Milling Through the Mire – Tesla Model S Drive Unit Woes.

rickard

We may have mentioned that the road to EV nirvana worldwide might be a tad longer than the press and His Eminence Herr Musk might lead you to believe. Stay with EVTV. We will get you there. We’ve seen this all before on the Internet rodeo in a past life. And trust me, it’s all going to happen – just not as quickly as you might want.

Electric Vehicles have had a mixed recent history actually. EVTV has kind of carved a niche out by doing strafing runs on bankruptcy auctions of the limp, the lame, and the dead strewn across the prairie of the Electric Vehicle frontier. We actually compiled a list of all the electric vehicle startups of the past five years who haven’t made the trip and we keep adding to it. It’s hard to remember them all.

The one bright spot in the firmament has been Elon Musk and Tesla Motors. We’ve profited personally and handsomely on just daytrading the stock of Tesla – sufficiently so to finance a P85 model sufficiently loaded to total $107,000 ducats. Many of our viewers have done likewise, and indeed some have also bought a Model S. We are unabashadly Tesla Fanboyz. But we do keep our eyes wide open and take our duties regarding keeping our viewers informed quite at the top of the priority heap. And the rule is if you have a problem sufficiently serious that it gets its own name, you indeed have a problem. In this case, it is the Milling Noise problem.

And not everything always goes 100% smoothly.

fijpkj0y4hw2ktwpxfii

zuvyxjkdw2jn3ne0k0pn

Some Tesla owners are experiencing difficulties. This one stepped on the wrong pedal at the wrong time – at the Freemont service center as it turns out. It’s a sign of the times.

We had an interesting conversation with a near fanatic Tesla Model S owner. Joe Robe lives in Florida but has already attended several Model S gatherings on the WEST coast. He drives his Model S cross country and in just a few months of ownership has driven over 40,000 miles. He’s also on his THIRD drive train and second battery pack.

Much is made of Tesla’s Supercharger network and it is indeed a plus. But Joe needed to stop in Cape Girardeau on his way to St. Louis from Nashville, enroute to a Model S gathering in Monterrey. He contacted us by e-mail and I responded that we did indeed have a real Tesla charging station and indeed three of them here in Cape Girardeau and he would be welcome to stop and charge.

BshFgVVCAAAsv01.jpg-large

He picked a kind of an inconvenient time to do that, right during dinner celebrating a friends birthday at 9:15 on a Saturday evening. But we went down to the shop and let him pull his white Model S into the shop and plug it into our Tesla charger, which is a well fed model that can do full current. We gave him the Azure Dynamics eTransit Connect to take to the hotel and he found it most delightful to drive.

We met the next morning for a tour of the shop and found that Joe was a software engineer and a bit of a teutonic Cherman perfectionist. He told a tale that is becoming quite familiar – inoperative charging stations across the land. He was delayed in Nashville at a Nissan dealer’s who’s charge station would run for just a few minutes, overheat, and shut down.

But Joe is a resourceful guy. He cleaned out the intake grill on the charger and got it up and running. Indeed he had to fix several charge stations on this trip from Tampa.

He also showed us a Chademo to Tesla charge adapter that is supposed to be a state secret at Tesla but is available actually. It’s $1000 and available “soon” but he had one.
Chademo_US_shot_02_1024x1024
He reports that there is nothing wrong with Chademo and he was easily able to charge his car adding 140 miles in range in a one hour charge.

But he also reports that with over 40,000 miles on his Tesla, he is on his THIRD drive train and SECOND battery pack. The battery exchange was actually, according to Joe, simply a fuse in the battery pack that had blown. But the Service Centers aren’t allowed to access the inside of the battery pack. Tesla shipped an entirely new battery pack to the local service center and they swapped it out in a matter of minutes.

The drive train is slightly more ominous. Tesloids already have a name for it, the MILLING NOISE. We admitted we only had 4800 miles on our Model S and Joe assured us – you will have to replace your drive unit at 8000 or 10000 miles.

There are also reports of a “klunk” sound when going from acceleration to regeneration and back. But the majority of failures involve a grinding or “milling” noise that increases over time.

After our conversation, Brain and I looked about the Internet and Brain came up with two very interesting reports because they are so public. Motor Trend Magazine has a Tesla Model S and they report that they had to have a drive unit replaced. More ominoous, Edmunds reports they are on their FOURTH drive train. This is a bit disingenuous. THey are actually on their third drive unit. One that was delivered to the Service Center was rejected for a broken electrical connector.

Rumor has it that one Service Center employee admitted they were changing 3 to 5 drive units per week.

Tesla features a 4 year, 50,000 mile warranty. For $4000 you can extend that another four years and you might want to take advantage of it. Some commercial buyers are not eligible for the warranty at all and at least one was dinged $15,000 for the drive unit replacement.

The drive unit itself is kind of an integrated unit consisting of the Tesla manufactured copper rotored AC induction motor on the left, the power electronics on the right, and a single speed gear reduction/differential unit in the center. It features a 9.73:1 gear reduction. No one appears to know where this gear drive comes from. Tesla used Borg Warner and an eGearDrive on the Roadster. From the looks of it, this Model S gearing has almost the identical cast aluminum type of case and it is entirely out of the question that Tesla is making their own gearboxes. But they are so secretive about it that there is so far no way to confirm that Borg Warner is a supplier of this unit. But to produce a gear drive unit like this in the numbers required, we would be astounded if it was any other but.

model-s-motor

Tesla_Motors_Model_S_base
We actually sell the eGearDrive of Borg Warner quite similar to the one on the Tesla Roadster. We’ve kind of developed a following among Ford Dealers attempting to support the eTransit Connect as they apparently have a quite high failure rate. I would venture a guess that this is the heart of the TEsla Model S drive unit problem. The radial load on the motor just shouldn’t be sufficient to point to a motor bearing, though the Model S does appear to use a very unusual open lubricated bearing. But milling noises and klunks are classic transmission type failures.

IMG_3988
And in one sense, any wonder. Our P85 motor is capable of 600 Newton Meters (435 ft lbs) using 305 kW of power. Applied instantaneously at moving 4675 lbs of car, plus people meat, forward, the forces are very very large. More importantly, and I was quite called to task by Otmar, who is no longer discussed at EVTV, for my characterization of the very small time frame of electric motor torque application. It remains actually pretty accurate.

