There is a safety issue with your LED indicators - they are not failsafe. If the LED fails, then it will indicate it is safe (switch is off) when actually it is not safe (switch is on). You should either invert the LED (so it lights when it is safe/high voltage is off) or add a second (or bicolor) LED which lights when it is safe.
Better yet replace the latching switch, or disable the latching mechanism, so that the plunger has to be pushed in for the device to stay energized, then add a latch to the plunger. Alternatively if there's too much air leakage for that to work, you can add a physical mechanism that is joined to the switch an moves a painted flag inside the case. The most foolproof solution would likely be to have the indicator bulb in series with the switch tho, that way the indicator bulb MUST be on in order for the circuit to be energized.
That’s exactly why I value UA-cam comments so much! They often provide valuable suggestions, important insights, and great ideas. During the construction, I did consider that the LED could pose a safety risk if it's defective. However, I decided against making the circuit more complex for two reasons: 1. When I turn on the power supply, I press the pneumatic switch to check if the LED is functioning correctly. If it doesn't light up when I press the switch, I know there's an issue. 2. The safety switch shouldn't be activated anyway if the power supply is not in a safe state. So, even if the pneumatic switch is on, it shouldn't be a problem to activate the safety switch. Nevertheless, you're absolutely right, and it would have been better to make the system fail-safe. I like the idea of a second LED (e.g., green) indicating when the device is in a safe state. I might add that later on.
@AdvancedTinkering the downside of the second bulb is that you then have to carefully consider the driver circuit for it, as it can't have a fail mode which leaves the bulb on. If it fails in the on state, and your caution bulb circuit fails in the off state, you're back to a dangerous condition. Your toggle idea makes sense, however the point of safety systems is to mitigate human error, i.e. if you're distracted or excited and forget to do it. Another thing about it is that light colors are rather subjective, for instance on the hv tools I work on red (amber really) indicates main power is off (and thus the hv is off) and green indicates power ready (and thus hv is on), however the breakers in the panel have a red flag to indicate energized and a green flag to indicate de-energized haha. You'd need labels.
When I was in uni we did a lab on high voltage/insulation. One of the topics covered was insulation coordination, in particular dealing with insulators made from multiple dielectric materials. When you have an insulator with multiple dielectrics, the electric field is concentrated in the material with the lowest relative permittivity. This means that if your insulator has defects like air bubbles, the electric field field will be concentrated in them, and potentially be high enough to ionise the air, damaging the insulating material and eventually leading to complete breakdown of the insulator.I suspect this is likely to happen with any 3d printed parts, unless you can make them free of air pockets.
I've bean 3d printing functional parts (prototypes and the like) for years, and I do not believe the FDM process is capable of making a part that's free of air bubbles. I spent some time prototyping pump heads, even printed solid they would weep (slowly leak) when pressurized. However, SLA printers definitely are, I think this project would have worked perfectly with some SLA prints.
I suspect the resin printed insulation would not have the same problem. However we can still improve the performance of printed parts. You can likely try using different materials than PLA but also try using the annealing method that powdered salt. You print at 100% and then place the part in a container filled completely with the finely powdered salt and bake in the oven at the glassing temperature - please look up the exact temperatures for the exact materials before attempting. The part shrinks a tiny amount but becomes completely solid, gets way better mechanical properties and likely also improved performance in this regard.
@@johnmadden9613 I could see the salt causing some issues given that the salt is made of sodium and chlorine ions but I'm not entirely sure about that to be honest
You could also try the salt annealing method. The PLA part is placed into an oven-safe container and it gets filled with fine table salt, which is then tightly compacted around the part. Then it's placed into an oven where the PLA (partially) melts, while the compacted salt keeps the plastic in the original shape. CNC Kitchen did some nice tutorials and testing of this method.
That was actually one of the methods I had considered. I saw CNC Kitchen's video on it a while ago and thought it could make a significant difference. Still, I wouldn't trust FDM-printed components when it comes to high voltage.
Please do the 3d print hv test! I’m planning on making a flyback with printed parts, and it would be very useful. Also, your power supply is the best looking one I’ve ever seen!
Thank you! If I make a video about different materials, printing parameters, and post-processing methods, I will also test PP. However, I can't promise when I'll get to it. Regardless of the results, I would never trust 3D printed parts as insulation for high voltage. A single defect in the layers is enough to reduce the insulation effectiveness to practically that of air. I would always recommend using epoxy resin with the appropriate dielectric strength as the actual insulation.
@@RiehlScience Possibly, that could be a method that works, especially with a very thin resin. However, in many cases, it might be simpler to design the parts so that they can be filled with epoxy resin. That's how I made the high-voltage connector for the power supply (shown in my last video on the power supply). I would be very interested in how SLA-printed components behave compared to FDM-printed ones. Unfortunately, I don't have an SLA printer to test that.
@@RiehlScience Hi, I did that, you can see that bubbles exit at all points where there is infill unterneath. Even at 100% infill there is still some air that comes out of it. I would sugest that you use PLA only for the structure, such as outside walls and standoffs, and leave the insulation to the epoxy. I used very thin PLA standoffs so the epoxy could get into all cracks under vacuum, and it worked out ok up to 30 kV.
@@AdvancedTinkering I wonder how much of the issue with the PLA prints is the filament color. Black plastic should have fairly high carbon content (from the pigment), I would be interesting in seeing performance of different colors.
That pneumatic switch looks like it uses a regular form factor microswitch. You should be able to carefully remove it and replace it with a momentary switch. This is a much safer option as there's still a chance of accidentally energising the the HV supply if the indicator light has blown.
Very good point! It would probably make sense to use a momentary microswitch and keep the air in the tube compressed through another mechanism (a three-way valve, for example) in case I need to run the power supply continuously.
Unless the high voltage supply is going to be fully integrated into whatever it's being used for then some form of "dead man's switch" is absolutely essential. This is especially true if it's going to be used as an interchangeable supply for multiple projects. Basically, if it's possible to energise with the high voltage exposed it should require some form of continuous human input to keep it going.
A magnetic latching switch could be solving some safety concerns, when powering off the device the electromagnet loses its capability to hold the switch "on" and it resets Its used in planes aswell
I and a few other people were thinking about this fail safe issue, seems like you know the perfect solution! Thanks for sharing your expertise with that!
I wonder if the chosen color had anything to do with it. To dye something black they usually ad carbon. So maybe try it with a natural filament without dyes. That being said, I think you're on to something with the air channels, maybe try to over extrude aggressively to minimize air inclusion.
I suspect that different colors perform differently. But since the dielectric strength even of black PLA should be relatively high (assuming no defects), I suspect it is related to small channels between the layers. I experienced a breakthrough with white PLA at a voltage it should have withstood with consistently fused layers. I'll test over-extrusion, thanks for the tip!
@@AdvancedTinkering I saw a demonstration where for a 600V rated cable, the white/red insulation held up to over 1kv but the black broke down just a little above 600V. so there could be some significant difference at high voltage.
@@AdvancedTinkering Maybe getting transparent filament would work well with the over-extrusion: Thomas Sanladerer has made a video on printing transparent (and not just milky-translucent) parts, which also boils down to over-extruding to remove all air bubbles. It would also help to see if a particular part has bubbles or not, without needing to test it with HV :)
7:00 another approach might be to use a resistive divider to divide the DC power supply output to the display so it displays n x 10kV. You might also be able to adjust its trim-pot so that when 10.7V shows up on the measurement line, it reads 20.0
@@AdvancedTinkeringIn a DIY/OS Birkeland-Eyde Reactor maybe if you replicate it? (Granted plenty of projects as-is, but a really well built one would be AMAZING!)
