$4000 Electric Car Inverter Teardown.

[@electrifyingvids3545]

Seems like a decent inverter. You can see where corners were cut to save cost, but over all, it seems to be a well designed inverter. Nothing too fancy, just something that can get the job done.

[@Noughtta]

Im curious if the foam provides enough protection against mechanical damage to them, My only gripe is the lack of silicone or at least hot glue on the caps but it otherwise looks decent.

[@D7UN4Channel]

What about the hot temp ?​@@Noughtta

[@Noughtta]

@@D7UN4Channel silicone RTV is extremely temperature stable and will last for years. It’s what’s used inside the industrial inverters I work on. So it seemed a bit odd to me to not see any.

[@D7UN4Channel]

@@Noughtta I see..

[@pootispiker2866]

@@Noughtta They didn't cut corners by not copying designs and leaving out bits that seem important.

[@meatwalker]

I wish all the teardowns been like that! Only correct and clear information, no blah-blah talking 👍

[@uiopuiop3472]

only in ohio-oh, only in ohio-oh now there comes my rap shong abot it: rewrote by mom Im gotta rappin like Ohio, spitting tracks like Ohio Getting your girl like Ohio, call me the rizzler in Ohio On the mic, Im a killer, flow so cold like Ohio Call me the rizzler in Ohio, watch me fly, O-O

[@andyandy6739]

本人吹き替えなのがすごくいい。しかも今まで人に吹き替えてもらってたのが謎になるレベルに流暢。

[@xxportalxx.]

Maybe it was cheap enough to have someone else do it, and he was just busy/lazy. I wouldn't blame him lol

[@BlackXeno]

@@xxportalxx. I think was an AI generated voice, text to speech.

[@thecatofnineswords]

@@xxportalxx. If you have a big enough channel, and sufficient income, it's good to share the proceeds - supporting fellow creatives.

[@Kaizzer]

So, is this really himself? I wish I were so fluent! At first the voice sounded like that of @Asianometry :O

[@picobyte]

I'm still amazed by how much current those tiny 'transistors' can handle. Not only in this tiny converter but also in our grid, running at 800kV DC/AC 380kV bidirectional converting at kiloamps.

[@frommarkham424]

Semiconductors are awesome

[@Rubacava_]

There are even much more efficient mosfets than the ones used here. Mosfets might be the greatest human invention

[@picobyte]

@@Rubacava_ The market for high voltage is very volatile. No one talks.

[@omniyambot9876]

and IGBT's eats current on steroids

[@mikehughesdesigns]

Problems start when designers take "Absolute Maximum Rating" values as a target...

[@chrisw1462]

@12:00 It's pretty simple. The buss bars are the DC supply to the inverter MOSFETS. The two thinner ones are the Negative supply. The positive supply only has One set of buss bars, so they are double the thickness.

[@JohannesChan-s8v]

The mosfets wired in parrelel actually should share current decently well because parrelel mosfets are naturally self balancing - if one heats up more than the others it will exhibit a higher conduction resistance, so more current will be off-loaded to the other mosfets.

[@whyargon]

facts

[@hanslepoeter5167]

Thanks. Some comments : It is in general possible to place power mosfets in parallel and many power mosfets have an internal structure based on fets in parallel. Furthermore rdson has a positive temperature coefficient so when one heats up the cooler ones process most of the current. Any series resistors help to maintain the balance and that might be the reason for the smaller busbar. As for the electrolytic capacitors, lifespan is about halved for each 10 degree celcius up in temperature ( rule of thumb ... might apply more or less ). I guess that's a very limitting factor for the life span of this inverter. The film capacitor is a way better option. Great video .. make more .....

[@pongpitwipasuramonton4092]

Great for me that you have English-speaking version. I guess the pink sheet attached on one side of the rubber is a thermal interface material. It's used for sharing heat between caps and for easier removing heat out from the caps, to lengthen the caps' life.

[@zaprodk]

Spot on.

[@marcoaurelio4903]

There is a reason for many US brand components in this inverter. If this board is an early revision, they make small batches, so relatively low volume western components in china are abt. the same price. Chinese 1st party sellers dont like to sell less than around 10~20k. Some manufactures start with small batches with western components and change te BOM to chinese made ones when scale up. Probably all those components are chosen bc. they have drop-in replacements for it. The TMS320 for examples have many drop-in clones. Majority of inverters use cheap TI DSP clones.. Dont subestimate those eletrolytic caps, this things are harder to make than us think. that's why nippon chemicon, and murata are market leaders. Eletrolytic caps looks like simple devices but very hard to make it right.