When you accelerate an ICE engine, there are a couple of things going on – none of it very direct. Stomp on the pedal in your brand new Vette and what you really do is open up the flow of intake air and fuel into the carburator or more likely fuel injectors. This causes the motor to increase in RPM but in a kind of logarythmic growth curve. At first quite slowly and then at an increasing rate. Along the way, if you have an automatic transmission, you have the further spoolup of the torque converter until it reaches lockup. The result is a kind of controlled application of the power ramping to maximum over the span of 500-700 msec.

To put that in perspective, the CODA UQM CAN module I’m working on would have sent 70 torque commands to the inverter in that time span. It would be of no advantage, but quite easy, to spool up an electric motor in 10 to 20 milliseconds. Worse, you could apply the entire 600 Nm with the FIRST TURN of the shaft – actually at stall.

Understand that this is entirely uncharted territory. Tool steel facing sudden shock forces actually does change its crystaline structure – I characterize this as gelatinize much to Otmar’s mirth. It reforms quite quickly. But I’ve held the results in my hands several times. The most mystical of these was a one inch diameter shaft on a puny 30 kw MES DEA motor. We had made an adapter for their involuted spline from 4041 hardened tool steel. The shaft of this adapter was less than 3 inches long. BUt we have one with a 3/4 twist in it that I marvel at every time I pick it up. It just defies belief.

Of course, the result of an inverter mishap. And no doubt, like all inverters, the Model S has certain rampup rates, usually termed “slew rates” built into the software to ramp up the application of torque. But the Model S is striving for that 0-60 in 4.2 seconds. It weighs 4675 lbs. The resulting forces are simply non-trivial.

ICE cars often have acceleration rates that are quite similar and even higher torque values – in some cases MUCH higher. So how is this new. Again, the only difference is the slew rate or ramp rate of the application of power. NO production car on the planet has the instantaneous take off of the Model S. You can beat it in the quarter mile. But no ICE engine car would be able to even come close in the first 500 feet. It’s not even theoretically possible.

So while we have heard grousing about fender bolts and some angst over steering knuckles, charging issues, and bluetooth, why hasn’t this been more widely discussed in the media?

There is kind of a natural ad hoc conspiracy of silence on the topic. Tesla is as described, one of the more secretive organizations on the planet, making our CIA and NSA look like a hotbed of whistleblowers. But realistically, most Tesla owners are Tesla stockholders as well. There ARE some discussions of it in the Tesloid forums. But we found one post reprentative. One Tesloid with 87,000 miles on his allowed that he had indeed replaced the drive unit at 40,000 miles, but didn’t think it significant to mention as he hadn’t had a single problem since. Best car he’s ever owned. Etc. etc.

The DRIVE UNIT is equivalent of the ENGINE AND TRANSMISSION AND DIFFERENTIAL on any normal production car. And it is apparently post warranty at $15,000. That kind of bears mentioning.

So beyond the stock price, this rather goes to the value of the car. If it becomes widely known that post warranty at 50,000 miles, you rather inevitably face further drive unit replacements at $15,000, we are looking at a resale value on a four year old $107,000 Model S somewhere north of about $5700. That’s pretty valient depreciation.

I have mentioned a number of times that one of the herculean obstacles to electric vehicle adoption is the value proposition – how much of a premium above the cost of a similarly equipped gasoline car will electric drive command AFTER we are all acculturated to and knowledgeable about the advantages of magnetic drive.

What we’ve never really discussed at any length is the other end of the stick. Resale value. This actually IS the 800 lb gorilla in the room. Almost everyone is taking kind of a wait and see view of electric cars and I think it is on this issue more than “range anxiety.” EVERYBODY viscerally understands the concept of batteries even if it comes “batteries included”. Nissan has just attempted to address this issue by announcing a purportedly reassuring $5495 replacement cost on their batteries. They aren’t actually offering it and you can’t actually get it but they have actually announced it. And it is in an attempt to address the thoroughly one sided statistics on Leaf leases versus Leaf purchases. Nobody is buying. Everybody is leasing. And Nissan is dutifully reporting each lease as a sale. It is not. It is typically a two-year lease at that.

The battery is kind of a known and understood problem inherently part of electric cars. But if it became known that Tesla had a problem with essentially the entire drive train – and a $15,000 issue at that, what would THAT do to resale value? And so I sit on a $107,000 car, that if you knew the troubles its seen you wouldn’t buy on a bar bet for pretzel money, should I mention it? Who wants a car where you replace the engine four times for each tire replacement?

Given Tesla’s stock volatility, you might consider a few PUTS to protect your shares or maybe get out of them until this is resolved.

Interestingly, I got a call from CAMERON at the Tesla Service Center in St. Louis WHILE I was writing this. He was calling to REMIND me that my annual service inspection is due. We worked out next Tuesday as a fine day for an inspection. I really wasn’t wanting to take a day right now and drive up to St. Louis for an inspection, but we have been having the steering knuckle noise so I supposed I must. I was astounded to learn they intend to come PICKUP the car at 9:00AM Tuesday, take it to St. Louis on a truck and do the inspection, then return the car THAT EVENING or perhaps the next morning. The service being offered is just unbelievable.

I did ask Cameron if they were replacing any drive units these days. He admitted they were – people were reporting a “whirring noise” as he characterized it.

This video describes a Tesla owner who’s car was in the Service Center for a purported 66 days. They sued Tesla for the cost of the car plus $18k in legal fees. Tesla settled for the requested amount. But the Tesla apologists and fanboyz have attacked the lawyer as some sort of scumbag. Had we had a car in for service for 66 days, he would be a hero at my house.

So your mileage may vary. I’m astounded at how GOOD the Tesla service really is. And thoroughly alarmed that they may be facing a recall of the entire fleet for drive unit issues – a non-trivial expense.

One thing is fairly certain. They can’t just keep installing new drive units in Teslas every 10,000 miles. I have to believe there is some panic in Freemont these days, trying to come up with a solution to a new drive unit that is much more durable to offer as a replacement. They have 35,000 cars out there now and how they are going to “fix” this and survive financially is a little muddy.

Worse, I don’t think there is really very much “wrong” with their drive train. If it’s the gearbox, about the only game left that can deal in that technology and those numbers is Getrag. And they do NOTHING quickly. My fear is they have run into a wall of hubris that cannot be re-engineered.

