Anyone using this incredible EMI debugging tool?

[udhay_cit]

I brought this incredible EMI debugging tool before quarantine for my company R&D lab usage. This is extremely useful for EMI troubleshooting and to localize the weak area.



I suspect that this is working under the principle of Jim Williams, Avalanche Pulse generator circuit. Unfortunately I can't cut open the enclosure to see the internal circuit, because this is belongs to our company.

I can see one IC chip (10 or 14 pin) and few inductors through the battery insertion back hole. Can't able to see the IC part number.

We had lot of struggles to troubleshoot the signal line surge & EFT problems earlier, but now this tiny marker shape tool helps to find the weak areas with in few second (yes in seconds, not even minutes).

Earlier we used microchip microcontroller for our products, but we changed to Cypress and ST ARM controller for better pricing and we totally eliminated all the use of microchip microcontrollers.
Do you know, after started to use this tool I immediately found that the Microchip is the only best performing controller (among the three brand I have mentioned) for harsh environment. I realized most of my EMI problem is only due to my ST and Cypress microcontroller and no problem with my protection circuits.  I verified by just replacing the microcontroller alone from ST to microchip and the entire project clears the surge test for 4KV. Earlier it fails at 1.75KV with ST ARM.
(FYI, we have 100% inhouse test facility for ESD, surge, EFT, voltage dip & drops, pulse magnetic field, CE & CS, vibration, thermal shock, etc..., So I can clearly verify and compare the test result of this EMI debugging tool with accredited Lab test equipment. Trust me this is one of the best tool I have ever used and it cost me 1,500 EUR)

So I'm designing an avalanche pulse generator for to clear my doubt. I can accept any suggestions or advises from this forum.

This tool has polarity chancing and single or burst mode output option(I hope that the single mode is to simulate the surge and burst mode for EFT) which is not available at avalanche pulse generator. But I'm always using the burst and positive polarity option for troubleshooting.

Udhay

[TheUnnamedNewbie]

This seems like a silly thing. No shit, if you introduce a kV high-bandwidth pulse a few tens of micrometers away from a high impedance line you get strange things happening. It seems to me that unless you have some really, really specific crazy application where that might happen, this is not a very good real-world analog.

[udhay_cit]

Here voltage is not a matter. Actually I can measure the induced voltage by the probe and it doesn't cross the 15V, but he rising edge is less than 1nS.

How you can say that this is only for specific application. I'm not here to advertise their product and trying to make one for my personal usage. Others also get benefit from this.

I have this tool with me and it is helping for my EMI troubleshooting in "real world".

You may aware of ESD gun, and as you said you are injecting 15KV of high bandwidth pulses to "few tens of micrometers away from a high impedance line" or even contact with the open metal parts. So this is not a zombie thing.

[T3sl4co1l]

So you're injecting that noise just anywhere into the circuit?  How often do you have cables bringing in EFT and whatnot, that are just anywhere?

That's the point, it's not representative using a super up-close near field test when the real offenders in practice are cables placed at some distance.

Or if you have outside cables running directly over your clock oscillators and whatnot, well, I guess that's your problem...

Tim

[JohnG]

Where I used to work (one of the last large corporate R&D labs), there was an EMI expert in our group who saved this company many, many millions of dollars. If you think I am exaggerating, how much do you think it costs to have a brand new major kitchen appliance factory delay production because you cannot pass certain tests related to EMI susceptibility? It can easily cost well over 1M USD every day of delay.

He used a similar idea, but substantially more sophisticated, to locate susceptible circuit points, sometimes down to particular pins on an IC. On a big IC like a microcontroller, it is not always the pin you might guess.

So, I have seen this type of test in action, and seen is used to solve big problems quite a number of times. I have also seen it used to find brands of controllers, or sometimes particular lines of controllers within a brand, that had high susceptibility, much in line with the OP. While I left this company some years ago, I have seen this method used to successfully troubleshoot problems, and often a little C or RC at the right pin is enough if the factory is waiting. The trick is finding that pin.