[@RminusOR]

I'd also think R&D happens with higher cost western components as their documentation/support and dev environments are well established and good. Once the kinks are ironed out, you can "deploy" on cheap drop-in replacements.

[@Chinhnguyen0497]

@@RminusOR true. they need to build up the reputation. But these cars like illegal to moving on the road, just a slight turn at 50 mph can send you to the hospital. But this is enough for them to make a lot of money in china

[@hog-hog-hogwarts]

I'll support also this channel from Japan with respecting for both Ichiken and Denki Otaku :D

[@Im_eyzi]

Regarding the bus bar, I believe the center of the board will become warmer over time. As a result, the resistances of all three sections will be similar if the middle bus bar, which is warmer, is made thicker.

[@pcgamer43pc72]

All things considered they did a good job for the design. I was expecting worse

[@ForgottenLore]

I agree, they cost cut on the electrolytics but the design is good value.

[@JojoJoget]

Good design, cheapened parts. Typical engineering things after they pass the designs to the costing department

[@jhonatancock2302]

The thing is, unless they control very closely their supply chain, some of those IC might be fake.

[@Chinhnguyen0497]

they invest in one thing making the other worse. cannot be judged by just an inverter

[@vex123]

Excellent overview! You gave just the right amount of detail while keeping the content interesting and enteraining!

[@LtenooHate]

please give us more of this type of content! much appreciation for such work!

[@FireboatJim]

Great job. Well presented!

[@bassobalalaikka636]

Great video, thank you and your professor. One thing I noticed about the power board: there is room for double amount of mosfets, which might suggest that there is possibility of making larger power version, even the gate resistors are mounted, maybe with larger size heatsink or even water cooling plate..

[@helikopterelidojosa5479]

I like your new translator. I like your meticulousness.

[@UsefulAlien]

Excellent description, very clear and understandable, thank you!

[@fritz194]

Brilliant... These TI controllers are used in lots of inverter stuff, lots of the functionality you would normally have to realize with discrete logic and circuitry is already inside including protection and hardware comparators - even if their operation is not trivial, deriving from an existing product is not that big effort. They are also available in multicore configuration driving highly integrated and protected solar converters. It would be interesting which switching frequency is realized... The controller has advanced ADCs which can run synchronous to switching frequency, the CLA coprocessor can preprocess the acquired samples and offload the core. Its a proprietary TI 32 Bit architecture with FPU, the CLA coprocessor can also operate floating point values. Weird Harvard architecture... perfect.

[@JK-zx3go]

Great video, I have seen similar topology in fire damaged BMW inverters.

[@RixtronixLAB]

Creative and informative video, thanks :)

[@TopViewFar]

Great work done! Good video!

[@BillDemos]

Wow! Very high quality content! Immediately subscribed!

[@brandonbehc]

You could talk about the topology of the gate and power driver MOSFET paralleled H-bridge to be activated simultaneously and how it makes it to have no noise

[@jamier.6634]

yes... I was hoping for a bigclive style reverse engineering around the H-bridge and drivers.

[@chasemartin4450]

I was *not* expecting to see Infineon MOSFETs on a low-cost Chinese inverter like this!

[@timhooglandyt]

If they're not fakes lol

[@chasemartin4450]

@@timhooglandyt You're probably right...

[@davypeleman3672]

Even infineon mosfets would not justify that price. What a ripp off. This is around 200 dollars in components imo., ok matbe 500 :) at most; Did you see that film cap, that's a good component for such application. Electrolitics, hahahaha will be very durable (sarcasm) But in these times everything is meant to last some years so yeah. Consumers are getting ripped of more and more.

[@stevelupton2533]

@@davypeleman3672 the whole car costs $4000, not just the inverter!

[@itabiritomg]

2:50 notice that the first capacitor from bottom to top is already bloated!

[@test-rj2vl]

Very good video. Easy to understand. Clear speech. I like that you were thinking what to say beforehand instead of coming up with words live.

