lab power supply

lm317http://www.ortodoxism.ro/datasheets/nationalsemiconductor/DS009063.PDF

improvements

ASCII ART: Sense ! ! ---------- --+-/\\/-+-- ! \\ /e R2 ! ! --- ! ! R1 ! R3 ! +-/\\/\\- --+--/\\/\\------+ ! e\\ / ! ! --- ! ! ! ! ----+-/\\/\\/--+--! LM317 !-----+------------ R4

R1 makes the transistor's E-B drop less important so temperature has less effect on the performance.

Moving the R3 connection does the same for the other transistor at no cost in parts.

You have Q24 flipped.

What are ypou trying to do with Q26?

lm317http://www.ortodoxism.ro/datasheets/nationalsemiconductor/DS009063.PDF

improvements

when r108*imax it becomes approximately equal to the Q26 Vbe, Q26 is turned on ,it short circuit R109 and then switch off Q25 and it limits therefore the output current http://www.ortodoxism.ro/datasheets/fairchild/LM7805.pdf pag.23 fig.13

what do you think about of the current regulator? you have some idea to avoid to use a double potenziometer for the current adjust? we could be taking advantage of the advanced operational U2A

this is the updated scheme:

lm317http://www.ortodoxism.ro/datasheets/nationalsemiconductor/DS009063.PDF

improvements

currenthttp://www.ortodoxism.ro/datasheets/fairchild/LM7805.pdf

scheme:

Sorry MooseFET I have mistaken to copy! excuse me! the problem is that in Italy it is late night :D and I am distracted of nature now I understand why the regolator did not seem correct! many thanks!! I have corrected the scheme from many errors this is the new one link

lm317http://www.ortodoxism.ro/datasheets/nationalsemiconductor/DS009063.PDF

improvements

currenthttp://www.ortodoxism.ro/datasheets/fairchild/LM7805.pdf

scheme:

I don't like what you have done with R4. Transistors should not be uses such that their HFE matters. You want an emitter resistor and not a base resistor. The has an indirect effect on the current limit issue.

I worry that the current limit circuit will overshoot, ring and do other bad things. When the current limit is not active, the LM358 will have its output as high as it can manage. This is bad.

A diode or the like on the LM358 so that it never tries to take its output all the way up would help to make the current limit engage more quickly.