It now seems obvious that the "odd" things that were occurring, such as the ripple increasing when a small series resistance was added before the smoothing caps, and also the somewhat-high ripple levels compared to what could possibly be achieved, were being caused by regulator dropout effects. i.e. The +/-30V regulators' input-to-output voltage differentials were too small.
I will modify the power supply, either to have a lower output voltage or to use a higher-voltage transformer, and will post an updated version as time permits.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
I will modify the power supply, either to have a lower output voltage or to use a higher-voltage transformer, and will post an updated version as time permits.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
New downloadable PSU Spice simulations:
I have made two "new-and-improved" regulated power supply models and have re-done my "LT-Spice Stuff" webpage, to include them, and have deleted the previously-discussed power supply versions from that page.
The new power supplies' downloadable spice simulations are at:
http://www.fullnet.com/~tomg/gooteesp.htm
There are GIF images of the schematics, there, as well.
One is a quad regulated PSU; +/-28V @ 5A max, and +/-18V. Its output ripple voltage ranges from 72 uVpp to 112 uVpp, for 1A to 5A output currents, with static (resistive) loads. It is powered by two 30VAC RMS sources (either dual secondary windings or two separate transformers).
The other one is a dual regulated PSU; +/-22V @ 4A max. Its output ripple voltage ranges from 33 uVpp to 94 uVpp, for 1A to 4A output currents, with static (resistive) loads. It is powered by two 25VAC RMS sources (either dual secondary windings or two separate transformers).
The second (+/-22V) one includes (and uses) a spice model of a toroidal power transformer, made directly from actual measurements, as described on the same webpage. (In this case, it was a Hammond 180L50 model, with both primary and secondary windings in parallel, giving it a rating of 120V --> 25V @ 4.8A.) Also included is a simple model of the AC Mains, as well as snubbers for the transformer and for the smoothing capacitors.
Both power supply simulations model many parasitic effects, including capacitors' ESR (Equivalent Series Resistance), inductance, and leakage current, and PCB traces' or wires' resistance and inductance.
A convenient star ground simulation scheme is used, including simple parasitic impedance elements, to make it easier for the user to begin to experiment with the sharing of various ground conductors by return currents, and to investigate the effects of different lengths and sizes of PCB traces or wires.
There is nothing very novel or different about these PSU designs, except perhaps that they use simple MOSFET-based soft-start (inrush current limiter) circuits to enable the use of extra-large capacitors on the main regulators' adjust pins, to provide very low output ripple. But the main purpose in presenting complete, working spice models was to enable others to more-easily begin to use spice simulations to investigate (and "tweak") power supply and other circuits' behaviors, including such things as star grounding and parasitic impedances, and, for example, the effects of allowing the wrong currents to share a conductor.
OT: The same webpage also includes a complete simulation model for a DC Servo that I designed for an audio-frequency amplifier, using a differential integrator and a "somewhat-unique" DC Accurate third-order low-pass filter, for which I don't know the name of the topology.
(Aside: "Somewhat unique" is bad usage, I know. Something is either unique or not. There is no "almost" unique.)
All comments are welcome.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
I have made two "new-and-improved" regulated power supply models and have re-done my "LT-Spice Stuff" webpage, to include them, and have deleted the previously-discussed power supply versions from that page.
The new power supplies' downloadable spice simulations are at:
http://www.fullnet.com/~tomg/gooteesp.htm
There are GIF images of the schematics, there, as well.
One is a quad regulated PSU; +/-28V @ 5A max, and +/-18V. Its output ripple voltage ranges from 72 uVpp to 112 uVpp, for 1A to 5A output currents, with static (resistive) loads. It is powered by two 30VAC RMS sources (either dual secondary windings or two separate transformers).
The other one is a dual regulated PSU; +/-22V @ 4A max. Its output ripple voltage ranges from 33 uVpp to 94 uVpp, for 1A to 4A output currents, with static (resistive) loads. It is powered by two 25VAC RMS sources (either dual secondary windings or two separate transformers).
The second (+/-22V) one includes (and uses) a spice model of a toroidal power transformer, made directly from actual measurements, as described on the same webpage. (In this case, it was a Hammond 180L50 model, with both primary and secondary windings in parallel, giving it a rating of 120V --> 25V @ 4.8A.) Also included is a simple model of the AC Mains, as well as snubbers for the transformer and for the smoothing capacitors.
Both power supply simulations model many parasitic effects, including capacitors' ESR (Equivalent Series Resistance), inductance, and leakage current, and PCB traces' or wires' resistance and inductance.
A convenient star ground simulation scheme is used, including simple parasitic impedance elements, to make it easier for the user to begin to experiment with the sharing of various ground conductors by return currents, and to investigate the effects of different lengths and sizes of PCB traces or wires.
There is nothing very novel or different about these PSU designs, except perhaps that they use simple MOSFET-based soft-start (inrush current limiter) circuits to enable the use of extra-large capacitors on the main regulators' adjust pins, to provide very low output ripple. But the main purpose in presenting complete, working spice models was to enable others to more-easily begin to use spice simulations to investigate (and "tweak") power supply and other circuits' behaviors, including such things as star grounding and parasitic impedances, and, for example, the effects of allowing the wrong currents to share a conductor.
OT: The same webpage also includes a complete simulation model for a DC Servo that I designed for an audio-frequency amplifier, using a differential integrator and a "somewhat-unique" DC Accurate third-order low-pass filter, for which I don't know the name of the topology.
(Aside: "Somewhat unique" is bad usage, I know. Something is either unique or not. There is no "almost" unique.)
All comments are welcome.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
Re: New downloadable PSU Spice simulations:
Hi Tom. The gif image format is better (smaller file size and less compression artifacts) than jpg for images such as the schematics you're posting.
gootee said:
Hi Tom. The gif image format is better (smaller file size and less compression artifacts) than jpg for images such as the schematics you're posting.
Re: Re: New downloadable PSU Spice simulations:
Thanks, Brian! I think I can do that.
- Tom
BWRX said:
Thanks, Brian! I think I can do that.
- Tom
- Status
- Not open for further replies.