[@kaumohlamonyane272]

Amazing teardown and break down man. I am impressed by the quality of your content. Another reason why the car inverter components might be spaced a bit further could be the higher operating voltage of the car inverter.

[@tellucas]

I guess passive cooling works because this thing will only see it's rated power for a few seconds at a time. Just a few more dollars and a fan could have been fitted as backup cooling.

[@Jergling]

I can't believe how small those power transistors are. I've seen larger Silicon IGBTs in e-bikes.

[@jacquesdebeauregard5935]

the limiting factor for the size is the heat that your package can output. With proper board copper layout, this number can be astounishing. Some packaging allows top and bottom coolling and so have a much bigger current capacity

[@garydunken7934]

Great work with the tear down. The mosfets look low profile. Perhaps they are sufficient for 25kW output.

[@boxoflazuardi9421]

real otaku with master electrical engineering

[@sajalsnanda6209]

such a great and informative video

[@horoki9060]

I had no idea there was another channel for Ichiken.😊

[@d2sfavs]

incredible how they made this thanks for explaining

[@santhoshnadella6928]

fantastic explanation,waiting for more stuff like this ! thank you brother

[@bass9112]

great video! glad the algorithm showed me this!

[@ChundomanX]

you won a subscriber, amazing content

[@dinkc64]

Thank you for tearing down and explanations, very nice presentation.

[@cghn8997]

Great video, well explained, excellent job.

[@Debraj1978]

This guys knows this entirely. Subscribed.

[@inothome]

Great tear down. I'm going to make an educated guess that thicker busbar is due to the two rows of FETs mounted back to back and not much thermal area on the board to dissipate heat. With that larger, what looks like aluminum, busbar it helps remove heat.

[@lloydtshare]

350 amps it's for current

[@inothome]

@@lloydtshare Why only one bus is larger than? The others will also carry the same amount of current and they are smaller. Which was the whole point on why he was questioning the one substantially larger bus. If you look at where the single larger bus is, the FETs between phases are closer than anywhere else on the board. So, needs more heat dissipation.

[@lloydtshare]

@@inothome Looks like caps to me and not even touching the DC input bus bar, the 3 phase ac output bus bars are thick so they can output more current than the input bus, also they won't remove heat they will create heat because of resistance so attaching any FETs is a really bad idea

[@inothome]

@@lloydtshare Look at 11:58, he mentions it numerous times. It's a larger bus, not caps and the bus will pull heat from the board. The FETs sink their heat in to the board and with the heatsink and larger bus it will help dissipate that heat. If you look at the two opposing rows of FETs, the larger bus bar has the FETs arranged closer together. Vs the other arrangements where the legs of the FETs point towards each other and allow for more heat dissipation without an added larger bus.

[@lloydtshare]

@@inothome looks like it's for more current on v phase as it has more mosfets around that bus

[@uiopuiop3472]

i love cost effective evs and their parts

[@siberx4]

I was going to remark that using standard electrolytic capacitors was a cheap-out choice for something like this because they age over time (especially in hot/adverse conditions) and are more likely to fail than other types, but I guess in a car this cheap they're willing to accept that eventuality. The pink sheet is definitely designed to be thermally conductive, although it might also provide some vibration dampening as a side effect. I doubt the sheet conducts all _that_ much heat out of the capacitors in practice because of how thick it is, although it is better than nothing. MOSFETs connected in parallel tend to give somewhat balanced results, because MOSFETs have a positive resistance coefficient with temperature. If a MOSFET is carrying more current it will heat up more, which will increase the resistance along the channel, so more current will flow through the other paralleled MOSFETs. This is reasonably self-stabilizing, so no runaway should occur where one MOSFET is handling all of the current. Of course, this is not perfect so you can't run your transistors right at the bleeding edge of their current ratings if you're relying on this to ensure stable operation.

[@DrHouse-zs9eb]

Texas Instruments and many western manufactures? The sanctions dont seem to work at all :D

[@WesleyLeeYang]

they probably have more stock than the US and the rest of the world combined

[@peterduxbury927]

Thanks for such a great explanation of the Inverter for use in a small EV Car. For me, it was interesting to see how the 22 - Electrolytic Caps were used, instead of one larger type. I also wondered if the Inverter had any thermal shut-down protection circuit, and also any (replaceable) Internal Solid-State Fuses for protection. In your opinion, would the failure point of this Inverter be those 22 Capacitors - as they dry-up and get old? I am not an expert, but Electrolytic Capacitors do seem to have a limited life, and their replacement could greatly extend the Inverters' life. Greetings from Australia.

