Post by: flyglas on September 14, 2020, 08:35:10 PM
does somebody have access to a (perhaps broken one) rogowski current probe for use with an oscilloscope?
I cannot find any pictures from the inside of such a probe (integrator part of the probe) anywhere in the internet.
I would be very happy if someone can open a probe and take some pictures.
Post by: Hydron on September 14, 2020, 09:02:19 PM
It's one thing winding a skinny coil, another to do it such that it doesn't pick up all sorts of garbage from E-fields etc yet still is flexible and high-bandwidth (the obvious ideas like a shield tend to work against these).
Post by: johnf on September 14, 2020, 09:03:43 PM
Post by: davekni on September 14, 2020, 09:31:47 PM
Also tried a similar test on my DRSSTC H-Bridge output. This time I added a non-shorted turn of copper foil around the coil for shielding. Worked fine, but I didn't keep it. My CT already has a scope tap, and digital integration isn't convenient for real-time viewing.
Post by: johnf on September 14, 2020, 09:47:08 PM
I built a tiny current transformer to go over the source lead of a TO220 fet (actually a 5 pin PowerIintegrations SMPS device) to actually see what was going on
I had already tried a 0.01 ohm resister to scope over but introducing the resistor caused other effects and masked my problem.
I used the core out of a ISDN modem that had common mode chokes on the data ports i pulled the core out and wound with 38awg wire 200 turns on it and high voltage varnished it up along with a 1.0 ohm burden resistor. Net effect was this causes 0.000025 ohms of impedance on the source lead which in my case was low enough to go back to Power Integrations for them to ponder over. I ended up making a second transformer and sending it to them for their use in their lab
Post by: johnf on September 14, 2020, 10:55:54 PM
[ Invalid Attachment ]
Post by: johnf on September 14, 2020, 10:57:36 PM
[ Invalid Attachment ]
Post by: johnf on September 14, 2020, 10:59:25 PM
https://www.digikey.com/product-detail/en/pulse-electronics-network/PE-65855/PE-65855-ND/3533394
Post by: flyglas on September 15, 2020, 06:41:58 PM
https://patentimages.storage.googleapis.com/85/26/79/fd153583c08c2a/US6614218.pdf
https://patentimages.storage.googleapis.com/b4/12/0a/f6a3234c6cc7e7/EP1073908B1.pdf
Post by: Hydron on September 16, 2020, 09:38:14 PM
Sometimes a very cheap simple solution works fine, especially if there's a way to calibrate. I once needed to probe output current of a large H-Bridge. Wrapped 10 turns around the end of a felt-tip marker and wedged that between the two wires of the output cable. Coil ends connected to a scope. Captured waveforms and integrated with software. Used a known 800A waveform for calibration. It was then very useful to capture high-frequency details of the current waveforms that weren't visible previously.Hmm, will need to give the similar simple setup I tried and wasn't happy with another go given I have a better scope these days.
Also tried a similar test on my DRSSTC H-Bridge output. This time I added a non-shorted turn of copper foil around the coil for shielding. Worked fine, but I didn't keep it. My CT already has a scope tap, and digital integration isn't convenient for real-time viewing.
As for small CTs, I spied these guys in a power supply teardown once:
[ Invalid Attachment ]
(from https://www.tomshardware.com/uk/reviews/corsair-ax1600i-psu,5406-3.html)
Can also be seen here: http://www.jonnyguru.com/wp-reviews/CorsairAX1600i/DSC_2384.jpg
I never had any luck finding a way to order these unfortunately - they look super handy! Any lead on them would be most welcomed.
Post by: flyglas on September 17, 2020, 07:43:00 PM
Yesterday i bought a rogowski coil PEM CWT from ebay. Perhaps the unit will satisfy my thirst for knowledge.
Post by: Hydron on September 17, 2020, 10:26:57 PM
Post by: flyglas on September 18, 2020, 02:49:05 PM
I paid 174 EUR including shipping from Italy to Germany. That is a fairly good price.
--> PEM CWT 6R
Post by: Hydron on September 19, 2020, 11:34:01 PM
I did see this internal picture, may be of interest as you wait for your unit:
https://gmw.com/wp-content/uploads/2019/03/CWT_Fuse_Replace_Instructions_GMW.pdf
Is not showing the cable entry from the coil (goes in the hole at the top), but there isn't a whole lot going on in there! Would be good to see and compare to the inside of the one you have coming.
I am definitely interested in how the coil is made, but unfortunately this is difficult to do non-destructively unless you have an x-ray machine handy.
From the info I can gather from PEM publications etc, some of their probes look to be shielded, some not, so there may be quite a range of performance in the presence of high dV/dt.
Post by: flyglas on September 25, 2020, 09:41:10 PM
Half of the tiny PCB is a power-supply and the other half is the integrator and output driver.
