Encapsulation for ATEX

[48X24X48X]

Hi guys,

I'm new when it comes to conformal coating encapsulation on components and PCB. The ATEX certification lab ain't recommending a specific brand and model of conformal coating encapsulation to be used as they do make money when you fail the testing. Does anyone here is familiar with this and has a regular brand or model to go to? I wanted to apply conformal coating encapsulation a LTE module which is the only component on the board that couldn't conform to the thermal ignition requirements under temperature class T4.

I have heard of HumiSeal 1A33 but not sure if there are better choices. Any suggestion would be great!

[mikeselectricstuff]

Does the LTE module have any exposed components in the RF section? If so then conformal coating may change the dielectric characteristics and affect its operation.
It's a long time since I did any ATEX stuff but I don't recall conformal coating being a valid way to deal with surface temperature issues, only to reduce creepage distance requirements.   

[48X24X48X]

Hi Mike,

My mistake, it should be encapsulation as I have mistakenly use that term. That is one part that I have concern on the RF although the ATEX certification engineers said "many of our clients" did this with the wireless devices. I guess that would need some testing to see how bad it affect.

[floobydust]

I have some experience with this as far as Zone 0 and 60079-11.
The potted modules were so much hassle. The LTE modules required a silicone caulk to first seal them up around the can, before the final potting. Encapsulant does change the RF characteristics a lot, shifts tuning to the point they don't work if it gets inside. I think it's the antenna output stripline that gets cratered. If you have a bias-T then that will get detuned as well. Sometimes the high-band just doesn't work.

I can dig for what encapsulant we were using. It was a two-part oven cure at 60°C. Had to pass the poke/hammer test and high temperature requirement. Problems with the thermal expansion coefficient because of the product's wide ambient -40°C to +60°C. It would shear and crack SMT parts on the boards. So had to first pre-coat the board with a soft silicone encapsulant, and then pot it with the hard encapsulant.

Then we had Manufacturing bungle the antenna connector, or mix it up with GPS antenna connector. This resulted in your basic unusable potted blob. RSSI would be very low but it worked on a test fixture. Only found out after a batch of 200 modules were not working and then tossed in the garbage after X-rays revealed that mistake.
You couldn't do a F/W upgrade because the connector for that was potted, and USB is not I.S. unless you workaround. Even then F/W upgrades sometimes needed access to special pins on the cell module, they could get bricked.
It was so costly and gross, I came up with a way to design out the use of potting compound and sucessfully put the product through HazLoc approvals.

The cert lab will not reveal what encapsulants meet the standards because some other customer paid to find that out, and why should you benefit from their expense. It's just awful to be paying many dollars repeating tests others have done. If you choose the wrong compound, thousand $ wasted.

[48X24X48X]

I have some experience with this as far as Zone 0 and 60079-11.
The potted modules were so much hassle. The LTE modules required a silicone caulk to first seal them up around the can, before the final potting. Encapsulant does change the RF characteristics a lot, shifts tuning to the point they don't work if it gets inside. I think it's the antenna output stripline that gets cratered. If you have a bias-T then that will get detuned as well. Sometimes the high-band just doesn't work.

I can dig for what encapsulant we were using. It was a two-part oven cure at 60°C. Had to pass the poke/hammer test and high temperature requirement. Problems with the thermal expansion coefficient because of the product's wide ambient -40°C to +60°C. It would shear and crack SMT parts on the boards. So had to first pre-coat the board with a soft silicone encapsulant, and then pot it with the hard encapsulant.

Then we had Manufacturing bungle the antenna connector, or mix it up with GPS antenna connector. This resulted in your basic unusable potted blob. RSSI would be very low but it worked on a test fixture. Only found out after a batch of 200 modules were not working and then tossed in the garbage after X-rays revealed that mistake.
You couldn't do a F/W upgrade because the connector for that was potted, and USB is not I.S. unless you workaround. Even then F/W upgrades sometimes needed access to special pins on the cell module, they could get bricked.
It was so costly and gross, I came up with a way to design out the use of potting compound and sucessfully put the product through HazLoc approvals.

