Hi, first time commenting. Years ago as a young engineer we developed small motor protection devices called polymeric PTCs. We used carbon spheres instead of carbon fiber but saw an effect similar to what you observed. At lower voltages and lower temperatures the plastic acts as an insulator. As the voltage increases a tunneling effect can be observed. As the temperature of the polymer increases, the current flow reduces. Loading the polymer with higher levels of carbon will produce a low resistance device at room temperature that switches to a high resistance in a sigmoidal curve. Hense, a motor protector. The trickle current keeps the temperature high until the circuit is opened. The the device will cool and reset to a low resistance again.
I was hoping more would turn out to be conductive. The cost of conductive filament is pretty high, and it would be nice to be able to print static dissipative enclosures and such for sensitive things
There are a few filament brands that do deliver a static dissipative filament but I never checked out any of these. I only use special coatings or even metal tape to get dissipative, static or even RF shielding.
Carbon Black (often of colorant) can impact RF. We utilize different variants of carbon black additive into injection molded boxes for radar sensors on autonomous haul trucks. Sheet metal boxes originally used were causing to much interference issues and giving faulty readings.
@itigg 's comment about the Carbon Black colorant is spot on. Seen it many times in industrial situations with all sorts of materials. Last one was on a newly installed machine that had a mechanism fed with 2500vdc that was isolated from the machine. It was pneumatically raised and lowered against a plate to apply the charge to another part. It was faulting out for high current. I got called to troubleshoot it and seen the installer had used black polyethylene tubing for the air lines. Told them they needed to change to clear tubing. Got a bunch of "Your full of BS". I took the black lines off and said "Turn it on now". I walked off while they said "Dang, it works now...."
Does things to CO2 laser cutting as well, black acrylic is just that little bit more difficult to cut than any other colour. White on the other hand cuts beautifully for some reason.
I guess similar would apply to Titanium white (one of the most common white pigments you'll find). A cursory google-fu shows that titanium dioxide is a semi-conductor. Cutting behaviour is probably down to energy and light absorption/dissipation properties of the material
@@konzetsu6068 Yeah, black needs a bit more power to cut because the IR light get absorbed more than the other colours. With white it's more about edge quality, often comes out really nice (flame polished). Can get the same finish on clear if the phase of the moon is right. On black? No way.
This was interesting to see. I think it may be mostly down to chance. The plastic itself is very high resistance, but the carbon fibres are lower resistance. Since the carbon fibres are distributed randomly they could align into a chain, close enough together to conduct, which might be what you saw with the Bambu labs PAHT-CF. I would also think that the longer the individual pieces of carbon fibre, the higher chance it ends up conductive, since you have to align less fibres then for the same distance. If Prusa uses recycled carbon fibre, it could be that they don’t chop or grind it up as much as the others, so the fibres are left longer. Longer fibres are better for mechanical properties, but worse for clogging. Bambu lab recommends a 0.6 mm or bigger nozzle for the PAHT-CF, which might mean that it’s fibres are longer and more likely to clog. It seems the two things that should effect it is fibre length and fibre content, higher of either probably means it is more likely to conduct, but I wouldn’t want to rely on them being non conductive. It would be interesting to see an extended test of some of the filaments to see if it was just chance that the PAHT-CF conducted and to see if some of the other filaments would conduct too. It would be very interesting to see you use filament with glass microspheres in it, they are supposed to have similar effects to carbon fibre but are supposed to be insulators too since the glass spheres are highly non conductive, so they may be an option for situations where you definitely do not want any conductivity. Glass fibre, glass microsphere and Kevlar (aramid) might be good to try. Edit: nylon with glass fibres is commonly used in injection moulding, sometimes at least 40% glass fibre by weight and it is used for all kinds of things, including power tool housings. It seems much more widely used than carbon fibre so maybe conductivity is a possible reason for that? The resistance for spectrums PA6 GK10, which is one of the glass microsphere filaments has its resistance in its data sheet, the insulation resistance is above 10^12 Ohms and the surface resistance is the same. Spectrum also do a PA6 CS20 FR which is a fire retardant nylon with ceramic spheres (where the CS in the name comes from), it also mentions its surface and insulation resistance but it is only 10^9 Ohms. Spectrums PC-PTFE also claims above 1e12 ohms.