[@ernestuz]

Otaku-san, you've been added to my subscriptions. It's hard to come by channels like yours.

[@Inexactbeetle]

Thank you and thank you professor!!😁❤

[@Zhaymoor]

I have learned from the video, thank you.

[@firasgh871]

هل استخدامهم لتقنية cree mosfet يعطي مردود أفضل بالحجم والحرارة

[@friedbrainz7998]

Mashallah bruv, these cars make great bombs.

[@jeanious2009]

What’s crazy is I only see 500.00-600.00 worth of components/pcb/aluminium. How companies can get away with ripping people off is beyond me.

[@somenygaard]

Its what we often see in instances where there is a single supply source. The customer is already captured and there is basically no other option to keep their already purchased car from becoming a useless item.

[@jaro6985]

huh? its way less than $500 of components as the whole car cost $4k... you misread

[@lbgstzockt8493]

You do realize there is a lot more to selling a product than juts the BOM cost right?

[@JawwadHafeez]

Plz provide details of the controller part as well, like the TI and ST micro controllers and the driver as well

[@Danndanny55]

So if i use power jack to move pallet of potatoes, does this such controller use them sensors to demand more amps? move something light motor is chill.

[@MatthewTaylorAu]

imagine the parts from this being available to the rest of the world to convert existing cars to electric. much of the EV on road costs are due to certifcations. engine swaps can be certified cheaply, there are thousands of local mechanics capable of swapping drivetrains etc. optimistic? maybe. Doable? yes?

[@jamier.6634]

Wrecked Tesla's seem to be a popular source of parts for EV swaps in the US

[@josephmcconnell7310]

Great video! I would be very interested if you do a similar teardown to various UPS units or other battery backup systems.

[@peterxyz3541]

I wonder if I can use this to drive a 3 phase lathe or mill by using batt back up system

[@quademasters249]

Probably easier and cheaper to put a UPS on a VFD. Most lathes don't need huge currents.

[@ronblack7870]

@@quademasters249 ups usually has small battery often lead acid . they advertise kw capacity but not the actual amount of kwhrs they hold. they are not meant for long term. primarily to power electronic equipment allowing orderly shutdown if long power outage,.

[@quademasters249]

@@ronblack7870 It depends on the UPS. Some UPS's can even mount auxiliary batteries. How large a battery were you thinking about hooking to this inverter?

[@jimsvideos7201]

The drivetrain from one of these vehicles seems convenient for reuse in electric boat conversions.

[@dimzen5406]

Two current sensors is usual to industrial inverters, just a few exceptions like Danfoss use 3 sensors

[@aandreoli8961]

Outstanding video

[@woodduck]

A tip I have learned for reading difficult product numbers it to keep the ic and microscope stationary while moving the light source around at different angle and distances. If this doesn't work I found you can angle the pcb at 30° away from you and angle the light source. You will see the engraving will light up eventually.

[@Bordberti]

es gibt Hybridmodule wo die ganze Technik schon eingebaut ist. sehr kompakt. IBCs heisen die.

[@ERROR-413]

Sehr gut! Danke

[@SomeDudeInBaltimore]

Looks like they left open pads on the board to have the option to add more or less MOSFETS to adjust current-handling capability. Still, I definitely would have gone with IGBTs. This circuit looks like something you'd see on Instructables when searching "how to build an inverter".

[@brandonbehc]

I think the only reason they do parallel mosfet and caps for power is the low cost. I wouldn't drive that car after watching this

[@aibi5532]

​@@brandonbehc no one reported any issues with these cars

[@MaxVandenbussche]

IGBTs would be far less efficient and have slower switching frequencies. They are better used at higher voltages but at 96V IGBTs make no sense here.

[@brandonbehc]

@@MaxVandenbussche Inverters work at low frequency

[@pingwinvonjelen6586]

​@@brandonbehc there is no way to handle that current with single transistor. Paralelling is a must

[@mostafaTobar]

thanks for all this information

[@conoroneill6696]

Interesting video, what's the reason for having 6 MOSFETs in parallel instead of having just 1?

