measure ripple on a voltage divider / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

measure ripple on a voltage divider

1 2 next

Move Thread

LAN_403

Andre

Thu Jun 20 2013, 04:53PM

Registered Member #6921 Joined: Wed Sept 26 2012, 07:47PM
Location:
Posts: 109

I'm trying to measure the ripple on a 100KV voltage multiplier, according to this formula it should be about 96V

after building the divider what size cap should I put on the end of the divider, to get a true ripple reading? so I can read the ripple with out reading the noise,

Proud Mary

Thu Jun 20 2013, 09:03PM

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

So far as I can see, you haven't shown us enough of the terms of calculation for anyone to comment on a Vripple 96V or whatever.

Let's set that on one side for a moment, and imagine a much bigger ripple, say 1 kV ripple, superimposed on your 100 kV DC. I don't think Â± 1% would be very easy to measure with much confidence in the hundreds kilovolt regime, nor is it easy to see what practical applications would require such accuracy even were it possible at an affordable cost in parts and labour.

First determine what the application requires, what ripple it will tolerate, and so on, and then design a circuit that will fulfill those needs.

Andre

Thu Jun 20 2013, 10:56PM

Registered Member #6921 Joined: Wed Sept 26 2012, 07:47PM
Location:
Posts: 109

hahhaha Proud mary you always switching the topic, I was not talking about the multiplier, I'm talking about the Voltage divider to measure ripple, is ok thanks for trying anyways, you can google to find out what a voltage divider looks like, also my voltage divider is 1:10000

Mattski

Fri Jun 21 2013, 02:12AM

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

Where are you planning to put this capacitor? You can't just hang a capacitor at the end of the multiplier since it will reduce noise there but also reduce the ripple. If noise is a problem in your measurement then you will want to design it so that it forms a low-pass filter with a corner frequency well above the ripple frequency if you want to filter noise without affecting ripple. Some scopes also have reduced bandwidth modes for this purpose.

Since you're trying to measure a 10mV ripple I imagine that it should be fine on most scopes without any filtering. Since you're trying to measure such a small ripple you'd want to AC couple the signal.

No need to be condescending to Proud Mary, her point is valid which is that your application probably doesn't require 0.1% ripple so you don't need to expend too much effort making that measurement. Instead just expend enough effort to make sure that the ripple is below whatever spec you actually need.

GrantX

Fri Jun 21 2013, 05:13AM

Registered Member #4074 Joined: Mon Aug 29 2011, 06:58AM
Location: Australia
Posts: 335

Andre wrote ...

hahhaha Proud mary you always switching the topic, I was not talking about the multiplier, I'm talking about the Voltage divider to measure ripple, is ok thanks for trying anyways, you can google to find out what a voltage divider looks like, also my voltage divider is 1:10000

I think Proud Mary is talking about the divider. You're asking for a 1:10000 probe to measure a 96V ripple of a 100kVDC multiplier. That means you are trying to build a probe to measure +/- 0.096% variation at 100kV. I think that's asking for an accuracy beyond the abilities of an affordable commercial HV probe. I'm not sure how accurate those cheap CRT TV repair probes are, but I bet they only have an accuracy around 1% to 5% - 10x to 50x times worse than your final goal.

Sulaiman

Fri Jun 21 2013, 12:15PM

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

I would use the resistive probe to measure the dc voltage
and a separate capacitive probe to look at the ripple,
all that's required is a capacitive divider of lower division ratio.
e.g. 100:1 could be 1pF (high voltage) to 99pF ('scope + cable capacitance + fixed capacitance to make 99.00 pF)
the low value capacitor could be purchased, or constructed
e.g. two 100mm diameter disks separated by 100mm is about 0.7 pF

Andre

Fri Jun 21 2013, 03:08PM

Registered Member #6921 Joined: Wed Sept 26 2012, 07:47PM
Location:
Posts: 109

very interesting Sulaiman, about the 100:1 is a 1pf cap.
I was putting the cap at the end of the voltage divider, are you saying it should be on the high voltage side?
I been searching on google but I have not found information of what size cap should I use,
with no cap is a lot of noise, with cap I reduce the ripple.
I would think the caps in the voltage multiplier would reduce the ripple with more stages.

I'm taking some data cause I'm building the power supply for a laser (I do NOT need advise on lasers)

Sulaiman

Fri Jun 21 2013, 05:56PM

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

Yes, the 1pF ( or other low value, high voltage ) capacitor
is from the C-W multiplier to the 'scope,
hence the 100mm dia discs with 100 mm separation for eht

I am surprised that you are 'scoping "noise";
although the voltage thermal noise may seem high, it is via a very high resistance, so it's the current noise that you may be seeing.
but at such high voltages the noise should be negligible.

As mentioned above, a low-pass filter can reduce noise,
a capacitor across the 'scope input forming a first-order filter at
frequency = 1/(2.pi.R.C)

R= resistance looking into probe from 'scope
in parallel with the'scope 1 MOhm
C= capacitance of probe from 'scope end, including cable
+ 'scope input capacitance (e.g. 15pF)
+ any added capacitance

to 'see' the ripple the cut-off frequency should be at least 7x ripple frequency (7x is debatable!)
but being a first-order filter the noise voltage will be reduced by half for every octave above the filter cut-off frequency.(-20dB/decade)

FIRST
get a piece of wire about the length of your probe/divider
clip it in your normal 'scope probe (x10 if avail.), remove earth clip
(better than waving around everywhere !)
see what is picked up ... may be surprising.

Proud Mary

Fri Jun 21 2013, 07:43PM

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

Isn't 1 pF really too small as the upper arm of the divider? If it is too small, then parasitic capacitances and variable proximity effects will make up a siginificant part of the total value of C1. Small, not easily accountable, changes in C1 will result in a much larger change in divider capacitance. In a nominally 1pF:100pF divider, an increase from 1 PF to, say, 1.25 pF will cause the divider ration to fall from 100:1 to 80:1

If C1 and C2 are now set at 10:1000, then the effect of having 0.25 pF drift on C1 is reduced by an order of magnitude.

Sulaiman

Fri Jun 21 2013, 08:16PM

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

You could have a 1ft. diameter disc etc.
but the stray capacitance problem will not go away
hence the 'scope probe connected to a length of wire.

I guess if I actually wanted such a probe (I've already got a P6015)
I'd use many series surface mount R//C (e.g. 100x 1MOhm//100pF )
but shielding would still be required, for 100kV a few inches diameter
and a long surface-tracking didtance support at the 'hot' end or very stiff central rod/pcb etc.........

Direct capacitive pickup to a normal 'scope probe seems to give a fairly reliable waveshape pickup, just no absolute reference.
Good for smpsu, rf circuits, flybacks etc. as a 'working' test.

1 2 next

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