Hey, great video!! I've built a prototype of HV PSU to run an experiment for simple fusor. It has reached only 10-15KV and was already scary as hell. Yours is a great example how to do it the right way:) It would be really interesting if you can try ABS and nylon for HV parts, but OTOH it's the FDM as technology that will spoil the part by leaving tiniest air gaps. Likely it's doesn't go through the plastic itself. To confirm that a resin printed part with same dimensions is needed. Another option would be to put the part into very fine salt, tamper it down very tightly, like for sand casting, and re-melt it in oven. The process won't release air bubbles, but will break air traces between extrusions. Also you should randomize perimeter start point to avoid having seam like on your part, to avoid creating a line of air channels.
Unfortunately no... I feel so bad that I neglected the Stern-Gerlach experiment. But unfortunately, I need a picoampere meter, which I can't afford at the moment. I hope you enjoy the other projects too! At least I can say that the project for which I need the high-voltage power supply fascinates me just as much as the Stern-Gerlach experiment.
Ich kann dir "Corona Dope" von MG Chemicals als Isolationsschicht empfehlen. Auch innerhalb der 3D-gedruckten Teile als zusätzliche Schutzlage. Sehr effizient.
In meinem letzten Video zum Bau den Netzteils zeige und benutze ich Corona Dope zum isolieren der Anschlüsse am Zeilentransformator ;) Es zum Beschichten von 3D gedruckten Teilen zu verwenden ist eine sehr interessante Idee. Das werde ich testen.
@@AdvancedTinkering Hmm, irgendwie habe ich den Teil im vorherigen Video übersehen. Muss deine watch- und session-time wohl durch nochmaliges anschauen etwas erhöhen. 🙂
if you carefully dissasemble the switch, you can remove the lever inside that provides the toggle function(push once for on, push again for off), making it momentary
I didn't think about that, thanks for pointing it out! But probably, I would rather buy a new momentary microswitch to be sure. Then I would just need to design a mechanism to keep the air in the tube compressed in case I want to run the power supply continuously.
Try gyroid infill with a missing top or bottom layer so you can then epoxy the infill rather than printing solid. gyroid infill is nice as it doesn't create internal walls, you can pour something in to it and it will fill up evenly.
Don't worry about word pronunciation. We nativw English speakers butcher our own language bad enough. In fact, I knew exactly what you meant when you said it. BTW, you could use fiber optic ethernet and create a protocol that requires several commands in sequence to activate the power supply.
As an American truck driver, I think you should keep the chrome. It’s good for at least a couple extra foot pounds of torque. I also approve of the addition of chicken lights to indicate that the high voltage side is active.
There is a good chance the color of the PLA will have a massive impact on it's dielectric strength. Black colors will often have carbon black added which is extremely conductive. White pla will have titanium dioxide which is far less conductive. Virgin undyed pla would possibly have the best dielectric strength of any PLA but that's hard to say. I'd love to see a follow-up video testing various 3d printer materials and their corresponding dielectric strengths.
I would try with resin printing. That works really well to form air tight bodies. I bet the arcing problem was not caused by 12mm of PLA, but the dozens of little air channels inside the layers and the infill. Try creating a gas / vacuum tight seal with a filament printer ... it will not work even at 100% infill. There is no way FDM printing gives you a 100% solid body. Your test demonstrated this: Failure along the seam and along a z layer. That is also the reason I went against building a PEEK filament printer. Better buy high quality resins and calibrate to your printer, and get really nice dimensional accuracy.
I absolutely agree. Since PLA actually has a relatively high dielectric strength, I'm convinced, that tiny air channels and bubbles inside the plastic (maybe in addition with absorbed water) cause the breakdown. I'm sure SLA printed parts would perform a lot better. I unfortunately can't test that, since I don't have a SLA printer.
You could use fiber optics to control the whole thing still. Control the power supply internally with 0-5v signal, and charge a capacitor up to 5v via PWM. Have a discharge resister parallel to the capacitor. So if you run a 50% duty 5v PWM signal in, the capacitor is charged to 2.5v, and this is used as signalling for the power supply. Turn off the PWM and the capacitor is discharged, and the HV is turned off. Have the circuit open a relay below say 5% duty cycle (iunno 0.2v?) to prevemt stray charge in the capacitor from trigering the HV on. Fiber optics are used in dual resonant tesla coil builds quite frequently, as any interference can cause the bridge to explode. Its just PWM and frequency for a DRSSTC tho not a voltage input. Anyway, just a thought for you.. Fiber receiver part# HFBR-2412TZ Fiber transmitter part# HFBR-1414TZ The connector type is "ST" or "ST-ST fiber optic cable"
Would you say you`ve changed the final amount of current available, by using a different power supply? always very nicely done project and neat work space ! the new project looks exciting!
i have used PLA parts for hv transformer bobbins and some other stuff. Very interested to see some tests with different materials/slicer settings in the 10-20kv range. wonder if resin prints are better
I always used white 3D prints around HV. Rule of thumb has always been to avoid "Black" items because sometimes the pigments contain carbon. Its the same reason white PVC pipe works well as a "chicken stick" for poking around HV, and black ABS pipe will just conduct it and shock you. The second you make a burn mark on PVC pipe though, that spot is now conductive, and it'll branch out like fractals. Fun stuff. But yeah. Im always surprised how strong circuits are against noise. I currently run my inverter driver exposed ( no metal shielding/case ) a couple feet away from 150KV and it doesn't even care about all the noise coming off that. I've ran arduinos ( TEENSY chips actually ) in metal shielded cans around HV without problems as well, and usually did fiber-optic cables ( TOSLINK ) when I wanted a safe way to remote fire a HV power supply.
Do you use the 3D printed parts as insulation for high voltage? And if so, at what voltage? I also experienced a breakthrough of white PLA during the project at a voltage that a solid piece of PLA should have been able to withstand. That's why I suspect it may be due to small defects in the 3D printed parts. I guess resin prints would perform a lot better. Good to know that the microcontrollers are relatively resistant to electromagnetic noise. Thanks for the information!
@@AdvancedTinkering I have use white PETG for high voltage standoffs, in which the print is just being used to hold a HV terminal up in the air. Since the print isnt conductive, it doesnt try to snake-down the bushing and arc to ground. But this is basically, If the arc can arc 60mm, I'm running that 60KV through HV wire ( Rated for >75KV ) up through the center of the 3D printed bushing, and at the top of the bushing, I have my contact. That bushing would be much longer than the arcing length, like say 120mm, to simply hold the HV up off the ground. So, in this case its not being a "insulator" in the sense of running a bare wire through it, but I tend to call ceramic high-voltage bushings ( like you see on top of transformers ) Insulators. If you're running bare wire through the center, and trying to make the plastic print keep that voltage in, then I could see it being hit/miss. Since like you said, impurities, if air-gaps exist they could encourage corona, and potentially break through. Sorta like how a 3D printed PLA vase although it looks solid, can still leak water. If water can leak through the layers, air/corona/electricity likely can too, if I had to guess by that logic. I guess ideally HV wire is nice for that stuff. But. Maybe another approach could be casting if needed? Like resin-cast somehow? Would work similar to the potting material. I think as you said, Resin-Prints probably would work, since its a solild material with no impurities. Could maybe look into 3D printing a form to cast silicone in. Silicone is often used for HV, such as super flexible high-voltage silicone wire
I would be *very* interested in seeing if another color of PLA might work better with high voltage. My thinking is that black plastic is very commonly made black with the addition of graphite, so there's potentially some conductive material in the PLA. That being said, I think the other person who mentioned that the issue might be air pockets ionizing and causing damage to the PLA.
