Chriull Posted July 8, 2019 Author Share Posted July 8, 2019 (edited) 4 hours ago, Arbolest said: Also, from the way it appears in the photo it seemsĀ like it might have actually come OUTĀ of the packages of the blown MOSFETs, which means it almost certainly happened post-failure Would be my best guess to. A bit could be the "glue" or whatever reflecting the flash, but some pieces definitely look like molten metal. This should come from the ?inside of the mosfets?, the vaporized mosfet leg and some vaporized solder. So all "post mortem stuff" 4 hours ago, Arbolest said: That all makes sense to me and tells me why and how the MOSFETs might have failed, but not the underlying reason, It's overheating by beeing thermaly insulated from the heatsink. @Phil McLaughlin made a nice test series shown in his youtube videos (i put a link to both in the first post of this topic). Without a proper cooling this Mosfets can just disspate a couple of Watts without melting - while EUC operation they have to disspate more something like 20-30 Watts, maybe even a bit more. 2 hours ago, Planemo said: I always thought that the casing of the mosfet was a common ground (also linked to the centre pin) so it would make no odds if the metal back came into contact with the heatsink, ground, negative or whatever. Also the reason why the fixing bolt doesnt need to be insulated either. Thats certainly how I have always wired fets. Maybe GW wire them differently... The mosfet drain (pin 2) is mostly connected to the metal plate of the mosfet. Since in EUCs the mosfets are used for 3 H-Bridges (for the three motor coils) the drains of the "upper" mosfets are at battery voltage, the drains of the "lower" mosfets are connected to the corresponding motor coil. So different potentials and a massive short if not insulated from the heatsink. Even if all were at battery voltage one would want to insulate the 100V(84V,67.2V) from the heatsink.... Edited July 8, 2019 by Chriull 3 Quote Link to comment Share on other sites More sharing options...
Arbolest Posted July 8, 2019 Share Posted July 8, 2019 (edited) 30 minutes ago, Chriull said: It's overheating by beeing thermaly insulated from the heatsink. @Phil McLaughlin made a nice test series shown in his youtube videos (i put a link to both in the first post of this topic). Without a proper cooling this Mosfets can just disspate a couple of Watts without melting - while EUC operation they have to disspate more something like 20-30 Watts, maybe even a bit more. Sorry, maybe I should have clarified here. I know that the MOSFETs on the Nikola boards are failing due to over-heating because of the hot glue used to hold the thermal pads in place during assembly, but what I meant by "the underlying reason" in that postĀ was more like "the root cause of this particular failure". I'm interested in what caused the MOSFETs on this board to blowĀ (which uses the 247 package MOSFETs, unlike the 220s on the other Nikola boards) because based on the poster's description of the incident it doesn't seem like it had anything to do with themĀ overheating. It sounds like one or both of themĀ shorted. Edited July 8, 2019 by Arbolest 1 1 Quote Link to comment Share on other sites More sharing options...
Chriull Posted July 8, 2019 Author Share Posted July 8, 2019 40 minutes ago, Arbolest said: Sorry, maybe I should have clarified here. I know that the MOSFETs on the Nikola boards are failing due to over-heating because of the hot glue used to hold the thermal pads in place during assembly, but what I meant by "the underlying reason" in that postĀ was more like "the root cause of this particular failure". I'm interested in what caused the MOSFETs on this board to blowĀ (which uses the 247 package MOSFETs, unlike the 220s on the other Nikola boards) because based on the poster's description of the incident it doesn't seem like it had anything to do with themĀ overheating. It sounds like one or both of themĀ shorted. Ups - sorry too from my side - i did just look at the pictures and not get, that is a new case of a new replacement board.... Maybe one can see something once @Jim Martin posts the pictures of the heatsink and the thermal pads... Somehow the mosfets have "something" on them - one sees this "pointy pattern" reflecting from the flash, most "again" on the blown mosfets?! - or is this just that the had the best angle for reflecting? Ā 1 1 Quote Link to comment Share on other sites More sharing options...
Planemo Posted July 8, 2019 Share Posted July 8, 2019 1 hour ago, Chriull said: The mosfet drain (pin 2) is mostly connected to the metal plate of the mosfet. Since in EUCs the mosfets are used for 3 H-Bridges (for the three motor coils) the drains of the "upper" mosfets are at battery voltage, the drains of the "lower" mosfets are connected to the corresponding motor coil. So different potentials and a massive short if not insulated from the heatsink. Even if all were at battery voltage one would want to insulate the 100V(84V,67.2V) from the heatsink.... Thanks for thatĀ explanation, helps a lot.Ā I have never used fets in either the quantities or voltages that eucs work at. 1 Quote Link to comment Share on other sites More sharing options...
Jim Martin Posted July 8, 2019 Share Posted July 8, 2019 Here is a picture of the thermal pad. 2 Quote Link to comment Share on other sites More sharing options...
Jim Martin Posted July 8, 2019 Share Posted July 8, 2019 10 hours ago, Marty Backe said: That really sucks. This highlights the fact that I don't believe Gotway tests the boards in any meaningful way. If they did my fan would be working and your board wouldn't have fried. Now on the other-hand, boards do fail and have been known to fail when the wheel is first turned on. KingSong, Gotway, Ninebot, etc. It happens. You just got unlucky. My board didn't even have the jumper wire so no way they could have tested it.Ā I had to do some research to make sure it needed the jumper wire and had to create my own (as seen in the image).Ā Well, I will see what Jason says, hopefully he can get me another one to try.Ā I guess I'm down for a couple more weeks.Ā I've had this wheel since day 1 and ridden it once 1 1 Quote Link to comment Share on other sites More sharing options...
Popular Post pico Posted July 8, 2019 Popular Post Share Posted July 8, 2019 (edited) 11 hours ago, Marty Backe said: That really sucks. This highlights the fact that I don't believe Gotway tests the boards in any meaningful way. In my daysĀ Ā QC PASSED meant: burn-in at maximum load at maximum operating temperature for 48 hours. Ahhh... progress... Edited July 8, 2019 by pico 3 1 Quote Link to comment Share on other sites More sharing options...
Marty Backe Posted July 8, 2019 Share Posted July 8, 2019 26 minutes ago, pico said: In my daysĀ Ā QC PASSED meant: burn-in at maximum load at maximum operating temperature for 48 hours. Ahhh... progress... QC stickers are meaningless when it comes to Gotway. I'd love to how they test the boards and/or wheels before shipping. I doubt there's anyĀ burn-in involved. 1 Quote Link to comment Share on other sites More sharing options...
Popular Post esaj Posted July 8, 2019 Popular Post Share Posted July 8, 2019 29 minutes ago, pico said: In my daysĀ Ā QC PASSED meant: burn-in at maximum load at maximum operating temperature for 48 hours. Ahhh... progress... "Agile development" 5 Quote Link to comment Share on other sites More sharing options...
Eliran Posted July 8, 2019 Share Posted July 8, 2019 I also sell Rolexes and other hi end watches as a hobby. Rolex (for exemple), tests their watches for 1 (ONE) year each watch before they go to the streets. Then you'll ask me: bu6r Rolex makes very few watches a year, so, they can have a very detailed and afficient QC. And I'll answer, No, they make around 800.000 watches per year.Ā Quote Link to comment Share on other sites More sharing options...
Patton250 Posted July 8, 2019 Share Posted July 8, 2019 4 minutes ago, Eliran said: I also sell Rolexes and other hi end watches as a hobby. Rolex (for exemple), tests their watches for 1 (ONE) year each watch before they go to the streets. Then you'll ask me: bu6r Rolex makes very few watches a year, so, they can have a very detailed and afficient QC. And I'll answer, No, they make around 800.000 watches per year.Ā Ā Rolex makes more watches than anyone else every year. Ā About 1.5 million. ToĀ put that in perspective AP makes 35,000 watches the year. ALS makes 5000. Patek Ā around 75,000.Ā Quote Link to comment Share on other sites More sharing options...
Eliran Posted July 8, 2019 Share Posted July 8, 2019 (edited) 21 minutes ago, Patton250 said: Ā Rolex makes more watches than anyone else every year. Ā About 1.5 million. ToĀ put that in perspective AP makes 35,000 watches the year. ALS makes 5000. Patek Ā around 75,000.Ā Perfect.Ā So, I think we have other things in commom , Edited July 8, 2019 by Eliran One lacking word 1 Quote Link to comment Share on other sites More sharing options...
