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 13 14 15 16 next last

Move Thread

LAN_403

Sulaiman

Tue Sept 07 2010, 05:44PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

I use a p6015 hv probe - very good.
e.g. ebay

Proud Mary

Tue Sept 07 2010, 06:29PM

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

Patrick wrote ...

I also figuered out how to make a high Quality Factor 50pF variable cap, with high reliablility.

Wouldn't an ordinary 50pF ceramic type with silver-plated brass vanes be good enough, Patrick? Front panel mount, or chassis mount trimmer type with screwdriver slot?

Don't forget the temperature compensating parallel capacitor/s to limit thermal drift if you use one of these - the silver-plated brass vanes do expand and contract so!

I've also got some very high quality 1nF 350V 5% silver mica capacitors with lo-Z flat strip leads for RF, and could spare a couple of them for your 2nF measuring cap if 5% is good enough - or I've got some oustanding 1.6nF 350V 1% silver micas in enamelled copper jackets if you could change your input cap to match.

Is 2nF a preferred value? I didn't think it was.

I've also got some 50pf 350V 2% silver micas which I could let you have.

It wouldn't be any bother to put it all in a jiffy bag and send it to America by airmail if any of it is any use to you. Let me know.

IntraWinding

Tue Sept 07 2010, 10:32PM

Registered Member #2261 Joined: Mon Aug 03 2009, 01:19AM
Location: London, UK
Posts: 581

Patrick wrote ...

Do you IntraWinding, know anything about bode plotters? poles and zeros ?
Analog signal anaylsis is not something im fully sure of yet.

I can remember doing Bode plots by hand as a student. It was just a plot of amplitude (or gain) versus frequency and a plot on the same graph of phase versus frequency. Or at least that would be the ideal. In practice we used rules of thumb to construct approximate graphs from straight lines. I am no expert on poles and zeros, so I can't help you, but we mainly used the Bode plots to see if an op-amp feedback scheme would be stable.

For your 'op-amp free' design hopefully the amplitude part of a bode plot would just be a flat graph showing the bandwidth of your probe, cutting off near DC and also at some high frequency. The phase part of the plot will indicate the phase errors that occur within a factor of 10 or so of the cut off frequencies.

A 'bode plotter' is an virtual instrument (since I don't think they were ever (generally) available in hardware) which plots the gain and phase of a circuit against frequency.

I'd expect the plot to look something like this (ignoring the scale on the Gain axis), but with the upper cut off at a higher frequency

Hope that's of some use.

And please correct me anyone if my memory's failed me

Patrick

Tue Sept 07 2010, 11:04PM

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

IntraWinding wrote ...

I'd expect the plot to look something like this (ignoring the scale on the Gain axis), but with the upper cut off at a higher frequency

that top graph is alomost exactly what i see. that bottom one is what i kept getting

Sulaiman wrote ...

I use a p6015 hv probe - very good.

i have used these before and others have recomended them, they are good. but i need to go past there max rating.

Proud Mary wrote ...

Wouldn't an ordinary 50pF ceramic type with silver-plated brass vanes be good enough, Patrick? Front panel mount, or chassis mount trimmer type with screwdriver slot?

trimmer is best with screwdriver slot.

Proud Mary wrote ...

Don't forget the temperature compensating parallel capacitor/s to limit thermal drift if you use one of these - the silver-plated brass vanes do expand and contract so!

eleaborate on this point please? do you mean a fixed in parallel with a variable?

Proud Mary wrote ...

I've also got some very high quality 1nF 350V 5% silver mica capacitors with lo-Z flat strip leads for RF, and could spare a couple of them for your 2nF measuring cap if 5% is good enough - or I've got some oustanding 1.6nF 350V 1% silver micas in enamelled copper jackets if you could change your input cap to match.

Oooooo!

Proud Mary wrote ...

Is 2nF a preferred value? I didn't think it was.

1.6 nF is fine the 2 pf was calculated. I will probably use 1.5nF caps,for my first calibration test, so 1.6 nF should just mean a simple adjustment.

Proud Mary wrote ...

I've also got some 50pf 350V 2% silver micas which I could let you have.
It wouldn't be any bother to put it all in a jiffy bag and send it to America by airmail if any of it is any use to you. Let me know.