The cert lab will not reveal what encapsulants meet the standards because some other customer paid to find that out, and why should you benefit from their expense. It's just awful to be paying many dollars repeating tests others have done. If you choose the wrong compound, thousand $ wasted.

Thank you so much for sharing such a detailed experience.

I read yesterday an article on Electrolube that mentioned it is not the resin that attenuate the RF signal but the filler material that used together with the resin. If the filler portion is either very little or not present at all, it *shouldn't* affects the RF signal. I'm not sure how true is that but I guess it will be a countless trial after this. They did mentioned some part number from the catalogue suitable for RF application. During the trial, the lab said I can use a power resistor (s) to represent the LTE module with the same exact power wattage expected. Of course if it pass the thermal ignition test, then I need to check whether the module still works and behaves as it supposed to be under the encapsulation.

Antenna and it's connector can get away from encapsulation if you can put a DC blocking capacitor (with > 1.5mm pad separation, yes it look silly when your other parts are 0402 and suddenly a giant 1206 appears) on the signal path.

The module is the only one I couldn't get it to conform to the surface thermal ignition but it seems more needs to be done from here onwards.

[floobydust]

A problem is you don't know how far into the module the potting compound has flowed. The striplines are designed for air and FR-4, and it can look good on low-band say 700-900MHz but fizzle out at 2.4GHz so you have to be really careful what band and signal strength is being looked at for testing.  MMCX connectors do not like being potted if it gets inside the pins (spring) had poor connections resulting from that. Figuring out what happened after potting, is pretty much impossible. If you can seal off the module perfectly, no manufacturing screwups, it could work. Using vacuum to eliminate bubbles would not work (for the entire assembly) without potting compound getting inside the air space of the modem.  The heat thinned it out enough to get rid of bubbles. Depending on your production volumes, Mixing and Dispense system can be needed.
I thought if things are housed in a plastic box, the potting compound can be lower density like silicones which can be chipped off if you needed to, as far as doing prototypes or troubleshooting.
There is a small part exemption for the thermal ignition test; you need two blocking caps (redundant) good for 1kV on the antenna. I had to get the manufacturer to declare total capacitance of the modem for ignition (spark tables) as well. It's only a 4V couple amps situation which is below the ignition limits for gases.

I don't know the major chemical manufacturers in your locale. In North America Dow and Henkel are the main players. You can contact their application engineering and get samples to evaluate as well.

[Tomorokoshi]

Without knowing system requirements, location classifications, installation methods, etc. an alternative to potting may be to use a flameproof enclosure to 60079-1 "d". For an off-the-shelf module where you can't design to "is" it might be a way to get started with the design, and move back to encapsulation "m" if there is a cost advantage. However, production volume might not justify that. Flameproof is nice because it may allow for easier replacement of the radio in case you need to change modules.

[48X24X48X]

A problem is you don't know how far into the module the potting compound has flowed. The striplines are designed for air and FR-4, and it can look good on low-band say 700-900MHz but fizzle out at 2.4GHz so you have to be really careful what band and signal strength is being looked at for testing.  MMCX connectors do not like being potted if it gets inside the pins (spring) had poor connections resulting from that. Figuring out what happened after potting, is pretty much impossible. If you can seal off the module perfectly, no manufacturing screwups, it could work. Using vacuum to eliminate bubbles would not work (for the entire assembly) without potting compound getting inside the air space of the modem.  The heat thinned it out enough to get rid of bubbles. Depending on your production volumes, Mixing and Dispense system can be needed.
I thought if things are housed in a plastic box, the potting compound can be lower density like silicones which can be chipped off if you needed to, as far as doing prototypes or troubleshooting.
There is a small part exemption for the thermal ignition test; you need two blocking caps (redundant) good for 1kV on the antenna. I had to get the manufacturer to declare total capacitance of the modem for ignition (spark tables) as well. It's only a 4V couple amps situation which is below the ignition limits for gases.

