The Road to Higher HP 100-200-500+ HP

Arlo1

1 TW
Joined
Apr 26, 2009
Messages
8,459
Location
Nanaimo
OK So I will start this way.... I have a 1969 Road Runner. I don't use it. Burning $100-$200 a night in fuel is not fun to me all it does is help some oil tycoon get more spoiled. I have done many things with this car DIY 16 injector EFI system where 8 run all spraying on the backs of the valves the time and 8 spray down the runner thoughts at Higher throttle positions for more HP a 4speed Auto with .69 ratio overdrive. I can get in the mid 20s for MPG (cdn. gallon = 4.54 litres) But Im done with gasoline. SO Either I spend a whole bunch of time working on it to sell it... Or I make a electric drive train for it.

Start. I have a bunch of induction motors. And I'm thinking about making them into PMAC motors see this thread. But the problem lies in the inverter. Although its purely amazing what a little mosfet can flow for amperage its still not enough. SO what I have thought about and apparently Mr Highhopes was thinking the same thing is splitting the inverter up to a 6 or 9 phase system.

Here is how I see it working. Lets just use a 6 phase as an example. Its not truly a 6 phase its more 2 separate motors in one case it will be 2 sets of three phase wires and non of either set will conduct to the other. Lets run though the thoughts as If the same motor was just going to be a 3 phase. Say I build the first motor for our Honda CRX. It will need ~150kw to make me happy. So lets start with trying to use 200 v mosfets. So maybe 150v under load. 150,000/150=1000 amps.
So we will need ~1000 or more phase amps for a 3 phase. But if we split it we could use 2 separate power stages for 1/2 the amps so 500amps or more....
Their is a couple benefits to this. 1 you will be taking sections of the motor and paralleling them to make 2 sets of phase wires which will give you 2x the inductance and resistance per set of phase wires. As well you can run a fuse to each power stage and if one fails it will still run depending on how you monitor current. This will be for the inverter because it only has to be designed for ~500 amps instead of 1000. Which as we all know causes the issues. Now a lot of oems are just upping the voltage and running the windings in series inside the motor. But I want to keep the voltage low as I can because higher voltage presents its own set of problems.

So I figure with the controller having 2 separate power stages it can have 1 doing the current measurement and phase voltage measurement and the other just a dummy powerstage where it just uses the same gate signals as the first one. The motor will need to be wound so the phase phases measured don't pick up stray measurements from the other section of the motor but I don't think that will be to hard.

if I can get this to work for the motor for the CRX at ~150kw then I will build the one for the Road Runner with even more power
 
Saving this post
 
Saving this post
 
IIRC, Sevcon has done 4x size6 controllers in master/slave configuration all feeding one motor with its windings setup to have 4 sets of phase leads out.

That would be 300hp capable if the motors are efficient.
 
I would go differently and take it to the extreme. I would wind every motor tooth seperately and drive it with its own power stage, for as far as it makes sense.

Example: lets say your motor has 15 windings. Conventionally you can wire 5 in series and end up with 3 phases. The voltages of the 5 windings in series are not exactly in phase, so instead of 5 times you only get something like 4.5 times the voltage (meaning the winding factor is 0.9. I would only wire phases together that have perfect winding factor.

I would run the thing with 15 power stages for max efficiency. The algorithm in the controller IC has been build for 3 but is capable of running any number higher than 2 phases, with FOC. Problem: a typical processor only has 3 (complementary) PWM outputs. So, you need an FPGA where you program in 15 PWM stages and an extra processor for the algorithm (maybe a Virtex4 would do). And you need 15 ADC inputs for all he current sensors.

Not difficult but an insane amount of work....

And... if you run 200kW you need at least 20kwh of batteries and even then it'll only run for 6 minutes...
 
Oh, and running 6 phases with the controller IC can be done, like you said, with inverting the signals to the 3 dummy phases.
I would not run without current sensors in the dummy phases... the current here should be the inverse of the current in the 3 master stages. So what you can do is run the phase wires for the dummy phases in the opposite direction through your current sensors, so that the output of the current sensors is the sum of both the master and dummy stage current.
 
