So I have a power supply I've been working on and in Ltspice it performs exponentially better than the jung super regulator. It is stable on the bench but I need special measures to measure the output impedance.
The sim shows 0 ohms @ 1hz and 162 nano ohms @ 20khz with some slight variation depending on virtual parts used and circuit alterations.
I'm inclined to believe the real life results will be somewhat in the same ballpark since the jung regulator simulated at close to published values.
But how do I even verify such small impedance?
Is there an inexpensive way?
The sim shows 0 ohms @ 1hz and 162 nano ohms @ 20khz with some slight variation depending on virtual parts used and circuit alterations.
I'm inclined to believe the real life results will be somewhat in the same ballpark since the jung regulator simulated at close to published values.
But how do I even verify such small impedance?
Is there an inexpensive way?
Jack Walton has done a complete measurement of several regulators including the output impedance, including the Didden/Jung regulator.
His article which details the measurement setup is available at Linear Audio.
Linear Audio | your tech audio resource
Jan
His article which details the measurement setup is available at Linear Audio.
Linear Audio | your tech audio resource
Jan
There are several ways to do it. Would suggest that you read the Keysight Handbook of Impedance Measurement.
The method I used is pretty straightforward and was described by Walt Jung in his 1995 articles -- (modified by me a bit) -- lightly load the output of the regulator with a non-inductive resistor. 25mA should suffice. using a sweep or programmable signal generator inject a small signal into the output via a 1,000uF electrolytic capacitor and 50 or 100 ohm resistor. the ratio of signal magnitude on both sides of the resistor will give you the impedance relative to the resistor value.
note that if you are using a bipolar pass transistor the impedance decreases with load as the transistor gain increases with current. don't kid yourself with a 1A load in a circuit which will only draw 25mA
i was able to measure down to a couple microOhms with this setup, you have to probe around the pass device for the lowest point node. I think that Walt probably did me better by several microOhms!
You don't need a voltage network analyzer, you can use an oscilloscope to measure the two voltages. The analysis should extend to 1MHz as the change of impedance will cue you to the regulator's stability.
The method I used is pretty straightforward and was described by Walt Jung in his 1995 articles -- (modified by me a bit) -- lightly load the output of the regulator with a non-inductive resistor. 25mA should suffice. using a sweep or programmable signal generator inject a small signal into the output via a 1,000uF electrolytic capacitor and 50 or 100 ohm resistor. the ratio of signal magnitude on both sides of the resistor will give you the impedance relative to the resistor value.
note that if you are using a bipolar pass transistor the impedance decreases with load as the transistor gain increases with current. don't kid yourself with a 1A load in a circuit which will only draw 25mA
i was able to measure down to a couple microOhms with this setup, you have to probe around the pass device for the lowest point node. I think that Walt probably did me better by several microOhms!
You don't need a voltage network analyzer, you can use an oscilloscope to measure the two voltages. The analysis should extend to 1MHz as the change of impedance will cue you to the regulator's stability.
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.