0

In all designs that I could find they used two LM317, one for CC and another for CV. Is there a reason why no one used a circuit like this with only one LM317?

schematic

simulate this circuit – Schematic created using CircuitLab

it's based on dave jones design, he used an LT3080.

Assuming we have a negative voltage for adjust pin and output voltage actually goes to 0V.

ElectronSurf
  • 2,185
  • 3
  • 24
  • 61
  • See my answer to [smartest-way-to-use-current-limit-using-lm317](https://electronics.stackexchange.com/questions/215418/smartest-way-to-use-current-limit-using-lm317/215474#215474) for an approach with two LM317s. Your "measure voltage drop" box is going to get complicated and there are better, more efficient ways to make a CV/CC PSU than using 40+ year-old technology. – Transistor Oct 27 '19 at 10:07
  • @Transistor I actually came across your post several days ago while researching (Thanks for writing such complete answer). By "more efficient ways" you mean using switching regulators? or there are modern versions of this **linear regulator**? – ElectronSurf Oct 27 '19 at 10:15
  • 2
    Your circuit is conceptually OK, but has its own issues. Using 2 x LM317 MAY be as easy. OR you could use NO LM317s and implement the CV with a pass transistor + opamp + reference with several advantages: Vout_min could be designed to be close to zero Volts with no negative supply, The minimum Vheadroom + V_I_sense of about 4V + 1.25V = 5V+ from 2 x LM317 would be avoided. || As shown your high side current sense needs a sense circuit whose Vcommonmode includes Vin+. This is not technically hard but either needs appropriate opamps or a current mirror. – Russell McMahon Oct 27 '19 at 10:27
  • 1
    I'd look at switchers but I'm a bit out of touch with the various devices. If you want to research some more older technology have a look at the LM723 from the same era. There are many designs for CV/CC PSU available and you may find the circuit analysis educational. – Transistor Oct 27 '19 at 10:37

1 Answers1

1

It is going to end up more complex than the circuit using two LM317- so the question might be more why you would use this circuit.

The LM317 has a lot of (minimum) voltage drop (3V minimum recommended) and two of them in series have twice that, so a lot of power will be wasted in the regulators. A sense resistor of 0.1 ohm will drop only 100mV at 1A.

You also need a negative supply voltage to turn the LM317 off but that may be necessary anyway if you’re making a “lab supply” that can be set to less than 1.25V.

Spehro Pefhany
  • 376,485
  • 21
  • 320
  • 842
  • I want to build it with a single LM317 to lower the voltage drop and heat, plus I can control it digitally this way. The complexity that you mention is because of the measuring the current at high-side of the circuit or there are some "things" that I need to be aware of? (other than minor circuitry around LM317 like capacitors, diodes etc..) – ElectronSurf Oct 27 '19 at 10:55
  • 1
    The high side measurement can be done easily with a modern measurement chip, not so easily with LM317 1970s technology. The only caution I see is that you might well get oscillation at the current limit rather than a smooth transition to CC linear mode as you have in a commercial lab supply. – Spehro Pefhany Oct 27 '19 at 10:58
  • What's your suggestions to reduce or eliminate(if possible) oscillation while it's in CC mode? (link to any article/tutorial is much appreciated). – ElectronSurf Oct 27 '19 at 11:05
  • 1
    Start by changing that transistor setup to reduce the current limit gain. Maybe use simulation before building anything. – Spehro Pefhany Oct 27 '19 at 11:07