Lebowski said:
Oh, and running 6 phases with the controller IC can be done, like you said, with inverting the signals to the 3 dummy phases.
I would not run without current sensors in the dummy phases... the current here should be the inverse of the current in the 3 master stages. So what you can do is run the phase wires for the dummy phases in the opposite direction through your current sensors, so that the output of the current sensors is the sum of both the master and dummy stage current.
That's a good idea but then Im going to be limited to a current by the current sensor. Now I don't think making a diy sensor is that hard so might not be a problem. I still don't see the problem of haveing the dummy phase un monitored. It will be 100% the same as the master phase.
 
liveforphysics said:
IIRC, Sevcon has done 4x size6 controllers in master/slave configuration all feeding one motor with its windings setup to have 4 sets of phase leads out.

That would be 300hp capable if the motors are efficient.
Thanks man. Do you have any more info on this?
 
dude - i am SO on board with this program. i've been researching this avenue for a few months off & on and am definitely intrigued.

i am more than capable of helping with or leading the hardware design of the power stage & layout but i lack skillz in software and have never rewound a motor (though as you know Ivan has posted a lot on how to do this). i can do also the analog/digital stage hardware design but it would be too much work i think (only soo much free time in the week). from my view, it seems the biggest problem will be to implement a multi phase FOC in a controller because that will be a unique solution algorithm with not much supporting documentation <--- This is where lebwoski chime's in and says "no problem" :D

the microcontroller isn't too bad because i think the modern DSP from TI have 6 hardware PWM outputs now and more than enough horsepower. one can always use an FPGA for gating if required, but i think the DSP all in one solution is possible. i had one picked out, i'll check my notes and get back to you for discussion.

it is possible to used a reduced set of sensors, 3 current sensors instead of 6 .. but .. you will lose on the reliability side and also will struggle to control during high accerlation which unfortunately are too important things for EV application. so .. i would plan to have two full up inverters with 6 current sensors, also measure phase voltage and DC bus voltage too but that's pretty cheap just takes board space cause of the high voltage. there would be just ONE controller (DSP).

as for bus voltage .. any voltage less than 600V is OK with me, i think it matters not at all from inverter point of view. it just depends on how you can get your battery pack voltage.. i'm not sure what the optimum would be. just thinking outloud here .. we know from ivan's forumn that #18 wire is the thickest you can reasonably hand wind a motor and for a typical EV you would want 50HP (right?) so there you have the current and the power now just solve for voltage? i looked into 5phase 6phase 9phase motors and what i think is the best choice is 6phase because it allows for 2 separte full up 3-phase inverters (well known how to construct) and gets you the double reliability. 5-phase would be an akward hardware design, 9phase is not much value for EV application and would lead to excessive hardware costs for little value gained over 6-phase option. my preference would be to use an induction machine because they are so easy & cheap to get and easy to modify for any voltage. i don't care too much if the power density is 20% less ... its a car (not a jet) i am willing to trade off volume for significant cost & effort reduction.
 
Two reasons I chose pmac over induction. There is no induction motor on the cheep or easily modded to make what I need for hp. You can get a copper rotor and oil cool the motor but... For the cost of a copper rotor I can make a magnet rotor.
Second I have lots of controllers for this.

Highhopes if we make a stator with 6 phases you can have 3 mirror the other 3 and not need 12 gate outputs.you will only need 6 gate outputs (normal 3 phase stuff) then the 2 seprate three phase sections will fire at 100% the same time and because the motor is mechanicly locked you will only need to sense one 3 phase section. For current sensors yes having all phases monotered is best but I think if built right you can make sure the second set of phases are 100% the same and it should flow the same current.
 
its probably too early to be arguing this point .. but in principle i agree with you. if the motor is operating at norminal steady state speed and system is relatively balanced there is no need for 6 sensors, only 3 .. heck you could use 2 with a "3 phases sum to zero" for the 3rd pseudo sensor.

then you accelerate or decelerate and your system has to be in control during this time of unbalance (during acceeration is when you find the unbalance) so measuring the 3rd phase becomes important becaues you will have zero sequence current during acceleration unfortunately.

still.. 3 sensors instead of 6 so not bad. but then you have a failure on one of the windings and it goes unnoticed because no current sensor on that winding. perhaps the failure is not too bad (high impedance) so it shows up as heavily unbalanced winding.. motor beggins to cog with no ability to control that phase so no ability to recover softly. still.. not too bad ... but then that initial failure has total short-circuit and .. sadly, no current sensor to detect so there is no shut down and significant damage occurs.

