Those back-to-back JFET/Depl.MOSFET limiters are best used with very high load impedance, normally connected to the OpAmp (or whatever) which also gets its DC bias through the limiter.
Then they are basically distortion-free.
I've used LND150s (as per Merlin Blencowe's article) which are rather high impedance, though, but can withstand 500V.
BSS139 is nice and is what I will use next, having much lower "on" resistance on the order of 100R total (for the whole cell) while limiting to a 10mA which simple (external) catch diodes to the rails can handle -- should be very low capacitance, of course, then the compensation of the (-)-input may be omitted when source impedance is low.
But the elephant in the room is ESD, you definitely need ESD protection (clamp voltage) to well below the rated max voltage. I broke LND150 instantly with one single ESD strike :-( I wonder why Mr.Blencowe did not mention ESD issues at all in in his otherwise excellent article.
I also had an eye on 100Ohm'ish 2SK880 JFETs which have 50V voltage rating but then overvoltage protection gets really critical.
As I am trying to figure out how to measure the PSRR on the unit I am developing for the regulators I am using for the power supply, I am testing how to setup the body plot of the a given trasnfer function, so fort the case I have setup a low pass filter and when I go and plot I get this nice wavefor:
UNFORTUNATELY, when I deselect DC coupled option I get this horrible waveform:
Which I have no idea why and where these peaks come from, they happen in conjunction with the autorange switching on to adjust the signal level. While you would think there could be a problem somewhere, the funny part is when I reverse the settings to do instead of output divided by input and calculating the body plot, but I do input divided by output as effectively the inverse of the trasnformation.....surprise!!
The autorange click same way but there is no glitch displayed. This is most likely a bug on the SW. Was this ever fixed on the AP2700 version?
UNFORTUNATELY, when I deselect DC coupled option I get this horrible waveform:
Which I have no idea why and where these peaks come from, they happen in conjunction with the autorange switching on to adjust the signal level. While you would think there could be a problem somewhere, the funny part is when I reverse the settings to do instead of output divided by input and calculating the body plot, but I do input divided by output as effectively the inverse of the trasnformation.....surprise!!
The autorange click same way but there is no glitch displayed. This is most likely a bug on the SW. Was this ever fixed on the AP2700 version?
right now I am using a simple low high pass filter config as show in the AP app note
However, I am getting a curve that is different from what I would expect, and I don't know if it has to do with the injection method, in fact due to the low impedance nature of the power supply, the injected signal is about 30-40mV pk-pk even though I have 500mV or so at the input 20ohm gen output.
I am considering the picotest J2120A line injector.
Do you know if there is a simpler way to build a line injector for the PSU bus voltage? I know signal transformers are used for many AC injecting on control loop, but they are not tolerant to any DC current, so to my knowledge, they are not an ideal candidate for voltage bus's application.
However, I am getting a curve that is different from what I would expect, and I don't know if it has to do with the injection method, in fact due to the low impedance nature of the power supply, the injected signal is about 30-40mV pk-pk even though I have 500mV or so at the input 20ohm gen output.
I am considering the picotest J2120A line injector.
Do you know if there is a simpler way to build a line injector for the PSU bus voltage? I know signal transformers are used for many AC injecting on control loop, but they are not tolerant to any DC current, so to my knowledge, they are not an ideal candidate for voltage bus's application.
Will post a schematic tomorrow. What is the quiescent current running through the V+ of the power supply?
thank you very much Duke for the explanation!!! So perhaps the artifact I see it a little bug on the APWIN perhaps which is not a big deal anyway. I am trying to measure the PSRR of the PSU module, I am having quite a bit of trouble without a signal injector.
Thanks a lot! I am having a lot of issue injecting the signal without loading the generator. The quiescent current is about 100mA.
I will look forward to seeing your schematic. I am completely stuck at this point without a signal injector.
BTW I have been reading a TI article, they are using a lot of extra HW to achieve the measurement.
https://www.ti.com/lit/an/slyt547/slyt547.pdf?ts=1725088745924
I am out now as soon as I go home I will look for the article. In fact this morning by doing this operation I have damaged the output of my tek test backup generator that I will now need to repair 🥲🤬
Last edited:
Thank you so much Duke I will connect with the LinkedIn group and look over the material you indicated.
I am going shopping before going home for a little small amp to use on the bench for injection test as you recommended!!!
Thanks again!
Duke, I got home a couple of hours ago, after retrieving a small amp at a goodwill. I have 220uF/20ohm between the PSU and the amp.
I tried to perform several measurements, but for some reasons, I am not getting consistent results. I will call it a day with testing since it's 11PM I will resume tomorrow morning when I am fresh.
BTW in AC coupling mode, I can still see those glitches I showed few posts above. It seems to be level dependent but not sure why in DC mode it's gone.
This works quite well, minimal parts. You don't want to suffer the unfortunate consequences of placing a large DC signal on the Analog Generator of the AP.
There is a way of using a gate driver IC to create the Vgs on the MOSFET. Otherwise you lose that voltage across the MOSFET. Choose a MOSFET with low-ish Vgs(th)
There is a way of using a gate driver IC to create the Vgs on the MOSFET. Otherwise you lose that voltage across the MOSFET. Choose a MOSFET with low-ish Vgs(th)