-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Jory, great info, thanks. I didn't know this cooling fluid. I'll see if I can get my hands on the other PEM.
I think the megapoles only have off the shelf parts on them and yes, they are identical, you can swap them between the phases. I did that when I was looking for an error on one of the megapoles.
The IGBT's are IXGX72 something (they were just under $10)... and are driven by an IGBT driver (TC4451).
No way
I can't do it to this signature Roadster
I had a similar idea already, but there are too less Model S in Europe and even wrecked, they are too expensive for playing around with them.
I sent CarboTech an email for the brakes. This is the reason for their statement on the phone:
I think the megapoles only have off the shelf parts on them and yes, they are identical, you can swap them between the phases. I did that when I was looking for an error on one of the megapoles.
The IGBT's are IXGX72 something (they were just under $10)... and are driven by an IGBT driver (TC4451).
No way
I can't do it to this signature RoadsterI had a similar idea already, but there are too less Model S in Europe and even wrecked, they are too expensive for playing around with them.
I sent CarboTech an email for the brakes. This is the reason for their statement on the phone:
PV4EV
Member
That's a curious comment for Carbotech to make.
Was this for Bobcat 1521's or AX6's ?
I ask because Carbotech will happily sell both types to Lotus Elise owners in the UK .. and yet we're all supposed to obey the same EU ECER90 directive . . .
Last edited:
Really surprised to see a Microchip part in a Tesla -- I've always thought of Microchip as a bit niche. A 12A IGBT driver. That's insane. I didn't know they made them that high current. I suppose, with huge gate charge comes huge driver requirements
These devices? IXGX120N60B http://datasheet.octopart.com/IXGX120N60B-IXYS-datasheet-11760468.pdf About $10 a piece 10kunits, so probably costing Tesla around $840 in IGBT per car, which is dirt cheap for what they are.
Each driver controls 7 IGBTs, 350nC/FET, Q=CV, C=23nF/FET (@18V drive), so it's driving 160nF @ 32kHz... that's a pretty heavy load there, explains why there's four of the ICs!
Marco, did you replace the driver when replacing the IGBTs? It's highly recommended to, because IGBT failure can damage the driver.
These devices? IXGX120N60B http://datasheet.octopart.com/IXGX120N60B-IXYS-datasheet-11760468.pdf About $10 a piece 10kunits, so probably costing Tesla around $840 in IGBT per car, which is dirt cheap for what they are.
Each driver controls 7 IGBTs, 350nC/FET, Q=CV, C=23nF/FET (@18V drive), so it's driving 160nF @ 32kHz... that's a pretty heavy load there, explains why there's four of the ICs!
Marco, did you replace the driver when replacing the IGBTs? It's highly recommended to, because IGBT failure can damage the driver.
Yes, I use them all the time in my home boards, but at work, we tend to use Linear, TI and Maxim (rarely.) Microchip usually pop up as an EEPROM. We don't use PICs at all (except in legacy designs, and we usually try to figure out how to replace it...) -- ARM STM32 for most, or STM8 for really low end, and Freescale i.MX for anything running Linux. Atmel is banned due to supply issues and a board where the customer waited 4 months because of no stock of a certain AVR.
Just a nice endorsement for Microchip there, surprised they did not attempt to market it. (IXYS mention they sell IGBTs to Tesla in their PR occasionally.)
Just a nice endorsement for Microchip there, surprised they did not attempt to market it. (IXYS mention they sell IGBTs to Tesla in their PR occasionally.)
Getting off topic, but I just had to share this because our Microchip rep sent me this TODAY (was PDF):
I've been in a worse situation with Motorola processors, where we had to redesign a board because they screwed up a die shrink cycle and when GM got wind of it they bought every part in existence. Texas Instruments is also known for "when the going gets tough, we go away." We actually use a lot of Atmel parts, some ATMEGA and some ARM, but we mostly use soft core processors in FPGAs (used to use Lattice but now Xilinx). We were forced to use a TI ARM on one recent project for driver compatibility reasons.
Now let's get back on topic!
I've been in a worse situation with Motorola processors, where we had to redesign a board because they screwed up a die shrink cycle and when GM got wind of it they bought every part in existence. Texas Instruments is also known for "when the going gets tough, we go away." We actually use a lot of Atmel parts, some ATMEGA and some ARM, but we mostly use soft core processors in FPGAs (used to use Lattice but now Xilinx). We were forced to use a TI ARM on one recent project for driver compatibility reasons.
Now let's get back on topic!
The car is at the body shop now. But they won't be able to begin working on it within the next week.
Well, more time to organize all the stuff for the homogolation to get it street legal in Germany.
Does anyone have any official data from the 1.5 Roadster, like gross vehicle weight, nominal power, etc?
Tesla uses a 72 A IGBT from IXYS, the IXGX72N60B3H1.
Yes, I replaced the driver, too.
Well, more time to organize all the stuff for the homogolation to get it street legal in Germany.
Does anyone have any official data from the 1.5 Roadster, like gross vehicle weight, nominal power, etc?
Really surprised to see a Microchip part in a Tesla -- I've always thought of Microchip as a bit niche. A 12A IGBT driver. That's insane. I didn't know they made them that high current. I suppose, with huge gate charge comes huge driver requirements
These devices? IXGX120N60B http://datasheet.octopart.com/IXGX120N60B-IXYS-datasheet-11760468.pdf About $10 a piece 10kunits, so probably costing Tesla around $840 in IGBT per car, which is dirt cheap for what they are.
Each driver controls 7 IGBTs, 350nC/FET, Q=CV, C=23nF/FET (@18V drive), so it's driving 160nF @ 32kHz... that's a pretty heavy load there, explains why there's four of the ICs!
Marco, did you replace the driver when replacing the IGBTs? It's highly recommended to, because IGBT failure can damage the driver.