I am a huge advocate of electric cars. I just love them and I never get over it. I will not rest until we replace 1.5 billion gas burners with 1.5 billion battery powered cars. But there are a few myths about electric cars that in the long term are counterproductive. One is that they never need maintenance. Joe Robe AFTER noting three drive unit replacements and one battery swap, assured me with an absolutely straight face that electric cars are “maintenance free.” I laughed out loud and called him on it and he very excitedly if not heatedly noted that those were all EXCEPTIONS and without these problems electric cars are maintenance free. Let me assure you Joe, they are NOT maintenance free. They are somewhat maintenance expensive actually. It is DIFFERENT from ICE cars. But it is not maintenance free.

And I fear in the case of Tesla that we have birthed a performance myth. Yes, they DO perform better than ICE cars and yes acceleration IS instantaneous. But we have 150 years of Darwinian evolution of the transmission based on a 500-700 msec ramp rate. The gearbox is what has plagued Tesla from the FIRST prototype they built, and I’m guessing it is very much the topic of conversation in Freemont right this minute. Electric vehicle performance and acceleration are not without limits. And it is my personal opinion that they have found the limit in the gearbox. The acceleration of 4675 lbs at a sufficient rate to move it to 60 miles per hour in 4.2 seconds is really not done guys. Sports cars at 3000 lbs do that with a 700 msec spoolup.

Our own Kevin Smith learned the hard way with Illuminati. He basically built an electric car in a corn field on a shoe string and had the Progressive Insurance X Prize essentially in the bag – until his transmission completely self destructed on the track. Again with a little 30kW MES DEA motor. He really DID have it won. Millions in prize money. For a guy in a barn who works at a government job. For lack of a transmission. If he does ever get over it, I won’t get over it FOR him until three years AFTER he gets over it.
DSCN1171-300x196
At this point, one thing Tesla has done right is attract a lot of very smart guys to California to work on this car. The solution to most technical problems is a function of desire, time, and money. But THIS problem could pressure the time and the money end of it. And I fear the ultimate solution is a 6 second acceleration rate – which I could live with just fine actually. I believe they will find the answer is a few lines of software code, but the answer will not be very satisfying to those who liked the performance claim on this car. It think the claim was unrealistic from the get go. Actually I did NOT find it unrealistic, I found it SURREAL. A light weight Lotus Elise? Sure. A 4275 lb sedan with seating for another 1000 lbs of people meat? Copper foil helmet award. Right out of the bag.

I’m pulling for them. But I think this particular issue is very very grave. And I don’t want the Al Gajda’s and Jeff Southern’s and others among our viewers to get the word later rather than sooner because I just didn’t want it to be so.

Our show this week is I believe the longest we’ve done at two hours and 48 minutes. That’s the length of two feature length films. We have a new quadracopter video camera intro and lots to catch up with as we took off for the Fourth of July. I was most pleased with a 30 second suprise at the end of the show – the first vehicle moving with my CODA/UQM Powerphase 100 GEVCU module.

[jwplayer streamer=”rtmp://s3einxnpkaij93.cloudfront.net/cfx/st/” provider=”rtmp” file=”news071114-iPhone.mp4″ hd.file=”news071114-1280.mp4″ image=”http://media3.ev-tv.me/news071114.jpg”  width=”850″ height=”522″ html5_file=”http://media3.ev-tv.me/news071114-1280.mp4″]

As my brother Thomas is marrying this week, we’ll miss another show for July 18th. I’m off to Point Clear Alabama to offer a toast. I face two younger brothers in a to-the-death golf match on Friday morning when I should be filming. I haven’t had a golf club in my hands for five years and they play like they don’t have a job to go to. At least I won’t be Elon Musk this week.

102 thoughts on “Milling Through the Mire – Tesla Model S Drive Unit Woes.”

  1. I have 10k miles now and have. It had any drive line issues. I did have to have th dash screen replaced…

    I will buy the extended warrantee.

    I am sure the right gear set and or bearings can solve this issue. EVs are he’ll on gear boxes. The tooth loading is worse at low speeds than high speeds. The instan torque at virtually zero rpm is the issue. It requires a higher than normal grade of steel in the gears…

    Just my opinion of course…

    1. The software will have a ramp-function where the torque is gradually increased. Mashing the accelerator does not equal 305 kW of power instantaneously. The car would be dangerous to drive if this was the case as the wheels would break traction and spin madly (like dumping the clutch). The problem lies with cyclic-stresses, (hard to calculate, virtually never quoted). The material-choice has to be mass-producible. They don’t want the drive-train to be expensive too as most of the cost of the car is currently in the batteries. Ceramic bearings and exotic alloys increase production-time and cost many times more than steel. I’m pretty sure the drive-train is in house because a 3rd party wouldn’t have access to the sort of loading required for a high performance EV.

    1. Yep the gearbox is the weak link…. However, if you put in the right gear set made out of the right material problems solved….

      It would be interesting to hear John’s Metrics thought as he is kind of in the middle of the material sciences needed to solve the issue…

  2. That is awful news. I wondered why the Tesla became so loud when I caned it. Was putting it down to the invertor but the sales guy told me they all sound different. Which rang alarm bells. Consistency being a realisable part of quality.
    I’ve been pondering a single speed transmission for a while on my 2~3 wheeler. The issue is the mix of noise, reliability, efficiency and size. I do not want my motor driving a chain for the initial reduction. Even if it is encased. 🙁

    1. Keep in mind that this is a noise issue, not a failure issue. The old muncie “rock crusher” transmissions from the 60’s come to mind. They were called that because they sounded like you are crushing rocks in the box. They would take all kind of abuse and rarely failed…

      In an ICE car you might not even notice the sound they are making. In an EV every little sound seems to be amplified…

      1. Actually not entirely Jeff. One owner was told the sound was normal. He kept driving it. You’ll be pleased to learn their 800 emergency service is quite good too and they came and picked him up when it quit entirely. This is a full metal jacket failure, giving warning early with lots of noise. The noise keeps increasing until it gets very very quiet.

        But the car won’t go then.

        Jack

        1. I do not know one way of the other, but so far the total failures are few. It may be that people are getting them replaced before they fail due to noise, so it is hard to tell but that would be the best bet….

          However, gear boxes do get much louder over time with normal wear.

          I have talked to people who have 20-45K miles and no issues, so we will have to wait and see the actual failure rate and average miles before failure..

    1. We’ve put 4800 miles on ours in a year. My wife mostly drives it 1.5 miles to the University. We did go to Kentucky Lake and back once.

