Linear regulated power supply design help

#12

Joined Nov 30, 2010
18,224
I'm glad we got that cleared up. The concept is valid. The accuracy might have to be polished with a voltage reference to stabilize the current limit command.
 

Thread Starter

PaulEE

Joined Dec 23, 2011
474
The design is going to use a high-end reference and the voltage and current outputs are going to be digitally-controlled. This particular portion of the design was what was hanging me up...once I get something functional, I can do all the polishing and testing...I feel a bit silly that I decided the idea wasn't right based on a computer. What on earth was I thinking!? ha.
 

crutschow

Joined Mar 14, 2008
34,281
I tried to run a simulation of a jfet amplifier stage with op-amp controlled DC feedback to compensate for the natural variance in jfet IDSS ranges. I had to build it one component at a time or the simulator couldn't figure it out. After more than an hour of adding one or 2 components at a time, it finally worked. I thought a simulator should work better than that.
So you could have built and tested the real circuit in less than an hour? You must be fast. ;)
 

Thread Starter

PaulEE

Joined Dec 23, 2011
474
Hi guys,
So far I've built the +\- supply and have a precision +5v reference buffered with a unity gain opamp and power transistor.

Next will be the differential volts and amps readings, followed by a real test of the circuit spice said wasn't working. Stay tuned :)
 
You'll find in practice that if you feed the current loop through the voltage loop it unnecessarily complicate the design and stabilization of the 2 loops. Better to have them separate and switch between loops with an analog OR like this. Of course this is just an illustration and in practice there is a current sense resistor in the negative rail and the rest. I have done this with discretes in which case the amps are open collector and no diodes are necessary. Cheers.
 

Attachments

#12

Joined Nov 30, 2010
18,224
So you could have built and tested the real circuit in less than an hour? You must be fast. ;)
I understand your position that you can simulate faster than I can build hardware. I can make a drawing in MSPaint faster than most people. :DHowever, I had already designed this circuit, tested it, built it, shipped it to the customer, and waited 10 years to see if it would fail. I was just checking to see if LTSpice could simulate a known good circuit instead of sitting there like a lump. Apparently LTSpice has difficulty with feedback loops, or beginners have difficulty with LTSpice, and that seems to be the problem this OP encountered.
 

Thread Starter

PaulEE

Joined Dec 23, 2011
474
I will look into the OR gating in more detail. After being up literally all night, I've got together the +/-40 V unregulated power supplies. For the sake of simulation, I need to step them down and use the resulting split supply to power the opamps, etc. I whipped up a few very, very quick opamp-transistor regulators for +/-20V and got as far as building the voltage regulation circuitry. That portion worked like a charm - while teaching me WHY to stick all the bypass caps near the amps! ;)

I use TINA SPICE and usually have very good results with it; this particular project caused it to either 1) give wrong results or 2) I didn't enter something correctly. 2) may be the case because I was very near the op-amp limits in the program as far as supply voltages, but even still, I'm surprised I didn't get closer than I did.

I will keep you all posted. Thanks very much :)
 

#12

Joined Nov 30, 2010
18,224
I guess I'm just not keeping up with the times. One of my beliefs is that electrons can not lie. Either you built the circuit right, or it won't work. Learning SPICE brings in a whole new set of opportunities to make a mistake. Still, we don't have all that many people come on here with SPICE failures. Some people figure it out, some don't.
 

crutschow

Joined Mar 14, 2008
34,281
...................................
Learning SPICE brings in a whole new set of opportunities to make a mistake.
................................
I suppose I'm beating a dead horse here, but the idea of SPICE is to detect errors in the circuit before you build it, not afterwards. :rolleyes:
 
Top