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 12 13 14 15 next last

Move Thread

LAN_403

Patrick

Mon Aug 30 2010, 10:04PM

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

This PDF has to do with probe design considerations ...SUPER IMPORTANT !!!
]the_secret_world_of_probes_oct09.pdf[/file]

This is no longer the current schematic.

Patrick

Thu Sept 02 2010, 06:12AM

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

OK final construction has begun, despite an unexplained simulation "abberation" or what ever you want to call it. I need help with this anaylsys. Look below. The blue line is the ideal 100kv+ to -40kV wave to be measured. the red is the measurement at the node 5 into the oscope.

This is not the most current schematic.

In the above screen grab you can see between the #2 yellow vertical cursor, and the #1 vertical teal cursor, a "step"; its approximatly 480mV. I cant explain why its here or its magnitude.

near the #2 yellow cursor, you can see a step down again about 480mV. then near the #1 teal cursor a step up about 480mV, i see what seems to be a offset appearing then disappearing, but i dont see a cause.

In between the yellow and teal cursor the red wave is consistantly lower then it should be by about 480mV, all the way up too and past the teal cursor. But most importantly there is a step to the right of the yellow cursor, I dont know what this means.

Myke

Thu Sept 02 2010, 06:35AM

Registered Member #540 Joined: Mon Feb 19 2007, 07:49PM
Location: MIT
Posts: 969

A few quick questions since I didn't see it mentioned:
The coax doesn't behave like a single inductor and capacitor since the inductance and capacitance is distributed along the line (making it a transmission line). Do you need to include the inductance and capacitance as long as the ends are properly terminated and such?
I'm not exactly sure why ground loops are bad but I heard they are. To remedy this, you would just wrap one of the probe leads around a ferrite toroid right (high impedance to the signal and a low resistance to DC)?

Patrick

Thu Sept 02 2010, 07:06AM

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

Myke wrote ...

A few quick questions since I didn't see it mentioned:
The coax doesn't behave like a single inductor and capacitor since the inductance and capacitance is distributed along the line (making it a transmission line). Do you need to include the inductance and capacitance as long as the ends are properly terminated and such?

I dont think i need to treat RG-59 line at 20" long as a transmission line when the bandwidth / wavelength im after is several thousand times longer then 20". When you approach 1/10 the wavelength then you must begin to do so. And at 1/4 to 1/2 you must do so.

I gotta get some sleep. ZzZZZzZZZzzzzZz

Mattski

Thu Sept 02 2010, 08:34PM

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

Not sure what to make of the step, but as a general troubleshooting idea, have you run a frequency sweep to get a bode plot of the frequency response of the probe/scope system?

You want gain to be flat across the desired bandwidth, and since you're testing with a square(-ish) waveform, that bandwidth will be higher than the frequency of your square wave.

And if I am thinking correctly the phase shift doesn't need to be flat but it does need to be linear to have a constant group delay across the entire bandwidth.

Patrick

Thu Sept 02 2010, 08:55PM

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

Mattski wrote ...

Not sure what to make of the step, but as a general troubleshooting idea, have you run a frequency sweep to get a bode plot of the frequency response of the probe/scope system?

Your right, i should do that now...i was a little tipsy last night, what with the boose and all.

EDIT: bode plot added:

Im not sure if I connected it right? Am I bode plottting the signal source or the CVD output (in the pic above)? I cant remember how to use this instrument, i think its been about 8 years. If i have it backwards ill be so embarassed.

in the above pic from far left towards the right:
-between 1 mHz (-156dB) to 28Hz (-80.3dB) is that 35 degree rise.
-then from 28Hz (-80.3dB) too 12Mhz (-81.7dB) its almost level but trending down towards neg dB.
-and after 12Mhz it rolls down (the curvy part) to 780Mhz (-107dB)

PS. Im not supposed to take signal analisys for another 3 semesters, so I dont have a complete understanding of all this, however lack of knowledge has never stopped me before.

Mattski wrote ...

You want gain to be flat across the desired bandwidth, and since you're testing with a square(-ish) waveform, that bandwidth will be higher than the frequency of your square wave.

yes

Mattski wrote ...

And if I am thinking correctly the phase shift doesn't need to be flat but it does need to be linear to have a constant group delay across the entire bandwidth.

yes, its almost flat but curves down in two slow steps, well rounded, which i think is OK. no sharp rises, falls, or cusps.

...

Fri Sept 03 2010, 02:18AM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

The bode plot shows a 'zero' at ~20Hz, and a 'pole' at ~20MHz. The slope on either side should be pretty dang close to 20dB/decade, although its possible you have either 2 closely spaced poles at the 10+MHz end, its hard to make out from the bode plot. There isn't any reason for the step to show up because of that, your simple pole's/zero's should give only exponential terms when excited with a square wave.

Perhaps try adding extra probes in to see where the step is coming from? I personally would expect it to be a simulation error of some sort.

Also, its personal preference, but I would recommend adding a large value resistor across the 2pf input cap, to simulate the leakage currents. This can also help the transient simulation from freaking out.

Patrick

Fri Sept 03 2010, 02:50AM

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

... wrote ...

Perhaps try adding extra probes in to see where the step is coming from? I personally would expect it to be a simulation error of some sort.

its not a simulation artifact, its real, its a feedback-harmonic type problem. when i racked my brain i thought of what i might have overlooked, sure enough, I should have made the RG-59's L and C more lossy with a 500 ohm series resistor...Presto! jitters and steps are totally gone!

... wrote ...

Also, its personal preference, but I would recommend adding a large value resistor across the 2pf input cap, to simulate the leakage currents. This can also help the transient simulation from freaking out.

Thats a great point I will do that. I will calculate the volumetric resistivity of the volume of oil around the ball-gap.

UPDATE: 12.7 Teraohms across the 2pF oil cap, and 10 Gigaohms across the 4pF film cap are calculated.

This is the most current schematic.
I will attempt to build this device in the next few days.

IntraWinding

Tue Sept 07 2010, 04:41PM

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

Here's a quick video of some experience building a 5KV X1000 probe
that might help inform your design.

How's the project going, by the way?

Patrick

Tue Sept 07 2010, 05:06PM

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

IntraWinding wrote ...

Here's a quick video of some experience building a 5KV X1000 probe
that might help inform your design.

yes, ive seen this before; it was useful.

IntraWinding wrote ...

How's the project going, by the way?

It was ontrack to be done in the next few days, however, a mishap killed my capacitence meter, so it will be a few more days before I get another. also, since ridding myself of the "op-amp idea" things seem much better, i got my sixth star too!

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

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

first previous 12 13 14 15 next last

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