      I would say I come a bit shy of pushing a car to the limit. I’m a controlled elegance guy. ALways looking for a glide and a slipstream. I suspect the Tesla drive train problem is exacerbated by guys who have to punch it all the time to demonstrate it for their friends. I don’t have to do that.

      So we may never have the drive unit problem – or we might. Hard to tell.

      As to transmissions, I recognized this problem from the beginning on the VW stuff. We get hardened gears welded to the shafts and the case ends and really everything you can get to make a VW transmission better. Instead of $800, they kind of run $2200 but it is worth it. To change one kind of requires disassembly of the whole car. Don’t ask me how I know.

      If I want a lot of noise and power, I have a Lear 24D that sounds like the seven demons of hell and climbs 7500 feet per minute. A car is not where I go for a power rush.

      But I understand the claim and the need for the claim and the desire for performance in such a car and at $107,000 you should have it too. But I think they have come up against a hard fact of EV life. All gear drives are designed for gasoline motors. And I personally believe it is not how much power, but how quickly it is applied.

      The Soliton 1 had a very easy to understand and configure slew rate and did quite a bit of power. Just playing around with it, left me feeling educated and wary.

      If you notice, we also always use a clutch. And on automatic’s use a torque converter. And I went into detail as to why in the episodes on the Cadillac several years ago. It’s ok to have things that are designed to slip, much as you have fuses, which are designed to blow. But I did turn a 6L80E transmission on the Escalade into a Claymore mine one day as you recall anyway. That’s a transmission designed for 400 horsepower and it is physically about 3 times larger and heavier than the one on the Tesla. The Escalade is 7777 lbs without ME at which point it glistens in four full tons of EV glory. So I definitely am included in Fred’s transmission club.

      Jack

  3. The de Haviland Comet was Britain’s aerospace pride until two mid-air disasters in 1954. Metal fatigue was the likely culprit. That is one reason why most airplanes were built in the State of Washington between then and 2000. And that was likely cumulative metal fatigue — the airframes had gone through several flight cycles before catastrophic failures. Not knowing enough about the molecular dynamics of torsional shock (which is what I call what JR describes) a working guess is that it is likely cumulative for the Tesla drive train as well.

    It appears Tesla drive trains notoriously broke in under 10,000 miles, not during one acceleration event. That brings Tesla’s pre-production, prototype testing into question. Two personnel shifts can put up to 1,000 miles on a car each day — 100,000 miles in about four months. This issue, both significant and seemingly widely encountered, should have been detected during normal prototype testing.

    A lagged acceleration software fix and/or a less easily torsionally stressed drive line seem the logical fixes, but a similarly simple fix (some of it being curved, not angle cut aluminum) would have “saved” the de Haviland Comet.

    History can rhyme.

  4. There is a solution to the single tooth loading all these transmissions are experiencing. A planetary reduction with 6 planet gears that output then drives a hyvo chain to the differential. This system would eliminate any single point loads. After that the next thing to fail will be the differential gears.

  5. Have a great time Jack. Beautiful opening Video BTW. I anxiously await another Ghia Update, regarding installation video of the Aftermarket Vintage AIr Conditioning!

      1. Thanks Jack! I enjoyed the update on the AC install in the Karmann Ghia. At some point I may have missed what you have decided to use for the AC compressor, and the motor/pulley arrangement that drives it. I think there was a reference to a “chinese Compressor” but I am not too sure what that means. I guess it is possible to direct drive a compressor, rather than use a pulley/serpentine. Or I guess you could drive it via pulley from the AC76 accessory shaft (though I doubt it since it is at the wrong end of the car). I do enjoy some of the closeup details of some of this stuff. I find it very illuminating and helpful when you have managed to do this in the past. Thanks again. And congratulations on the new Sister in Law!

        1. Robert:

          This is an electric airconditioning compressor that runs off the high pack voltage. It doesn’t run off the motor at all. It has a little 3 phase controller and a little 3 phase AC motor that in turn powers the compressor.

  6. To all,
    I’ve got a question about the far distant past. (EV time). I am trying to use the thermistor-based temperature pickup on a net gain warp 11 motor. I can’t find anything in their documentation that tells me the electrical characteristics of this thermistor or manufacturers part number or something like that.
    So the question at hand is, do any of you know the electrical/thermal properties of that thermistor. Or have any of you implemented a temperature indicator using the thermistor.
    Any help that you can provide would be greatly appreciated.
    Thanks,
    Al Gajda

    1. Al, it doesn’t look as though anyone responded to your question. I could be wrong on this, because I am not that familiar with the Warp 11, but if I recall correctly the Warp 9 has a therm-switch (Not sure what to call it exactly) and not a thermistor. Basically it just opens and closes at a set temperature so there are no real “electrical characteristics” for it. At least I believe that is true with the older Warp 9s. It’s either normally open or normally closed.

  7. Mighty Jack, love your blogs and videos. Another factor in the tranny thing is spinning it so much faster than than an ICE. I think you meant the Tesla would dominate for 50 feet, not 500 (although with its instantaneous electric acceleration head start it is still going to do very well to 500 feet, but not Veyron, GTR, etc. fast).

  8. Surely when you have cracked the canbus you will just be able to control the maximum torque applied to the Drive train. Tesla owners didnt buy the car to be a dragster

  9. Paulo and Celso,

    Excellent work on the “JLD505”! What a nice small package, considering everything the board is doing. Can I ask which micro-controller and current sense IC you are using? Is it a Freescale Cortex-M4, combined with an ACPL-C79A/B? With perhaps an ACPL-C87A/B in front of your voltage divider? It sure doesn’t look like you are using many components. Nice engineering.

    Keep up the great work — I look forward to more reports on what you guys are up to.

  10. Hey, way to go Jesse on the drone piloting! Great “heads up” miss of the power wires coming in towards the shop in the last part of the video…well done!

    Very sorry to hear about the Tesla Model S drive train woes. I hope this doesn’t become too common of a thing. I totally understand what you are saying Jack when you talk of moving such a hulk of a car in such a hurry.

    This situation with the Model S reminds me of my problem with my Electric Ox utility tractor. Both of the gearboxes in my tractor have been plaguing me for a few years now. My problem is that the shafts in the gearboxes have been snapping. Out of the three shafts in each of the gearboxes the output shaft has been breaking the most and it’s almost 1″ in diameter! This is all in a 800lb tractor with 2 – 48v, 1.1kW motors on 40:1 reduction drives.

