Need some help - really stumped with this AC Capacitor issue

[^zilla^]

Hi all,

I've been struggling with an issue for a while and have finally admitted that I am in over my head and need some help :)


here's the situation.

I have a string of variable capacitors; all in parallel. let's say for this example that there are 100

V  -------------------------------------------
     |   |   |   |   |   |   |   |   |   |   |   |   |   |
     =  =   =   =   =  =  =   =   =  =   =  =   =   =
     |   |   |   |   |   |   |   |   |   |   |   |   |   |
     ------------------------------------------
    1    2  3  .   .   .  .  . . .   . . . .               100

each cap is connected to it's own ground, but can be measured at any given time. The injection voltage is small - 5V AC with a frequency that can range from less than 100 kHz to over 300 KHz.

The phenomenon that I am observing is this:

At "high" frequency, if we set all of the caps to the same value C1 and measure all of them there is a gradient in the voltage drop from 1 to 100 (left to right)

At "high" frequency, setting all of the caps to C1 and measuring capacitor 100 will give a different reading compared to grounding capacitors 1 thru 99 and only having cap 100 set to C1

the same test at "low" frequency sees almost all of these effects disappear.

Obviously this has something to do with the frequency of the injected signal and coupling between capacitors. I've tried to model this using some circuit LT spice, but wasn't able to replicate what I'm seeing empirically.

Does anyone have any ideas or suggestions?

many thanks

Tim

 

[Carl Pugh]

The simulator adds some inductance or resistance.
This is the only way that there can be a difference of voltage on the capacitors.

 

[NascentOxygen]

The phenomenon that I am observing is this:

At "high" frequency, if we set all of the caps to the same value C1 and measure all of them there is a gradient in the voltage drop from 1 to 100 (left to right)

At "high" frequency, setting all of the caps to C1 and measuring capacitor 100 will give a different reading compared to grounding capacitors 1 thru 99 and only having cap 100 set to C1

the same test at "low" frequency sees almost all of these effects disappear.

It is unclear whether you are observing this with actual capacitors, or on a simulation. If real components, can you describe your wiring connecting the capacitors. Also, the Earth wiring.

Your "high" frequency being 300kHz, and "low" frequency 100kHz?

 

[sophiecentaur]

That circuit will have some series L or R between the capacitor nodes and so it will constitute a Transmission Line, fed at one end so you will have a standing wave on it. The Simulation has inserted some arbitrary values for the R and L but you could insert some of your own and see what happens.

 

[^zilla^]

hi all, thanks for the replies.


To clarify,

I observe the phenomenon with the actual capacitors, but the simulation cannot replicate it.

"something" happens in real life when going from the 'low' frequency of 100 kHz to the 'high' frequency of 300 kHz.

if it matters, the string of capacitors is about 2 meters long.

each capacitor is connected to Earth via it's own switch - this way i can turn them all on, all off, or only turn on individual capacitors to measure the voltage drop.


sophiecentaur: i was under the impression that since the length of strip is significantly less than the wavelength of the AC voltage that it wouldn't act as a transmission line?


Thanks

Tim

 

[NascentOxygen]

if it matters, the string of capacitors is about 2 meters long.

What value is each of your caps? You say they are variable capacitors, so are they tuning capacitors, with each being, say, 10-300pF?

 

[^zilla^]

less. ones to tens of pF

 

[NascentOxygen]

less. ones to tens of pF

In which case there is more capacitance in your wiring and switches than there is in the "capacitors" themselves. The equivalent circuit is nothing like what you think you have. Whatever you are using to measure the voltage is drawing as much current (if not more) from the circuit than is going through each capacitor!

Have you calculated the impedance of a 10pF capacitor at 100kHz? Compare this with the input impedance of your voltage measuring apparatus.