I don't know the major chemical manufacturers in your locale. In North America Dow and Henkel are the main players. You can contact their application engineering and get samples to evaluate as well.

Yeah, there is a pin #1 indicator hole on the module which I'm afraid the resin will probably get in. It's either I can ask the module manufacturer if they willing to have 1 without or pre-seal it with something. I think the RF part needs retesting. I have an on-board antenna and an auxiliary diversity SMA port. Thank you the links, I will check them out.

Without knowing system requirements, location classifications, installation methods, etc. an alternative to potting may be to use a flameproof enclosure to 60079-1 "d". For an off-the-shelf module where you can't design to "is" it might be a way to get started with the design, and move back to encapsulation "m" if there is a cost advantage. However, production volume might not justify that. Flameproof is nice because it may allow for easier replacement of the radio in case you need to change modules.

I think if the enclosure is never going to be opened at the site (zero maintenance), it is considered valid.

[geggi1]

Why dont you go for a Ex-d in sted of an Ex-m or Ex-e. By using an Ex-d you can put almost anything in the enclosure as long as the heat radiated is less than a certain wattage. By going for an Ex-d type of enclosure you get away with plenty of the troublesome stuff related to ATEX. You will need some kind of enclosure for your product and its simpler to have the Ex protection in the enclosure. Look at the websites of Cooper Crouse Hinds and similar companies or contact them about your requirements.

[mikeselectricstuff]

Yeah, there is a pin #1 indicator hole on the module which I'm afraid the resin will probably get in. It's either I can ask the module manufacturer if they willing to have 1 without or pre-seal it with something. I think the RF part needs retesting. I have an on-board antenna and an auxiliary diversity SMA port. Thank you the links, I will check them out.

There will always be a hole in an otherwise-sealed  soldered-down shielding can to relieve pressure due to gases released by fluxes during soldering. If it's a single hole in an accessible place, then simply putting a label over it should be enough to stop encapsulant going in

[pj]

Hi all,

Just wanted to join the conversation as I am going through a similar experience. I'm looking to get a modem with external antenna approved for zone 0 (Ex ia IIB T4 Ga). We are going to encapsulate pretty much everything except the external antenna. We have been working with a testing lab for 4 months on this so far, it's very difficult, I feel like we will go through the whole process and then right at the end they will tell us it's not possible to achieve intrinsic safety.

I understand the testing labs have to be impartial and can't offer design advice however they really seem to be screwing us around. For example I sent the lab 10 samples of our proposed battery, they have known for 4 months what battery we plan to use but waited until they received the samples before telling us they are unsuitable because they are protected batteries. 

Have other people had similar experiences with their labs?

</rant>

[48X24X48X]

I have the encapsulation compound chosen but will avoid conformal coating by swapping some components to larger size (couldn't care how it look like in the crowd). Chosen a solution from Electrolube but they are back ordered until end of this month.
Batteries if you can select those already have reports readily available, that will save you money. The one we chose from Tadiran, doesn't have. So, we sent 10 samples to them for 2K quid to test! The battery is the 1st thing you look at before proceeding down the rabbit hole as the remaining components design specifications depends on it.

[pj]

That is good to know thanks, how would you determine which batteries already have reports completed already? Would it depend on the lab, so different labs would have different reports or is there a centralised database somewhere?

I'm also looking at the electrolube products, in particular the RF one.

[cburgess]

We manufacture an ATEX certified low power radio transmitter. Potting is a nuisance but it does deal with the surface temperature issues. It also excludes the explosive atmosphere which simplifies the assessment process and can give a bit more latitude with the circuit design.

I can't vouch for the RF compatibility because we use an antenna outside of the potting - but our units work just fine.
We use the Acc Silicones SE3000. The potting represents a significant proportion of the overall unit cost.
We use a spot of compatible RTV silcone around the ufl antenna connection before potting. The SE3000 mixture is extremely thin and will wick into everything. It cures as a soft rubber - so it only works for us because we are encapsulating into the case.