$1500 is pretty dear, though, but I bet they do not sell many.

Just my (highly devalued) $0.02,
John

[Weston]

Oh wow, those are obscenely expensive. Circuity is probably a pretty basic pulse generator, I bet the plastic case is the most expensive part of the BOM.

Its a purely qualitative tool, so its not like you are getting anything from a brand name / reputable source. This seems like it would make a good project in one of those electronics project magazines / something I would buy off Tindie if it was sub $50 or so.

[thm_w]

Where I used to work (one of the last large corporate R&D labs), there was an EMI expert in our group who saved this company many, many millions of dollars. If you think I am exaggerating, how much do you think it costs to have a brand new major kitchen appliance factory delay production because you cannot pass certain tests related to EMI susceptibility? It can easily cost well over 1M USD every day of delay.

He used a similar idea, but substantially more sophisticated, to locate susceptible circuit points, sometimes down to particular pins on an IC. On a big IC like a microcontroller, it is not always the pin you might guess.

So, I have seen this type of test in action, and seen is used to solve big problems quite a number of times. I have also seen it used to find brands of controllers, or sometimes particular lines of controllers within a brand, that had high susceptibility, much in line with the OP. While I left this company some years ago, I have seen this method used to successfully troubleshoot problems, and often a little C or RC at the right pin is enough if the factory is waiting. The trick is finding that pin.

But the first step would be injecting whatever pulse (ESD gun) into the exterior of the circuit no? To determine if the circuit is actually susceptible from the outside or not.

As mentioned above, if you have some "weak" spot but its an internal node that is far from any external input/buttons/wires I can't see how it would matter.

edit: or maybe we are talking about radiated EMI as well, which would be harder to test in house.

[Elasia]

Fancy... i usually just pull out a bag of change and the used igniter out of my former gas grills.. been doing it that way for a long time.. neat trick one of my profs showed us well over a decade ago

Here is an article about it

https://www.edn.com/an-emc-troubleshooting-kit-part-2-esd-immunity/

[Neilm]

I have used equipment that looks similar. It does have a selection of antenna to allow broad areas to be affected, or concentrate the noise for affecting specific traces. I have found it a bit limited but invaluable in certain applications.

[udhay_cit]

First, this tool is not for radiated emission or susceptibility test. For RE & RS you have some other tools like near field probe. Long time back I had a chance to use an excellent radiated measurement tool from "emxpert" which very expensive and not available anymore. I've given the YouTube video link below.



Even though the EFT and surge signals are injected from outside cable or connector, finally the victim is internal circuit components. Most of the time the problem arises due to poor PCB layout and lack of shielding which creates a weak point. You can use various equipment's to identify EMC immunity of your product, but you can't pin point the weak point inside the PCB. I can consider this tool to identify the internal circuit weak point.

1500 euros is pretty expensive, but this the fate whatever instrument or tools in the market in the name of "EMI/EMC" will cost very expensive. This is Ok for a corporate R&D lab, but this is unimaginable for hobby or individuals.

Still i"m not sure about what is inside, but I assume that the circuit may similar to avalanche pulse generator.

[Karel]

https://www.amazon.com/Milty-5036694022153-Zerostat-Anti-Static-Blue/dp/B0033SHDSS/

[nctnico]

I'm using a modified gas ignitor for checking ESD. Cost me a few euro. It is handy to get a feel on the ESD sensitivity of a product. But be sure to keep all other equipment away. My old HP 6002A PSU went bezerk when I started zapping a circuit near it.

Maybe having a lower voltage version would be nice. It shouldn't be hard to put something together. You can get nice sharp pulses from letting a bipolar transistor avalanche.

[2N3055]

I'm using a modified gas ignitor for checking ESD. Cost me a few euro. It is handy to get a feel on the ESD sensitivity of a product. But be sure to keep all other equipment away. My old HP 6002A PSU went bezerk when I started zapping a circuit near it.

Same here. It is amazing how much havoc that thing can create...

