Hi power inverter for Nissan leaf motor. Dyno's 302.3hp p15
- Thread starter Arlo1
- Start date
Lebowski
10 MW
Ok, thanks... i use a totally different method and am not familiar with the usual terms...
P.s. i am not using the method described in the Microchip AN's
P.s. i am not using the method described in the Microchip AN's
Lebowski
10 MW
What would also be cool to test once is the sensorless from standstill with full torque, from the thread where I demo this with the drill...
Some time in a hopefully not too far future it would be amazing to see Lebowski's OR Axiom's software running on the Axiom hardware (all open source), so that a direct comparison with mutual learning effects would be possible. Something like the battle of the titans!
Or, Lebowski, don't you want to write your algorithm in C and add to VESC or Axiom? One advantage is that the FPU in the ST uC-s makes the calculations easier. Another possible advantage is the nonlinear BEMF (flux) observer in the VESC, that is very effective. I didn't dig much into the VESC code, but earlier I tried the same observer on my setup. In standstill if the load torque is high or the motor is blocked, it tends to oscillate, but in normal operation it very stable. This combined with your true sensorless position detection in standstill and low speed would be the winner.
marcos
1 kW
- Joined
- Nov 19, 2016
- Messages
- 348
I think I asked Lebowski if he wanted to try his algorithms in our board, but we code in C and he codes in assembler.
I agree it would be cool, and we also have the fpga available for custom functionalities that he could monetize as optional features.
I agree it would be cool, and we also have the fpga available for custom functionalities that he could monetize as optional features.
Lebowski
10 MW
I don't remember being contacted by you Marcos... I was once contacted and told "we need you to this and that" which is a sure fire way for me not to be interested. Even at work I don't accept that type of commanding...
C.... when I look at C I always have the feeling it is very inefficient. Lots of declaring and defining stuff, you really have to search to find a bit of code that actually does something. I find Assembly is simple as they are all the same, all have just the same basic operations (add, subtract, moving data, jumps etc). Plus in C you don't really know what it does on the processor, it may do very inefficient things.
C.... when I look at C I always have the feeling it is very inefficient. Lots of declaring and defining stuff, you really have to search to find a bit of code that actually does something. I find Assembly is simple as they are all the same, all have just the same basic operations (add, subtract, moving data, jumps etc). Plus in C you don't really know what it does on the processor, it may do very inefficient things.
marcos
1 kW
- Joined
- Nov 19, 2016
- Messages
- 348
Ugh, I'm sorry if that came the wrong way at that time. If you ever would like to give this hardware a try I'd happy to give you a walk through the coding workflow, at least the workflow I use which of course may not be the best.
The code is not ready to support paid options as it hasn't been a priority, but I think its doable with some groundwork.
You can code this thing is assembler, just type ASM { }and there you go, but the instruction set of an ARM core I think is rather lengthy to learn.
A friend of mine that is a linux dev guru told me he tried to beat the compiler-generated asm with his own asm and when -O2 optimization was enabled the human and compiler performed roughly the same. Then he tried -O3 and the compiled asm crushed his. So that and the portability steered me away from asm.
C++ on the other hand... yeah it can generate very bloaty code if you are not super careful.
I wonder if the FPGA can fit a dsPIC-compatible core, that way some of your asm would work, right? I know it can fit an risc-V core because I sinthesised one, and ran a test program compiled with gcc, but it doesnt leave much free logic gates, even for a decent pwm peripheral.
The code is not ready to support paid options as it hasn't been a priority, but I think its doable with some groundwork.
You can code this thing is assembler, just type ASM { }and there you go, but the instruction set of an ARM core I think is rather lengthy to learn.
A friend of mine that is a linux dev guru told me he tried to beat the compiler-generated asm with his own asm and when -O2 optimization was enabled the human and compiler performed roughly the same. Then he tried -O3 and the compiled asm crushed his. So that and the portability steered me away from asm.
C++ on the other hand... yeah it can generate very bloaty code if you are not super careful.
I wonder if the FPGA can fit a dsPIC-compatible core, that way some of your asm would work, right? I know it can fit an risc-V core because I sinthesised one, and ran a test program compiled with gcc, but it doesnt leave much free logic gates, even for a decent pwm peripheral.
zombiess
10 MW
Any particular reason you want to run 17" rims vs 15 or 16"?
InterestedinEVs
1 kW
- Joined
- Sep 21, 2010
- Messages
- 309
How wide are you looking to go?
InterestedinEVs
1 kW
- Joined
- Sep 21, 2010
- Messages
- 309
So...325's? Drag only? I'd be really surprised if you had grip issues at even double your last hp with decent race rubber.
InterestedinEVs
1 kW
- Joined
- Sep 21, 2010
- Messages
- 309
Arlo1 said:
Fair point, sir! Continue on!
My weight transfer calcs say you should be about 70/30 rear, so that's a good plan.
Got some 16" rims for $100 for now... Guy thought they were 15" and swears he had a 15" tire on them I was like oh ok....
I will likely run 225 tires all the way around or 245 in the rear and 225 in the front. Once I have traction problems I will go biggger likely a 17" 275 or 315 in the rear. Or 15" drag slicks.
Anyways these will let me finish the build I will try to leave a bit of space for bigger rubber in the rear
I will likely run 225 tires all the way around or 245 in the rear and 225 in the front. Once I have traction problems I will go biggger likely a 17" 275 or 315 in the rear. Or 15" drag slicks.
Anyways these will let me finish the build I will try to leave a bit of space for bigger rubber in the rear
Attachments
Lebowski
10 MW
Man, 250c, they'll be empty in 14 seconds !
InterestedinEVs
1 kW
- Joined
- Sep 21, 2010
- Messages
- 309
If you found another good deal on 15's, there are some really, really good 245-15's these days, though they are mostly meant as AX and road race tires, not drag. Still, pretty sticky.