Oh, your too good to me, set them aside but dont send them yet, I need to make some measurements with HV first before I take all your stuff.Thanks for the offer, if i get this too work i may send one finished CVD to you.

Mattski

Wed Sept 08 2010, 05:15AM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

IntraWinding wrote ...

Patrick wrote ...

Do you IntraWinding, know anything about bode plotters? poles and zeros ?
Analog signal anaylsis is not something im fully sure of yet.

[...]
A 'bode plotter' is an virtual instrument (since I don't think they were ever (generally) available in hardware) which plots the gain and phase of a circuit against frequency.
[...]

A network analyzer is basically a bode plotter that can measure forward and reverse gain, plus reflections. And various companies made/make spectrum analyzers which can control a swept signal generator to create a "gain only" bode plot. The common man's way of doing it is to drive your circuit with a signal generator, manually sweeping the frequency, and measure the input and output voltage with a multimeter/oscilloscope. If you use the oscilloscope, you can measure phase as well. But when you have a system like this voltage divider with so much loss, and which will be affected by parasitic capacitance, it can be challenging.

The way that bode plots work is your gain is a complex-valued ratio of polynomials such as (x^2+2x+1)/(x^2+5x+6), and any polynomial can be factored into roots by solving for the zeros. In my above example we have (x+1)(x+1)/(x+2)(x+3). The the "zeros" are the roots of the numerator (-1 and -1), and the "poles" are the roots of the denominator (-2 and -3). And it just so happens that you can pretty easily plot pretty accurate gain and phase of the complex valued polynomial versus x when you follow the Straight line bode plots rules. If you pass a root then the term with that root gets bigger, so going past a zero means the gain increases, going past a pole means it decreases. Now you may have noticed that I gave examples where x<0, where x is frequency. That's okay, negative frequency is basically the same as positive frequency, I don't really recall the details, but if you had a system with all a non-negative pole frequency then it would mean it is an unstable system.

Patrick

Wed Sept 08 2010, 05:42AM

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

ok Mattski, thanks for that it helps. ill look at some books at my csu library.

Proud Mary

Wed Sept 08 2010, 10:55AM

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

Patrick wrote ...

Proud Mary wrote ...

Wouldn't an ordinary 50pF ceramic type with silver-plated brass vanes be good enough, Patrick? Front panel mount, or chassis mount trimmer type with screwdriver slot?

trimmer is best with screwdriver slot.

Proud Mary wrote ...

Don't forget the temperature compensating parallel capacitor/s to limit thermal drift if you use one of these - the silver-plated brass vanes do expand and contract so!

eleaborate on this point please? do you mean a fixed in parallel with a variable?

Oooooo!

Proud Mary wrote ...

Is 2nF a preferred value? I didn't think it was.

1.6 nF is fine the 2 pf was calculated. I will probably use 1.5nF caps,for my first calibration test, so 1.6 nF should just mean a simple adjustment.

I have no shortage of them, so when you're ready, just send your delivery address in a PM and I'll send them on by airmail. Just give a few dollars to a Pakistan Floods Disaster Relief charity appeal when you get them.

The tempco C in parallel with VC is something essential in, for example only, stable RF oscillator design - but may not be so important in your application if a very critical capacitor adjustment is not necessary. If you can, do without it, because the Q of the Ag/ceramic trimmers is very high, and you wouldn't want to go bringing it down with lots of baggage fitted to it.

Patrick

Wed Sept 08 2010, 08:31PM

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

ok on the pakistan flood, will do. also i wont use temp comp caps until i know if there needed.

Patrick

Fri Sept 10 2010, 04:50AM

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

my new meter says 3.5pF and ive made changes to increase accuracy, so i think i belive this number. also this meter has half the uncertainty in error, or so it claims.

if anyone would be kind enough to volunteer a high accuracy calibration method id be grateful. (i.e. = -+1% @10kV )

Patrick

Sun Sept 12 2010, 03:27AM

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

The X and Y caps from ATX power supplies are in the 1, 2, 3 and 4 nF range. These or Proud Mary's silver caps may be needed, Pakastan flood donation done Proud Mary.

first previous 13 14 15 16 next last

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