 amps or so, is the correct?

[cyw1984]

Use a higher voltage IC or a separate power supply for the IC.


Just add a zenner, see attachment.


Why do you need to offfset adjust? Any error in the output can be taken care of with a small trimmer resistor.


You could also power the IC from a zenner diode shunt regulator to get round the voltage problem but this won't work with my MOSFET circuit, however I can solve this problem too.

I chose a MOSFET because it's simpler, will work out cheaper, has a lower drop out voltage (not important here though), will have a lower quesant current and a far better transient response.


Even though you haven't considered limiting the current at this stage it's still very important to know the magnitude, this will give you an idea of the componant ratings the design techniques you should use. By looking at the componats you've used and your origional design I assume you want the output current to be around 10 amps or so, is the correct?

I want the output current to be around 2 to 3 AMPS.....Your 's enclose can the output achieve to 0-30V???
Also Use a higher voltage IC ................which IC should be suitable on it?? And which  you used??

[audioguru]

Hi nearly old man Staigen,
You are absolutely correct! I was way off with my selection of high-value "turn-off" resistors for the output transistors in my "fixed" circuit. I have now chosen lower value resistors based on what is inside a TIP120 darlington.

I used a 1k resistor in series with the collector of the 2N3906 to limit its collector current and dissipation if the project's output is shorted. It will still provide enough drive to the output transistors to blow a fuse if the output is shorted.

I think the circuit can deliver 30V at 1.5A with a 34V supply to protect the 36V max-rated opamp.

It would be fun to try to stop this extremely high gain circuit from oscillating. ;D


Hi Alun and Cyw,
I don't think a single 2N3055 or power Mosfet can dissipate more than about 60W even on a pretty big heatsink. Therefore the max current is only 2A at low output voltages. That's why I proposed using 2 output transistors in our 0V-30V modified project for 3A output, and using 3 output transistors on our modified for 5A project.

[Guest Alun]

Hi Alun and Cyw,
I don't think a single 2N3055 or power Mosfet can dissipate more than about 60W even on a pretty big heatsink.

That's why I'd use a large MOSFET like the IRF9131 or IRF9531 they can safely dissipate 75W and the dropout voltage will be only 0.9V at 3A, while yours will be at least 3.6V ;D

[audioguru]

Hi Alun,
Those little Mosfets can't dissipate 75W safely, in practise. They are rated for 75W if their case is held somehow at 25 degrees C. You can't find a heatsink that is infinitely large to do the job of perfect cooling, so your heatsink will get hot and you must follow the power derating curve. You might be able to allow a dissipation of 50W if you use a big heatsink plus a fan. With a supply of 31V, the max current would be only 50/31 = 1.6A.

A 115W-rated 2N3055 would be able to dissipate 60W with a big heatsink and without a fan. Since they have a metal case, their chip is allowed to go to 200 degrees C vs only 150 degrees C for the plastic case.

I like 300W-rated Mosfets that will do a good job, but they are expensive. Actually, I haven't seen P-channel 300W Mosfets, only N-channel. ;D

[Guest Alun]

Just parallel them then - it's also far easier to do this with MOSFETS than BJTs.

[audioguru]

Yeah, When you parallel Mosfets then the hottest one increases its resistance so it allows the cooler one to carry more of the current. ;D

[Guest Alun]

Here's an idea to make the current limit programable.

U3 is a differential amplifier with a gain of R4/R3, this measures the volagate across