Hey WizardTim. I work for a company in the United States that makes high voltage diodes - quite similar to this one. To give you a bit of feedback on the forward voltage drop, two main points: 1) HV diodes tend not to be a typical PN junction since they hold off about 1kV per wafer, they therefore are thicker and have a different resistance profile because of it. This makes each diode wafer usually more in the 1V - 2V per wafer range, though this does depend some on the reverse recovery rating of the diode. 2) Keep in mind that the rated VF is the absolute maximum for the diode, so there is likely some margin there. For instance, on a similar diode we manufacture, a 45V VF rating usually means the diode tests (on the bench) more in the 35V range. Feel free to comment back if you want more details or have any other questions about high voltage diodes or high voltage diode manufacturing. I would be happy to provide more information.
@WizardTim
2 жыл бұрын
Thanks for the info! Since posting this video I measured the diode with a PSU at some currents: 15.5 V @ 1 mA (0.70 V/die) 21.2 V @ 5 mA (0.96 V/die) 24.8 V @ 10 mA (1.12 V/die) So as you say they’re just rather generous with their max VF rating. This is kinda surprising to me as the power dissipation will be almost half, but at 20 kV @ 10 mA (200 W), 0.2 W is 0.1% so I guess it doesn’t matter that much.
@danield5939
2 жыл бұрын
@@WizardTim Yeah, it is a little surprising that it was about half of what their rating was for the parts. Looking at the datasheet that google provided, I don't see any graphs for VF vs IF which is what I usually point people towards instead of the max of the part. For our diodes, we have a VF rating, but if you look at a typical VF vs IF graph, you can see that at the rated current, the diode typically has about 75% of the VF rating as voltage drop. You are right about the 0.2W not being much with regards to the overall circuit (20kV and 10mA), but for a device small like this, 0.2W of power dissipation could be significant depending on your heat path. We have a similar part rated at 16kV (datasheet here if you are interested: voltagemultipliers.com/wp-content/uploads/2018/10/M25UFG_M50UFG_M100UFG_M160UFG.pdf) that could have internal heating of 13°C at 0.2W dissipation. Though that isn't much, it will change the IR, VF, and TRR characteristics of the diode. Anyhow, food for thought. Good video. We cross section parts sometimes in a similar way to look for damage of our diodes, but I am not patient enough to do it by hand so we have a system for sanding and polishing.
@poormanselectronicsbench2021
2 жыл бұрын
Great view of the inner construction! It's funny, if you read descriptions on using diodes, they often say NOT to string diodes in series to raise voltage ratings for numerous reasons, usually describing a cascading failure scenario due to one junction breaking down earlier and passing the current to a weaker one, BUT, that is the exact way they construct high voltage diodes as shown by your careful dissection and viewing. Granted, more care in matching the forward breakdown junction is most likely put into these, but there still will be variables that will make no construction a "perfect balance" throughout its manufacture.
@teresashinkansen9402
6 ай бұрын
Almost all HV "diodes" are made like this, you should see the diode strings used in megavolt powers supplies for particle accelerators like dynamitrons, its the same just thousands of them. I think the key for successful series diodes is to greatly under rate it compared to the single diode specs. Say a single diode has the specs of 1000V 1A peak and a recovery time of 20ns but you want a diode string to rectify say 10kv 100ma so you put like 30 of those diodes in series, that could theoretically yield a 30Kv 1A diode string but you only rate it as 15Kv peak 500ma 100ns recovery time.
@hullinstruments
Жыл бұрын
Man oh man I hope you make more videos someday! Your channel is top-notch surprised I just found you
@hullinstruments
Жыл бұрын
If you ever need help or have questions regarding resin and acrylic type stuff... Feel free to hit me up as I buy it by the 55 gallon drum. Usually over a dozen drums per year according to how busy orders get It's For use in making synthetic mother-of-pearl, Ivory, and other materials for the gun, knife, guitar, and many other industries.
@sabujpattanayek8799
3 жыл бұрын
someone get this man a sanding belt
@joels7605
2 жыл бұрын
Great video. You should buy a combo belt and disc sander. You can get them fairly cheap from your local gettin' spot. Even a low end one will last you a lifetime. Might save some skin on your knuckles.
@hernancoronel
Жыл бұрын
At 5:50 “Thank you KZitem” for ruining the final close up shots with overlays of other videos to sell. Thanks for the video Tim!
@WizardTim
Жыл бұрын
Good call, I changed it so the end screen only appears for the minimum amount of time now (5 seconds). I wish KZitem would let you turn them off in the player like back when you used to be able to turn off the old annotations. Many channels just "apply to all" the same end screen which can ruin the end of older videos from before end screens.
@signalworks
3 жыл бұрын
Excellent work. The amount of effort you put into this is amazing. (I posted this to HN last night, seems like it got into the second chance pool!)
@WizardTim
3 жыл бұрын
Thanks for posting it there, KZitem says 43.9% of views came from Hacker News. Really makes me reconsider how important posting on other platforms is for discoverability. I decided to Tweet the video from a new Twitter account to see if anyone finds it and what becomes of it in future. Also normally it takes the Algorithm™ a couple days to start sharing videos on this channel but this time it was put into peoples feeds instantly, not sure if there's any correlation with what I did differently this time or if the Algorithm™ just felt like it. Luckily this channel is just a hobby for fun so I don't have to rely on it for much although seeing it succeed certainly justifies the time spent on it.
@hullinstruments
Жыл бұрын
Also you can get dirt cheap diamond discs from China. For use in lapidary industry and elsewhere. Usually diamond lapidary discs can be hundreds of dollars each but the cheap ones from China are only a few dollars each and go up to crazy fine grits. From there you can even get diamond paste that goes up to 0.25 micron which is about sixty thousand grit I believe. Whereas the individual grit Pace would cost 20 or $30 for a tiny amount here... The knockoff versions from China or like $10 for the whole set of all the grits and indecent quantities. You could get half a dozen or so different grit discs for really cheap and they cut through ceramic, silicon, glass, whatever and leave behind a great finish. It's super fast and easy compared to sandpaper and can even be used with water or other fluid coolants.
@matt15d0p3
2 жыл бұрын
Dude. Your videos are awesome
@jaro6985
3 жыл бұрын
Just seeing the 1N4001 is really cool.
@h-primehiro1044
Жыл бұрын
Thank you thank you thank you very much nice job nice video i love that's
@arenaengineering8070
3 жыл бұрын
100ns switching time - may be it is silicon carbide inside. Silicon carbide single diode have an approx 2 V drop voltage. 2v * 22 crystal = 44 v drop voltage.
@WizardTim
3 жыл бұрын
I don’t think it is that as I didn’t pay much for them and the manufacturer doesn’t mention SiC which would be an important marketing point. Also, from what I understand SiC diodes are more suited for fast switching high power devices (similar to GaN components) like computer power supplies where switching losses account for a large amount of the losses. So far the best reason I can come up with is maybe they just use whatever dies they have lying around at the time to get to a 20 kV breakdown, just so happens I got one with 22x 1,000 V dies but maybe 40x 600 V dies is a possible configuration which would give 45 V forward voltage drop? Measurements I got (not very accurate just with a bench PSU): 15.5 V @ 1 mA (0.70 V/die) 21.2 V @ 5 mA (0.96 V/die) 24.8 V @ 10 mA (1.12 V/die)
@arenaengineering8070
3 жыл бұрын
@@WizardTim in hight voltage (1000v) hight current rectifier made in silicon carbide is hight cost. But in low current device's silicon carbide is ordinary material. Like as some LED contain silicon carbide crystal inside, but not information about this in datasheet for user.
@danield5939
2 жыл бұрын
@@WizardTim I agree, it is unlikely to be SiC. SiC is more expensive, and from what I understand it does not stack well enough (at this point) to be able to do high voltage this way. If it was SiC, the vendor would have advertised that since it is very much a selling point.
@taliesinbeynon
3 жыл бұрын
How much could this be automated? I'm imagining building a volumetric 3D image by photographing it at regular sanding intervals. That could be quite fascinating to see...
@WizardTim
3 жыл бұрын
I agree that would be super cool. I remember watching a KZitem video a few years ago where someone mentioned that idea and showed a prototype of a device to do exactly that which was fully automatic, however I don't believe they got it working well enough to reconstruct a volume. I have a feeling it was 'electronupdate' but I can't find the video. However 'mikeselectricstuff' showed you can do it on a larger scale with CNC cross-sectioning: kzitem.info/news/bejne/s36ntGp5jYGVrXY In industry you can also do non-destructive μCT scans of components and get very good resolution, unfortunately getting time on those machines is difficult and expensive.
@vnagaravi
2 жыл бұрын
Or you can cast all parts exactly same size and sand half of the part by setting distance limit and stop the machine at that point
@vdkaa
3 жыл бұрын
Maybe a small benchtop belt sander would be a good investment for the bulk of sanding this down?
@WizardTim
3 жыл бұрын
I do have access to a belt sander however I’m using an ‘ultra-clear’ epoxy designed for art applications so it has a glass transition at about 50 - 60 °C so I probably couldn’t go much faster but it would take a lot of the physical effort out of it. But I would also have to make the puck larger so I don’t sand off my fingers. I might give it a try if I do something larger. For this video it was about 15 min of prep for the epoxy, a couple hours to cure, 30 min of sanding and 30 min of cleaning up (excluding all the added time filming).
@RixtronixLAB
2 жыл бұрын
Cool video, thanks :)
@mernok2001
2 жыл бұрын
Nice video,I think microwave diodes are similar to this in their structure.I have used microwave diodes for several 100kHz to rectify a Tesla coil output without any problems.
@picanazo420
2 жыл бұрын
is it possible to use microwave diode for high frecuency? even that the datasheet says it works for 60hz?
@NiHaoMike64
Жыл бұрын
@@picanazo420 Many modern microwave ovens use switching power supplies, it's possible newer diodes are faster to be able to work for those.
@andr27
3 жыл бұрын
Nice done
@EvanCox10
3 жыл бұрын
Pretty cool, hope you are wearing some good respiratory protection though, like N100. Wouldn’t want to inhale any of the dust from the sanding, potentially lots of nasty stuff in there.
@WizardTim
3 жыл бұрын
I wasn't wearing a respirator but I did have a fan with dust filter sucking away some of the dust. I also have several laser dust sensors and the readings weren't too bad. My neighbor's wood smoker is worse.
@leonardmilcin7798
3 жыл бұрын
@@WizardTim It is not about the dust but rather about rare metals and crystals that are present there.
@Doctorlockpick
5 ай бұрын
What brand uv epoxy did you use?
@WizardTim
5 ай бұрын
Unfortunately I don't remember the exact brand I used in this video, but it's at least not a UV resin. It's an "ultra clear deep cast epoxy resin" which comes in two parts, the advantage of this type is it isn't very viscous so it flows easy around the part and can be used in reasonably large moulds without overheating. The downside is the ratio of the two parts is very sensitive and it takes several hours to cure.
@DudemcmanStudios
3 жыл бұрын
Would this technique work for looking at the internals of an IC?
@WizardTim
3 жыл бұрын
Yes it’s perfect for analysing the lead frame and die placement but I think it would be rather difficult to look at the actual circuit on the die as you’d have to sand down to the die perfectly flat and parallel and not sand too far, I might give it a go and see if it works but heat or acid would probably be easier.
Пікірлер: 38