    Now, in my search for a solution to this problem I have come across a very interesting drive from a company in the UK called Magnomatics. They have what is called a PDD ( Pseudo Direct Drive ) and it’s gearing is done magnetically. This link explains how a PDD works in a video :

    http://www.magnomatics.com/HomePageVideo.aspx

    I have talked to David Latimer of Magnomatics who is their Business Development Manager and he says they have a prototype version of a 22″ wheel drive system but their cost is presently prohibitive to people like me in the tens of thousands for a pair of motors and controllers.

    For somebody out there who was in the develpment stage of an electric tractor company, this might be the way to go…gear reduction without any gears! This might also be a solution for companies like Tesla with the drive problems they are experiencing. This would also be a nice fix for Jack’s “Noisiest Electric Lawn Mower” too. Let me know when you get yours and video how they work…LOL! 😉

    As always, keep up the great work!

    John

    1. Jack has showed the Magnomatics in one of the episodes or in one of his blogposts some time ago. I think he is well aware of them.
      Nevertheless they look like interesting drives. But they still need good bearings…….

      Regards
      /Per

      1. Hi Per, Thanks for the feedback on the Magnomatics PDD motor! I must have missed ( or forgot ) Jack’s comments about the motor…my bad.

        In the meantime, I have found ( I hope ) the solution for my particular problem. I have found a company here in Canada that makes a nice little AC hubmotor that should work in my situation. The motors are purported to have 250Nm (184ft lbs ) of torque per wheel and I will be ordering a pair of them. They have been tested and used in the ride on mower called the ZeroE, which was sold here in Canada through Home Hardware stores, and now they are being used as add-on kits to make golf carts 4WD! The company is called Silver Wolf Motors Inc. http://www.silverwolfmotors.com and there are some pretty cool videos on thier site showing how these motors work. ( plowing snow with a golf cart, pulling cars, driving through mud, etc. )

        I had a nice chat with the CEO, Bill Jager, the other night and he seems to think that I will be able to adapt them for my use. The cost is $2000 CDN/pair plus shipping and taxes for anybody who may be interested. You get the two motors plus the two controllers and the system works with an electronic differential for great turning traction. I forgot to ask if a wiring kit comes with the setup but I will find out more about that when I place my order. I hope to get a set of these before the summer is out and I will post how thing work out.

        Jack may be interested in a set for his mower or better yet, his fancy lithium/solar powered golf cart!? Could also make a nice product offering in the parts section of the site.

        Jesse, your not finished working on that golf cart yet! I feel you will soon be installing a set of these motors…lol!

        I can’t wait to see if these will work on my tractor!

        Cheers,

        John

  11. Hey Jack

    Once again your blog is very timely. I hadn’t heard of the Tesla milling noise issues, possibly because we’ve been spending lots of time dealing with some milling noises of our own. We’ve got the 40 kW MES DEA motor and Tim 600 controller, almost the same as your original Mini’s set up. When you mentioned the tesla noise, the 435 ftlb torque rating and the 8000-10,000 mile range they sounded familiar…and gave me a sinking feeling. Our motor is only rated at about 160kw, but it’s torque is rated at 420 ftlbs, we have 9600 miles on the car and have been chasing different ‘milling’ noises in the drivetrain all year. As soon as we find one problem and fix it another one crops up. So far in the last three months we’ve found and replaced: two bent rims, one front drive tire with flat spots, two bad CV joints, one bad ceramic wheel bearing(front drive wheel), two warped rotors, two bad calipers, a ripped CV boot that lead to finding a tooth on the transmission drain plug which lead to a cracked spider gear in the differential.
    And now something else has started a milling whirring noise…I’ll find out this weekend but I think the new outer CV joint I just replaced has gone bad.
    We’re also getting a banging or slapping noise upon the regen engaging. When I had the Trans open and the differential apart to replace the spider gear I noticed a lot of wear on the axle output gear on both the load and regen sides of the teeth. I think it’s only a mater of time before all the noises go away all at once, right after one really big noise….I think you you have a recording of it somewhere.

    So this fall, looks like I’m adding some custom hardend differential gears to go along with the custom drive gears already in the trans…I hope it’ll wait until fall anyway.

    Then we’ll see where the weak point has been moved to next.

    Still, $15,000 for a new Tesla drivetrain doesn’t sound bad compaired to that $5,000,000 clutch I had to replace….and I’m sorry, but you’ve still got more than three years to go until you get over that one

    1. Tesla may have made rod for their own back with such impressive performance figures. As jack says, its a lot of car to start moving quickly.

      There have been a few examples over the years of the (negative) effects of trying to suddenly put big loads through transmissions, one that come to mind is the Abrams M1, through its development it had few transmission problems, but as it was issued to units, gearboxes began to fail, the cause was the ability to put the transmission straight from Drive into Reverse whilst still moving forward, the crews used it as part of a firing maneuver.

      The M1 weighed 54 tonnes, no gearbox can stand that stress, the fix by Chrysler was to put in an interlock, to stop reverse selecting before the tank had stopped, it worked fine, BUT the Army hated it because it stopped them using “their” vehicle how they wanted to.

      The Nissan GT-R had its fair share of tranny issues, with gearboxes running at £16K it made buyers of used cars very wary, but they still sold well, and the issues are apparently resolved now.

      Soft start systems for electric motors are very common, I suspect Tesla may “upgrade” their firmware to incorporate a fix and soften the start loads that will no doubt knock a few seconds of the 0-60 time.

      Time will tell if this will happen, I certainly have no special knowledge but its seems a logical thing to do.

      PS Our thoughts in the UK go out to our neighbors over the water in the Netherlands, and the rest of the countries affected by the loss of Flight MH17

      Mark

      1. Being a former M1 driver, gunner, and loader, I can tell you that it weighed 60 tons, not 54. And, there was no interlock, you could select reverse while rolling forward in neutral. Fun machine though! I would love to convert one into the ultimate dune buggy.

        1. Hello Hans, sorry my mistake I was referring to the M1 rather than the M1A1 which was heavier. The story came from Orr Kellys book “King of the Killing Zone” but its a while since I read it so I may have the details wrong but I hope the gist of it was ok? Good to get info from the horses mouth though, thanks! According to Wiki the M1 transmission prices out at $1/2M!! (but thats just Wiki!)
          mark

  12. Doug Ingraham

    Jack

    When I log into the EV database I can look at my build and upload new photos and make changes but I cannot see my vehicle in the regular listings. It does not show up in a search. I can filter just by my name and it gets zero results. I think there might be a problem. My car is #31. I added some photos a few days ago.