Some black PLA plastics use carbon for coloring. I'd really love to see a comparison between black PLA and white/transparent PLA, and maybe transparent PET-G tossed in there too.
I made some epoxy covered HV-electrodes, where the housing was 3D printed and also the electrodes with conducting PLA. When i vacuum degased the thing, it bubbled very long, so even thick PLA has quite some air left in it. I also would not use black PLA, as they can have carbon particles in it as a coloring, i had much better results with white PLA. With that it survied 30 KV at 10 mm electrode distance.
black materials are sometimes dyed with carbon dust. Try using bright color fillaments. Also, try dipping PLA into the acetone, it cannot dissolve PLA, but it will dissolve surface layer of the plastic, therefore making it airtight, this can improve HV properties, by eleminating microscopic gaps in print
The little voltmeter can be hacked to show the output voltage instead of the power supply voltage. If turning the pot doesn't allow it to go up all the way, there should be a resistor that you can change that gets you in the ballpark range.
the carbon black used to make black PLA makes it break down much easier. the same thing happens with black insulated wires. have you tried using white PLA, or just other colors?
I also experienced a breakdown with white PLA filament at a voltage that a flawless, solid piece of PLA should have withstood. However, I didn't compare two identical PLA parts made from white and black filament. That would indeed be very interesting. I agree with you that pigments in black plastics often lead to a lower dielectric strength.
I took a very low tech approach to remote control and bought a Christmas lights controller. It’s basically a radio controlled relay on a pigtail with an outlet. I think the remote is like 49 MHz or something like that. I have had zero issues with interference making the remote unreliable. Just food for thought. For safety I put a power strip after the remote relay so that I can “arm” the system with the switch or “disarm” it so that if for some reason the remote got activated it would still be off.
People saying “try different colors”: it’s not the plastic that’s breaking down. FFF will always leave air pockets between lines, even at 100% infill. It’s that air that’s breaking down. Even a microscopic air pocket will let the field through. As soon as it does, it gets hot, melts the plastic around it, and just widens the channel.
2:28 all you have to do is add a relay that holds itself closed while powered. If power is lost, the relay drops and you have to press the "power enable" button again to make it hold itself. This is how power tools and machinery have to work nowadays and it just takes a single relay with 2 contacts.
The problem with that method is: how do you connect the "power enable" button to the HV supply without an electrical connection. Which means I would be back at an IR remote. Or I would need a momentary microswitch which is actuated by the pneumatic switch and then actuates the relay. Or am I missing something?
@@AdvancedTinkering We have at work insulation tester up to 6kV AC/DC. Maybe I can print out a test piece 10mm long (PLA/PETG) and measure the resistance/impedance. I have SLA but I dont use it now. Just to know for the future.
Haha 😂 yes, perhaps sktchy to control things with budget microcontrollers. But... Used a few Arduinos to control the main plasma heating of the JET fusion reactor as an "upgrade"
Can you try 3D printing PP (Polypropylene) filament, as it's known to have very good layer adhesion and is used for water tight prints, it would be interesting to see if it can handle high voltage, and if not how does it compare to PLA. It is also known to be hard to print. Very interesting video as always!!
I have used natural PETG for HV, but never at those voltages. I think the dyes in the plastic probably make the problem worse. I would love to see a test of this.
I absolutely agree that its the air pockets in the 3d printed parts causing the electrict breakdown, "My Prusa MK4 is now clearly better!" from CNC kitchen might be the closest I've ever seen to "air bubble free" FDM printing. Honestly actually the whole concept of insulative 3d printed parts seems like something he'd be interested in testing, might be worthwhile to reach out to him. SLA printing on the other hand probably just works ? I'd absolutely love to see some tests done with it to confirm it because I can't see a world in which it doesn't work
Try print remelting, it should get rid of most bubbles, and you can use fine calcium carbonate sand as the reflow media, wich wont afect surface resistance
Haha I enjoyed the outro, silent letters really are a pain 😅 Does German have any words with them? Also I appreciate the sneak peek of the project this was made for. I think that looks a bit like an x-ray tube so my guess is you're going to give x-ray crystallography a try 🤔
I believe there are a few words with silent letters, but it's more of an exception. I can't think of any off the top of my head. Perhaps. At least, X-ray crystallography would be a very interesting project ;)
@@AdvancedTinkering ah then it must have been the French influence that gave English so many 😁 Ones like knife and knee have always baffled me. If you're not going to pronounce the K why not just spell them without it? 🤷♂️ Haha, I do agree, very interesting indeed. I've always wanted to learn more about it so hopefully someday I'll get to see someone give it a try 😉
Hi, my 1st thought when you said "security switch" was, pneumatic... On the other side, you might want to consider printing a pneumatic security pedal switch, in case you get arc-ed, you fly 1 meter in the air (got arc-ed at 27kv when working on CRT tv long time ago)... Do you intend using this power supply in your vacuum chamber to evaporate tungsten/molibden by ionization? I wonder how plasma looks at 10e-6...
I suspect black PLA could be significantly worse than other colors. Carbon black is a fairly common black dye, and it being conductive could make a difference especially at those voltages..
I am pretty sure that the problem is the carbon black pigment in the PLA. I think Marco Reps did some testing of white vs. black cables. PLA is rather for high voltage applications actually, have a look at the Blog of Kerry Wong. He measured a dielectric strength of 47,5 kV/mm in an uncompromised part. DOI 10.1088/1757-899X/461/1/012091 is also an interesting read. Cheers
Yes, I researched the dielectric strength of PLA before building the power supply, and it does indeed seem to be relatively high. However, my practical experiments have shown that 3D printed parts with multiple layers do not come close to reaching these values. I was already familiar with Kerry D. Wong's blog post. Interestingly, he also uses black PLA. Since the PLA layers for his tests were only 0.18 mm thick, I suspect that there was a very good fusion of the PLA. Once you have many layers, there might eventually be a compromised spot where an arc could form. That's why I suspected defects between the layers. But it's a very interesting question. I think I will test a few different colors and post-processing methods.
@@AdvancedTinkering The layer-height is very interesting. Usually thicker layers mean stronger parts. Maybe printing hotter will help because it increases fusion? Indeed very interesting
Try the same HV test with a clear or natural colored PLA, the pigments in plastics, especially black and gray, often contain conductive materials like carbon. In general, black and dark gray plastics should be avoided around high voltage because of that. I doubt the other filament types would have much effect on the insulating properties of the parts, although maybe something like PETG or TPU would be more air-tight than something like PLA or ABS.
Just a fyl. Not sure about other colors, but black polymers are usually pigmented with carbon black. Carbon black is somewhat conductive. Low percentages of carbon black are not meausrably conductive with a normal multimeter but are conductive enough to dissipate high voltage static electricity. 10^5,6,7,8 ohm range depending on pigment load, but your voltages are high...