[@JeromeDemers]

you forgot to mention the controller board that all components are assembled on one side and the other is full of test point. You can see Curtis inverter use aluminum casting and the automobile use CNC extrusion aluminum for heatsink.

[@ForgottenLore]

Frankly for the cost the vehicle, the design seems to be of good value, I am surprised they didn't use slightly better electrolytics but I suppose there are a lot of them.

[@aibi5532]

check the price of the vehicle

[@keithminchin1817]

36 Mosfets. Wondering how many gate drivers are used?

[@onGlobalproductions]

6

[@Alan96555]

Muy buen video, gracias por la traducción

[@TheRailroad99]

The TMS320 is 12 or 16 bit if I remember correctly

[@shaun2072]

Excellent video.

[@TheConnorGames]

Any chance we could get more details on how the board to board connections are made? I've been taught that creep in the PCB material will make a bolted joint like that loose preload over time, do they have some kind of spring on the fasteners to account for that? Also, are the connecting rings really aluminum? If so do they have some way of avoiding bad contact due to the natural passivation layer?

[@nielsdaemen]

11:56 I would expect them to be equal

[@younesthabet]

whats the topology of the inverter? clascial 3 phase H-bridge?

[@cadsonmikael9119]

Yep, regular 3 bridge topology.

[@АртемБургарт-ф5б]

What is the chip in the metal case on the second board?

[@AaronSchwarz42]

I like that you point out US made chips are more reliable even if slight more costly

[@typedef_]

*designed chips...* we all know where they're made

[@papukaiffa]

Fyi Texas and Onsemi are US based. Infineon is German and STM is Italy/French origin...

[@NaeimAlaagibnaeim]

Is there CANBUS communication between inverter and Motor and battery.

[@rilosvideos877]

What is the switching frequency of the inverter? Is it near pure sinus or rectangular AC output? How is speed/power controlled via amplitude of voltage or PWM like?

[@Kek5kopF]

Do you know why they used MOSFETs instead of say IGBTs?

[@systemui3609]

because mosfets are really good switches and do not consume any current

[@pulkittiwari4952]

Can you please share the product links as well? For example, which project uses this particular inverter (exact make and model).

[@eriklethdanielsen3968]

What gatedriver was there in the 2.one for fork lift?

[@we-are-electric1445]

Smaller electrolytic capacitors are also more likely to be more readily available hence cheaper and if one starts to fail there are many others to share the load. Electrolytic caps are a failure point so why risk the system go down ? Industrial inverters have been around a long time and so are pretty well proven. You might be overthinking the cooling - the inverter will have a track record and they will be pretty sure what its capabilities (and weaknesses) are and any alterations would be pretty well thought out and implemented already elsewhere. I haven't worked in power electronics for a long time but I would be surprise if the layout is not replicated across many electric car inverters. The car industry tends to be pretty conservative - read hanging on to old technology which works as long as it can 😊.

[@L2M2K2]

This. Like, I really don't understand the criticism for electrolytic capacitors at these voltages. If the 105 °C rated capacitors run at a reasonable temperatures, like around 55-65 °C under normal use, they will last over 10,000 hours if driving. Which, even at city speedw will be 300,000 km or more, much more than the expected lifetime of most other compnents in such a city car. The design looks clean, looks reasonable, and looks highly optimised for its use. A few fancy parts where they are needed (like the big film capacitors and the high-quality ICs) but not wasting money overengineering the bulk of it.

[@we-are-electric1445]

@@L2M2K2 The electrolytics do fail. Expecting temperatures to remain below 65 deg C may be optimistic in some climates where ambient is 35 degrees plus. You also have to remember stress is caused by the constant heating and cooling cycle not just a high absolute temperature. A lot of inverters used in cars have industrial origins so some robustness is designed in but a car is a very different use case and it can be difficult to tell how long something will last in an uncontrolled setting such as a car as opposed to an inverter being used in a factory machine which is monitored

[@piotrcurious1131]

usage of aluminum collars for high current interconnect in-between pcb layers is a bad engineering decision. Aluminum oxidizes over time and the contact will weaken, possibly melting the board or even causing electric arc and fire. Actually at 8:40 you can see the result of chemical reaction in-between copper and aluminum around the collars, spreading around the PCB. So even in a brand new inverter you get damage already there, and you can see that as contamination creeps around the PCB it will make the device broken beyond repair very soon. Remember both copper and aluminum are alloys, and the least noble metal will surface out and be leached with moisture. Losing alloy elements will also cause alloy to be brittle.