    Doug

  13. Pingback: The Geek’s Reading List – Week of July 18th 2014 | thegeeksreadinglist

    1. Actually quite marvelous. My sister Carole and her clan, and brother Andrew and his clan, all showed up as well so it was more or less a family reunion. Second marriage for both in middle age so it was relaxed and kind of done up for laughs in a really nice golf resort on Mobile Bay in Alabama.

      My golf game has suffered in the past few years. I’m going to have to work out something where I can play 9 holes at least a couple of mornings. I had the low hole score of the day ( a 2 on a par 3) and the high hole score of the day (13 on the next hole) back to back. Brother in law brought some Cubans. The four of us drank two cases of beer in 18 holes.

      Rickard family reunion. Picture a race riot under the influence of bathtub gin….

      But very pleased to be home and back to work of course…. we’re going to start cleaning up a bit this week in preparation for EVCCON.

      Jack Rickard

      1. Great to hear that. Actually as you get older being out of practice at golf is a bit of a badge of honour: shows that you have better things to do with your time.

          1. Dear Garn.
            What is the “it” that you do not comprehend?
            The enlightening blog post. The mention of golf or the bathtub gin laden race riot of a family get together. Or maybe the last post concerning Jetsons noises. Hmm. How about a donorcycle version. 🙂
            https://www.youtube.com/watch?v=W1CSdYsJIWQ

  14. I am glad that I’ve found this site, your taking apart the fanboys over at TMC is great. If there is a fatal flaw in the Tesla the word needs to be spread far and wide to prevent others from being sucked into a dark hole of perpetual repairs. A 100k flower box wasn’t what I need, I will be ponying up the 4k for the extended warranty, I just hope the guys are still in business to support the car when it’s my turn for what seems to be the inevitable failure of the cars drive train.

  15. I know nothing about transmissions. I have some silly, unanswerable questions whirling around in my head. I’m just wondering. Is the power restricted until the drive unit warms up? Would hard acceleration on a cold drive unit do more damage than when the unit is at normal, whatever that is, operating temperature? Do you think Tesla are compiling information from the data logs to investigate whether driving behavior could account/contribute for some of the failures? Is it just the P85s that this is a problem with? Does Tesla have a transmission Guy? Failures at 10 000 km. How could a design flaw go unnoticed?

  16. Jack,
    You’ve mentioned before that you’ve sold some eGear drives to car dealers that are having to replace broken ones.
    Do you know by chance what part(s) are failing?
    Primary gears?
    Differential?
    Or does it vary based on abuse?

    Why I ask is as the numbers of Ev conversions increase and we’re hearing/experiencing more gear failures and now possible differential failure, you having the foresight to start by using hardend steel gears and a Quafe differential, and now with increasing failures with the eGear drive, are the gears in the eGear drive made of hardend steel or are they cast? Also what type of diff is it using? Standard pinion gears or Quafe style?

    If not hardend steel and Quafe, then there may be a market for such replacement parts.

    Thanks
    Kevin

  17. Looks like Yehu is faster than Otmar. Yehu’s Tesla is running in a Samba.

    Ok it is only half a Tesla, half the batteries at least. Plus he has successfully tested the fuses. Yes, they do blow.

    I hope he has a video for us.

    Cheers
    Peter and Karin

    1. Yes, and what about his claim that Tesla modules will soon be available for anyone? Last I heard they were claiming to be supply-constrained.

      1. Doug Ingraham

        I think the modules Jehu is playing with were the ones for the Rav 4 or the ones that went to Mercedes. They are not the Roadster or Model S configuration and look like they have 2.6 AH 18650 cells instead of “The good cells”. These could be lots of things. We will find out eventually.

        Jehu had a little over 12kwh of cells in the Samba and he was overloading them. I think he needed at least twice that in order to not stress them with an HPEVS setup. Going to six modules (18kwh) would have been marginal at 650 amps but would work most of the time.

        Doug

  18. Show very good as usual this week (25th). Dave Bogard’s work on the batteries was really great. On the strength of it I’m working on heating arrangements for the batteries in my Civic (our problem is low temperatures: we are not generally troubled by temperatures beyond the mid twenties centigrade for more than a handful of days some years). Current plan is low power silicone rubber heating mats under an aluminium sheet in the bottom of the battery box powered by the mains whenever the J1772 is hooked up and cycled by a thermostat; plus an interlock that prevents charging if the battery box temperature is too low. The theory is that if you plug in at night charging will only start at a good temperature, and once it is done, the batteries will be kept warm until you unplug and drive off in the morning. If it works I’ll do some video.

    I’m gutted by the news that they are discontinuing the Ampera. I’m not surprised though as I’ve seen fewer than a dozen in the 15 months I’ve had mine. They don’t seem to have marketed them at all, to the point where I almost wonder if it is elaborate preparation for a story along the lines of “electric cars? We made some but no one bought them, so now we are offering clean [sic] diesel until batteries improve which will be in 2050 and therefore someone else’s cash flow problem.”

    It is a particular shame because, the Ampera is the only store bought EV that doesn’t look like a lunch box (apart from the Model S at more than twice the price)

    1. John,

      Are you by chance using the CANbus enabled TCCH? Ray and I have made great progess on the CANdo Temp module. It monitors temperatures of you pack and controls the charger over CAN. You can set your own minimum temperature threshold. Likely our next CANdo will be the CANdo Heat, a 220ac heater controller. Will have the CANdo Temp at EVCCON, and will be showing it off in Ray’s Miata.

      http://www.youtube.com/playlist?list=PL3HcqdYYXMNL9JY2xV-70Hp-OhaiEH9pz

      1. Many thanks for posting this Nabile. Yes I have been looking at CAN but not for this application: I was planning on using a simple thermoswitch for the heating mats

        1. That is the path I took. To bad results. As well, my situation was of some serious cold while charging outdoors. It turns out my thermoswitch warmed with the box while my cells actually remained colder. That singular point was insufficient in my case. I believe more sensed positions may have inhibited charge for me longer enough to warm the cells enough more to have avoided the damage.