Just buy a momentary micro switch and replace the latching switch in the pnuematic switch and your're fine. Or set up 2 hose barbs mounted on the exterior of the case, 1 with a momentary and 1 latching switch. Then move the hose between the 2 types of switche as needed for whatever experiment you are doing
What i can suggest is a pneumatic foot pedal valve attached to an air compressor that's linked to the switch so if god forbid you get electrocuted you can let go of the foot pedal and it shutoffs instantly, and also add a time delay with a buzzer to indicate it's starting as a backup i used to use this in my tesla coils where an air compressor and a foot pedal are used to actuate a compact cylinder that's linked to a switch so that you can always make sure it's not turned on and live and the switch is also interlocked with the main switch so that it's not allowed to function if the foot pedal is stuck closed upon startup
Amazing! Also, I built a 50-60kV power converter, and was wondering how you discharge the HV capacitors in the voltage multiplier? Do you just do it manually after every use? For my next version, I've been considering adding a motorized switch which automatically shorts the output when the power converter is off. That way, there aren't any residual charges. Any thoughts?
I'd imagine that it's a combination of the voltage breakdown rating for the plastic, if it even has one, and the microscopic gaps between the layers. I think that SLS might be best printing type and ABS possibly the best type of plastic for this application. However, SLS ceramic would, in my estimation, give the best insulating results.
Could it be that the black PLA uses a mildly conductive material (graphite?) to get the black colour? Black cables have been found to have slightly higher leakage currents sometimes
What about your 3D printed HV Connector? Is that safe? Besides lots of air pockets in FDM printed parts, black PLA is also somewhat conductive due to the pigment being carbon based. I don't know if resin printed parts would perform better for this application. Otherwise make the mould from sheets of PVC. Also for potting, look for dedicated HV Potting Compounds. Perhaps one that is silicone based.
The HV connector is double-walled and filled from the inside with an epoxy resin with a high dielectric strength. The only possibility of a breakdown would be if the high voltage arcs orthogonally to the layers over a distance longer than the spark gap in air. I documented the construction of the connector in my last video on the power supply. Nevertheless, I cannot guarantee that the connector is safe. I'm pretty sure that SLA (Resin) printed parts provide better insulation.
@@AdvancedTinkering Ah yeah, I remember now. Lets hope it holds up. In regards to a interlock on the lid, it doens't hurt to add it and should be fairly simple/cheap to do so. And as others have mentioned, consider fail safe states. I'm looking forward to see the progress and what the supply is made for. Stay safe!
One idea for 3d prints to stand up to high voltage: print a structure designed to be filled with epoxy. In particular, make sure the extruded lines do not touch each other while parallel within a region that is meant to withstand high fields, only at significant angles. This will give a good chance of not having air channels between the lines, because they always make point contacts, and if there are channels they will be short. I don't know if this will work, or if it does how well
That's a very good idea, and exactly the method I used to create the 3D printed HV connector for the power supply ;) The connector and the socket are double-walled and filled with epoxy resin from the inside. If you're interested, I documented it in my last video on the construction of the power supply.
another problem is you're using black plastic for your high voltage housings. Tesla-coilers know to avoid black plastic because it's often tinted with carbon black, and becomes leaky if not conductive at high voltages.
First of all 100% infill doesn't mean it's 100% filled. All those prints suck for high voltage because they always contain defects (trapped particles or gas bubbles). There is a reason XPE for high voltage cables is produced under cleanroom conditions. Defects will cause partial discharges that erode the material. In gas insulated lines there are traps for particles to avoid inhomogeneous e-fields. That brings me to point number second. It seems like everyone online uses random pointy electrodes. Inhomogeneous fields are obviously tougher to insulate. That's why there's a whole topic on how to control the field and avoid any inhomogeneous fields. You can't test anything with randomly shaped electrodes. There are standard testing procedures for gases, liquids and solids including electrode shape and voltage/time steps.
I used an X9C104 as a potentiometer. However, I wasn't really satisfied with the module. There were constantly minor issues, and I never managed to get them to run reliably. This was also the reason why I abandoned the idea of an IR remote control. But it could just have been my lack of skill or a cheap/defective module.
The pigments will indeed reduce the dielectric strength. However, I also experienced a breakdown with white PLA at a voltage that the plastic should have easily withstood if it did not have any defects. So, I'm quite sure that other colors printed with FDM printers may not provide sufficient insulation either. But it seems like some testing would be necessary to confirm.
@@AdvancedTinkering white PLA is more often than not dyed with zinc oxide and its concentration is relatively very high. If insulation is what you're going for, use clear plastic and also print it very hot and very slow to avoid air gaps.
Hey please i want a request to my question because i need a supply to do the electrospening I need a supply for a distance of arc between 15cm to 20 cm how i can do that with 47kv
Instead of a push to make, push to break, toggle switch, you ideally want a latching momentary switch with a solenoid. This way the switch goes back to "off" if the power is removed. Additionally I would suggest a green "safe to turn on" LED instead of an orange "unsafe to turn on" LED so that if the LED fails the system is still safe. Lastly you could add an interlock using a relay so it's only possible to turn on the high power side if the remote is in the off position.
That's coming very nicely. There's always something to learn! As many said the pressure switch is not ideal. Without modifying it you could use a washing machine pressure switch as photonicinduction did for his power supply-> ua-cam.com/video/XJVNT-LkCjA/v-deo.html minute 7.39. It offers all you need (and more) for cheap/free.
@0:12 - Well aren't you quite the underachiever! (I am making a joke here for all the sensitive folks.) 50kv is - mind-bending to have at your fingertips.
There is a safety issue with your LED indicators - they are not failsafe. If the LED fails, then it will indicate it is safe (switch is off) when actually it is not safe (switch is on). You should either invert the LED (so it lights when it is safe/high voltage is off) or add a second (or bicolor) LED which lights when it is safe.
Better yet replace the latching switch, or disable the latching mechanism, so that the plunger has to be pushed in for the device to stay energized, then add a latch to the plunger. Alternatively if there's too much air leakage for that to work, you can add a physical mechanism that is joined to the switch an moves a painted flag inside the case. The most foolproof solution would likely be to have the indicator bulb in series with the switch tho, that way the indicator bulb MUST be on in order for the circuit to be energized.
That’s exactly why I value UA-cam comments so much! They often provide valuable suggestions, important insights, and great ideas.
During the construction, I did consider that the LED could pose a safety risk if it's defective. However, I decided against making the circuit more complex for two reasons:
1. When I turn on the power supply, I press the pneumatic switch to check if the LED is functioning correctly. If it doesn't light up when I press the switch, I know there's an issue.
2. The safety switch shouldn't be activated anyway if the power supply is not in a safe state. So, even if the pneumatic switch is on, it shouldn't be a problem to activate the safety switch.
Nevertheless, you're absolutely right, and it would have been better to make the system fail-safe. I like the idea of a second LED (e.g., green) indicating when the device is in a safe state. I might add that later on.
@AdvancedTinkering the downside of the second bulb is that you then have to carefully consider the driver circuit for it, as it can't have a fail mode which leaves the bulb on. If it fails in the on state, and your caution bulb circuit fails in the off state, you're back to a dangerous condition.
Your toggle idea makes sense, however the point of safety systems is to mitigate human error, i.e. if you're distracted or excited and forget to do it. Another thing about it is that light colors are rather subjective, for instance on the hv tools I work on red (amber really) indicates main power is off (and thus the hv is off) and green indicates power ready (and thus hv is on), however the breakers in the panel have a red flag to indicate energized and a green flag to indicate de-energized haha. You'd need labels.
When I was in uni we did a lab on high voltage/insulation. One of the topics covered was insulation coordination, in particular dealing with insulators made from multiple dielectric materials. When you have an insulator with multiple dielectrics, the electric field is concentrated in the material with the lowest relative permittivity. This means that if your insulator has defects like air bubbles, the electric field field will be concentrated in them, and potentially be high enough to ionise the air, damaging the insulating material and eventually leading to complete breakdown of the insulator.I suspect this is likely to happen with any 3d printed parts, unless you can make them free of air pockets.