for an EV application i think a robust topology is critical to long term success and as any driver knows, nothing worse than a car that breaks down on the highway. with 6 independent phases with 6 independent sensors any ONE and almost any TWO phases can have a complete failure without the driver hardly noticing (just a "check engine" light). the software design could easily handle both if you are adamant about 3 sensors as i am about 6 by having a user selectable value. just have the other 3 sensors that are not in your embodiment equal to the 3 measured values you do have and let the controller continue on happily.

and just so we're on the same page, when i'm talking HP rating for motors i'm talking about continuous rating. i don't think it makes much sense to speak about peaks for 30 seconds or some such as you would if we were talking about an ICE.
 
There is no induction motor on the cheep or easily nodded to make what I need for hp
agreed when you are talking greater than 100HP (continuous). i just read your thread topic, yikes 500HP! wow.. how big is your car!? :)
 
So lets start with trying to use 200 v mosfets

if you look inside some of the other commercial EV controllers rated for 500A or higher that are using discrete mosfets, how many do they have in parallel per switch? this could give some idea what the limit will be.

my personal opinion is that 4 in parallel is the most you can do in a reliable way using a conventional gate driver and taking the time to match the mosfets before install. maybe you could do 6 but i think it would take a dynamically balanced gate drive (could be a fun project!). if you need higher current, then that is where the 6-phases comes in .. and if you need higher still, i'm thinking you'll have to go to a module. :?:
 
just kicking off some discussion threads..

for me, 100HP continuous is way more than enough to have fun in an EV. i think even 50HP is fine. .. so, for me, i like induction motor because at these power levels such a motor is super easy to find at motor re-wind shops. i already have a 10HP induction motor that i got for $100. it will need a re-wind for lower voltage (100V?) to better match a battery supply, so maybe by the time i'm done it will cost me $300 ? that's not bad for practically brand new 10HP.

anyway, what i wanted to say is that maybe it doesn't matter too much what motor you chose. the hardware design is practically the same except for the rotor position feedback. the software changes a lot i think. so for me its still a worthwhile projecct even to use PMAC because the hardware could be re-used on an induction machine though you would need to download the TI DSP with software to run that sort of machine (InstaSpin?) otherwise, pretty much the same.
 
A Road Runner DESERVES 500hp. If you're going to electrify it, do it right.

There's the Road Runner of the early/mid 70's that I'd really want, then there's the package for a '69 I'd settle for. Plus a custom job it kills me I could never run out and get.

72plymouthrr_1.jpg

1969RoadRunner.jpg

plymouth+roadrunner+van.jpg
 
Dauntless said:
A Road Runner DESERVES 500hp. If you're going to electrify it, do it right.

There's the Road Runner of the early/mid 70's that I'd really want, then there's the package for a '69 I'd settle for. Plus a custom job it kills me I could never run out and ]
Dad had a few Mopar muscle cars. His first was a 72 Road Runner.

Yup. I have made it pretty fast on a ok budget. But if I am going to keep it I need to electrify it. I grew up with this car. I bought it after I broke my legs and shoulder and was in the hospital for a month when I was just about 16 spent my 16th birthday in a wheel chair. Parents made me sell the 93 rm125 I was jumping over my buddies car so I said fine Im buying a muscle car like dads 1969 charger. Found my Road Runner when I was 17 and got it for $1600 worked as a janitor at my high school and put 850 a month into it while doing work experience in and extra hours at my cousins body shop to earn the labour for a pint job. Had it on the road for my 18th birthday and Grad! The local rich kids voted it the nicest car in school which meant a lot to me because I built it with my own hands while being made fun of for being a janitor lol. Everyone was envious when I drive it IM always getting the thumbs up and people wanting to ask me about it and do a burnout lol.
But my life has changed now. Since ES I have learnt there is more to life then material Possession. I don't car what I own or what it looks like I just want to build cool stuff and ride/drive it!

So.... I have been collecting Induction motors because yes I can make some good power as an induction motor but they are a cheep lamination goldmine. That's the one thing I found expensive/hard to find for DIY motors. The stator I might use is 10.25 ID and 5.125 long. I think it will work as a good PMAC motor with a max rpm in the 4000-6000 rpm range. IM off to a slow start with FEMM....
 
