Can anyone tell me what this RC network does

[gerrysweeney]

Hi,

I am working on my own PSU design and I have been looking at other implementations to compare and/or draw inspiration from.  One of the HP E364xA PSU's has a fairly classic differential amp input stage which is nothing special.  The gain setting resistors are bypassed (the four 39pF caps) to improve the response time which I understand.  What I don't understand is what the two resistor/caps highlighted in red boxes are for. I am guessing something to do with voltage or overload protection but I am only really guessing.  As the current path is through the 39pF cap, I am having difficulty understanding what role the 0.1uF cap has, it would clearly be a resistor at DC and a dead short at AC at any reasonable frequency. It would make sense that the resistor would act to smooth out any large transient voltage but I would have thought that the 0.1uF cap would counter that which is why my voltage protection theory seems a bit flawed.

As there are lots of smart people on this forum I thought I would ask to see if anyone can give me a "layman's" answer , a simple high level description would be great. 



Thanks,

Gerry

[Paul Price]

The circuit is obviously to stabilize the power supply under violent load changes, such as when the output is short circuited in or out of constant current/constant voltage operation.

What is intriguing here is the very small capacitor the .1uF is in series with. This means, that the circuit will only provide some small compensation to accomplish undershoot/overshoot recovery correction when the output sees a load dump after being  some while in constant current mode.

[gerrysweeney]

@Paul, thats what is confusing me, because of the small cap, it would appear that the 0.1uF and the 70K resistor might well as not be there.  I ma sure it must do something but it does not make much sense to me.

Gerry

[gerrysweeney]

@madires, thanks for the response. In isolation I understand that but does the 39pF cap negate that because if a 0.1uF has a high impedance at say 30Hz (just a notional value) then the impedance of the 39pF can is going to be really high, making the 0.1uF cap pointless?

Gerry

[madires]

@madires, thanks for the response. In isolation I understand that but does the 39pF cap negate that because if a 0.1uF has a high impedance at say 30Hz (just a notional value) then the impedance of the 39pF can is going to be really high, making the 0.1uF cap pointless?

I've deleted my post since it was redundant :-) My guess is that the RC network should decrease the total capacitance (in series with the 39p cap) at higher frequencies. At low frequencies (below the point at which the impedance of the 0.1µ cap starts decreasing) the resistor dominates. So it seems that the time constant of the RC network plus the 39p cap is changed over frequency. A high time constant for low frequencies and a low one for higher frequencies.

[Rufus]

It doesn't do anything the until parasitics of the components make a difference at ridiculously high frequencies.

The resistors R661 and R667 could be open or short without making a significant difference.

I expect they are filled in place holders for components where a much lower value capacitor and resistor could make a difference.

[gerrysweeney]

@Rufus, thats pretty much the conclusion I came to, and it would not be the first HP schematic I have seen with completely different values in the schematic than there is in the actual unit. 

Gerry

[skrap]

In the service manual for the E3633A/E3634A the same RC network is marked with "*Only for E3634A" with the values 33pF and 33k.

[gerrysweeney]

@skrap, thanks for the response, now those values would make much more sense, would you be willing to give me a copy of that manual, the one I have does not indicate that. Thanks

[skrap]

@gerrysweeney

http://ridl.cfd.rit.edu/products/manuals/Agilent/power%20supplies/CD1/service/E3634ser.pdf

Schematic - page 128

Relevant partslist:
Page 114: R109
Page 118: C301, C302, R301

I really like your power supply project btw, keep up the good work! :-)

[gerrysweeney]

Hi @skrap, thanks for the link, that really helps, it now makes sense. 

Thank you for the PSU project feedback, I am still working on the PSU, I have not given up  . I am have way through the 14th episode, am getting some good results in the basic response measurements, not quite as good as my HP benchmark but very good none the less. Its taking me longer now because I am at levels of detail that I am not so proficient or experienced so I am having to learn a lot to make progress.. will hopefully push the next article out in a few days time.

Gerry

[amyk]

It's a filter, you may try running a simulation of its response to get a better idea.

[Paul Price]

Rufus has got it right.

If it is a filter, it will only work to filter out mysterious forces invisible to us mortals.