One other thing to watch out for is the CTI requirements (Table 5 60079-11) if you've got anything above 10V. We found a lot of encapsulant manufacturers couldn't give a CTI value.

Getting something through ATEX certification is a whole world of pain.....

 

[EE-digger]

The last ATEX product I worked on a year ago is a small telemetry / logging device for various physical parameters (temp, pressure, etc.)

You will get no help from the lab on encapsulation or other material choices.  In fact, their scope of examination for the material is so narrow that in the end, our mechanical designer chose another material over the explosion rated Loctite product I had selected.

In the end, we only had one tiny fuse which needed to be encapsulated and the depth/thickness of that encapsulation tested and proven by our production documents (yes, your documentation is VERY important to passing ATEX).

If you have an antenna that is chip based, it will need to be impedance matched (i.e. tuned) with encapsulation in place.  This is tricky but I've done it on several Bluetooth (BLE) antennas.  It's very important to do this EARLY in design.

Find engineering help with RF background and VNA with appropriate probes, etc. to do the matching AND to look at effect of the material on transmission characteristics (radiated power).  Avoid metallic fillers but also, evaluate your plastics and encapsulant as well. 

Make sure it works for you.  If you are encapsulating a non-chip antenna, good luck.  Be sure the encapsulant does not pull the operating frequency range of the antenna in its band(s).

The ATEX lab is only going to look at electrical and thermal properties.  They did not even care that one encapsulant had a documented explosive rating vs the one we ended up with which did not.

To meet the depth requirements of the encapsulant, we had to make a small "swimming pool" ring around the part, with defines location, dimensions and height.  Again, production documents had to be sure the "pool" was filled to an inspectable height to guarantee thickness of the encapsulant.

"cburgess" summed it up nicely ... a "whole world of pain".  Much of it seems to make no sense and in fact, much does not.  But if you are going into a hydrogen (for instance) environment, it's good to know that someone did their homework where it matters.

Another thing on labs.  We worked with a new branch office of one of the leading ATEX test houses in the world.  They knew almost NOTHING about testing small ATEX rated batteries such as the small Tadiran lithium ion cylindrical cells.  This lack of some basic knowledge cost us a wasted $20,000 or more in fees which I believe we later gained some concessions on to help recover it.  But they also cost us time, which is always valuable.

[mikeselectricstuff]

Many years ago when I was doing ATEX stuff, ISTR around batteries you basically had to repeat the tests regardless of whether they had tested the exact same battery the day before for someone else.

After I'd left the project I heard that there apparently were cases of batteries that had passed in the past, which now failed due to changes in chemistry.
We were using 9V alkalines back then, but they later changed to 3.6v CR2 type lithium cylindrical cells as these were much easier to get approved due to the lower voltage.

[EE-digger]

We also had to submit the 10 battery samples as a prior poster mentioned.  Prior ATEX success is like a great rating on a product from Amazon.  It might be a good starting point or you may have just found 500 stupid customers and 3 who understood the product was crap.

[48X24X48X]

That is good to know thanks, how would you determine which batteries already have reports completed already? Would it depend on the lab, so different labs would have different reports or is there a centralised database somewhere?

I'm also looking at the electrolube products, in particular the RF one.

Some of the batteries are advertised as ATEX rated by battery manufacturer. Chances those have the report especially the short circuit current and thermal ignition test. These easily cost GBP2500 per test by the lab. Brands like Saft & Tadiran usually have these sort of batteries. And even if they are not advertised as such, chances one of their customer already used them in an ATEX product. At least you know if you are going to spend GBP2500, it will pass. I think most important is the short circuit current for battery selection. You will be in dilemma of getting enough current for the RF peak transmission and at the same time the need of getting the short circuit current to be bellow certain value. Else there's just too much of fuses at different level on the circuit as at times some components are just too hard to high one that has better thermal resistance or they just don't exist and at worst case, potting them.

Many years ago when I was doing ATEX stuff, ISTR around batteries you basically had to repeat the tests regardless of whether they had tested the exact same battery the day before for someone else.