Tesla uses a 72 A IGBT from IXYS, the IXGX72N60B3H1.
Yes, I replaced the driver, too.
djp
Model 3 Performance
Yes, I've finally been able to buy it, because he didn't find the error.
Thanks djp, that might help.
If anyone knows of a 1.5 Roadster that is street legal in Europe, please send him/her a link to this thread or as an alternative, I could give you my contact info by PN!! Thanks
I know there are some of them, but don't know one in person.
Thanks
@PV1
What error did it show in the salvage lot? Powertrain error? This was probably due to the damaged service disconnect. There's a small switch in the socket which detects if the disconnect is inserted properly.
The LED ring doesn't cause any error, it's just there as a charge indicator and doesn't give any feedback to the car.
BTW: The car got upgraded to Founders Series
Errors I saw were:
1. Backup power supply service required
2. Park lock problem. Vehicle may be free-rolling.
3. Check left front signal bulb
4. Check left rear signal bulb
5. Airbag system service required
6. Before starting, charge port door must be closed.
Plus a reminder message:
1. Engage hand brake when vehicle is parked.
When I looked at the car, it had about a quarter charge with 62 miles range. Later on, I believe after it was relisted, they pulled the service plug and was down to 1/8 charge and I believe 11 miles range. The coolant pump was cycling while I was there. The charge connector was intact, though the port was gone. They may have been able to charge it, but it's good that they didn't.
Awesome job in repairing this Roadster. Can't wait until it's back on the road.
I'll keep that idea in mind, if I get bored somewhen
Cool, thanks!!
I think #1 is still there. Error ID is 420. It shows up everytime, you take out the Service disconnect and insert it again. The error disappears, when you restart the APS in the service menu.
That's strange on #5, I never got an Airbag error massage even though there are no airbags installed, yet.
I think you wrote about the 62 miles somewhere at the beginning of this thread, which made me send someone to inspect the car
I told them to remove the service disconnect. Otherwise the batterie wouldn't have survived the trip to Germany. Was a bit surprised to see photos of it on the auction page later that day.
That's exactly where I ve been working on the last couple days, because the Roadster is at the bodyshop, now. The device works, but has to be verified and tested in the car. When it doesn't detect an external Control Pilot, it generates its own CP with regard to the attached cables current limit.
At first I wanted to put it into the PEM, but it needs an extra wire to the chargeport (to pass through the signal for detecting the charging cables current capability) and I don't want to drill an hole into the PEM or change the connector from PEM to the chargeport.
The chargeport connector has to be replaced with a J1772 or European Type 2 socket becuase it needs an extra pin there too.
No real progress on the car the last days. I'll probably take it to a different bodyshop. They didn't even start after having it there for two weeks now^^
Last edited:
I'm a little confused by this. If I understand you correctly, you are trying to generate your own pilot signal (CP) when there isn't one based on the attached cable's current limit. That's essentially what the UMC does. Now you are trying to make the UMC internal to the PEM. That will create a lot of problems. The first of which is that you can't use the stock charging inlet, but that's minor compared to the safety issues. You also can't use any of the standard 62196-2 charging cables. You can only ever use your own DIY charging cable. You could never sell the car legally because it would be out of compliance with 62196. It would also be difficult to charge from any public charging station because you would have to measure the incoming pilot and make sure your own CP doesn't exceed that. Too much room for error and not very practical.
I think Imore was referring to default charging at 230v with no pilot signal. The US spec roadsters are designed to charge at 15A if they detect 110v and the pilot signal is grounded, but I don't believe it will do this with 240v. I'm pretty sure the European Roadsters will charge at 13A 230v with pilot grounded. You can try it and see what happens. Ground the pilot signal and then supply 230v and see what happens but I suspect you will need the EU firmware.
I've been in a hurry earlier when I was writing the post, maybe my hands didn't write what my head thought
I basically migrated Martin Eberhards Foundry Charger to an Arduino Nano board and added some changes:
-Instead of an diode for recognizing the current capability of the adapter cable, I use an resistor between protective earth and the Proximity Pilot (1k5 = 16 A, 220R = 32 A and 100R for 63 A)
-Yes, the Roadster Connector has to be replaced by an EU Spec Type 2 connector or an J1772 connector because it needs an additional pin
-Probably the most important change and maybe an error source: I added a relay, which acts as a bypass of the "internal EVSE". When an external CP is detected, the Arduino only activates the relay and waits until the external CP disappears.
With the pilot grounded and at 230 V, the US Roadster 1.5 doesn't charge at all. Just for a test to figure out if theres another error in the PEM or if it's just the US Firmware, that prevented the Roadster to start charging, I put an function generator (±12 V, 1 kHz and with an 1k resistor in series) on the pilot line and the roadster started charging at 30 A. The 50 % duty cycle of the generator is recognized as an 30 A capable power outlet.
And that is just the case. There is no regular "European" software that tesla is giving out to a 1.5 owner here. A few of the options that you can chose, for instance break lights when regen slows down the car and TPMS yes or no, are selectable by a tesla mechanic only can not be saved with the 1.5 only for the 2.X. So is the the HID head light check. For the EU only 25 watt lights are allowed for the roadster (no mandatory wash or hight adjustment) so the software is adjusted for the 2.X only never for the 1.5. There are these little differences in Europe as explained by mr Tarantino. Just learned from the other thread that the mirrors are differed on the drivers side. Back in the day they told me that the were not.
The car is still at the body shop, but they finally started working on it and I'll probably get it back next week.
Does anyone know where one can find this sticker on the car?
Does anyone know where one can find this sticker on the car?
wiztecy
Active Member
I believe that's around the drivers side door. If you open the door I believe its on the body to the right (main body panel).
Last edited: