RX400H transaxle

[Canberra32]

"pretty much" and "approximately"
See now my "get a bigger hammer, motor, lever" is so simple the math is just < >

And for reference with I think it was 325hp, 18psi and 100kg stripped out the mx did 0-100 in a flat 3seconds...
I stripped teeth on two diff centers before I simply (not so simple fitting) put in a bigger one
However... After a defect notice and an epic emission test fail now it's a NA 1.8 with a pile if bits on the bench gave it back to the EX-wife due to its state of pinkness.
Everything about those little cars was designed to make it fast.
Other than the sloppy tunnel issue and a lack of space they are perfect for EV conversions buy them now people!!!!

[weber]

Great stuff Canberra32. For EV conversion I recommend the MX-5 NB series 2, 3 or 4, i.e. from November 2000 to 2005. These come with a stronger diff and better brakes than the NA or NB series 1. And they are much safer, having ABS, dual air bags and power steering as standard. While their stronger diff has a 3.636:1 ratio, this should still be OK for direct drive if star-delta (or series-parallel) switching is used.

The MX-5's lack of space does make fitting the batteries a challenge, but we have spent many months on that problem, and have solved it, provided you use 40 Ah cells. The new grey CA series from CALB look brilliant.

[Canberra32]

We had the NB 1.8
The one with the none pop up headlights for those out there less car savvy

[PlanB]

You would need to write the code for Tritium's Driver Controls Unit (DCU) to decide when to change, and to send the right commands to the WaveSculptor to make it change quickly and smoothly.

Not exactly a 'clicking black boxes together' pathway chief. I like that 3350+Nm your ABB gets from the 300A limit, why did you opt to keep the gearbox rather than go straight onto the diff?

[weber]

Not exactly a 'clicking black boxes together' pathway chief.

I assumed you would want to have _some_ personal input, and I had the idea you were a microcontroller guy. But Coulomb and I could always write that code for you.

I like that 3350+Nm your ABB gets from the 300A limit, why did you opt to keep the gearbox rather than go straight onto the diff?

It was very touch and go, whether we should lose the gearbox or not. In the end it was just a lower risk option for a first conversion. If it turns out OK to drive it in 4th gear all the time then we could go direct next time with the same (2-pole) motor. If we found we could drive it in 2nd gear all the time then we could go direct with a 4-pole motor. But you're always going to get better accelleration with a gearbox (or with star-delta switching), provided you can change fast enough. And that's part of the fun of driving a sports car. We also had the idea that it would be easier to sell if it worked exactly like the petrol version.

BTW, I love this MX-5 quote from Jeremy Clarkson:
"The fact is that if you want a sports car, the MX-5 is perfect. Nothing on the road will give you better value. Nothing will give you so much fun. The only reason I’m giving it five stars is because I can’t give it 14."

The Top Gear team have taken turns reviewing every model of MX-5 over the years and they just love them all. In one case referring to it as a sort of honorary classic British sports car. See
https://www.google.com.au/search?q=Top% ... %20youtube

[PlanB]

Yeah I can't find much else that comes in under the tonne. The Supras are nice (I want a hardtop) but I doubt I could get the donor below 1000kg even after you take out the 6 pot & gearbox, same goes for RX7s.

Boy those solar boys in Newcastle spared no expense on that Delorean @ $60k not counting the car! Love that EVO motor purr. I think I'm in the wrong line of work Dave.

I just shudder at the thought of multiple changeover contactors swapping star delta & TJs box not looking briefly into a dead short or something but thanks for the offer. Ross must be a genius to pull that off.

[Johny]

It really is just finding the space and maintaining two confurations in the controller. Star-delta contactor pairs are easily available that have mechanical interlocking to stop both closing together. There is no advantage if the controller can provide delta current though.

[weber]

There is no advantage [to star-delta] if the controller can provide delta current though.

I used to think that. Or at least I think I used to think that. But since my recent self-education effort, I now think that you can always get more torque out of a motor at low rpm by feeding it in star with the same current your VFD can feed it in delta. The only question is how rapidly it will heat up when you do this.

[PlanB]

I checked with the TJ on this Dave. The Sew is in delta by default, switching to star gives less torque but out to a higher speed than delta before it drops off. So I'm never going to get more than 5.1 x 240 = 1224Nm launch torque out of the Eurodrive direct to a differential. I'm just not sure it's enough, given Ryan's problems with hill starts on his Rx400H transxale at 900Nm, only 26% less.
In theory if I switched to star at the speed where delta torque has dropped below star torque, acceleration might improve but it's a lot of extra switching & code changing on-the-fly in the wavesculptor for a maybe.
Woody's 90kg/366Nm monster would give 1867Nm with a 5.1:1 diff but I think it's probably too big @ 90kg for a direct-to-diff coupling.

[coulomb]

I checked with the TJ on this Dave. The Sew is in delta by default, switching to star gives less torque but out to a higher speed than delta before it drops off.

Something's wrong here. Any three phase load gets more current through each element in star than in delta, since all the line current (instead of one divided by the square root of three times the current). So more current in the windings should result in more torque, but out to lower speed, since each winding gets only 1/sqrt(3) of the voltage when star connected. So that's why you start in star and switch to delta at higher speed.

I would think that the SEW drive, wound for 100V, is 100 V in delta; otherwise, if it was star by default, you could overclock it even more by connecting it in delta.

So the idea is to start it in star, even though that's nominally 173 V, then switch to delta after a certain speed.

I sure hope I got that right...

[ Edit: added "in star" before "than in delta" ]

[weber]

I checked with the TJ on this Dave. The Sew is in delta by default,

Yes. I was assuming that.

switching to star gives less torque but out to a higher speed than delta before it drops off.

I seriously doubt that TJ said that, although we all have our off days. As coulomb suggested, this is entirely back-to-front.

Starting off in star gives more torque but out to a lower speed than delta before it drops off. And the motor will heat up 3 times faster for the same VFD current. Tritium suggest the motor can take 300 A for 30 seconds in delta, so it should take 300 A for 10 seconds in star, which is still quite useful.

In theory if I switched to star at the speed where delta torque has dropped below star torque,

No. You start off in star, and it should switch to delta at the speed where star torque has dropped to delta torque. Actually, to minimise motor heat dissipation, it should switch to delta sooner if you back off the pedal such that you are asking for torque that can be supplied by delta. Like an auto gearbox.

acceleration might improve but it's a lot of extra switching & code changing on-the-fly in the wavesculptor for a maybe.

Acceleration will improve.

At the end of the Wavesculptor manual it says:
"18.5 CONFIGURATION COMMANDS
Commands to configure and calibrate the motor controller are also present.
Contact Tritium for a full specification if necessary."

Woody's 90kg/366Nm monster would give 1867Nm with a 5.1:1 diff but I think it's probably too big @ 90kg for a direct-to-diff coupling.

Our motor is 95 kg.

[Tritium_James]

Yeah I think I replied without reading it properly - as Weber said, I had the reply exactly back-to-front re star/delta.

You'd have a config for star, and a config for delta, and can switch with a simple CAN command, no problems. The only thing I don't know will work is switching configs with the motor already spinning - it should work, but we haven't tested it on an induction machine. It does work with BLDC motors.

[PlanB]

Starting off in star gives more torque but out to a lower speed than delta before it drops off. And the motor will heat up 3 times faster for the same VFD current. Tritium suggest the motor can take 300 A for 30 seconds in delta, so it should take 300 A for 10 seconds in star, which is still quite useful.

So does this mean I get 3X torque in star, 720Nm instead of 240Nm?

[weber]

Starting off in star gives more torque but out to a lower speed than delta before it drops off. And the motor will heat up 3 times faster for the same VFD current. Tritium suggest the motor can take 300 A for 30 seconds in delta, so it should take 300 A for 10 seconds in star, which is still quite useful.

So does this mean I get 3X torque in star, 720Nm instead of 240Nm?

No. Sqrt(3) times torque. Torque is proportional to current while rate of heating is proportional to current squared.

[weber]

There is effectively a piece of this thread in another thread. My bad. In it we look at the specs for Woody's 4-pole 115 V ABB motor, as a possible option for PlanB's conversion.

See the 8 posts starting with this one, but don't forget to come back here to reply.
viewtopic.php?title=how-to-convert-a-hy ... 661#p37910

[PlanB]

So TJS motor at 60kg & $4k for 240 x 1.732 = 416Nm with star/delta switching OR Woody's beast at 90 kg & $2k at 366Nm with no switching.

[coulomb]

So does this mean I get 3X torque in star, 720Nm instead of 240Nm?

No. Sqrt(3) times torque. Torque is proportional to current while rate of heating is proportional to current squared.

Well, torque is approximately proportional to current; you lose some because of the way vectors add, and some other small effects.

But it's close. I'd say you could confidently expect 1.65 times the torque (about 95% of sqrt(3)). For those that want a better figure than that, an ABB paper somewhere has a formula for it. It's been mentioned in a recent post. I'm not at the computer that would have it.

So a little under 400 Nm for the SEW motor, I'd say.

Gee, 1.65 x 350 ~= 575 Nm... oops!    Sorry, PlanB, you asked me not to mention that figure. Slip of the key   

[ Edit: (matching (my brackets)) ]

[PlanB]

I know this is probably a dumb question but I'll ask it anyway. If I reversed an MX5 diff so the input was pointing backwards could I remove the fuel tank & mount a 90kg motor under the boot? It would save all that jigsawing trying to get between the diff & the seats.

[weber]

So TJS motor at 60kg & $4k for 240 x 1.732 = 416Nm with star/delta switching OR Woody's beast at 90 kg & $2k at 366Nm with no switching.

Is the SEW really $4k?!

I know this is probably a dumb question but I'll ask it anyway. If I reversed an MX5 diff so the input was pointing backwards could I remove the fuel tank & mount a 90kg motor under the boot? It would save all that jigsawing trying to get between the diff & the seats.

You should come up here and crawl under an MX-5. It would answer so many of your questions.

The fuel tank isn't under the boot. It's immediately behind the seats, above the rear subframe which is above the diff. The muffler is under the boot, but removing it doesn't give enough height, including breakaway-angle, to fit a 132-frame motor. We just manage to fit a 210 mm high battery box there.

The only practical place to fit the motor for gearbox-free drive is where the gearbox is. I checked it last Friday and found that the motor would probably not project into the engine bay at all. This assumes you couple it to the tail section of the old gearbox. I think you'd need the small-diameter-flange option on the motor for it to fit.

Here's the drawing of our ABB motor. Woody's has the same dimensions. This is with the standard (large) flange.
http://library.abb.com/global/scot/scot ... 1.5003.pdf

[Renard]

So TJS motor at 60kg & $4k for 240 x 1.732 = 416Nm with star/delta switching OR Woody's beast at 90 kg & $2k at 366Nm with no switching.

Is the SEW really $4k?!

For the SEW I paid $2100 last August.

($2300 inc GST)

[PlanB]

Thanks Dave. I read that long thread a while back about using the back end of a gearbox & all that rigmarole with oil seals. I never did figure it out. Why wouldn't you just fit a universal joint to the motor shaft, hook the tail shaft onto that & be done with the gearbox altogether?

[woody]

It really is just finding the space and maintaining two confurations in the controller. Star-delta contactor pairs are easily available that have mechanical interlocking to stop both closing together. There is no advantage if the controller can provide delta current though.

