few questions regarding linear PSU

Thread Starter

gigabyte091

Joined Dec 15, 2015
17
You just noticed a bad design choice. Look at my blogs about voltage regulators. They show how to get to a zero output voltage and how to float the regulator. You can get to zero with an LM723, but it requires a dual voltage supply to get there. The OP-07 is an accurate chip, I just can't remember if it will work with a single supply.
I will definitely look at your blog
 

dl324

Joined Mar 30, 2015
16,943
I wanted to ask where is the best to put voltage adjust pot ? in my schematics i put pot across voltage reference and feed voltage to op-amps, but is this a good way ? pot isn't temperature stable element.
That's the usual way. Few components will not have a tempco. What is your application (need for temperature stability)?

You have a number of unnecessary resistors in your circuit. For example, what are R3, R5, R6, and R7 for? Pulling current through the base of Q2 will turn it on. R4 seems large.
upload_2015-12-16_8-15-13.png
 

#12

Joined Nov 30, 2010
18,224
Without looking at the links by Jony130:

R7 and R4 defeat the leakage current from collector to base at high temperatures. Look for ICBo @ Tj =Tmax. Assume that is the high temperature collector leakage current and set R7 and R4 to pass that current while keeping the base to emitter voltage below 0.5 volts. 10k looks reasonable. 220 ohms does not look reasonable. R6 and R3 don't even have a job to do, except maybe keeping Q1 busy, but Q1 will change in the final design.
 

Thread Starter

gigabyte091

Joined Dec 15, 2015
17
Temperature stability isn't so important, this power supply is going to be used as battery charger, and for powering 12V air pump, etc... Nothing sesitive.

As i said this is not final design and will be changed. R6 and R3 is there for current limiting tru Q2, also R3 is there for stability reasons, that information was pulled put from other forums.

What value of R7 should i put ?
 

dl324

Joined Mar 30, 2015
16,943
R6 and R3 is there for current limiting tru Q2, also R3 is there for stability reasons, that information was pulled put from other forums.
R3 and R6 are not necessary. The PNP/NPN combination you're using is a classical substitution for a power PNP. That circuit uses your R4.
What value of R7 should i put ?
I think it's superfluous.

Refer to post #26. Pulling current through the base of the PNP turns it on.

If you were using the PNP as an external pass transistor for a 3 terminal regulator, you'd need the resistor to bias the transistor on.
 

#12

Joined Nov 30, 2010
18,224
max output current will be about 2A
I know, i will add 2-3 BD911 in parallel, and BDX53C as a driver transistor.
Temperature stability isn't so important,
You already know that using only one pass transistor will get too hot, then say temperature stability is not so important?
OK. Try using only one pass transistor, do not follow my instructions in post #24 to calculate a resistor from base to emitter, and find out if it works. If the pass transistor has low leakage, a good heat sink, you live in a cold place, and/or there is always a load more than sufficient to use up the leakage current times the transistor gain, it will work.

I am accustomed to designing power supplies that work properly under all conditions of power line voltage, load resistance, and ambient temperature. You are not required to use my criteria.
 

Thread Starter

gigabyte091

Joined Dec 15, 2015
17
I mean it isn't important if pot or reference Voltage source change value with temp change too much, my psu is not going to be ultra super precise lab psu.

Ofcourse is important that transistors is kept as cold as possible. Im planning to use transformer with multi taps on it and change input volts to keep power dissipation as low as possible. You misunderstood me about temperature stability.

I'm more than Grateful for your advices and its helping me alot :)
 

dl324

Joined Mar 30, 2015
16,943
Is there any books or app notes on this topic you could recommend me ?
I don't think you'll find many app notes that go into detail on how to design a linear regulator from scratch; a text book will likely cover the basics and not much else. Standard procedure now is to use a commercially available part. What aspects do you want to understand?

Be careful of what you find on the net. The two that Jony130 posted may work, but the resistor values are odd. The PNG picture is a basic regulator with current limit and only has a few unnecessary resistors (like the 1K 1W in series with the regulator input and opamp power pin; just wasting power). The size of the resistor used to turn on the power NPN is too small unless making T1 work hard was intentional. It needs supply decoupling for the LM358, but may have been omitted for clarity.
 
Last edited:
Top