Electronics > Beginners
TL431 linear power supply
iMo:
I posted the LTSpice source - see above - you may download it and play with it as you like..
PS: you may want to install the "Bordodynov" library
http://ltwiki.org/?title=Components_Library_and_Circuits
not1xor1:
--- Quote from: imo on December 19, 2018, 10:43:33 am ---You have done a nice LTSpice simulation above - you may try with say 15V output, switching from 10mA -> 5A -> 10mA output current (with, say, 200ms period and 10us edges) and adjust the loop when required..
--- End quote ---
I got inspired by another circuit posted in a different thread by xavier60 (see below), but I used the dual voltage boost (common emitter) stage to tame open loop gain by local feedback, that is a sort sziklai with gain greater than one. Such circuit makes it much easier to compensate and variations of the compensation capacitor are not critical (at least according to LTspice) so you can choose between faster or more damped...and you even save a BJT :D.
I'll show the circuit later (I want to read the other messages in the thread before).
Here is xaveir60 circuit.
BTW the BJT in the current control feedback is another interesting trick, although I found it works much better, preventing overshots with no recovery delay, when used to avoid saturation by keeping the output of the voltage control opamp just a bit below (or above depending on the topology) the output of the current control one. I have to run more simulations to be sure it works in any condition though.
not1xor1:
--- Quote from: imo on December 19, 2018, 11:10:04 am ---The fact is when the control loop is "slow" an abrupt load change at the output will propagate "slowish". Thus it will not track properly.
If the user not1xor1 made the LTSpice source public, we could have played with it and finalize the efforts of the contributors..
--- End quote ---
I apologize for my laziness, I've little time and to post the source I should have chased lib files in my folders and made a zip as various models (lm358, tip41c) are not included with ltspice. :)
not1xor1:
--- Quote from: blackdog on December 19, 2018, 09:02:36 pm ---Hi,
Imo, try this if you want to test.
Remove Q3, R5, C1 and R7.
Use a diode between R1 (make him 100 Ohm) and the anode of the diode on the base of the BD243.
Place a 4K7 resistor between the base and the emittor of the BD243.
Make the currend source I1 10mA.
Make C2 220pF (play with it for good stability)
Make V3 30V
Connect V2 opamp power supply to V3, so remove V2.
Watch your opamp inputs, are they still connected the wright way in this test setup?
The problem with the schematic and many schematic of power supplys.
The average opamp has between 55 and 65 degrees phase margin.
And for a good stable linear power supply, you need at least 45 degrees phase margin, 60 is better.
If one or more transistors are behind the opamp, these transitors slow down the loop in which the opamp in the transistors is placed.
These transistors all eat a bit of the phase margin.
Below the 30 degree phase margin you come into the danger zone.
Please , please, please, do not disregard this information, because it always goes wrong if you do this.
It doesn't matter if you are developing a linear or a switching power supply or an electronic dummy load, this is always the case!
See my previous post with links to info on this topic from the TI website, first class info from TI!
Kind regards,
Blackdog
--- End quote ---
That has been discussed before and I also posted a couple of schematic diagrams showing that there is no need for a common emitter BJT in the loop.
I did not mind posting the simulation results, but both circuit have safe margin (70-90° or greater) as showed both in AC and transient simulations.
not1xor1:
here is the circuit with the local feedback. Unfortunately it has a really ugly short-circuit recovery overshot.
here is for comparison the previous circuit I posted, with just an opamp as voltage controller.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version