          Specifically, the damage included increased voltage drop under load, and an immediate need to remove two cells that were more damaged than the average. It so happens I had marked those two cells when I first put the pack together. I haven’t explicitly tested the rest for capacity on the bench, but I see the knee 5-8ah sooner according to the JLD.

          Look forward to seeing you at EVCCON!

          1. Many thanks Nabil. The thermoswitch I am planning to use has a probe on a flex lead that I should be able to bury in amongst the cells.

            Yes indeed – be good to meet again at EVCCON. Apologies for the extraneous “e” at the end of your name in my last post!

  19. Here in the UK the mayor of London has noticed that there is more to air pollution than CO2, specifically that diesel cars are spewing out stuff that is bad for the health of Londoners. He is proposing a surcharge on the congestion charge (a toll levied on vehicles entering central London) for diesels. This has hit the news and started a debate: most populons have been carefully trained to regard diesel as somehow “green”.

    This will I trust further encourage EV adoption.

  20. About a week from now I will be loading my car on the trailer and heading out to EV Mecca for my 4th EVCCon. Trying to get any last minute things done on the car so I am more comfortable letting other people drive it. I hope to see all of you there and give you the chance to drive my car or at least take a ride if you are more comfortable with that. Look for me, I will be #85 with the black RX-7.

    See you soon!

    Doug

      1. You’re quoting the Tesla Motors Club – a group pathologically positing nonsense in huge quantities. We’ve been there. They are not subject to any technical input.

        The “milling sound” appears to be in the “differential” according to Musk and they rather vaguely allude to a 50 cent shim. There are no 50 cent shims but I gather it is not a very serious problem. Some other noises were explaned by a cable rubbing on the drive train.

        I gathered from teh conference call that all is not in on this one. We’ll wait and see.

        Jack

    1. Doug, Looking forward to seeing you and your RX7 this year. I’ve decided to keep my hands off the evTD so I don’t screw something up before I load it on the trailer, but I do think I’ll drill some drain holes in the floorboards since I’ve had to bail it out the last two years.

      1. Good to hear both of you are bringing your cars. I rather missed the RX7 and always enjoy the TD.

        Just got my tow bar situated so we will be bringing both Rays Miata and my Bradley. I haven’t been able to get much done with my suspension, but have markedly better brakes. So hopefully autocross will fare better.

        Any other Miatas confirmed?

        1. Really appreciate you guys (and all the others) going to the trouble and expense of bringing cars. I haven’t got my Civic running yet but even if I had there are a few billion tons of cold salt water in the way

          1. Jack, et All,
            Isn’t the Model X four wheel drive? (and possibly future model S) It seems to me that spreading the load to four wheels from two will half(ish) the load on the drive train perhaps solving this problem. I know, I know, then increasing service and problems elsewhere, like front end drive components.

    1. Jack, I just caught your 8/1 show and the problems plasti-dipping the Doka. Since you’re going for a do-over, be aware that even with the solid-color dip, different-color substrate can still lead to different color results. The solid-color product is translucent to a degree though of course not as much as the clear/pearl mix. You might get the bed walls close but probably won’t match *exactly* the rest of the vehicle if the underlying paint is different. A couple of coats of white, or black dip, as a “primer” base over the entire job, will minimize that effect. White will help the yellow to pop also.

  21. Enjoyed the 1st August show as ever. Looking forward to EVCCON.

    I am one of the guys you referred to as having a Volt (well an Ampera actually but it is the same clockwork). You are absolutely right – I enjoy this car more than any other I have ever owned whilst at the same time share the horror of the over engineering; let me try to analyse why. I think firstly it works well just as an EV. If I am sitting next to anything driven by a piston engine at the lights, if he isn’t in the right gear, on cam and with the turbo spooled up when the lights change, he is toast. With no gears and plenty of torque available at all speeds (but especially round town) the car feels sharp and eager. If you have driven a good pure EV you will know the feeling.

    Given the current status of the infrastructure (thousands of petrol stations and very little fast charge) the series hybrid configuration is a fair compromise. I can drive to Edinburgh without planning anything, but even after the battery is exhausted, I still get the same EV driving experience, the same Impreza-flattening low speed torque, the same eagerness on the move, the same utter silence when stationary, the same efficiency in stop-start traffic. The engine is small and well buried, and not running a lot of the time so you even get much of the same quiet. Regen still works so it has good mileage even on petrol.

    The arrangement also means I do more EV miles net. If I had a Leaf and wanted to do a 35 mile round trip, I’d have to take a gas guzzler: 0 EV miles. With the Volt I drive there (and in summer half way back) on electricity and do the last 20 miles on petrol: 35 – 50 EV miles

    The Ampera (unlike the Volt!) also looks sharp and is admired by folk who have no idea what drives it. It has its faults though; the minor controls follow the over-engineered theme. The climate controls are a nightmare and the fiddly touch sensitive buttons on the centre console hopeless. The charger is weedy. It has just four seats. The boot (trunk) floor is quite high.

    I expect it will be the only hybrid I’ll ever own. When it wears out I hope there will be a low cost 200 mile range EV available and fast charge stations all over the country.

  22. The only answer for the Tesla is to have fluid drive – torque converter. No steel is going to survive a hammering from the impact of the motor.
    The Tesla is no different to the Edison Electric car.
    It’s carrying around the equivalent to about 5 litres of fuel.
    More Teslas get towed due to running out of fuel, than any other car.
    Look at the REAL problems with the car, with refueling and technical faults and you will see why it is car to be avoided.
    You are not changing your carbon footprint, and anyway, we need MORE CARBON. It produces bigger and better fruit and vegetables

  23. Pingback: Tesla Model S Drive Unit now has Infinite Mile Warranty | Electric Vehicle Reviews

  24. Well, the story of no ICE cars having to endure similar loads is a bit misleading. Take 911 turbo S with PDK transmission. It can make 100’s of launch control starts, which basically mean revving up the engine and the computer dumping the PDK clutch. I’ve done it and its violent, even on dry road all 4 tires spin for a moment, even when warmed and on good surface. And no, they’re not taving any issues. Most I’ve heard is almost 1000 such starts (with norma 911 Turbo) without maintenance needed.
    So I think it’s not a question od electric druve being more brutal, it’s just a question of design and materials. It’s just a question of price … Is Tesla willig to shell out (or can they, at this car price) for better and more expensive gearbox?