Nothing much in there.
My probe arrived during the week, but the probe is in a not working condition.
As long as the situation is not resolved with the seller, i will not dismantle the probe.
Post by: klugesmith on September 26, 2020, 04:46:12 AM
I have taken some steps toward winding a flexible Rogowski coil for the purpose outlined in "current transformer for narrow spaces" thread. Commercial ones for AC power metering come close, but all seem to have small end diameters bigger than 7 mm (0.28 inches). If the current to be measured is a sinusoid of known frequency, we can skip the voltage integration. I expect to get about 10 mV out, from 200 amperes at 60 Hz.
What dimensions are easy to see on little commercial probes like the PEM?
For any Rogowski coil, the sensitivity (in units of inductance) is simply u0 A N/L, where A is the small loop area, and N/L is the winding density in turns per unit length. Commercial ones with different size big loops, made from the same tightly wound flexible rod material, all have the same sensitivity.
Post by: flyglas on September 26, 2020, 10:57:58 AM
What dimensions are easy to see on little commercial probes like the PEM?
The smalles probe is the PEM CWT Ultra Mini with only 1.7 mm in diameter.
I have worked with this tiny probe a few years ago. In my application I have mounted the probe between a minibrick and a PCB on top of the minibrick.
Post by: Hydron on September 26, 2020, 05:27:37 PM
Did you notice much influence from nearby dV/dt transients when using the CWT Ultra Mini? That one isn't shielded (it looks like the only shielded probe they do is the CWT miniHF) which concerned me about it's use near fast switching transistor legs, but if interference from nearby voltages isn't such an issue then DIY versions become much more practical.
Post by: flyglas on September 27, 2020, 11:01:13 PM
I did not noticed the influence of nearby dV/dt transients with the CWT Ultra Mini.
We also performed double-pulse tests with unshielded probes.
Post by: Hydron on September 27, 2020, 11:48:35 PM
Do you know whether the coil is damaged, or if the fault is in the integrator box?
Post by: flyglas on September 28, 2020, 12:30:58 AM
If you wiggle at the output cable of the coil the output waveform sometimes becomes good. Therefore the integrator works fine.
So there have to be a faulty connection somewhere between the coil and the transition to the cable.
The next step is to open up the heat shrink at the coil and to dig deeper into the underlying plastic case (T-piece).
Post by: flyglas on September 29, 2020, 07:45:22 PM
I think the photos speak for themselves.
Post by: klugesmith on September 29, 2020, 08:08:22 PM
Can we see what the free end of coil looks like? I guess it's presently clamped in place at the middle of T shape, perpendicular to the output cable and fixed end of flexible coil.
Is there a schematic that shows what those resistors are doing inside the T ?
Post by: Hydron on September 29, 2020, 08:13:43 PM
I am also interested in any light you can shed on the coil construction and termination - it's not completely clear how the termination is done. Are the two termination resistors soldered together to one end of the sensing coil (with the other end being the red wire)? Or are they each going to a separate end (in which case, what does the red wire connect to)? Is the input to the integrator a normal piece of co-ax or is there something else going on?
Sorry for all the questions, the more I see about these the more it looks like a relatively simple DIY job to replicate!
Post by: flyglas on September 29, 2020, 09:12:52 PM
The coil itself seems to be constructed around the inner part of a coax cable (plastic tube with center conductor).
At the free end the red wire is connected to the center conductor. After many turns the red wire reaches the t-piece.
The termination resistor in the t-piece is connected between the center conductor and the red wire.
The red wire and one end of the termination resistor is connected to the outer conductor of the coax to the integrator.
The second resistor is a series resistor and connected between the center conductor of the coil and the center conductor of the coax to the integrator.
As cable between coil and integrator a triax cable is used. The outer conductors (copper braids) are conntected inside the integrator but not in the coil.
In the coil the outer most conductor is unconnected.
Post by: Hydron on September 29, 2020, 09:50:02 PM
http://hotstreamer.deanostoybox.com/TeslaCoils/Misc/RogowskiCoils-816epe99.pdf
and that the rest of the circuit is similar to the non-inverting integrator they present there.
I suspect the major integration job is done by IC2, with IC10 as an output driver, though without knowing what's on the bottom side of the board it's impossible to tell exactly how it's laid out.
Some more relevant info:
https://gmw.com/wp-content/uploads/2019/02/ias_2000_pem-1.pdf
(I have a big stash of papers, patents etc that I've saved whenever I've gotten interested in these, just never found the time to do much more than some quick and dirty testing though)
Post by: Weston on September 30, 2020, 05:24:48 AM
Post by: klugesmith on September 30, 2020, 08:39:53 AM
Let's talk in a few days after I have rolled my own flexible Rogowski coil, and materially supported your Bee probe.