Electrify something much much lighter. Batteries are still too expensive to electrify our 70's steel land barges.
 
wow.. very nice cars. even the van is cool ;)

i'm thinking that the 500HP rating you want comes from the idea that car probably had a 500HP ICE in it. ICE motors are rated in peak while electric motors are rated in continuous. so.. if you want the same 500HP ICE equivalent, then that would be an electric machine of some lower value.. just a guess but maybe 300HP? what would a 300HP PMAC weight? are you OK with that weight?

i'm trying to narrow the discussion down to power level, then voltage & current to get some ideas of the hardware necessary.
 
I thought John in CR had a hub motor that was 6 phase. In Solar panel design it has been popular to go to the use of microinvertors mounted on each say 200 watt panel of a 10 kw system.
with the huge expense of high power controllers I had been wondering for a while why you couldn't have one brain but all the high power elements of the invertor distributed down at the battery sub pack level that way you would be switching much lower voltage and currents.
 
HighHopes said:
wow.. very nice cars. even the van is cool ;)

i'm thinking that the 500HP rating you want comes from the idea that car probably had a 500HP ICE in it. ICE motors are rated in peak while electric motors are rated in continuous. so.. if you want the same 500HP ICE equivalent, then that would be an electric machine of some lower value.. just a guess but maybe 300HP? what would a 300HP PMAC weight? are you OK with that weight?

i'm trying to narrow the discussion down to power level, then voltage & current to get some ideas of the hardware necessary.
Let's keep it at a goal of 500hp for the road runner. I don't want it to be kinda like it used to be I want bad ass. You have to understand I run a tranny with a 2400 rpm stall that will flash 3000 on a launch and this makes for an epic start soke guys run much higher but I wanted to be streat able. When you combine the power of the motor at 3000 rpm and 1st. Gear being 2.45:1 pluss the final gear we will need it to be pretty serious to even keep up. I'm talking more hp then tesla... But one step at a time I'm on a budget but I love to design,build and learn.
 
hal2000 said:
I thought John in CR had a hub motor that was 6 phase. In Solar panel design it has been popular to go to the use of microinvertors mounted on each say 200 watt panel of a 10 kw system.
with the huge expense of high power controllers I had been wondering for a while why you couldn't have one brain but all the high power elements of the invertor distributed down at the battery sub pack level that way you would be switching much lower voltage and currents. Then I read this thread which seems to go some distance toward this idea. Is there any way to create and control ac power at the cell level of the battery say by combining a traditional fixed voltage cell with an integral capacitor that would store and increase the voltage over short time durations needed for phase switching
Would this allow you to forgo the expensive high power sections of your controllers and multiply the apparent voltage coming out of each cell. I wish I knew enough about this stuff not to have to ask!
You and lebowski have a point winding ever tooth separately will help make it easier to control. Hmmm the more phase wires we have the lower the amperage and the lower the number of fets in parallel we will need.
 
if you want 500HP continuous out of an electric machine + the batteries i think the weight will be more than the car.. ?
 
John in CR said:
Electrify something much much lighter. Batteries are still too expensive to electrify our 70's steel land barges.
Lmfao. Dude right now my road runner weighs 3500lbs with a heavy iron v8 and a heavy 4speed auto. The only peice of carbon finished is the rear bumper but plans for a front clip and dash from carbon fiber are in order. Plan is to make it work while battery tech advances.

But seriously so many people think old is heavy but its not the case most cars with v8s are well over 4000 lbs now adays even look at tesla.
Not to mention the lack of mandatory features mine has compared to the new stuff helping make it easier to work on and lighter weight.
 
HighHopes said:
if you want 500HP continuous out of an electric machine + the batteries i think the weight will be more than the car.. ?
Have you looked at nano tech Lipo? Or what warp factor 3 is using? I will calculate that tonight when at a computer.
 
just to put things into perspective.. the 100HP rated continuous Hyperpack from Infioneon costs $5K. so, 500HP would cost .. 25K? so if you build yourself maybe cost 10K? and that's if it works the first try? i dunno man.. seems like a big risk. the design & software part is not a big deal because its just on paper but eventually you'll want to give it a go. hope you have some deep pockets :!:
 
Back
Top