I've got some strange behavior with this power supply I'm building. It's the Maida topology with adjustability from 2V to 490V. The intent is to use this supply for tube electronic diagnosis and design. Here is the schematic for the part of the power supply relevant to my question:
Under the current test conditions, I've got the variac that's driving the front end of the power supply (not shown) dialed down to deliver 118V DC, which is applied to the node labeled "DC In". I can dial down the voltage on the regulator board to 60V output across a 10K load. At that point, the voltage applied to the mosfet gates is 75V, and everything is working just fine...until... I dial down the voltage adjust pot just a bit more which causes the output voltage to jump to 114V DC. At that instant the voltage across the zener (the one below the 181 ohm resistor) jumps from 15V (where it should be) to about 8V, and the voltage at the gates of the mosfets jumps to 118V. Dialing the voltage adjust pot back up just a bit causes everything to swing back into regulated mode (60V output). The pot setting is 5.1K at this boundary condition. Currently there is no fine adjust pot in the circuit--only the coarse adjust pot and the 1K resistor below it, and it's the coarse adjust pot that reads 5.1K under these conditions.
The question is why is it behaving this way? I am thinking I should be able to dial down the output voltage to about 8V DC. I speculate the gates of the mosfets suddenly become unhappy at 60V DC output, and can no longer control the drain-source voltage. But why?
Can anyone help me understand what is going on and what I might do to fix this?
Also: The external protection zener and diodes on the mosfets are not present in the circuit as tested because they are built in to the mosfet devices themselves.
Under the current test conditions, I've got the variac that's driving the front end of the power supply (not shown) dialed down to deliver 118V DC, which is applied to the node labeled "DC In". I can dial down the voltage on the regulator board to 60V output across a 10K load. At that point, the voltage applied to the mosfet gates is 75V, and everything is working just fine...until... I dial down the voltage adjust pot just a bit more which causes the output voltage to jump to 114V DC. At that instant the voltage across the zener (the one below the 181 ohm resistor) jumps from 15V (where it should be) to about 8V, and the voltage at the gates of the mosfets jumps to 118V. Dialing the voltage adjust pot back up just a bit causes everything to swing back into regulated mode (60V output). The pot setting is 5.1K at this boundary condition. Currently there is no fine adjust pot in the circuit--only the coarse adjust pot and the 1K resistor below it, and it's the coarse adjust pot that reads 5.1K under these conditions.
The question is why is it behaving this way? I am thinking I should be able to dial down the output voltage to about 8V DC. I speculate the gates of the mosfets suddenly become unhappy at 60V DC output, and can no longer control the drain-source voltage. But why?
Can anyone help me understand what is going on and what I might do to fix this?
Also: The external protection zener and diodes on the mosfets are not present in the circuit as tested because they are built in to the mosfet devices themselves.
Indeed. When troubleshooting regs, always, always, always stick a scope on the output.
I have had similar problems with a discrete version of the same topology. In the end it was the internal opamp doing the comparison between the (floating) reference and the voltage sample (referenced to the output), which latched up because of being forced to operate outside its common mode input voltage range.
When trying to regulate for a low output voltage, the sampled output is so low that it may go below the minimum common input mode, thus triggering a voltage reversal at the output of the opamp, creating a latched up condition, and forcing a maximum output (i.e. something within a few volts lower than the input to the regulator circuit). The situation might be quite catastrophic in some circuits.
When trying to regulate for a low output voltage, the sampled output is so low that it may go below the minimum common input mode, thus triggering a voltage reversal at the output of the opamp, creating a latched up condition, and forcing a maximum output (i.e. something within a few volts lower than the input to the regulator circuit). The situation might be quite catastrophic in some circuits.
A big cap on the output of the LT1084 does improve the situation, but there is still some instability (scope shows it). And when the instability is reached with the cap on the output, the "flop" out of regulation kills the regulator! Furthermore, a Zobel on the output (as Maida also used) makes the instability worse.
A bipolar predrop device (instead of the mosfets) is much more immune to this behavior. Additionally, I found that a non-variable voltage divider seems to improve things. Unfortunately, all of this is evidence that the LT1084 is pretty much a lousy choice for this type of circuit.
In hindsight, I should have started with the "stock" Maida circuit and tweaked from there, which is what I'm now going to do. Thanks for the replies.
A bipolar predrop device (instead of the mosfets) is much more immune to this behavior. Additionally, I found that a non-variable voltage divider seems to improve things. Unfortunately, all of this is evidence that the LT1084 is pretty much a lousy choice for this type of circuit.
In hindsight, I should have started with the "stock" Maida circuit and tweaked from there, which is what I'm now going to do. Thanks for the replies.
In any case I'm using the LM317 and it's working nicely at this point. Ultimately I moved away from the LT1084 because of the need for a big capacitor on the output for absolute stability, not to mention a cap on the output does not take kindly to accidentally shorting the output lead to ground.
This is simply not correct. As a matter of fact, the example in the original Maida paper IIRC is over 300V. Perhaps you could give some more information on what you are trying to do?
https://www.google.it/url?sa=t&sour...iL-dnxl8RtMER-Uag&sig2=ki2-1omGPN6Y4Elv4Zqozw
What's the original paper do you refer?
Can you please post a link?
I simply want to make a maida regulator @250Vdc and @50mA.
Oops, I stand corrected. You did link to the original paper, and it is a 160V example he gives.
No matter, we use higher voltages all the time for tube circuits. You can search on "Maida" in the tube forum. Here is one I built a while ago:
http://www.diyaudio.com/forums/tubes-valves/75268-maida-regulator-these-results-expected.html
No matter, we use higher voltages all the time for tube circuits. You can search on "Maida" in the tube forum. Here is one I built a while ago:
http://www.diyaudio.com/forums/tubes-valves/75268-maida-regulator-these-results-expected.html
That's not a Maida regulator. There's no device to protect the LM from a short.
This circuit seems to me to be very interesting.
This is expecially true after I read this excellent explanation of how Maida regulator works:
LM317 Maida Style regulator - Electrical Engineering Stack Exchange
Actually it is not so difficult at all (after I read that...).
I'll do further investigations.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.