Engine and Transmission Tech

I've been thinking about the control scheme for this transmission, and I'm currently leaning towards an industrial programmable logic controller (PLC) such as this one:

[news.thomasnet.com]

This particular example is built for harsh environments, submersible to 3 ft, has 4 inputs and 6 outputs, 4 of which are pulse width modulated (PWM) programmable with an integral feedback loop and runs on 8-48vdc. Pretty near ideal, it may have a big brother with a bit more capacity. (The A650E has 7 solenoids)
I'll need to get more info on the control scheme for the tranny before deciding on a controller, but PLC's have been around for at least 30 years and are pretty refined and versatile. There's no reason I can think of that there wouldn't be an off the shelf item that would be perfect for this application.

JB

Michael came through for me on the AA80E at last, and I have it in the lab on an engine stand with the pan and filter off to get a good view of the inner workings. It's a rather beefy unit, significantly larger than the A650E as you would expect from the increased horsepower and torque ratings. It uses a different and slightly larger bellhousing pattern and a larger bolt pattern for the torque converter as well with 10mm bolts instead of the A650E's (oddball) 9mm fasteners. Common hardware bolts fit just fine but hardened bolts will be used. The VIN tag confirmed that it came from a 2008 IS-F.

My local Lexus dealership has been extremely helpful in assisting me with my efforts. Although their parts prices are understandably rather high, their attitude and helpfulness is well beyond reproach and they have gone the extra mile every time I've been down there. That's the kind of help you rarely find and it's really appreciated when you do. Anyway, one of the most significant discoveries was that the IS-F uses a separate transmission control module (TCM). This unit controls all 9 (!) of the shift solenoids and inputs all of the sensor information from the transmission. However, it receives commands from the engine control module via CAN bus, and the ECM tells it when to shift and what gear to select. If there is no CAN bus input it uses only 1st and 3rd gears, from my cursory review of materials just received. What this means in practical terms is that the TCM must communicate with a CAN bus capable unit such as either the stock ECM, one of the later versions of Megasquirt with suitably modified code, or possibly a CAN bus capable PLC. It's also possible that a suitable PLC could control the transmission directly. I have some studying to do to determine which options are the most feasible, but at present I'm getting quotes for the stock TCM, ECM, and related wiring harnesses.

JB

Nice and narrow. I'd love to see that behind a 4.6/4.9 EFI MGB.

I have experience with programming, even down to 0's and 1's, but there would certainly be a learning curve with this application. I suppose a semi-manual paddle shifter would be one compromise, but a fully automatic option would be a better baseline. Throttle position and manifold vacuum sensors would get you started. Is shift firmness adjustable on this transmission?

Yes, at least 8 of the 9 solenoids use PWM signals and they use the linear characteristics of the solenoids to control engagement, so shift firmness is a simple matter of changing the PWM slope. I'd need a PLC with 8 or 9 PWM capable outputs (one is inverted, apply decreases line pressure) and probably at least ten inputs, some digital, some analog. I have all of the solenoids identified as well as truth tables for on/off conditions so basically we'd be looking at PWM slope and delay/overlap times for a simple manual shift, plus any programming for automatic functions. At least that's my take on it at present. Where are you located Paul?

JB

Jim, I'm in Idaho. Meridian, Idaho, and right on the Boise, Idaho border.

You sound like you have experience with controllers. I have had exposure to them so there's little mystery, but I lack practical experience. Mostly disk drive electronics from the late 80's and early 90's. My background was originally electronics, starting with radio/television. That evolved into digital electronics, especially disk drive controller and host bus adapter (GPIB/HPIB, ST506/WD2010, SCSI, etc), but some PC server stuff, too (ISA/EISA/MCA). Obviously, programming became important after that. I still do some stuff with low-level interfacing, but it tends to be more at a host command level - TCP/IP but more specifically SNMP/PML via UDP.