Many simple products for niche applications, well designed and produced, command lots of money.
It helps if vendor is on Approved Vendor lists. Selling accessories for six-figure instruments is like options on six-figure automobiles.
For hands-free, solder-free probe positioning at home:
Post by: klugesmith on October 02, 2020, 06:08:07 AM
Here is picture of a baby's first step. Expected to grow into an open-able Rogowski coil for measuring currents at 60 Hz, without much concern for instrument risetime and waveform fidelity.
Is the commerical coil by PEM wound directly on exposed dielectric of a coaxial cable? For my project that seems like needless trouble, and I have no coax with a suitable dielectric OD. Finished coil must fit through a 7 mm gap. So I'm using the jacket from a multi-conductor cable, slipped over a straight steel mandrel. Later it will slip over a coaxial cable dielectric. The relatively thin, soft sleeve might help to prevent bunching of the small turns, on inside of the big loop.
This practice winding has 27 turns of 34 AWG magnet wire, made by twirling the mandrel by hand.
How do you all like to strip insulation from fine magnet wire?
Post by: flyglas on October 02, 2020, 04:38:51 PM
I do not exactly know the material of th dielectric inside the coil.
On some of my pictures you can see white material with approximately 4.5 mm diameter.
I like to strip fine magnet wire with a soldering iron using much solder melting the insulation layer away.
Post by: Hydron on October 02, 2020, 05:06:36 PM
"The Rogowski coil is manufactured as one continuous wire which is wound around a plastic tube, then the conductor returns down the tube centre."
If doing the same, I suspect it will be important to have a close fit of the wire in the tube to ensure it is as centred as possible.
Flyglas - did you get any pictures of the rear of the integrator PCB? Am trying to compare the implemented circuit vs what is written in papers/patents.
Post by: flyglas on October 02, 2020, 08:59:48 PM
Post by: davekni on October 03, 2020, 04:57:35 AM
I had an email from PEM stating the following regarding one of their much older probes:
"The Rogowski coil is manufactured as one continuous wire which is wound around a plastic tube, then the conductor returns down the tube centre."
If doing the same, I suspect it will be important to have a close fit of the wire in the tube to ensure it is as centred as possible.
I'm wondering if it really matters about centering of the return wire. If the bandwidth is high enough to need constant impedance of the "coax" of the return wire, then I can see. But that high a bandwidth seems unlikely. Otherwise, any field picked up by wandering of the center wire position would be tiny compared to all the turns of the outer coil.
For simple construction, how about some soft rubber high-voltage wire? That has thick insulation with a center conductor. Wind it with magnet wire and solder one end to the center HV wire. My guess is that the biggest error will come from any non-uniformity in winding pitch along the wire.
Yes, as long as the magnet wire insulation isn't the high-temperature (typically crosslinked polyimide enamel of some sort) type, soldering is the best. Red and green magnet wire is fine. The gold/copper colored enamel is usually this high-temperature version. For fine wire as 34AWG, I use careful sanding with fine-grit paper. Other options include chemical stripping and welding instead of soldering.
Post by: plasma on October 03, 2020, 06:54:25 AM
Post by: Hydron on October 03, 2020, 11:20:37 AM
Post by: klugesmith on October 03, 2020, 06:22:34 PM
Its purpose is so the Rogowski coil has zero instead of one net turn encircling flux that's perpendicular to the big loop. That cancels the largest stray field sensitivity.
Plasma, you don't need to choose between left and right hand twists. Can put both in the same space, joined to each other at one place and brought out on the other side of the "break". Never seen that on a flexible RC, but I made two rigid RC's that way.
Both have 60 turns and calculated sensitivity of around 37 nH. One is wound on a thick flat nylon washer. The printed circuit version went in waste area of a 6 mm thick board I designed at work about 10 years ago. I have about 10 of them somewhere; does any pulse power experimenter here want one to play with?
I like the kynar idea, to save work in a proof of concept coil. But my 60 Hz sensitivity estimate (order of 10 mV from 200 A) is for small loop diameter of about 6 mm (constrained by gap between current conductors) and winding pitch of 6 turns per mm (150 per inch). As with close wound 34 AWG magnet wire, more than twice as dense as 30 AWG wire-wrap wire. Maybe it's premature to contrive a geared lead screw to feed the wire at constant pitch, like I guess is done for some Tesla coil secondaries. In this case we want the turns a bit looser than touching each other, to allow for bending without bunching.
Mod edit: removed partial duplicate post above this one.
kluge edit: Sorry 'bout that. Pictures seemed to disappear so I put 'em back.
