Magnetic coupling and current transducer

Very nice, and McMaster has them in stock. Thank you.

i

I have some experience in high current connections and measurement, and I generally agree with most of the advice presented here. We use DC shunts for AC measurement, and they work fairly well at 60 Hz. However, there is some effect of the routing of the conductors into the shunt. This may not apply to DC, but it will affect any AC components of the current. I have seen errors of 3-5% from routing the conductors into the ends vs the sides of the shunt blades. This was for shunts of 800 A 100 mV, and 2500 A 50 mV, and currents up to 15,000 amperes AC.

Although most of the shunt resistance is in the temperature-stable Manganin fins, the rest of the shunt is copper or brass alloy which has a tempco of

0.6%/DegC or more, and may contribute about 20% of total resistance. If one shunt in a parallel connection runs hotter than the other, this could affect accuracy.

As far as bolted connections of bus bars, some things that help are: (1) use high grade or stainless steel hardware that can be torqued for high pressure, (2) apply a coating of grease or "No-Ox" to the mating surfaces after cleaning them, (3) make sure the surfaces are smooth, (4) use thick hardened wahers to spread the bolting force over a larger area, and (5) possibly consider silver plating the mating surfaces to reduce resistance. The Belleville washers are good, but I don't think they will provide enough pressure to make much difference on 1/4" bus bar.

I am wondering why you need to use something like an LEM current transducer, which I think is a Hall effect device, which needs an accurate excitation voltage, and requires zero adjustment. Usually these CTs are used to reduce burden loading which a resistive shunt causes. The millivolt signal from just the single 300 ampere shunt should give you all the current waveform information you need. A good instrumentation amplifier can give you relatively noise-free readings down to about 0.1 mV. I use a 1000 amp, 100 mV shunt in my Ortmaster to measure currents down to 50 amperes full scale, and the noise level is less than 1 ampere. It will also measure short duration pulses up to 6000 amperes. I am replacing my old hardware with new boards that use a PIC for the A/D and a serial port to Windows software that reads the current and can display (and analyze) the waveform, at 2500 samples per second. The old boards connect to a parallel port, and the software must use MSDOS to work properly. I could probably send you an old (working) board for the cost of shipping, and you can download the software from my website.

The main problem with a shunt is the direct connection to the measurement circuitry. In my application, the recloser test set generates up to 500 VAC, but the shunt is solidly grounded, and the common mode voltage is generally less than one or two volts. I have about a hundred units working in the field, and only occasionally have I seen units with the front ends burned up. This probably happens if the ground was improperly installed.

Let me know if I can help. When will we see the photos of your 2" arcs?

Paul

I see. Were the errors dependent on current?

Here, I hope that since both are 50 mV, both should change similarly with increased current.

Thank you, looks like there is an art to making very tight connections. I may just do it somehow first, and then will come back and redo it again.

it is.

You see, my SCR firing system kind of requires current measurement to be on the negative side, same for voltage. Measurements should be close to ground. On the other hand, for a welding application, it is a really bad idea to ground the negative side (both should float).

I did just that -- ground the negative side -- for the first tries. I used a current shunt for feedback into the SCR firing board. It worked and I could lay welding beads. But I know that it is not right and I should decouple negative side from ground. That's what I am trying to accomplish.

That's very interesting. My issue with such amplifiers that I looked at is that they do not provide enough isolation. AD629 provides isolation to

270 volts, whereis in my welder (which will double as plasma cutter, see below) the bus voltage could be as high as 300 volts.

So, I am using a proper voltage transducer and current transducer to provide isolation.

Regarding welding and plasma cutting: my transformer can be wired for

66VAC (parallel delta) or 220VAC (series wye). I can explain how, it involves having two secondaries on each of the legs of 3 phase.

I am doing this rewiring using contactors. (it is already working). If I turn on three contactors, I get 66VAC. If I turn them off and turn on another two contactors, I get 220VAC 3 phase.

Yes, that's right, but in my case, I want to disconnect negative from ground.

I am just a little afraid to use my camera with these arcs, I have a junk helmet, but I am not sure if it is safe for the camera. I will try though, as my welder is still working with the old version of the setup (direct input from shunts and negative grounded).

i

Lem sensors are in most cases good for 3X the rated current. Well...Sorta... That is usually a peak requirement before the amplifier saturates. I do not think they will handle 3X continuous. If you got a few of the 50amp parts then you can put them in parallel and sum the output voltages.

You can make a shunt yourself some kind of material that isn't so dependant on temperature. Like copper is near 0.4%/c. I think maybe they make em out of brass, but please don't quote me there. If you need isolation use an Agilent serial optocoupler solution - I think It's Avago now - They're getting to be as bad as Mot with the name changes.

Another potential solution for you would be to use an Allegro Microsystems part. In some cases I will mount a hall effect sensor next to some current carrying tracks on a pc board and no need for a flux concentrator like a gapped toroid or tape wound. I met with allegro last week and they have some very interesting parts they call LEM killers.

I can poke around my toybox or intercept a used one on the way to the scrap pile. Your cost would be shipping. Unfortunatly a very good friend of mine just recently left Lem, but I'm sure I can hook you up if time isn't a concern. I am interested in your welder cause I play around with welding too. I am strictly buzz box now, but I have a giant diode bridge and some inductors kicking around. I design large inverters for a 9-5 so maybe one day I will build a slick welder like yours.

Sorry if I repeated the thread at all. I read it with only one eye open.

regards, Bob N9NEO

Ignoramus19197 wrote:

When I said "look in my toybox or intercept..." I meant a 300amp LEM sensor with a large window.

regards, Bob N9NEO

Yzordderrex wrote:

Thanks Bob.

I could not find any high enough rated current transformers at acceptable prices, hence my interest in current shunts. Summing outputs is kind of a pain, since they are referenced to ground (middle point between + and - supply leads, I suppose).

If you have a CT that is capable enough -- if that is what you referred to when you mentioned your scrap pile -- I would be most interested.

Thanks!

igor

That would be just awesome. Thanks Bob!

igor

I don't think you want a current transformer if you are looking at DC. No?

regards, Bob N9NEO

Ignoramus15879 wrote:

LEM makes "current transducers", they use Hall effect (whatever it is) and work for both DC and AC. I definitely would not have much use for a regular "current transformer".

i

[snip]

I have three LEM LF 50-P's (±70A). I'm pretty sure they use *both* a CT *and* a Hall effect device.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.

That's interesting. Thanks.

By the way, mostly to Bob, I found 300 amp current transducers this morning, I bought them from CTR Surplus for not too much (really). These guys are very good. They know that moving stuff fast is more important than getting top prices.

i

[snip]

The error seemed to be a percentage of the current value, so the effect was linear. The waveform changed in one case because the load (an old electromagnetic circuit breeaker) was a saturable inductance, but the true RMS readings of the test set and the shunt matched within about 0.5% after a scaling factor was applied.

[snip]

It's a good idea to check voltage drop at the connection, and also check for hot spots, which are a sure indicator of excessive resistance. A good bus bar connection is something like 20 uOhms, or 6 mV at 300 amps.

I suppose the reason for that is because the item you are welding may be (or should be) connected to ground, and there could be a lot of current flow through the ground rather than the return lead. It is probably still a good idea to put a resistor from the return lead to ground, so that leakage currents from the transformer will not create a shock hazard. Perhaps even use a two-pole circuit breaker where one side will trip on excess ground current, and the other side shuts down your main supply. Sort of a GFCI.

Actually the *instrumentation* amplifier I use is an AD620, which is not an

*isolation* amplifier. It uses a differential input that can handle common mode voltage of several volts. I'm sure there are isolation amlifiers with much higher voltage ratings, but they are probably expensive. Your LEM CT is a good choice.

Paul

I have also dealt with CTR Surplus for large IGBTs, SCRs, rectifiers, capacitors, and other high power devices. They are very good and I have them in my Favorite Sellers list.

Paul

That's not bad (that it is linear).

This is all beside the point now, as I was able to find a 300 amp current transducer for $6 (bought two just in case).

correct.

Hm, I do not understand what is the danger with leakage currents, I am sure that I am missing something.

Yes, I think, why bother with circuits when I can just get a proper CT for a few bucks, with all gotchas properly resolved by LEM.

As I said, I bought a 300 amp CT today, you can see a similar one here:

I will, therefore, not have ANY issues whatsoever related to paralleling current shunts, in fact, I will not need a current shunt for current feedback. (note that I have another current shunt attached to the ammeter gauge, that was on this CyberTIG originally, that I will keep, but it is not part of any logic).

i

Yep, they are great and I sometimes check their auctions in search for cheap stuff that is not properly described. They are very good.

i

Sweeet! Buying used for homebrew is the best deal if your boss isn't cooperating by handing over the goods.

Closed Loop LEM sensors use a coil wound over the gapped toroid. The coil is used to counteract the amp-turns and drive the flux to zero. So the Hall Device and the focusing core are actually operated near zero flux. The current in the coil is viewed across a resistor for the output. They are a little more expensive than the open loop, but better performance.

There is also a hall effect and coil in the Tektronix current probes, but not sure if the same principal. I think Edwin Hall was the guy who first discovered the Hall Effect, and that was way back in late 1800s IIRC.

regards, Bob

Ignoramus15879 wrote:

te:

a