I've bean 3d printing functional parts (prototypes and the like) for years, and I do not believe the FDM process is capable of making a part that's free of air bubbles. I spent some time prototyping pump heads, even printed solid they would weep (slowly leak) when pressurized. However, SLA printers definitely are, I think this project would have worked perfectly with some SLA prints.
@@TinkerLynx Printing clear PLA proves that very well. No part is clear but cloudy, unlike a section of spooled filament.
I absolutely agree. I think SLA printed parts would perform a lot better.
I suspect the resin printed insulation would not have the same problem. However we can still improve the performance of printed parts. You can likely try using different materials than PLA but also try using the annealing method that powdered salt. You print at 100% and then place the part in a container filled completely with the finely powdered salt and bake in the oven at the glassing temperature - please look up the exact temperatures for the exact materials before attempting. The part shrinks a tiny amount but becomes completely solid, gets way better mechanical properties and likely also improved performance in this regard.
@@johnmadden9613 I could see the salt causing some issues given that the salt is made of sodium and chlorine ions but I'm not entirely sure about that to be honest
You could also try the salt annealing method. The PLA part is placed into an oven-safe container and it gets filled with fine table salt, which is then tightly compacted around the part. Then it's placed into an oven where the PLA (partially) melts, while the compacted salt keeps the plastic in the original shape. CNC Kitchen did some nice tutorials and testing of this method.
That was actually one of the methods I had considered. I saw CNC Kitchen's video on it a while ago and thought it could make a significant difference. Still, I wouldn't trust FDM-printed components when it comes to high voltage.
Please do the 3d print hv test! I’m planning on making a flyback with printed parts, and it would be very useful. Also, your power supply is the best looking one I’ve ever seen!
Thank you!
If I make a video about different materials, printing parameters, and post-processing methods, I will also test PP. However, I can't promise when I'll get to it. Regardless of the results, I would never trust 3D printed parts as insulation for high voltage. A single defect in the layers is enough to reduce the insulation effectiveness to practically that of air. I would always recommend using epoxy resin with the appropriate dielectric strength as the actual insulation.
@@AdvancedTinkering I wonder if you could submerge the part in epoxy and place it in a vacuum chamber to fill the cracks.
@@RiehlScience Possibly, that could be a method that works, especially with a very thin resin. However, in many cases, it might be simpler to design the parts so that they can be filled with epoxy resin. That's how I made the high-voltage connector for the power supply (shown in my last video on the power supply).
I would be very interested in how SLA-printed components behave compared to FDM-printed ones. Unfortunately, I don't have an SLA printer to test that.
@@RiehlScience Hi, I did that, you can see that bubbles exit at all points where there is infill unterneath.
Even at 100% infill there is still some air that comes out of it.
I would sugest that you use PLA only for the structure, such as outside walls and standoffs, and leave the insulation to the epoxy.
I used very thin PLA standoffs so the epoxy could get into all cracks under vacuum, and it worked out ok up to 30 kV.
@@AdvancedTinkering I wonder how much of the issue with the PLA prints is the filament color. Black plastic should have fairly high carbon content (from the pigment), I would be interesting in seeing performance of different colors.
That pneumatic switch looks like it uses a regular form factor microswitch. You should be able to carefully remove it and replace it with a momentary switch. This is a much safer option as there's still a chance of accidentally energising the the HV supply if the indicator light has blown.
Very good point! It would probably make sense to use a momentary microswitch and keep the air in the tube compressed through another mechanism (a three-way valve, for example) in case I need to run the power supply continuously.
Unless the high voltage supply is going to be fully integrated into whatever it's being used for then some form of "dead man's switch" is absolutely essential. This is especially true if it's going to be used as an interchangeable supply for multiple projects. Basically, if it's possible to energise with the high voltage exposed it should require some form of continuous human input to keep it going.
A magnetic latching switch could be solving some safety concerns, when powering off the device the electromagnet loses its capability to hold the switch "on" and it resets
Its used in planes aswell
I and a few other people were thinking about this fail safe issue, seems like you know the perfect solution!
Thanks for sharing your expertise with that!
That was hilarious at the end. 😂
Love your channel man, keep it up.
I wonder if the chosen color had anything to do with it. To dye something black they usually ad carbon. So maybe try it with a natural filament without dyes.
That being said, I think you're on to something with the air channels, maybe try to over extrude aggressively to minimize air inclusion.
I suspect that different colors perform differently. But since the dielectric strength even of black PLA should be relatively high (assuming no defects), I suspect it is related to small channels between the layers. I experienced a breakthrough with white PLA at a voltage it should have withstood with consistently fused layers. I'll test over-extrusion, thanks for the tip!
@@AdvancedTinkering I saw a demonstration where for a 600V rated cable, the white/red insulation held up to over 1kv but the black broke down just a little above 600V. so there could be some significant difference at high voltage.
@@AdvancedTinkering Maybe getting transparent filament would work well with the over-extrusion: Thomas Sanladerer has made a video on printing transparent (and not just milky-translucent) parts, which also boils down to over-extruding to remove all air bubbles. It would also help to see if a particular part has bubbles or not, without needing to test it with HV :)
Black PLA is indeed conductive to some extent.
7:00 another approach might be to use a resistive divider to divide the DC power supply output to the display so it displays n x 10kV. You might also be able to adjust its trim-pot so that when 10.7V shows up on the measurement line, it reads 20.0
oh wow, that's amazing! looking forward to the application for that much raw power!
Thank you!
@@AdvancedTinkeringIn a DIY/OS Birkeland-Eyde Reactor maybe if you replicate it?
(Granted plenty of projects as-is, but a really well built one would be AMAZING!)
Hey, great video!! I've built a prototype of HV PSU to run an experiment for simple fusor. It has reached only 10-15KV and was already scary as hell. Yours is a great example how to do it the right way:)
It would be really interesting if you can try ABS and nylon for HV parts, but OTOH it's the FDM as technology that will spoil the part by leaving tiniest air gaps. Likely it's doesn't go through the plastic itself. To confirm that a resin printed part with same dimensions is needed.
Another option would be to put the part into very fine salt, tamper it down very tightly, like for sand casting, and re-melt it in oven. The process won't release air bubbles, but will break air traces between extrusions. Also you should randomize perimeter start point to avoid having seam like on your part, to avoid creating a line of air channels.
Aha, is that the Stern-Gerlach setup I spy? 👀 Those were the videos that introduced me to your channel!
Unfortunately no... I feel so bad that I neglected the Stern-Gerlach experiment. But unfortunately, I need a picoampere meter, which I can't afford at the moment. I hope you enjoy the other projects too! At least I can say that the project for which I need the high-voltage power supply fascinates me just as much as the Stern-Gerlach experiment.
Ich kann dir "Corona Dope" von MG Chemicals als Isolationsschicht empfehlen. Auch innerhalb der 3D-gedruckten Teile als zusätzliche Schutzlage. Sehr effizient.
In meinem letzten Video zum Bau den Netzteils zeige und benutze ich Corona Dope zum isolieren der Anschlüsse am Zeilentransformator ;)
Es zum Beschichten von 3D gedruckten Teilen zu verwenden ist eine sehr interessante Idee. Das werde ich testen.