Patton250 Posted July 8, 2019 Share Posted July 8, 2019 3 minutes ago, Eliran said: Perfect.Ā So, I think we have other things in commom , Ā Yes. Itās an incredibly expensive hobby though. LOLĀ Ā Quote Link to comment Share on other sites More sharing options...
Smoother Posted July 8, 2019 Share Posted July 8, 2019 (edited) 7 hours ago, Arbolest said: Ā I was examiningĀ the high resolution versions of the photos and noticed some weird silvery stuff around the left two MOSFETs on the upper row in the first picture.Ā Well spotted my vote is glue reflecting light as mentioned above byĀ @Chriull But since you forced me to examine the picture more closely, it does look like all of them have some amount of this woven pattern effect; but none as much as on the two blown fets.Ā It might just be possible that the heat caused the transfer of material (more where hottest) rather than the other way round as we are speculating. And the materiel is simply the insulation sheet partly decomposing from being outside its design temperature range. Edited July 8, 2019 by Smoother 1 Quote Link to comment Share on other sites More sharing options...
Popular Post mrelwood Posted July 8, 2019 Popular Post Share Posted July 8, 2019 In the photos it looks like there are remains from molten solder, but Iām not sure wether there is any on the underside of the mosfet. The plastic shell of the mosfet can melt though.Ā But besides that, all mosfets have something glossy on them. If the thermal pads are of the sticky sort, I would think that there is not enough adhesive to stick to the mosfet surface. And since the wheel was powered on for only a few seconds and not stressed, I donāt think that the other mosfets had time to get very hot. Silicone doesnāt reallyĀ melt. I donāt know what kind of temperature would be required forĀ it to break down, but at least a blue butane lighterĀ flame doesnāt do much at all. Iām tempted to call this a short. Wether it is from a (unnecessary) thermal paste, a blob of solder, an untrimmed or deformed component leg, or something else, I donāt know. What I do know is that Gotway is having a humongouslyĀ hard time to fix an issue that shouldnāt have happened in the first place. 1 7 Quote Link to comment Share on other sites More sharing options...
Popular Post Nils Posted July 8, 2019 Popular Post Share Posted July 8, 2019 18 hours ago, Jim Martin said: I don't really know WTF to trust anymore with this damn NIKOLA.Ā So got the new board installed, powered it up, running great, fan kicks in and I'm thinking awesome wheel is working great and my fan works.Ā I'm rolling it back and forth and SNAP, CRACKLE, POP!Ā Wheel shuts down and what is that SMELL!Ā Did this new board just blow some MOSFETS?Ā Surely not!Ā I took the board out and opened it up.Ā Damn, something really bad happened! Oh man, sorry for your troubles The manufacturing problems really seem to line up, don't they? So we had glue gate, and while one instance do not a gate make, are we seeing the beginning of a conductive thermal paste gate or something? Hopefully not, but there are obviously some things to address in the assembly chain for Gotway, and the last thing they need at the moment considering the issues so far. Once again it goes to show though that design is one thing and manufacturing/assembly another. The use of smaller MOSFETs was one concerning issue when the original Nikola problems arose, but actual problems so far have been due to the incorrect use of glue while assembling. While we don't have any definitive answers as to much the TO-220 can handle, this at least goes to show that bigger MOSFETs is not a guarantee for anything in the face of other issues such as bad assembly. My own replacement board with the TO-247 MOSFETs should be in shipping now, but quite frankly I'm really not that eager to put it in when it arrives. It's always nice to have a board ready if something should happen, but better to have one working board than ten unproven ones. Regardless, at this point, I'd not put the new board in without first opening it up and inspecting the thermal pads. 2 2 Quote Link to comment Share on other sites More sharing options...