After I'd left the project I heard that there apparently were cases of batteries that had passed in the past, which now failed due to changes in chemistry.
We were using 9V alkalines back then, but they later changed to 3.6v CR2 type lithium cylindrical cells as these were much easier to get approved due to the lower voltage.

Yes, the higher the voltage, the harder to pass the test.

[Gribo]

Tadiran is Saft these days. I managed to have an ATEX (iib, T4) certified Li-polymer pack, no potting except for the fuse, series resistor and the voltage limit Zener diodes. The design had a 4.2V supply.

[tszaboo]

Many years ago when I was doing ATEX stuff, ISTR around batteries you basically had to repeat the tests regardless of whether they had tested the exact same battery the day before for someone else.

After I'd left the project I heard that there apparently were cases of batteries that had passed in the past, which now failed due to changes in chemistry.
We were using 9V alkalines back then, but they later changed to 3.6v CR2 type lithium cylindrical cells as these were much easier to get approved due to the lower voltage.

For ATEX, it is sufficient to have the proper test reports for the batteries, the NoBo might accept them. They don't have to though. It depends how much money they want from you. For IECEx it is different, test is to be repeated.

Tadiran is Saft these days. I managed to have an ATEX (iib, T4) certified Li-polymer pack, no potting except for the fuse, series resistor and the voltage limit Zener diodes. The design had a 4.2V supply.

Are you willing to share the battery MF and part number?

[Gribo]

It was a Kokam small cell, Sadly, Kokam is no longer making small cells. The PN was SLPB544032, it had a carbon anode, so no leaks, no ignition. There are now equivalent parts from other manufacturers (Varta) that have similar certifications.
The trick was Zener diodes for voltage limit, 2x 4.7V Microsemi SMBJ5337 (obsolete, leaky) and 375mA fuse for the current limit. Today I would use 2 5.1V Zener diodes, and a smaller 315mA fuse.

[voltsandjolts]

I noticed the IS ATEX IIC T4 certified Keller Leo Record Ei uses one 2.2Ah Tadiran SL760 AA 3.6V Lithium Thionyl Chloride primary cell which is not potted and is user replaceable.

The cell has rated nominal discharge current of 2mA with continuous discharge currents of 100mA and a pulsing current draw of 200mA.
I've never tried shorting it out, but to me it sounds capable of generating an energetic spark...but I must be wrong of course

[tszaboo]

I noticed the IS ATEX IIC T4 certified Keller Leo Record Ei uses one 2.2Ah Tadiran SL760 AA 3.6V Lithium Thionyl Chloride primary cell which is not potted and is user replaceable.

The cell has rated nominal discharge current of 2mA with continuous discharge currents of 100mA and a pulsing current draw of 200mA.
I've never tried shorting it out, but to me it sounds capable of generating an energetic spark...but I must be wrong of course

For batteries, the testing is to determine how hot they get in case of an internal short circuit. That's the T4 part. And maybe they can create a spark, but in case the spark energy is low, it is not going to ignite the gas most of the time (which is in the order of parts per billion in this case).

It was a Kokam small cell, Sadly, Kokam is no longer making small cells.

Thank you. Unfortunately, the newer cells are quite problematic, since the goal for cell development is to lower the ESR, and to make the battery lighter.

[voltsandjolts]

Ahh, OK, so energy in the heat source or spark must be far below the Minimum Ignition Energy rating for the specified environment.
How do you calculate the energy in a spark from a battery or capacitor shorting?
For a capacitor do you just use the stored energy 0.5*CV^2 ?
For a battery?

[tszaboo]

Ahh, OK, so energy in the heat source or spark must be far below the Minimum Ignition Energy rating for the specified environment.
How do you calculate the energy in a spark from a battery or capacitor shorting?
For a capacitor do you just use the stored energy 0.5*CV^2 ?
For a battery?

There are specific rules defined in IEC 60079-11 that define the applicable conditions, if a circuit or a component or certain voltage and current is safe or not.