@@AdvancedTinkering Hmm, irgendwie habe ich den Teil im vorherigen Video übersehen. Muss deine watch- und session-time wohl durch nochmaliges anschauen etwas erhöhen. 🙂
War auch nur kurz und fast beiläufig erwähnt. Kann man schnell verpassen :)
Aber ich beschwere mich natürlich nicht, wenn du es erneut schaust ;)
if you carefully dissasemble the switch, you can remove the lever inside that provides the toggle function(push once for on, push again for off), making it momentary
I didn't think about that, thanks for pointing it out! But probably, I would rather buy a new momentary microswitch to be sure. Then I would just need to design a mechanism to keep the air in the tube compressed in case I want to run the power supply continuously.
Try gyroid infill with a missing top or bottom layer so you can then epoxy the infill rather than printing solid. gyroid infill is nice as it doesn't create internal walls, you can pour something in to it and it will fill up evenly.
Don't worry about word pronunciation. We nativw English speakers butcher our own language bad enough. In fact, I knew exactly what you meant when you said it.
BTW, you could use fiber optic ethernet and create a protocol that requires several commands in sequence to activate the power supply.
As an American truck driver, I think you should keep the chrome. It’s good for at least a couple extra foot pounds of torque. I also approve of the addition of chicken lights to indicate that the high voltage side is active.
As an Italian, if you paint it red it will turn on faster.
There is a good chance the color of the PLA will have a massive impact on it's dielectric strength. Black colors will often have carbon black added which is extremely conductive. White pla will have titanium dioxide which is far less conductive. Virgin undyed pla would possibly have the best dielectric strength of any PLA but that's hard to say. I'd love to see a follow-up video testing various 3d printer materials and their corresponding dielectric strengths.
I would try with resin printing. That works really well to form air tight bodies. I bet the arcing problem was not caused by 12mm of PLA, but the dozens of little air channels inside the layers and the infill. Try creating a gas / vacuum tight seal with a filament printer ... it will not work even at 100% infill.
There is no way FDM printing gives you a 100% solid body. Your test demonstrated this: Failure along the seam and along a z layer.
That is also the reason I went against building a PEEK filament printer. Better buy high quality resins and calibrate to your printer, and get really nice dimensional accuracy.
I absolutely agree. Since PLA actually has a relatively high dielectric strength, I'm convinced, that tiny air channels and bubbles inside the plastic (maybe in addition with absorbed water) cause the breakdown.
I'm sure SLA printed parts would perform a lot better. I unfortunately can't test that, since I don't have a SLA printer.
Hey, if you want to turn that pneumatic switch to a momentary switch, there usually is a toggle latch lever that you could removed
I love that you used an RF television remote control to mimic using an IR controller on your device. *cheers*
I'd really like to see how annealing the PLA would help, if you get it hot enough to the point where it's hot enough to reflow and fill any voids.
that breakout arc on the 3d printed part looks interesting
You could use fiber optics to control the whole thing still.
Control the power supply internally with 0-5v signal, and charge a capacitor up to 5v via PWM. Have a discharge resister parallel to the capacitor. So if you run a 50% duty 5v PWM signal in, the capacitor is charged to 2.5v, and this is used as signalling for the power supply. Turn off the PWM and the capacitor is discharged, and the HV is turned off. Have the circuit open a relay below say 5% duty cycle (iunno 0.2v?) to prevemt stray charge in the capacitor from trigering the HV on.
Fiber optics are used in dual resonant tesla coil builds quite frequently, as any interference can cause the bridge to explode. Its just PWM and frequency for a DRSSTC tho not a voltage input.
Anyway, just a thought for you..
Fiber receiver part# HFBR-2412TZ
Fiber transmitter part#
HFBR-1414TZ
The connector type is "ST" or "ST-ST fiber optic cable"
I didn't know that. Thank you very much for the part numbers!
@@AdvancedTinkering no problem 👍
Would you say you`ve changed the final amount of current available, by using a different power supply? always very nicely done project and neat work space ! the new project looks exciting!
i have used PLA parts for hv transformer bobbins and some other stuff. Very interested to see some tests with different materials/slicer settings in the 10-20kv range. wonder if resin prints are better
I always used white 3D prints around HV. Rule of thumb has always been to avoid "Black" items because sometimes the pigments contain carbon. Its the same reason white PVC pipe works well as a "chicken stick" for poking around HV, and black ABS pipe will just conduct it and shock you. The second you make a burn mark on PVC pipe though, that spot is now conductive, and it'll branch out like fractals. Fun stuff. But yeah. Im always surprised how strong circuits are against noise. I currently run my inverter driver exposed ( no metal shielding/case ) a couple feet away from 150KV and it doesn't even care about all the noise coming off that. I've ran arduinos ( TEENSY chips actually ) in metal shielded cans around HV without problems as well, and usually did fiber-optic cables ( TOSLINK ) when I wanted a safe way to remote fire a HV power supply.
Do you use the 3D printed parts as insulation for high voltage? And if so, at what voltage? I also experienced a breakthrough of white PLA during the project at a voltage that a solid piece of PLA should have been able to withstand. That's why I suspect it may be due to small defects in the 3D printed parts. I guess resin prints would perform a lot better.
Good to know that the microcontrollers are relatively resistant to electromagnetic noise. Thanks for the information!
@@AdvancedTinkering I have use white PETG for high voltage standoffs, in which the print is just being used to hold a HV terminal up in the air. Since the print isnt conductive, it doesnt try to snake-down the bushing and arc to ground. But this is basically, If the arc can arc 60mm, I'm running that 60KV through HV wire ( Rated for >75KV ) up through the center of the 3D printed bushing, and at the top of the bushing, I have my contact. That bushing would be much longer than the arcing length, like say 120mm, to simply hold the HV up off the ground. So, in this case its not being a "insulator" in the sense of running a bare wire through it, but I tend to call ceramic high-voltage bushings ( like you see on top of transformers ) Insulators.
If you're running bare wire through the center, and trying to make the plastic print keep that voltage in, then I could see it being hit/miss. Since like you said, impurities, if air-gaps exist they could encourage corona, and potentially break through. Sorta like how a 3D printed PLA vase although it looks solid, can still leak water. If water can leak through the layers, air/corona/electricity likely can too, if I had to guess by that logic.
I guess ideally HV wire is nice for that stuff. But. Maybe another approach could be casting if needed? Like resin-cast somehow? Would work similar to the potting material. I think as you said, Resin-Prints probably would work, since its a solild material with no impurities. Could maybe look into 3D printing a form to cast silicone in. Silicone is often used for HV, such as super flexible high-voltage silicone wire
I would be *very* interested in seeing if another color of PLA might work better with high voltage.
My thinking is that black plastic is very commonly made black with the addition of graphite, so there's potentially some conductive material in the PLA.
That being said, I think the other person who mentioned that the issue might be air pockets ionizing and causing damage to the PLA.
Some black PLA plastics use carbon for coloring. I'd really love to see a comparison between black PLA and white/transparent PLA, and maybe transparent PET-G tossed in there too.
Dude, where are you at? Your content is excellent, the people want more! Lol
Good work. Keep it up
@9:25 🤣 I wondered if you'd realized at some point.
Unfortunately a bit too late :D
I'm extremely interested in various printed material's insulative capabilities. PA12 comes to mind.
I'd like to see vapor smoothed ABS/ASA in a 3d printed test
I made some epoxy covered HV-electrodes, where the housing was 3D printed and also the electrodes with conducting PLA.
When i vacuum degased the thing, it bubbled very long, so even thick PLA has quite some air left in it.
I also would not use black PLA, as they can have carbon particles in it as a coloring, i had much better results with white PLA.