Glass fiber is significantly cheaper than carbon fiber. Also glass is significantly less irritating for some reason. You could probably get the same durability with significantly less CF than GF but it probably would be harder to make sure it is evenly distributed.
Interesting video. I use these tools at an aircraft manufacturer who builds the fuselage almost entirely of composites. Between the layers of composites there is a metal mesh for static dissipation and lightning strikes. During final checks we use the ohm meter in various locations such as static wicks, cowlings, flight controls, etc to check resistance. If not in limits we use the insulation resistance tester at 1000v to improve the readings. Every time I set the tester to 1000v I always said to myself that the amps is what kills you, not the volts. Never was completely convinced and did not want to find out.
As someone who really doesn't know much about the technicality of this video it comes off as extremely nerdy. An extremely deep dive into the unnecessary testing of Carbon Fiber impregnated filament and its relationship with Conductivity...It was pretty great. Thanks for another awesome video.
Nice test, I think it's cool that you decided to do a video like this even tho you didn't expect anything from it. I also have wondered about printed cases for things like the ESP32 blocking the radio wave signals...
This is the exact video I have been searching for! I'm building an automotive under hood fuse box that PA-CF has the best mechanical properties for, but was concerned about the conductivity as the circuits are in direct contact with the material. The only thing that would have made this video better would have been testing pre and post annealing of the samples along with saturation, would probably change the results quite a bit. I think a part 2 is in order!
Thanks for the video. Reminds me of the time I had a very well known company that makes reproduction rubber components for vintage cars make some spark plug boots for an old engine. They arrived, looked beautiful. I assembled some, started the engine, and it ran awful. Misfiring badly. I could hear and see sparks jumping off the new boots. I called the company, and an engineer realized they had unintentionally molded these with a conductive rubber. 🤦♂️ I didn't know there was such a thing before this happened. They made new boots from a more appropriate rubber, and all was well.
At work we use 3D printed plastic enclosures before the tooling for moulded production versions is available. We had a problem where the prototype products were readily failing flash test, which was down to the 3D printed enclosures being conductive. Doubt it was a carbon fibre filament, but it was a black colour which consistently gave issues. Going back to the supplier, their best guess was if the parts had absorbed moisture? I wondered if it was the use of carbon black in the colour of the filament? Never did really get to the bottom of it, just learnt that 3D printed components can be conductive and shouldn't be relied upon to insulate in important situations. We could measure it with a multimeter, didn't need the flash tester at a couple of kV to see the problem, but that was what drew our attention to it. Printed parts a different colour seemed ok.
Yep carbon black, lots of people here mentioning having the same experience. White timing belts on a (to be honest faulty) moving roof motor, not causing any issues, the black ones shorting between the faulty and the good motor....
That was interesting. At least for anything higher voltage, I would have had a gut feeling avoid CF filaments just to be safe. Sometimes a tiny conduction can burn tiny bits of insulator and it may fail in time. I would never have guessed a measurable conduction at low voltage like Brymen! Yet another video that I learned something from. That seems to be the case practically every time on this channel.
Excellent video! Thank you for the thorough investigation. If you're feeling particularly adventurous, you could also test corrosion effects from mating CF filled prints to metals. It's a big no-no for CF lay-up parts without insulating well
That was indeed a surprising result. From your previous video I did remember (correctly, I just checked) that there is no direct electrical contact between the printed plastic and the traces on the PCB. The only thing that might have been an issue is the clearance on the larger pads the big screw nuts connect to but putting some clearance on these would have been an easy job. Years ago I have been talking to FormFutura engineers about the conductivity of their CarbonFil and possibly creating a dissipative filament. They though that since the carbon fiber is embedded in the plastic, they think conductivity - or the lack thereof - is not something they can guarantee. I think the conductivity may even vary across different badges of the same material, depending on the carbon particles, plastic type and possibly even on slight variations of process parameters during production.