Lukas83 Posted July 8, 2019 Share Posted July 8, 2019 19 hours ago, Rywokast said: what in the F#CK!?!??!!! why is it missing screws? did you remove those?? are you planning to take it apart and inspect it? this is extremely disappointing........ eehm... my guess is that the big brown capacitor made the short directly on the pins of the mosfets. The capacitor is only coated and under the coat they are electrically conductive. This would explain why the both outa mosfets blew up. Look at the pictures... Quote Link to comment Share on other sites More sharing options...
wheelr Posted July 8, 2019 Share Posted July 8, 2019 (edited) 1 hour ago, Lukas83 said: eehm... my guess is that the big brown capacitor made the short directly on the pins of the mosfets. The capacitor is only coated and under the coat they are electrically conductive. This would explain why the both outa mosfets blew up. Look at the pictures... I don't know, the capacitor looks intact. Though if you lookĀ on the side where the capacitor is connected, one can observe an irregular pattern in the board. I'm not sure if this isĀ part of its design or a defective PCB. Edited July 8, 2019 by wheelr 1 Quote Link to comment Share on other sites More sharing options...
Mrd777 Posted July 9, 2019 Share Posted July 9, 2019 All in all Iām still (somehow) really trusting my 100volt Nikola. Ā I ordered the speaker mod, so I will open it up this week and also examine my board as I replace the speaker. Ā I will remind everyone that this wheel is truly a blast, and the acceleration is fabulous, so letās not give up on the Nikola just yet. Gotway will work this out. 3 Quote Link to comment Share on other sites More sharing options...
Popular Post Marty Backe Posted July 9, 2019 Popular Post Share Posted July 9, 2019 12 minutes ago, Mrd777 said: All in all Iām still (somehow) really trusting my 100volt Nikola. Ā I ordered the speaker mod, so I will open it up this week and also examine my board as I replace the speaker. Ā I will remind everyone that this wheel is truly a blast, and the acceleration is fabulous, so letās not give up on the Nikola just yet. Gotway will work this out. Jason is still missing in action and my wheel continues to be sidelined. Frustrating to see the various owners having a blast with theirs while mine collects dust. 2 3 Quote Link to comment Share on other sites More sharing options...
Mrd777 Posted July 9, 2019 Share Posted July 9, 2019 6 minutes ago, Marty Backe said: Jason is still missing in action and my wheel continues to be sidelined. Frustrating to see the various owners having a blast with theirs while mine collects dust. Hmmm. Iām sorry to hear this Marty, I thought Jason was on it.. 2 Quote Link to comment Share on other sites More sharing options...
Marty Backe Posted July 9, 2019 Share Posted July 9, 2019 2 minutes ago, Mrd777 said: Hmmm. Iām sorry to hear this Marty, I thought Jason was on it.. Last I heard from him was July 3rd. I think he's so busy keeping EWheels running that he's not keeping it running, if you know what I meanĀ Ā It's been weeks since he got his big wheel shipments in but his e-mail auto-reply makes it sound like they just came in. When you're running a business but have to resort to e-mail auto-replies, you know that you're losing the battle. Let's hope it turns around for himĀ But it's not his fault that I'm having the Gotway problems. It'll get resolved one of these weeks ... Good to hear that your wheel is still spinning and you're still lovin itĀ 2 Quote Link to comment Share on other sites More sharing options...
Lilvodka Posted July 9, 2019 Share Posted July 9, 2019 Excuse my ignorance of electronics. Ā WhenĀ I fit my new board, is there some way I can test it without riding - leaving it on all day leaned up against a wall for instance? Quote Link to comment Share on other sites More sharing options...
atdlzpae Posted July 9, 2019 Share Posted July 9, 2019 That would be one way to do it... But I think that the best way would be to gear up and limit yourself to 10km/h for your first ride so that you can jump off if it cuts out. Quote Link to comment Share on other sites More sharing options...
pico Posted July 9, 2019 Share Posted July 9, 2019 (edited) 43 minutes ago, Topkek said: Excuse my ignorance of electronics. Ā WhenĀ I fit my new board, is there some way I can test it without riding - leaving it on all day leaned up against a wall for instance? Ā The way the Kingsong testersĀ do it. Between two ramps. Making violent tic tocs. You hold onto the ramps... and use wheelog to check the temperature... Use Wheelog to check the temps. If you hear a warning at low temperature, say before 50 deg C (correct my valueĀ those that have the wheel), you STOP. Not the video I was thinking about, but you get the idea of the ramp... Ā Edited July 9, 2019 by pico 2 1 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.