    1. The difference between the Tesla and all other internal combustion motors is the PWM operation of the electric motor. It produces a lot of equivalent HAMMERING as the shaft rotates and and this is just like hammering a coal chisel. The end turns into a mushroom of burrs.
      The only solution is a fluid coupling or fluid drive or an automatic transmission. My automatic transmission has done 340,000 km and they claim the fluid doesn’t even have to be replaced. That’s a vast difference to 10,000miles.

      1. Colin:

        From whence and where did you obtain such nonsense? First, Tesla doesnt’ use a PWM controller at all. It is a 3 phase space vector output to a 3 phase induction motor. There is no “hammering” at all – it is totally smooth power. Gasoline powered engines actually DO use EXPLOSIONS to drive a shaft and it is quite smooth as well.

        There is no equivalent hammering going on in either case. The only difference the drive train would experience between the two is that an ICE engine has a spool up time and only developes maximum torque once it spools up to a specific RPM. Electric motors can apply full torque at stall in a few milliseconds.

        Jack RIckard

        1. Ok. I take back the PWM hammering, but there must be something that is totally destroying the teeth in such as short period of time. Maybe even the accelerating torque is so great the oil film is not being maintained or the pressures are greater than the materials can withstand.

      2. From one Colin/Collin to another: PWM has nothing to do with it. Technically a three phase controller does still do PWM but, as Jack said, under space vector modulation. This still entails pulse width modulation in the signal to the IGBT modules but it is done in a very math intensive way. Anyhow, the PWM frequency is very high – something like 10-20 kilo hertz. Just to start with that’s far faster than the hammering caused by an ICE. But, to make matters even better there are several factors improving this situation. 1: The PWM into the IGBT does not cause nearly as large of current or voltage spikes as you might imagine because of the inductance of the motor windings. This acts to limit the rate of change for current. The end result is that current and voltage follow a fairly close approximation to sine waves. 2: Magnetism does not immediately collapse either so the magnetic field in the motor is not literally pulsing at 20kHz or anything like that 3: The motor has mass so any forces acting on it act on that mass and do not immediately show up at the output shaft with their full strength instantly. All of this conspires together to make three phase motors actually spin extremely smoothly. Having said all of that, there is some intrinsic cogging that can occur with permanent magnet motors. Wouldn’t you know, that’s exactly what Tesla uses. Still, I’m sure Tesla’s trouble is exactly what Jack said: the electric motor has extreme torque at 0RPM

    2. JakaR

      We do a thousand “starts” in a couple days normal driving. Take any transmission you like, they evolved to service ICE – not electric motors. And the only difference experienced at the output shaft is that ICE requires a spoolup, producing very limited power at first and building to a quite narrow actually “peak” at usually a quite high rpm. This usually involves a delay of 200 to 700 ms. Electric motors can and do often apply 100% torque at stall within a few milliseconds.

      If your 911 turbo S did a launch control start every time the car accelerated from a stop, it wouldn’t last 3 weeks.

      Jack Rickard

  25. Pingback: Kvarnen - Tesla Club Sweden

  26. Pingback: Tesla Model S Drive Unit now has Infinite Mile Warranty | Selfservice Future Technology

  27. I am obviously not an expert on the problems discussed with the gearbox – since, as stated in the article, confidentiality concerns restrict information from Tesla. However, there are some common sense fixes other than exotic metals and such. Number ONE would be to ‘fix’ the unrealistic EXPECTATION of instant acceleration ! I think the price of the vehicle ALONE would indicate that there are enough ‘ecology-minded’ people willing to sacrifice to make the change away from burning hydrocarbon fuels. They would surely be willing to accept a more moderate acceleration rate as long as the top speed was well within any highway needs. This is actually much MORE easily accomplished with electronics ( SCR controls ) than in traditional fuel burning engines. This controlled acceleration would mainly be needed at starting from a stop – – AND the acceleration curve could easily be very steep and still prevent damage due to the physics involved in overcoming inertia. I fear the typical hubris of hierarchical corporations, especially ones so sensitive to adverse publicity, as car sales is, is to blame. This ALSO is probably due to the underestimation of the sincerity of their market ! Given: the need to wean ourselves from Big Oil, the environmental concerns, and the human health advantages, it is just silly that they wouldn’t present this small change as yet ANOTHER example of responsible living ! After all, wouldn’t this extend battery life? Wouldn’t this make the vehicle more reliable and longer lasting? – – – A simple modification between the accelerator and the battery feed is all that is necessary – – along with the MUCH more difficult change in corporate thinking. I am SURE that their base market > the environmentally concerned consumer and those worried about the continued availability of ‘cheap’ oil, would be much more accepting of this simple fix.

  28. When some wrench happy people own them and start tearing into them themselves the problem will get solved.
    Hotrodders have always been and remain to this day superior to factory engineers in solving problems and improving design.

  29. The shape of the gearbox shows it must use a belt or most likely a drive chain similar to what is in a 4 wheel drive transfer case.
    If the Tesla engineers are using something other than steel?
    does anyone remember how when GM tried using Nylon teeth in camshaft gears to make them quiet how they would need replacement every 20 or 30k miles?

    Flat link chain stretches and the pin holes elongate in a relatively short time compared to a roller chain arrangement.
    it’s also likely they are using a flat link chain and the “chain stretch” is resulting in the chain rubbing on something or causing misalignment of the teeth of the chain with the cogs in the gears.

    The slapping or clunk between power and Regen is most likely no different than the clunk in a worn out transfer case when rocking a 4×4 from drive to reverse and back to drive

    1. Where do you get all this? Must use? There is no chain. No belt. No nylon teeth. The slapping or “clunk” was caused by software. It’s gone. Smooth as silk now. This motor turns at 16000 rpm and Elon’s little bobble head doesn’t even bobble.

      Where do you get all this?

      jack

  30. Pingback: Would you buy an electric car? - Page 4

  31. Wow! I stumbled across your news and it presents an intelligent insight into the EV world of Drive Train Troubles (DTT)! I’ve done some work in the past using electric motors with vibratory conveyors; one of the most aggressive uses of motors. It should, of course, come as no surprise that Variable Frequency Drives, (VFD), or simply Inverters are commonly used in this industry. Controlling the ramp and deceleration are a routine thing, spinning the motors to desired speeds, and indeed, passing through troublesome speeds where natural frequency of vibration is to be avoided. As you pointed out, a tweak in the software to avoid over zealous acceleration. Launch Mode is only needed to impress the rest of the world!

Leave a Comment