@@Clough42 the manufacturing process might allow you to prevent conduction, for example introducing slight gaps that interrupt carbon connections. On a very small scale like 1mm it might not work as they could still be contiguous but beyond that statistically impossible. However that might also introduce weak spots depending upon each print. You want fibres somewhat overlapping as that allows the force to be shared. If you had a single layer for example without any carbon fibres then it acts like a kink in a space elevator. I read a few papers on how the length of carbon fibres affect composites. I did this after reading far too many online comments dismissing CF filament since they aren't contiguous and therefore don't benefit from carbon bonds. The research was interesting, the whole thing is incredibly complex. Depending on the manufacturing process used you have massive variation in fibre length but more importantly fibre alignment. The fibres can have a narrow average angle cone making the material very strong longitudinally. In those cases longer fibres are significantly beneficial. Many are overlapping along the same direction and sharing load. If you have nearly randomly angled fibres then shorter fibres are no longer disadvantaged. You want them all overlapping, lots of small fibres create more overlap than a few long ones. Hazarding a guess, this type of thing could drastically affect conductivity. Even if there is a conduction path through carbon fibres end-end, a single microscopic, short fibres and barely contacting chain will have massive resistance. A mix of random orientation long+short fibres could produce numerous conduction path ways through the shorter fibres with the long ones interconnecting them all. Forcing consistent conductivity seems impossible. It would need long molecule carbon chains like unbroken nanotubes or graphene which we can't manufacture that would be interrupted during printing anyway.
I think though, that when you're a home hobbyist, you have less tools and time to really get to the bottom of things, so no point introducing potential causes for magicpixy, occasional problems that are hard to solve when you don't have to.
A really interesting investigation. I personally expected more than 2 filaments to be (barely) conductive. But most being non-conductive to even high voltage meters, that's surprising to me. I also didn't know yet that voltage limits max. measurable resistance either.
Perfect timing! I just ordered some Extrudr DuraPro ABS explicitly because it is electrically isolating (as per the spec). I assumed that CF filaments would conduct, which is why I even looked at a different filament. Cool!
20 years ago we were using carbon fiber heaters that were used to evenly heat a 6'x16' polypropylene panel. I also understood that many composite aircraft use carbon fiber on the leading edges of wings for de-icing. Just for kicks' I just took some carbon fiber 2" tape and hooked an ancient pot on it and had the whole 5-foot or so glowing red and pr.producing a fair bit of heat
One of the reasons I follow your channel is you test things, you do the math. I would do that too if I had the slightest idea where to start. In this case, having watched the video a couple of times and looking at my test with 110v mains, the difference must be in the volume of carbon fiber in the filament vs the ribbon.
Very interesting! Some carbon, like that of a pencil core has varying conductivity depending on the mix of graphite and whatever filler (originally clay, then maybe some polymers). The whole issue revolves around microscopic contacts within the carbon particles, such as graphite flakes. The contact probability is science way beyond my abilities to tackle. All I can imagine is that there is some non-linear or maybe even abrupt limit in the carbon content when the carbon to carbon CHAINS start forming. Keep in mind that just contact between only two carbon particles is not sufficient for overall conduction. But here comes something from my distant past. I once built a control (or monitoring) system for manufacturing paper insulated phone cables. Paper is quite good insulator, provided it is dry. So the manufacturing process included heating the cable before a lead jacket was extruded over it. That happened in in a vacuum chamber. Rather than just guessing when the cable was dry enough, some measurement was desired. There were two things that I implemented. One was measuring the resistance increase in one ore more wire pairs. The other was measurement of the insulation resistance between some wire pairs. And that is where we made at first a rookie mistake. The wiring between the chamber and the measuring station was PVC insulated, and it was leaking to the tune of some tens or hundreds of megaohms. We had to change the connecting cable to polyethylene type, after which valid measurements of the cable dryness was achieved - several tens of gigaohms. Later on, I have had a Fluke 8060 handheld digital multimeter that had in addition to the usual ohms resistance measurement, also a conductance measurement capability down to nanosiemens levels, which was handy for sorting some 1N4948 diodes for low leakage, to be used in a timer circuit. Even later, I purchased a couple special instruments, a Keithley Electrometer with sub-femtoampere capabilities, and an HP4329A high voltage, high resistance meter, up to 2*10^16 ohms. Both meters were surplus items at flea markets and I have needed them maybe once or twice.