[@leancarv]

Excelent video.

[@georgechikomwe6549]

Hello great video do you have schematics for those inveters especially for forklifts for sale please assist

[@bisnetodeportugues]

🇧🇷.Muito boa Explicação...👍

[@electroumit]

Thanks.

[@mcskifter]

Where can I buy some?

[@AltereggoLol1]

Can't believe that little thing handles more current than my entire house sytem

[@Simsol487]

Which inverter topology is used?

[@boxoflazuardi9421]

next video, scope the sine wave

[@FilterYT]

Well done, thanks.

[@brunonikodemski2420]

Some comments. I'm not seeing any EMI suppression on these boards, so how does this car allow radios, bluetooth, GPS, or any other incoming radio signals from being disrupted by these high-speed and high-current circuits. Even if the box is solid, that noise will come out on the cables. Second, this design appears to be modeled based on thermal considerations, which we also did. The development boards were probably IR mapped, and then appropriate parts placement was made to distribute the heat. The incoming busbar flat plate arrangement does provide some medium frequency EMI suppression, but it is clearly inadequate low band radio. Per prior comments, electrolytics tend to have lower ESR as they heat up (until boiling over) so there would be some thermal matching in low power operation. Film capacitors have low ESR only if they have thick sheets, and require a large flat surface. Their volume factor is usually inconvenient of other parts have high heights. We did some of equivalent fork-lift and cargo-carrier designs, and the overall design is often limited by the space allocation within the original equipment. We made electronics to fit the application. That's why they paid us those big dollars.

[@papukaiffa]

This system is quite low voltage so switch times can be slowed as well. Output connection goes straight to motor which is shielded and the wires could be also shielded.

[@fritz194]

The current sensors somehow seems to me of the Rogowski coil type... Maybe they are hybrid with an additional Hall sensor.

[@arduinoguru7233]

Could this inverter be purchase separately?

[@bobsoft]

Similar layout I have seen on Electric Forklifts (Hyster)

[@deineroehre]

Most Inverters fail due to semiconductor blowing up, so choosing an european Brand like Infinion seems legit and in the end cheaper - if you have to swap all the inverters due to failure, it will cost you a lot of many, especially if you sell many of them. If you demand the same quality standard like in Europe (or with lower standards, but still rather high like in america compared to chinese cheap mass-production standards), the prices between european and chinese products are basically the same, especially with products which don't need that much man-power for the manufacturing.

[@sammyd7857]

It is not a 20kw inverter. Inverter rating is at constant output. So it is 14 kw. But going by the size it would be lucky to do 10 kw

[@rust941]

Can someone explain whats the difference between this controller and smaller vesc based controllers? Some of them are rated for 100V 200A (continuous) but are of small tablet size. Why is this one so big if both are air cooled?

[@jaro6985]

I think the VESC is unlikely to meet its rating at higher ambient temperatures (55C+). Also way less capacitors so may not be reliable and drive the motor as smoothly. But the parts its using (FETs) are more expensive, which can explain the size difference somewhat.

[@onGlobalproductions]

Vecs will heat up quickly if doing low rpm high torque output, This is the area where capacitors and body diodes do their work

[@jaro6985]

@@onGlobalproductions But 100A at low RPM and 100A at high RPM would have similar losses, they would even be higher at the high RPM. So is the spec legit or no?

[@L2M2K2]

Which small VESC does 200 A truly continuously under actual use? The biggest VESC model “100/250” (absolute maximum peak voltage of 100 V and “continuous” current of 250 A) for example has even its power connectors only rated for 90 A continuous... (The 250 A seems to be derived from internal transistor ratings for infinite cooling, or something similar.) So, it probably wouldn't like actual 90+ V battery pack either with all the voltage spikes. Basically, the actual realised continuous power is likely much less. Not much less, but then again the size is also not that much less.

[@TonyWeirPD]

If someone had given me that assembly and said "guess the power rating", I think I would have said 2-3kW, perhaps 5kW peak. I can't help wondering how long this unit will last - particularly those Al-Cu compression fittings.

[@papukaiffa]

What's wrong with the Al-Cu connections?