Small advantage of star in this case:
because you have more current available you could overvoltage / overcurrent the motor to get more torque at low speed (at the cost of efficiency while you are doing this)

[woody]

Thanks Dave. I read that long thread a while back about using the back end of a gearbox & all that rigmarole with oil seals. I never did figure it out. Why wouldn't you just fit a universal joint to the motor shaft, hook the tail shaft onto that & be done with the gearbox altogether?

In my case the gearbox output shaft provides:

1. Rear mount for the engine/gearbox
2. Mechanical speedo Drive output (1000 rotations per mile)
3. Sliding yoke for the tailshaft to allow for the diff moving up and down (live rear axle).

I haven't decided which way to go: (3 looks good)
1) long tailshaft with expansion + electronic speedo drive/internals
2) attach the back half of the gearbox to the motor somehow and keep the current tailshaft and speedo
3) just strip the internals of the gearbox out and leave a straight through shaft (i.e. 4th)

I don't think I can keep the clutch (especially)or gearbox (unmodified) as the emotor torque is about 4 times the petrol motor torque.

[weber]

Thanks Dave. I read that long thread a while back about using the back end of a gearbox & all that rigmarole with oil seals. I never did figure it out. Why wouldn't you just fit a universal joint to the motor shaft, hook the tail shaft onto that & be done with the gearbox altogether?

You really need to look under an MX-5. Although I'd be surprised if there aren't enough photos of the underside of an MX-5, between us and Ian Hooper and all other "Miata" conversions on EV Album, to enable you to grok the whole PPF (power plant frame) thing. Here's one of Ian's, with his motor and adapter in place of the gearbox



The PPF is a truss beam that connects the gearbox casing to the diff casing, running alongside the propshaft. The gearbox does not mount to the chassis anywhere. Its casing forms a structural component between the engine and the PPF. The whole drive-train, consisting of the engine, gearbox, PPF and diff, bolts together into a rigid assembly that is hung from 4 points -- the two engine mounts and the two mounts on the "wings" attached to the diff.

When eliminating the gearbox, the motor will need two beams that go forward from its feet to the original engine mounts, and it will need to somehow connect rigidly from its flange back to the PPF.

That's the motor casing. Now consider the shaft.

The distance between the motor and the diff will vary plus and minus a few millimetres as the car goes over bumps etc. and the chassis flexes. That's why the propshaft always has a spline somewhere that gives a sliding joint. As with many vehicles, the MX-5 has a male spline on the gearbox output shaft and matching female inside the propshaft. It also has a bush inside the tail end of the gearbox casing, that forms a plain bearing with the outside of the propshaft that supports it and stops it flopping about on the spline, while still allowing it to slide in and out of the gearbox by those few millimetres.

You can't attach the propshaft or its uni-joint rigidly to the motor shaft or there will be extreme end-thrust loads onto the motor bearings and diff bearing (unless you add a sliding joint somewhere else in the propshaft).

You could design and manufacture an adapter to go between the motor and the PPF, containing a bush or bearing to support the propshaft, as Ian Hooper does here (also having to lengthen the propshaft): http://zeva.com.au/conversion_blog.php?post=12

Or, as Ross Pink did on his eVan, and I suggest on the MX-5, you could make life simpler and unbolt the tail section from the gearbox and use a simple adapter plate between it and the motor flange. And rigidly couple the sawnoff gearbox output shaft to the motor shaft (as it appears Ian might have done).

Don't worry about the oil-seal thing. That's a minor consideration. Ross Pink got around the problem by using grease instead of oil. Alternatively, it can be dealt with by a simple drainhole drilled in the adapter plate, vertically down from the hole where the motor shaft passes through it.

[PlanB]

Tnks for the pic, that PPF is unusual. Well at least to me it is but then the last time I crawled under rear wheel drive cars was 40 years ago, I guess the world moves on.
All this mechanical stuff is reminding me why I went for the Lexus transaxle in the first place. I wish I could get my hands on something more EV friendly mechanically.