At first i was thinking what a bullshit video but after seeing what you did to prove a point, I am impressed, the time it took to set up everything must have been at least 20 hours plus video editing... Thanks again for your time. We learned a few things today...
Good experiment. Another factor's going to be humidity and how the material absorbs moisture, especially nylon, again probably only an issue at high voltages.
What about resistance change vs. applying pressure to the material? This might be our topic for the next video :D :D :D I always like your videos, keep going James.
I like this type of video and this type of response to what I'm sure many tubers feel are annoying comments. It's always good to see someone with a good grasp of the many to one relationshipstthey have so they can see things from the perspective of the millionth person to say something rather than blaming them or feeling bothered. As for the tests,I really have to wonder if this test really was all that conclusive or if the semi random alignment of the carbon fiber as it was being laid down just gave any given one of those tests some chance of being somewhat conductive. So I guess it, to me, seems like it's likely that the commenters were right, and to me, sometimes is good enough to not plan designs around a assuming there is no conduction, so good on the commenters and cough for putting this up on RU-vid for people to have any actual evidence to back their intuitions up or correct them.
There are static dissipating filaments. I have printed some parts form 3DXSTAT ESD PET. I could not get the finish print to show conductive on my basic meter. The customer had more specialized equipment to test it. He said it was would conduct and dissipate static. He was pleased.
Loved your video about potential space heaters. As the saying goes, "Jamie want big boom." We'll keep hoping. :) Also, it's spelled, "Burgundy". Take care!
In composites, long, skinny fibers offer improved strength and stiffness compared to spherical particle reinforcement. Your electrical tests are a reasonable way to evaluate fiber aspect ratio. Spherical particles would be insulated on all sides by polymer matrix, with almost zero particle-to-particle contact. A circuit along high aspect-ratio fibers would have far fewer insulating junctions, and a greater likelihood of fiber-to-fiber contact, or near-contact. Assuming the recycled carbon fibers are aerospace grade, we can also assume they have high quality and aspect ratio, so it's not surprising that they result in measurable electrical conductivity. They should also result in increased thermal conductivity, but the meager difference would be even more difficult to measure.
An interesting side note, they are using carbon fiber infused in concrete as an ACTIVE LOAD CELL. This was patented as an automated way to measure potential failure of concrete bearing structures w/o the continuous need for an eyes on site inspection. Apparently as the concrete undergoes changes in the amount of load it’s under (contraction or expansion) the carbon fibers will undergo a measurable change in their resistance (higher compressed areas lower the resistance [of the carbon fiber in the area] while areas undergoing relaxation or tension will see an increase in resistance). While it wouldn’t do a lot of good to run a seamless run of carbon fiber throughout the concrete structure, placing precast smaller load cells embedded strategically in areas of concern alongside a group done at constant intervals would go a long ways to monitoring the health of aging concrete structures. The fact that they are able to measure these slight changes in resistance with a high enough degree of accuracy brings renewed interest in utilizing more than just its strength but its electrical properties as well.
After the initial breakdown at high voltage - a pathway is carbonised and will stay conductive to a lower voltage after initial breakdown of the insulating polymer between.
Another aspect: nylons particularly, but all plastics to some degree, absorb moisture. I presume these came more or less straight from the printer and have not absorbed any humidity. So as they absorb moisture over time the conductivity will probably go up too.
I really don't think that will be a problem. Even if you graphed the different base plastic moisture absorbancy, pure water doesn't conduct, and atmospheric water isn't going to have a lot of conductive salts dissolved in it to begin with. There "shouldn't" be any conductive materials readily dissolved by water in the part, either. If anything, moisture getting in might push the carbon filler further apart, reducing conductivity. If anything, these samples would act more like a capacitor than a resistor. But, assumptions and speculation are worthless, as shown in this video when everyone, literally, was wrong! Including the man himself. It would be interesting to test these at higher voltages, like with a Tesla or even a MOT (with appropriate safety systems and interlocks).
@@dogsarebest7107 Not able to quickly find out what MOT you are referring to. The most common answer is Ministry of Transportation which refers to car tests done in the UK.
i wonder if this is less to do about CF and more about what impurities might be in the plastic (perhaps from recycling or just purity of supplies). If true, you might see something like these even on some non-CF filaments. Seems like an awesome science fair project!
The aircraft industry had concerns about carbon fibre and lightning. The concern is more the resistance causing heating and explosion of the material. That is at the other end of the spectrum. I think the problem with CF in aircraft was solved by including copper wire in the buildup.
Curve tracer or SMU? That is, it would be nice to see an I/V characteristic curve, even if you have to fake it, stepwise. Be safe! Like a curve tracer, probably best to use AC and not DC.
Pretty thorough. However, I am printing parts for Tesla coils these days, and having some CF PA in storage, I am just drooling at it, but at 15kV primary, or 300-400kV secondary, it would not be very wise to use it. Regardless of how much they would conduct at 1000V, or less...
Great test ! Now, what about RF at low and high voltage ?? And, how about a nuke-ro-wave test... To see if the material changes properties ?? Thanks, this was really good info !!
I recon the conductivity of some of these is 1 of 2 (or even both) things. At higher voltages the resistance of the plastic breaks down and might create a path of "altered" plastic that is now conductive. Or (and/or) the carbon fibers randomly line up in a print so you have a continous path from stud to fiber to fiber, etc, to stud. Now I have to admit the first example of pathing usually happens only "well" above 1kv, hence almost all domestic/lv industrial installation goods, tools, ppe being rated up to 1kv. Thus I suspect the conductivity you saw where the fibers in allignment. But that is a suspicion on my end based on observation not measurable facts.
I was at the South East Electric Funfly when Carbon fiber DLG tangled with the power lines. The glider lost. It took down power for half the event site. There is a video from a drone that happened to be pointed in the right direction. The glider flat out exploded. Ken
I am going to guess the length of the chopped carbon fibers is a huge contributor to how/if adjacent layers and walls manage to make contact, and also within the same continuous extrusion as well.
Well now I have to take this knowledge and apply it to nickel plating 3D parts. If I raise the voltage enough to make up for the resistance, I should be able to have it start plating. As it starts pointing the conductivity will go up and I'll just have to put a current limit on it... But that means I should be able to grow plating onto the plastic parts directly
I can tell you that the 30% carbon filled nylon we used to use for backup rings had significant conductivity. Now it was filled with graphite, not carbon fiber but it shouldn't matter.
Graphite has aligned Pi bonds that are very conductive across the planes. Carbon fibres are not always graphitised enough to make micro sheets or tubes of Pi linked C=C bonds
I think this may also be affected by how the object is sliced, as well as just randomness. My guess is that you got conductivity when the printer happened to deposit a single fiber across the test points.
Hi James, Plastic are prone to absorb moisture from air, what about shelf duration of filament ? It's also worth noting that the two type of plastic that conduct electricity are Polyamid !? Thank you for your videos learn a lot. Geeting from france
I'd bet you'd see interesting effects if you tested at higher frequencies. I'm guessing some of the conductivity is dependent on fibers randomly touching within the material, as the frequency increases small gaps will act as capacitors and it'll likely turn conductive more easily. As I'm sure James knows, Ohm's law is only V=IR at DC, in general it's V=IZ, where Z is the complex impedance R+jX. None of the testers shown here can measure at particularly high frequencies.
