Capacitive Voltage divider, for O-scopes (10,000:1) / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Capacitive Voltage divider, for O-scopes (10,000:1)

first previous 3 4 5 6 next last

Move Thread

LAN_403

Proud Mary

Thu Jul 01 2010, 02:55PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Patrick wrote ...

Proud Mary, as i will post the math later, the result of thermal expansion amounts to no significant change of dimension, plastic, metal, and ceramic. like +0.008 to -0.002 inches at the gap, and the gap will be 1.000 inches, so the capacitence which varies as 1/x > 0 of the distance will see very little capacitence change from even very much rod movement, and 0.008 inches is not very much. therefore the capacitence change by mechanical means will not be a significant factor though it will be identified and noted.

Oh well, if it's worth a footnote....

Carl Pugh

Thu Jul 01 2010, 04:16PM

Registered Member #1064 Joined: Tue Oct 16 2007, 05:04PM
Location:
Posts: 42

There was a question on why the amplifier was clipping.
The clipping may have been caused by U1 + input floating.
An oscilloscope is not a reliable way to allow OA input bias current to flow.

Proud Mary

Thu Jul 01 2010, 04:40PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I forgot to say that there will be a number of frequencies when the Ls and Cs in your probe start singing backwards and forwards in a ringing duet.

Haefely Test AG address this problem in a free download here:

A New Type of Voltage Divider for the Measurement of High Impulse and A.C. voltages
by K. Feser, Haefely Test AG

klugesmith

Thu Jul 01 2010, 04:42PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Good eye, Carl!
Any decent circuit simulator would refuse to analyze a node with no DC path to ground. So the C1-C2 node must get a DC level defined by the op-amp model's input bias, leakage, or clamp currents. Try adding a 100K bias resistor to ground.
Also: can't tell the absolute voltages in your clipped "scope" trace. Are you sure the OA output is not getting too close to negative power supply rail?

You can buy ready-made instrumentation amps with all the matched ratio resistors built in. Example: LT1101, which is just a bit too slow for you (down 5 dB at 50 kHz); I bet you can find faster types. But you don't need an instrumentation amp; what would be the problem with conventional amp with, say, 10K input R? Then low frequency rolloff point would be 2 kHz, and error at 50 kHz would be much less than 1%.

Patrick

Thu Jul 01 2010, 07:48PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

i have less than 220$ left and my professors have been laid off, and the remaining ones have had a 10% pay cut. so it looks like im back to making stuff myself. also for next semester i will get even less from the cowards in sacremento. CA will be bankrupt in about 9 minutes anyway, idiots!

OOps! <embaressed> carl pugh and klugsmith are right, let me put another path for the output. also the negative supply rail does not cause the clipping i already thought of that and that wasnt it.

from C1,C2 node through 100kohms to ground, and the dam thing works, with no loss of bandwidth! clipping solved.
dam, i just got schooled on that one.

Proud Mary, footnotes in science and engineering are like bronze and silver medals, ill be happy if i get one.

klugesmith

Thu Jul 01 2010, 08:18PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Patrick wrote ...
... for next semester i will get even less from the cowards in sacremento. CA will be bankrupt in about 9 minutes anyway, idiots!

Should the CA legislature demonstrate bravery by more aggressively reducing services, increasing taxes, or borrowing money?

Patrick

Thu Jul 01 2010, 08:26PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

yeah klugesmith they have too make cuts now , and i will be on the short end of that stick, had they not spent money like wild eye maniacs over the past 30 years, cuts in services now wouldnt be needed. but now the cuts are needed, and no bailouts for CA, this is theft from other states who arnt so bone headed.

Klugesmith,i was aware of the LT1101 family IA, but they aree slow, also im not making a differential measurement so i dont really need an all up instrument amplifer with 3 OA's, one OA should be enuff, also as we all know two basic things,

first, no amplifier exists which is perfect. noise will be added.
second, therefore one amp in series with another must be even worse.

Antonio

Fri Jul 02 2010, 01:44AM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

You canÂ´t connect a capacitive divider directly to an input of an operational amplifier. Without DC bias any small input current will charge the capacitive node and saturate the amplifier. Look at what is the simulated DC voltage at the input of the amplifier. The simulation may be not precise in this situation, depending on how is the model used for the amplifier.

Patrick

Fri Jul 02 2010, 02:59AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

abc

above: top trace is red (final output) and slew rate returns!, blue HV to be measured, green is divided voltage from c1,c2 and green trace is input towards the op-amp.

This is no longer the planned schematic.

above schematic as it is scheduled to be built. i included R4 50 ohms, as the Sam Goldwasser article shows, is this to simulate the impedence of the BNC RG6 line? or to do impedence mathcing to the scope?

below: 100kvac, 100khz, sine wave, same circuit as above.

Antonio

Fri Jul 02 2010, 03:32AM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

The 50 ohms resistor doesn't have any effect in this position. You can connect the 10 kohm resistor to node 9 instead of node 4, leaving the 50 ohms resistor in series with the output to have a precisely controlled 50 ohms output impedance in your device. Note that you are attenuating the signal and amplifying it again. You can use this to obtain lower cutoff frequency, by increasing C2 while keeping R6, at expense of some noise and lower upper cutoff frequency.

first previous 3 4 5 6 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob