Staggered 50uH and 250uH inductor design for LISN

[CopperCone]

But I just still put in in parallel with a 50ohm resistor?

[floobydust]

Sims show large inrush currents, if powered-up at Vpk line.
120VAC mains over 2.5Apk for very short time ~100nsec to charge the 0.1uF cap.

The transient-limiter 1206 resistors+ LPF cap can get walloped pretty hard.

The Tekbox protection scheme seems wrong, 5V TVS in parallel with GDT- which will never light up until the TVS pops...
GDT appears ineffective, needs over 100usec to ionize and too slow to help.

I'm not sure of options to ensure the input resistors don't get damaged. Spike average power is 0.5W and 120Wpk. ESD9L5.0 < 1pF

[Jay_Diddy_B]

Floobydust and group,

The Tekbox 5uH LISN is a low voltage LISN, typically used for 12V Vehicles, 28V avionics etc. So it unlikely that the gas discharge tube will ever fire.

The transient limiter in my line voltage LISN is styled after the HP11947A transient limiter. The manual can be found here:

https://www.keysight.com/upload/cmc_upload/All/11947-90006.pdf

The schematic is:




The (partial) BOM:



The HP 11947A is obsolete. Similar products are available:

http://www.leobodnar.com/shop/index.php?main_page=product_info&products_id=275

Leo is a member of the forum and famous for his fast pulse generator:

https://www.eevblog.com/forum/projects/yet-another-fast-edge-pulse-generator/msg1251589/#msg1251589

Regards,
Jay_Diddy_B

[floobydust]

Using a GDT gives impressive specs, who keyed up into this...
"Maximum Input Level: Continuous: 2.5 W (+34 dBm), Pulse: 10 kW for 10 µS, DC:±12 V"

For the mains inrush, the HP11947A transient-limiter through-hole 1/2W RN65 resistors can take some abuse.
I'd prefer something stronger than 1206 chip resistor like MELF MMB0207 1W, 225W impulse and good past ~400MHz.
Or maybe several clamp-diodes like LL4150 adds a pF. I think max. input 50mW is 2.24Vpk so two series strings of four maybe.
The A-band (10-150kHz) LISN has bigger 0.25uF caps that would have very large inrush.


For the 250uH inductor, I could not find much >10A off-the-shelf with ferrite cores. Abracom ATCA-08-251 10A DC toroid. Bourns 1140-270K-RC 270uH 12.4A DC.

Instead, looking at using mains line filters, power entry modules, they have a CM choke which is easy to find at 250uH and 15A.
But very little for DM attenuation.

[dazz1]

... I am a little reluctant to share the 50uH PCB because of liability....

I am not a lawyer but I struggle to see how you could be held liable for supplying a PCB.  A PCB is a component that can do no harm or damage. Someone else has to add a collection of components to it before they might be capable of damage or harm.

Has a PCB designer/supplier been sued somewhere?

[CopperCone]

Ok, I assembled the coils and capacitors on a PCB.

If anyone is curious, my shitty 50uH coils wound in 4 identical sections were an approximately 1.5% match at 100KHz. The spacing between the coil sections were about 1 inch. You can trim via mutual inductance, by moving them apart or seperate from each other. By squishing the live connected coil a bit (I wound the coils on wax paper over a PVC pipe, which I then bound with a line of super glue, then removed from the forming pipe to place on the actual inductor structure), I was able to tune it to about 0.5% match. Already alot better then my capacitors, which only had a 2% match, so I left it at that.

I suppose I can tune it up a bit better, it would be nice for common mode response I guess. I would need to get some low value high voltage capacitors to fix that spec up...

I think someone said the spec for a LISN is about 10% match between the sections though?

[CopperCone]

What value of power would you use for the resistors that break up the circuit in parallel with the coil?

If you use 270V, and the value is 460 ohm each, you get the following:

Simulated with a 80mOhm resistance in the 250uH coil, and 80mOhm resistance in the secondary coil (higher because of all the solder joints despite being much shorter), and 5 ohms of series resistance with the 7.5uF capacitor. The parameters of the primary capacitor (2uF) can be ignored I think.

0.01Ohm Load - Short Circuit - 60mA, 470 Ohm 4x - so approx 2k parallel.
Peak Power - 1.6kA for the load, which is approximatly 7 watts, across 4 resistors, so 1.8 watts per resistor, meaning you should use 3watt resistors. This means something shorted out real good. Possible even?

3 ohm load - 80Amps flowing through coil, 3.4mA going through the resistors, 18mW per resistor, so 1/4 watts is more then acceptable

7.5 ohm load - 35 amps flowing through coil.. 1/4 watt parallel resistor more then acceptable. The 2.65mm wire inside can handle it, if hooked into a 40 amp 2 phase outlet. Would need a special connector on it though.


Now, I assume this beast should be fused.... so slow blow fuses? I don't know what the dynamic resistance of these guys is.

As you can see the theoretical spread is pretty large. What would be a good practical value of wattage to use? What is the expected peak current from a reasonable outlet? How should a resistor be rated (I don't really see transient power specs on resistors).

Do I need to go with 3Watt resistors to account for the possibility of the load shoring out? Or is this completely insane? How do you derate average resistance for a pulse that lasts in the microseconds? I assume its pretty nonlinear with respect to the gauge of wire used to manufacture the resistor.

What kinda average current that the resistor cares about will a shorted outlet put out before the circuit breaker/fuse kicks in? I think the resistors can be damaged by hot spots.. but the time scales of these overloads ellude me, and the overload duration of the resistors cannot be easily found.

Getting a kiloamp from a 20amp home outlet seems kinda unlikely, no matter how short the pulse is, but then again its a spark gap so . 

My hunch says I should just put like 0.75 watts of dissipation per resistor to tolerate stuff.   

[dazz1]

The inductor is Wurth 744 332 0680. Here is a picture of the packaging, from Digikey:




Jay_Diddy_B

I have just read my way through this thread.  I like all this modeled and measured data.  My old Professor used to say it ain't true if it ain't measured.

I note from the inductor data sheet that although the inductors are rated to 13A, the inductance starts to fall away at about 8A.   I don't see this as a problem at all.  I would just rate the LISN at a max current that is within the defined performance curves defined in the standard.    It is relevant to me because I intend to test equipment rated to 10A. 

[Jay_Diddy_B]

Hi,
You can choose inductors for the current rating that you desire.

We can consider the impact of inductance changing with current.

This is the curve for the Wurth 744 332 1000 Inductor:



It is 10uH with zero current and 7uH with 10A.

We can then build two models, and two additional models to create the CISPR 16-1-2 +/-20% limits:



If you look at the results the LISN meets the CISPR 16-1-2 specifications, with 10A, except at a very small area between 150 and 180kHz where there is a minor discrepancy.

Remember that I am plotting LISN impedance, in use you are measuring transmission from the DUT port to the Output with a low impedance source. The LISN impedance will have very little effect on the accuracy of the EMC measurements.



You have to decide the current rating of the LISN you want and choose inductors accordingly.

May aim was to make a small portable LISN rated at 5A which 600W at 120 VAC and 1.2kW with 240 VAC.

Regards,

Jay_Diddy_B

[T3sl4co1l]

Remember that rectifier-input devices draw more peak current.  Peak to RMS ratio may be 4 or more!

This is especially troublesome for the LISN, because those peaks are also where you get the strongest signal: the diodes are conducting, carrying EMI from the EUT to the LISN.

Result is, you could be missing some dB's at the low end, missing the switching fundamental by as much.

Tim

[CopperCone]

Has anyone had a look at the ground isolating inducotor in the tekbox schematic?

Its specified at 1.8 milihenry. Atb60hz its reactance is 600 ohm.


Should multiple inductors be used in a cascade here? You are pretty much forced to use ferrite.

Should this guy be a torroid? I am getting very confused by the field interactions with this ground inductor.

Do you need the heavy wire gauge here? Lets say your load has hv dc output earth ground referenced. If there is a short you might blow up a thin gauge ground isolation inductor right, then have a malfunctioning chassis at hgih potential.

Also, since the reactance is 600ohm at 60hz, a hot ground would be potentially still pretty dangerous unless you have a gfic. You cant really rely on it to blow a fuse.

I am worried that the gfic is pretty noisey and its kinda an off requirement to use one if its not built into the equipment. Especially dangerous since the gear being tested is experimental and is more likely to have wiring problems.

Ideas? I thought maybe a mov or low voltage gdt across the 1.8mh inductor that could pass enough energy to blow the input fuse? Mov is a nonlinear junction and may mess with mixing emissions.

What if like high voltage high frequency energy gets on your 1.8mh isolated gnd? Can a mov or gdt even shunt that reliably? If its the waveform output of dcdc converter before filtering, so say a 300v or more pwm signal thats passing through a live potential heatsink at frequency between 1khz to 20MHz or more.. Idk where the limits of switchin frequency are when it comes to high voltage dcdc converters or where it goes from heart attack hazard to slightly lessdangerous flesh cooking hazard

In a lab setting at home you might be adjusting the equipment tosee output waveform at different operating conditions could be hazardous

[floobydust]

I find that ground inductor uncommon and would not implement it. There are other means to look at CM noise only. Some reference to it being a holdover from old German VDE 0871 standard. It should be a 16A inductor.

ETS-Lindgren:
EARTH LINE CHOKE SWITCH
The safety ground isolation choke selector switch switches the 1.6mH earth line choke IN and OUT of the safety ground circuit. The ground choke is designed and manufactured with sufficient capacity to conduct the maximum current rating of the Model 3816/2 and at no time is the safety ground of the unit compromised. The earth line choke avoids a double RF ground connection (safety ground and measurement ground) in the conducted emissions test setup."

This is another ref. to it, I think:
"Whenever the EUT dimensions are such that the protective earth conductor is long enough to show a significant impedance, or be close to ¼ of a possible wavelength, or the enclosure has poor conductivity, the test will be performed using the non-fused, built-in artificial protective earth."

[CopperCone]

But what about its high reactance and the fact that it will compromise safety? Will a 90V or less GDT in parallel with it work?

I want it to blow the mains fuse if there is high voltage on any thing with a metal enclosure due to a short that can fry you in short order.

I know it wont help much with a current limited short to chassis, but it would still make me feel alot better. MAYBE a partial short could be covered by the ground from a spectrum analyzer... but a BNC cable is no saftey ground for 16A!!!!!! (more like 40 for what I built).

I am worried that the measurement ground, may not be connected at the time (i.e. powered up before the SA/EMI reciever is connected.

And I don't really want a high current going through my spectrum analyzers ground. Or you might wanna hook up a battery powered instrument to it, like a hand held SA or hand held MM (for true rms) or hand held oscope, so you don't even have the option for a shitty saftey ground through a BNC/SMA cable (horror to ground 20A with a thin rg-whatever spagetti cable). It could explode and give you thermal burns!

I don't feel comfortable with that level of saftey and I would like to find some kind of better solution. I don't even know how to setup a GFIC to measure current difference between the oscilloscope/SA ground either, accounting for the split ground impedance through the 1.8mH inductor. It sounds like this is not really a existing solution.

[T3sl4co1l]

But what about its high reactance and the fact that it will compromise safety? Will a 90V or less GDT in parallel with it work?

I want it to blow the mains fuse if there is high voltage on any thing with a metal enclosure due to a short that can fry you in short order.

You might want to recalculate the reactance of that part.

Tim

[CopperCone]

rofl how did i fuck up an order of magnitude there

But you still can get some kind of a short occurring at high frequency. The calculation I made is for 60KHz.

Still think it needs some protection.

For clarity/sanity I am interested in IGBT based induction heaters that run off mains. Kinda like this guy:
http://www.instructables.com/id/30-kVA-Induction-Heater/

Especially since I am interested in things like water cooling. I know getting hot RF on the chassis is kinda unlikely but it bothers me.

[CopperCone]

does anyone have this :
https://en.wikipedia.org/wiki/Electrical_injury#/media/File:IEC_TS_60479-1_electric_shock_graph.svg  with the frequency axis on Z?

[T3sl4co1l]

Eugh, that instructable is a mess, no protection, no feedback, one slip and you're out a whole set of transistors.

Tim

[CopperCone]

I don't plan to build that, I wanted phase shift control of power and optical driver and maybe some kind of circuit to track resonance as it works but thats just an example

that circuit is bootleg in more ways then 1

and it has no interlocks on the chassis or anything normal for something so high powered, leak detector, etc.

[floobydust]

The mains LISN leakage currents due to the 8+uF caps are a huge danger. Just mentioning for the noobs. Lifting (RF) PE on the EUT I wouldn't do, lovely tank circuit with Y-caps.

Warning labels on LISNs are also strongly recommended for:
    Their lethally high earth-leakage current
    The need to maintain two independent protective earth connections at all times
    Their use only by authorized and trained personnel

Taken from In Compliance Magazine, Guide to Testing Conducted Emissions

[CopperCone]

Yea but its clearly designed in.

I wish I had a higher power RF source to test with MOVs and GDT to see. My best RF source is 50Vpp

Hmm maybe a step up transformer.

[CopperCone]

Now that I thought about it, I swear I saw a teardown video with a MOV put across an isolation transformer for a HV power supply or something like that, so the unit would ground itself if the chassis voltage became high. I can't remember where though.

May have been a capacitor that looks like a MOV, can't find any record of either though.

[dazz1]

Remember that rectifier-input devices draw more peak current.  Peak to RMS ratio may be 4 or more!

This is especially troublesome for the LISN, because those peaks are also where you get the strongest signal: the diodes are conducting, carrying EMI from the EUT to the LISN.

Result is, you could be missing some dB's at the low end, missing the switching fundamental by as much. 

Tim

I don't think this is necessarily a deal breaker.   It should be possible to apply corrections to the measured values.  This may require recording the current.

As Jay said earlier:

I haven't tested one of the single-layer solenoids normally found in commercial on LISNs on a VNA. I don't expect it will perform all that well.
If I was winding a single layer solenoid, I would space the winding from the coil former and I would ensure that there was some space between the turns to lower the turn-turn capacitance.

Regards,

Jay_Diddy_B

So I think the next step would be to measure the performance of a real air-cored inductor to see what effect all of the actual parasitics have on performance.  The hypothesis to test being that ferrite with current induced inductance roll off will provide better performance that a large air-core with parasitics.  I'd do it if I could but I don't have the test equipment.

[fcb]

The one thing I haven't seen so far on this thread, and in my view (& experience) something that is essential - a discharge mechanism on the power input.

I built a LISN a few years back for my own lab, when I finished using it I switched it off at the wall socket and pulled the plug.  Whilst handling it to put it away I touched the plug pins and got a real belt off it from the internal capacitors (obviously switched it off at peak cycle).  Since then I fitted a mains relay inside that disconnects the L&N input and shorts the internal L&N to earth via a pair of 7W 4K7 resistors.

[CopperCone]

The tekbox schematic has 30k going to ground on both L and N

[floobydust]

With a mains LISN, the priority is RF signal integrity, not safety. There is no fuse, no on/off switch, no power on indicator lamp, and huge internal (10uF) capacitance.

IEC 61010 requirement is that after disconnecting power, plug pins shall not be hazardous live after 5 seconds.
Hazardous live is defined as 70VDC dry, 35VDC wet conditions. This sets the highest bleeder resistor value.

240VAC (339Vpk) down to 35VDC in 5 seconds, 10uF I calculate 220k ohms, and 0.26W steady state.
30k ohms is 0.7 seconds, and 1.9W steady state at 240VAC. A 5W wirewound resistor like venerable Yageo SQP rated 700V I would use. Lindgren is 39k ohm.


An on/off light would be great, but neon generates some RFI and an LED would radiate switching hash from the rectifier diode(s), inside the LISN enclosure.
If you could have a very RF quiet mains LED arrangement, I would definitely incorporate.