With that it survied 30 KV at 10 mm electrode distance.
Hi! For high voltage 3D print I think is better resin and a very good UV cure.
black materials are sometimes dyed with carbon dust. Try using bright color fillaments. Also, try dipping PLA into the acetone, it cannot dissolve PLA, but it will dissolve surface layer of the plastic, therefore making it airtight, this can improve HV properties, by eleminating microscopic gaps in print
Some in-depth knowledge on 3d printing for high-voltage purposes would be good to know yeah!
The little voltmeter can be hacked to show the output voltage instead of the power supply voltage. If turning the pot doesn't allow it to go up all the way, there should be a resistor that you can change that gets you in the ballpark range.
the carbon black used to make black PLA makes it break down much easier. the same thing happens with black insulated wires. have you tried using white PLA, or just other colors?
I also experienced a breakdown with white PLA filament at a voltage that a flawless, solid piece of PLA should have withstood. However, I didn't compare two identical PLA parts made from white and black filament. That would indeed be very interesting. I agree with you that pigments in black plastics often lead to a lower dielectric strength.
I took a very low tech approach to remote control and bought a Christmas lights controller. It’s basically a radio controlled relay on a pigtail with an outlet. I think the remote is like 49 MHz or something like that. I have had zero issues with interference making the remote unreliable. Just food for thought.
For safety I put a power strip after the remote relay so that I can “arm” the system with the switch or “disarm” it so that if for some reason the remote got activated it would still be off.
People saying “try different colors”: it’s not the plastic that’s breaking down. FFF will always leave air pockets between lines, even at 100% infill. It’s that air that’s breaking down. Even a microscopic air pocket will let the field through. As soon as it does, it gets hot, melts the plastic around it, and just widens the channel.
2:28 all you have to do is add a relay that holds itself closed while powered. If power is lost, the relay drops and you have to press the "power enable" button again to make it hold itself. This is how power tools and machinery have to work nowadays and it just takes a single relay with 2 contacts.
The problem with that method is: how do you connect the "power enable" button to the HV supply without an electrical connection. Which means I would be back at an IR remote. Or I would need a momentary microswitch which is actuated by the pneumatic switch and then actuates the relay.
Or am I missing something?
I believe SLA ceramic should work just fine for HV.
Yes, SLA printed parts should perform a lot better. Unfortunately I do not have a SLA printer to actually test it.
@@AdvancedTinkering We have at work insulation tester up to 6kV AC/DC. Maybe I can print out a test piece 10mm long (PLA/PETG) and measure the resistance/impedance. I have SLA but I dont use it now. Just to know for the future.
Haha 😂 yes, perhaps sktchy to control things with budget microcontrollers. But... Used a few Arduinos to control the main plasma heating of the JET fusion reactor as an "upgrade"
I'm so terrified of HV that just seeing you touch it with both hands makes me shudder. You know more than me though, I just have respect haha.
High voltage is indeed one of the most intimidating things I have worked with. It can behave in ways that are very challenging to predict.
Can you try 3D printing PP (Polypropylene) filament, as it's known to have very good layer adhesion and is used for water tight prints, it would be interesting to see if it can handle high voltage, and if not how does it compare to PLA. It is also known to be hard to print. Very interesting video as always!!
I have used natural PETG for HV, but never at those voltages. I think the dyes in the plastic probably make the problem worse. I would love to see a test of this.
I absolutely agree that its the air pockets in the 3d printed parts causing the electrict breakdown, "My Prusa MK4 is now clearly better!" from CNC kitchen might be the closest I've ever seen to "air bubble free" FDM printing. Honestly actually the whole concept of insulative 3d printed parts seems like something he'd be interested in testing, might be worthwhile to reach out to him.
SLA printing on the other hand probably just works ? I'd absolutely love to see some tests done with it to confirm it because I can't see a world in which it doesn't work
Yes certain 3D plastic is generally very bad especially 60 kv. Cool video!
Try print remelting, it should get rid of most bubbles, and you can use fine calcium carbonate sand as the reflow media, wich wont afect surface resistance
Haha I enjoyed the outro, silent letters really are a pain 😅 Does German have any words with them?
Also I appreciate the sneak peek of the project this was made for. I think that looks a bit like an x-ray tube so my guess is you're going to give x-ray crystallography a try 🤔
I believe there are a few words with silent letters, but it's more of an exception. I can't think of any off the top of my head.
Perhaps. At least, X-ray crystallography would be a very interesting project ;)
@@AdvancedTinkering ah then it must have been the French influence that gave English so many 😁 Ones like knife and knee have always baffled me. If you're not going to pronounce the K why not just spell them without it? 🤷♂️
Haha, I do agree, very interesting indeed. I've always wanted to learn more about it so hopefully someday I'll get to see someone give it a try 😉
Hi, my 1st thought when you said "security switch" was, pneumatic... On the other side, you might want to consider printing a pneumatic security pedal switch, in case you get arc-ed, you fly 1 meter in the air (got arc-ed at 27kv when working on CRT tv long time ago)... Do you intend using this power supply in your vacuum chamber to evaporate tungsten/molibden by ionization? I wonder how plasma looks at 10e-6...
I suspect black PLA could be significantly worse than other colors.
Carbon black is a fairly common black dye, and it being conductive could make a difference especially at those voltages..
You could try re melting the pla case in salt to mitigate the micro air channels. It's a lot of post-processing, though it might not be worth it
I am pretty sure that the problem is the carbon black pigment in the PLA. I think Marco Reps did some testing of white vs. black cables. PLA is rather for high voltage applications actually, have a look at the Blog of Kerry Wong. He measured a dielectric strength of 47,5 kV/mm in an uncompromised part. DOI 10.1088/1757-899X/461/1/012091 is also an interesting read.
Cheers
Yes, I researched the dielectric strength of PLA before building the power supply, and it does indeed seem to be relatively high. However, my practical experiments have shown that 3D printed parts with multiple layers do not come close to reaching these values.
I was already familiar with Kerry D. Wong's blog post. Interestingly, he also uses black PLA. Since the PLA layers for his tests were only 0.18 mm thick, I suspect that there was a very good fusion of the PLA. Once you have many layers, there might eventually be a compromised spot where an arc could form. That's why I suspected defects between the layers.
But it's a very interesting question. I think I will test a few different colors and post-processing methods.
@@AdvancedTinkering The layer-height is very interesting. Usually thicker layers mean stronger parts. Maybe printing hotter will help because it increases fusion? Indeed very interesting
Try the same HV test with a clear or natural colored PLA, the pigments in plastics, especially black and gray, often contain conductive materials like carbon. In general, black and dark gray plastics should be avoided around high voltage because of that. I doubt the other filament types would have much effect on the insulating properties of the parts, although maybe something like PETG or TPU would be more air-tight than something like PLA or ABS.
Just a fyl. Not sure about other colors, but black polymers are usually pigmented with carbon black. Carbon black is somewhat conductive. Low percentages of carbon black are not meausrably conductive with a normal multimeter but are conductive enough to dissipate high voltage static electricity. 10^5,6,7,8 ohm range depending on pigment load, but your voltages are high...
You should try mixing Na K with Hg I've never seen it done before.