Maybe having a lower voltage version would be nice. It shouldn't be hard to put something together. You can get nice sharp pulses from letting a bipolar transistor avalanche.

Agree.

[JohnG]

But the first step would be injecting whatever pulse (ESD gun) into the exterior of the circuit no? To determine if the circuit is actually susceptible from the outside or not.

As mentioned above, if you have some "weak" spot but its an internal node that is far from any external input/buttons/wires I can't see how it would matter.

edit: or maybe we are talking about radiated EMI as well, which would be harder to test in house.

Not susceptible, or at least, passes the test - no problem, ship it!

Otherwise, anything radiated can become conducted, or visa versa. It's a question of amplitude and size of the circuit structures relative to wavelength, which affect coupling. Even very inefficient coupling can be a problem, since interference sources can be very large. For example, a spark usually involves many kV, and currents can be very high (>100 A) and impulsive. Depending on the source, it can ring for some time after the spark is over. The pin to a micro is typically high impedance with a low threhold voltage several orders of magnitude smaller (i.e. 60 to 80 dB smaller or more).

Don't forget that the IC is not just a lump of semiconductor, either. You can have conductors on the IC that can form a loop with PCB traces outside and make a nice antenna, etc.

Cheers,
John

[Elasia]

I'm using a modified gas ignitor for checking ESD. Cost me a few euro. It is handy to get a feel on the ESD sensitivity of a product. But be sure to keep all other equipment away. My old HP 6002A PSU went bezerk when I started zapping a circuit near it.

Same here. It is amazing how much havoc that thing can create...

Maybe having a lower voltage version would be nice. It shouldn't be hard to put something together. You can get nice sharp pulses from letting a bipolar transistor avalanche.

Agree.

lol the igniters are my favorite toy... so much carnage for a couple of bucks

[JohnG]

I'm using a modified gas ignitor for checking ESD. Cost me a few euro. It is handy to get a feel on the ESD sensitivity of a product. But be sure to keep all other equipment away. My old HP 6002A PSU went bezerk when I started zapping a circuit near it.

Same here. It is amazing how much havoc that thing can create...

Now imagine if you had one of these ignitors connected to spark gaps via long wire connections, in a multi-compartment metal cavity filled with wires, slots, and holes that couple the cavities together. Throw a microcontroller in there, which controls the flow of highly flammable gas. It's called a gas range.

Interestingly, it could be close to another appliance with a microcontroller that minimizes bacterial growth in your food. The power cables for both might even run right next to each other for many meters. Better make sure the set temp is what you are getting.

By the way, sourcing will be screaming to your boss because you want two (!) layers on the PCB instead of one...

Cheers,
John

[shakalnokturn]

I recently discovered that one of my DMM's that used a voice-coil type of beeper had neither freewheeling nor snubber, a beep causes peaks over 70V on a track less than 2cm away from the ASIC.
No obvious issues here, however cutting costs and ignoring well-known good design practices are the way to go for a poor result.

[mk_]

a beep causes peaks over 70V on a track less than 2cm away from the ASIC.

whats the problem with 70V with 2cm distance to a ASIC? I even don`t see a problem if there are only 2mm distance if they didn`t route that signal nearby very sensitive parts of the analog subsystem....

[Berni]

Seams like a nice tool for pinpointing the problematic area on a board after you have determined there are interference issues.

But its way too expensive for what it is. Id guess inside is just a capacitor that is repeatedly discharged into a coil by a oscillator. For the E field version replace the capacitor with a xeon flash lamp igniter transformer to turn the sharp high current spike into a sharp high voltage spike. Most likely just 10 USD worth of parts inside. Might be worth making an open source design for one of these.

But when it comes to microcontrolers the most common cause of a crash or lockup from interference is the reset line picking up noise. These are typically just pulled up rather than actively driven, so they are easy to influence electronically. Very brief pulses can on that line can reset the whole MCU. Perhaps the difference in immunity you observed is that perhaps some MCUs need a longer reset pulse for it to actually reset it.