I know this is going to sound incredibly specific. But one day you are going to take us on the ride to reverseengineering the Firmware of an odd shoptool to extend its functionality because one of its behaviours is driving you nuts. I am already thrilled for that day.
The static dissipating petg from polymaker and 3dxtech are conductive under the surface resistance test. They do specify higher processing temps for layer adhesion though
I got the bambulab CF filaments because I wanted to print ESD safe boxes for my components. But was disappointed in that I could not measure any conductivity at all. I then bought Fiberlogy ESD-PETG-Filament which turns out to have easily measured resistance. Just sad that it is very expensive.
I would expect the Bryman to use a constant current source for measuring resistance. That provides voltage directly proportional to current. VOMs and VTVMs used constant voltage sources, why is why they have non linear scales with zero at the fat right.
I'm now curious if a strength Vs. conductivity trial scenario could be constructed. Maybe someone would be willing to try using carbon fiber filament shown to be conductive and then evaluate conductivity from batch to batch. If two batches show consistent variations in conductivity do they also show correlated differences in measured strength? Another similar experiment could be constructed using one batch of filament with variations in deposition settings.
would i use CF filament for a 40kv plug? no, would i use it for a project with wireless connectivity? probably not, something i would need to test would i use it for 60v? yes, without concern
Excellent and informative test! Thank you so much. I am however wondering if some of the apparently non-conductive materials might break down and conduct or arc at higher dielectric voltages and currents? I've worked quite a bit with high-voltage switchgear underground cables and insulators and have observed that might be effective insulators at low to medium voltages (medium being 5kv and below) would breakdown at higher dielectric potential voltage. Higher currents also seem to effective dielectric insulator breakdown. Arching might be another potential issue, as the the carbon-fiber might be more susceptible to arcing carbon tracks which could lead to permanent arcing. Be interesting to see how these materials hold up to a Hi Potential 'Hi-Pot' tester, or something like a 10,000volt neon sign transformer with terminals set within and beyond the arc gap, and see if the arc tends to be drawn to the material.
Chopped CF plasticis in the same space as CF resin composite it doesn't exhibit meaningful conductivity in a macro scale to be useful but in a micro scale it is conductive enough to mess with RF signals.
I wonder if the Prusament would have shown surface conductivity if the surface was roughed up with an abrasive, but I guess that wouldn't really be the surface anymore.
I think what you're seeing is by random chance enough carbon fibers are in close enough proximity to conduct. I imagine if you printed enough samples of each filament you would get some that fail with every brand since the arrangement of the fibers within each layer are random. The lower the fill percentage of carbon fiber the less likely this will happen but probably still possible.
Is there any chance this is related to moisture in the material? Could you dunk all the samples in a bucket for a while and see if they all then conduct? If yes can you repro the same with non CF materials? Perhaps the CF granules just get you to the point of nearly conducting and then various little differences just take you over the edge? Size of granule, moisture, something else?
It’s almost like when you chop up a conductive material and surround it’s random orientation in an insulating material it isn’t likely to make a circuit; Is it ideal, no. Will it likely work fine, probably depending on use case.
In conducting, there is resistance. Where there is resistance there will be heat. Heat + PLA = Melt Even if it is possible, I don't think it would be a good idea for long term use.
I wonder if the phenomenon only happened with nylon due to its hygroscopic qualities. I'm no material scientist or anything near that, but I wonder if a combination of absorbed moisture in the plastic and the carbon fiber threads could have caused better continuity than the other polymers with much lower hygroscopic qualities.
I know you live in a dry desert, but considering that the two conductive specimens are PA based, could moisture play a role in this result? Maybe you could try to oven/dessicant dry those parts and test it again. Just for science.