Just buy a momentary micro switch and replace the latching switch in the pnuematic switch and your're fine. Or set up 2 hose barbs mounted on the exterior of the case, 1 with a momentary and 1 latching switch. Then move the hose between the 2 types of switche as needed for whatever experiment you are doing
love to see different 3d platic that is anield in salt or over air
What i can suggest is a pneumatic foot pedal valve attached to an air compressor that's linked to the switch so if god forbid you get electrocuted you can let go of the foot pedal and it shutoffs instantly, and also add a time delay with a buzzer to indicate it's starting as a backup
i used to use this in my tesla coils where an air compressor and a foot pedal are used to actuate a compact cylinder that's linked to a switch so that you can always make sure it's not turned on and live and the switch is also interlocked with the main switch so that it's not allowed to function if the foot pedal is stuck closed upon startup
Definitely test more 3d printed stuff
Amazing! Also, I built a 50-60kV power converter, and was wondering how you discharge the HV capacitors in the voltage multiplier? Do you just do it manually after every use? For my next version, I've been considering adding a motorized switch which automatically shorts the output when the power converter is off. That way, there aren't any residual charges. Any thoughts?
PLEASE try PETG! Carbon fiber PLA I think would be obviously bad, but worth a shot. TPU might be good????
I'd imagine that it's a combination of the voltage breakdown rating for the plastic, if it even has one, and the microscopic gaps between the layers. I think that SLS might be best printing type and ABS possibly the best type of plastic for this application. However, SLS ceramic would, in my estimation, give the best insulating results.
Could it be that the black PLA uses a mildly conductive material (graphite?) to get the black colour? Black cables have been found to have slightly higher leakage currents sometimes
What about your 3D printed HV Connector? Is that safe?
Besides lots of air pockets in FDM printed parts, black PLA is also somewhat conductive due to the pigment being carbon based.
I don't know if resin printed parts would perform better for this application. Otherwise make the mould from sheets of PVC.
Also for potting, look for dedicated HV Potting Compounds. Perhaps one that is silicone based.
The HV connector is double-walled and filled from the inside with an epoxy resin with a high dielectric strength. The only possibility of a breakdown would be if the high voltage arcs orthogonally to the layers over a distance longer than the spark gap in air. I documented the construction of the connector in my last video on the power supply. Nevertheless, I cannot guarantee that the connector is safe.
I'm pretty sure that SLA (Resin) printed parts provide better insulation.
@@AdvancedTinkering Ah yeah, I remember now. Lets hope it holds up.
In regards to a interlock on the lid, it doens't hurt to add it and should be fairly simple/cheap to do so.
And as others have mentioned, consider fail safe states.
I'm looking forward to see the progress and what the supply is made for.
Stay safe!
By the way, it's just "newmatic" the p is completely silent. ❤
Ah, just saw the end of the video 😂
Aceton vapor smoothed ABS/ASA might have a slightly better insulating property
One idea for 3d prints to stand up to high voltage: print a structure designed to be filled with epoxy. In particular, make sure the extruded lines do not touch each other while parallel within a region that is meant to withstand high fields, only at significant angles. This will give a good chance of not having air channels between the lines, because they always make point contacts, and if there are channels they will be short. I don't know if this will work, or if it does how well
That's a very good idea, and exactly the method I used to create the 3D printed HV connector for the power supply ;) The connector and the socket are double-walled and filled with epoxy resin from the inside. If you're interested, I documented it in my last video on the construction of the power supply.
@@AdvancedTinkering Haha, I saw it; I just forgot!
another problem is you're using black plastic for your high voltage housings. Tesla-coilers know to avoid black plastic because it's often tinted with carbon black, and becomes leaky if not conductive at high voltages.
First of all 100% infill doesn't mean it's 100% filled. All those prints suck for high voltage because they always contain defects (trapped particles or gas bubbles). There is a reason XPE for high voltage cables is produced under cleanroom conditions. Defects will cause partial discharges that erode the material. In gas insulated lines there are traps for particles to avoid inhomogeneous e-fields. That brings me to point number second. It seems like everyone online uses random pointy electrodes. Inhomogeneous fields are obviously tougher to insulate. That's why there's a whole topic on how to control the field and avoid any inhomogeneous fields. You can't test anything with randomly shaped electrodes. There are standard testing procedures for gases, liquids and solids including electrode shape and voltage/time steps.
Could you try different colour PLA? Black pigment often times is graphite.
I also had a breakdown of white PLA. But I never tested the difference in detail.
I will do that!
I'd suggest replacing your mechanical safety switch with a magnetically held relay to ensure the whole thing fails safely.
I'm looking for a digital potentiometer for a project, what model did you buy?
I used an X9C104 as a potentiometer. However, I wasn't really satisfied with the module. There were constantly minor issues, and I never managed to get them to run reliably. This was also the reason why I abandoned the idea of an IR remote control.
But it could just have been my lack of skill or a cheap/defective module.
I am trying to measure the output current for a similar power supply. Any thoughts on that? Would you plan to add this?
It's not PLA, it's carbon-based black dye.
This stuff is minimally conductive and there's no surprise it conducts 60kv.
The pigments will indeed reduce the dielectric strength. However, I also experienced a breakdown with white PLA at a voltage that the plastic should have easily withstood if it did not have any defects. So, I'm quite sure that other colors printed with FDM printers may not provide sufficient insulation either. But it seems like some testing would be necessary to confirm.
@@AdvancedTinkering white PLA is more often than not dyed with zinc oxide and its concentration is relatively very high.
If insulation is what you're going for, use clear plastic and also print it very hot and very slow to avoid air gaps.
9:35 Yeah English is like that.
Magnetron sputtering 💯
Have you tested other colors of PLA? Black migh be carbon-dyed...
Add lables PLEASE
I don't know why I completely forgot about this, even though I have a label printer. I will do that right away!
PETG? with a concentric infill?
Hey please i want a request to my question because i need a supply to do the electrospening I need a supply for a distance of arc between 15cm to 20 cm how i can do that with 47kv
test glass fiber filled nylon filament
I would never use Black anything for HV. Black usually means carbon filler. always use white or clear.
I imagine resin printed parts would have less air channels, and better insulation.
Nanovia makes a high voltage filament but it's expensive $100 for 1 pound.
Instead of a push to make, push to break, toggle switch, you ideally want a latching momentary switch with a solenoid. This way the switch goes back to "off" if the power is removed.
Additionally I would suggest a green "safe to turn on" LED instead of an orange "unsafe to turn on" LED so that if the LED fails the system is still safe.
Lastly you could add an interlock using a relay so it's only possible to turn on the high power side if the remote is in the off position.
Those are great suggestions! Thanks!
@@AdvancedTinkering no problemo, used to work on HV laser power supplies in a past life
Are you building a fusor?
Is that a microfocus tube, or are you just happy to see me? :D
It is not the plastic, it is the color. Black is carbon, which conducts electricity. Try white PLA.
in my opinion never use pla xD i started after the first roll pla to petg as basic standard and its alway the better case after years^^
Nifty
Is this outputting 60 kv DC or AC?
DC with quite some ripple since it is the output of the voltage multiplier.
That's coming very nicely. There's always something to learn! As many said the pressure switch is not ideal. Without modifying it you could use a washing machine pressure switch as photonicinduction did for his power supply-> ua-cam.com/video/XJVNT-LkCjA/v-deo.html minute 7.39. It offers all you need (and more) for cheap/free.
9:12 EDM?
this video felt violent, especially the noises
@0:12 - Well aren't you quite the underachiever! (I am making a joke here for all the sensitive folks.) 50kv is - mind-bending to have at your fingertips.
Aww, your English is cute! It's "new-matic" the "P" is silent, because English is a nightmare, not logical, and exists to frustrate everyone.
9:35
Nice, but be careful with your ears.