Here's link to old 4hv post from which I scraped those thumbnails. It has more pictures, theory, and measurements. https://4hv.org/e107_plugins/forum/forum_viewtopic.php?p=1&id=82880#post-82880
Post by: plasma on October 03, 2020, 06:27:26 PM
I'm thinking it will be non linear, but you guys explained its only one turn, so wouldn't be that large.
Post by: Hydron on October 03, 2020, 06:59:57 PM
https://www.ti.com/tool/TIDA-01063
Post by: Weston on October 03, 2020, 08:46:41 PM
10 mV/ 200 A is probably more than enough sensitivity unless you were planning on feeding it directly into the scope input.
Post by: plasma on October 03, 2020, 10:03:25 PM
Post by: klugesmith on October 03, 2020, 10:20:47 PM
What's the noise level at microphone input to a PC audio "card"? Digitized voltage signal could be low pass filtered, averaged, and (lest we forget) integrated offline.
Just remembered the commercial model 916 fluxmeter box. It does voltage integration, is plenty fast for 60 Hz, and has offset voltage low enough for "DC" measurement. Invites Rogowski coil tests on DC buses whose current can be switched on and off.
https://highvoltageforum.net/index.php?topic=819.msg5555#msg5555
This chat has served as reminder that fancy winding (34 AWG magnet wire, wound beautifully, maybe with mechanically controlled pitch) is probably unnecessary for stated objective, DIY flexible coil for a single application.
Would be a briefly engaging squirrel-chase.
Post by: flyglas on October 04, 2020, 04:21:24 PM
Post by: klugesmith on October 04, 2020, 06:16:24 PM
Picture shows 50 turns in 6.1 mm, for winding density of 8200 turns/m (208 turns/inch). Estimated coil diameter is 1.15 mm outside and 0.91 mm inside. The larger red rectangle is 1.7 mm high, matching the Ultra Mini jacket OD.
Using coil mid-diameter, computed sensitivity is 8.55 nH.
Anybody seen that parameter given in literature from PEM, or other commercial RC maker?
My test of closely-wound AWG34 magnet wire got 5900 turns/m (150 turns/inch). Can probably get 2100 turns/m with plain old wire-wrap wire. In a 5.5 mm coil that would give us 63 nH. On 200 A at 60 Hz, 4.7 mV.
Post by: flyglas on October 04, 2020, 07:49:17 PM
I also have done the math (with your values) and calculated the same mutual induction as you.
The mutual induction value is plausible.
With a integrator time constant of 1.7 µs (R = 1700 Ohm and C = 1.0 nF) we have a sensitivity of around 5 mV/A.
Post by: flyglas on November 06, 2020, 08:37:32 PM
A few days ago I received PCBs for a rogowski coil integrator.
Additionaly I wound a small coil onto the insulator PTFE of an RG178 coaxial cable.
The insulator has an diameter of 0.86mm and the used wire is 0.1mm. The windings are wound without space between the windings.
As a protective layer a heat shrink is on top of the wire. The total diameter of the coil is only 1,75 mm.
At the moment I only tuned the low frequency range (active integrator with opamp).
A test at 50 Hz and around 6 A peak shows the same respones as my professional coil.
Both coils are designed to have 5 mV/A output signal.
The next step is to tune the high frequency range (passive integrator).
For this task I designed a pulse generator (16 µs pulse with 150 A peak and flat top, rise and fall time around 800ns).
I found the calibration procedure/waveform in an official calibration document.
Post by: klugesmith on November 08, 2020, 03:25:03 AM
It shows that 6 amps at 50 Hz gives a useful signal, even with the tiny area of minor loops in the coils.
Have you looked at the sensitivity to test currents routed outside the R.C. loop, or threading the loop backwards?
Will we get to see any pictures of your helical winding operation?
What is the low frequency corner of the active integrator circuit?
Post by: flyglas on November 08, 2020, 02:56:58 PM
Have you looked at the sensitivity to test currents routed outside the R.C. loop, or threading the loop backwards?I have build a current pulse generator (150A, ~750ns rise and fall-time, 20µs pulse width) for adjustments to the high frequency passive integrator.
With this puls generator and two turns of wire, I have tested the influnece of nearby current pulses.
Depending of the position the influnece is well noticable (see pictures influence_diy_1...3).
With the commerical unit the influence was also noticable, but almost only at the t-piece (see picture influence_prof_1).
The resumee is: Keep away with currents from the t-piece.
Will we get to see any pictures of your helical winding operation?This was no big deal. I removed the inner conductor of the PFTE insulator and replaced it with the 0.1 mm enameled wire.
At the open end I applied a tiny ammount of epoxy glue. After curing the insulator was twisted between to fingers to wind the wire around the insulator.
Everything is held together with the heat shrink.
What is the low frequency corner of the active integrator circuit?I did not tested it yet. Simulation shows good results down to 1 Hz.