What I guess you'd be doing is creating algorithms to change PWM signals to control shift points and shift firmness. Sounds easy on paper but there's so many corner cases to account for. Sudden changes in throttle position, sudden changes from acceleration to braking and back, downshifting, and so many more.

I have a knack for that stuff, but there was never much demand for it where I worked. It was always more programming.

Paul, that sounds like just the sort of expertise we need on this project. If you're willing to help on it I'm pretty sure there will be a market demand for a controller, as nobody is offering one. It's something we might be able to offer to the British V8 crowd at the very least. I haven't really looked at the PLC's lately but I expect they rely heavily on cobol and related languages and graphic programming interfaces were becoming popular last time I checked. I have a busy weekend planned but if I can I'll start looking at what PLC's are being offered on ebay.

JB

There's more research to be done but the nano/pico/etc type of PLC looks promising in terms of size and capability. They commonly use a "ladder" type of programming logic. I'll post more information as I get it.

JB

Jim and clan, is it possible, TRD-Toyota racing developement, could lend assistance ?

Jim,

Came across this south African Co. that makes ECU's for auto trans, also very cheap with the dollar/Rand ex, rate.

[www.spitronics.co.za]

Love the developments on your blown 340, at a lazy 80 BHP per Litre you've got to be looking at 450+ BHP!

Regards,

Kevin

Mr. Blackwood, perhaps you could chime in ? Specifically the V16x, V160 and V161, built by "Getrag' for Mark 4 Supra use, thru 2002'. Reportedly,"much" stronger than R154. ratios: 1=3.821, 2=2.36,3=1.685,4=1.312,5=1,6=.793. Perhaps your, auto trans supplier, has a deal on these ? Anybody? Thanks, roverman.



Edited 1 time(s). Last edit at 06/05/2010 08:41PM by roverman.

I've about decided that Michael specializes in top end sports cars, but never having seen his yard I really don't know what he does or doesn't have. Bill Young might be curious to check it out as he's in Bill's area but it's too far for me to go just for curiosity.

Anyway, Art I am totally in the dark on those transmissions you mentioned so can't really help you there. As far as TRD, who knows? But right now I'm trying to get a large diameter boring tool for the mill made up so that I can finish the bellhousing adapter. If I get that done, and the engine in the car before the V8 meet I'll be doing right well, because first I have to bore the center hole of the adapter which is about 13-1/2" diameter and center it on the SBB crank. That is a cut-and-fit, measure, repeat until centered, type operation and with no micrometer dials on the mill it's an interesting challenge. Heck, just cutting the hole round and roughly centered is sort of a big deal since I also have no downfeed on the quill yet and the rough opening is a good 1/2" off center. So it's not quite in the same league as doing the job with a rat-tail-file, but doesn't miss it by all that much.

JB

Jim, can you throw up a sketch ? Sounds like you need a rotary table or a face plate for your lathe ? Lathe will be easier/faster, if you can "swing-it". LOL. Did you have the blank water jetted/roughed out ? Rotary tables are very versatile. I suggest vert/horizontal, much more possibillities. Ok, looks like your lathe isn't a gap-bed, so your plate woun't swing. In leiu of vernier dials on the mill, I suggest a couple of mag bases with adequate travel indicators, to read table and quill movements, (actually more accurite than dials). Good Luck, roverman.



Edited 1 time(s). Last edit at 06/07/2010 08:49PM by roverman.

Too big for the lathe, and my rotary table is too small also, but the real problem is that it has to locate off the dowel pins. Not having a print for the BOP bellhousing pattern, I had to transfer the holes from my 2004r case and rough out the center hole with the porta-band. That worked fine but now it's time to finish the center hole. Once that's done I'll clamp it to the AA80E case and dial it in on the input shaft and then drill that pattern, and then a little finish work and it's done. As for the lack of dials, that doesn't really bother me. Standard 5 turns/inch ballscrews, no backlash, and I can just about dial in the right number of thousandths blindfolded. With a pair of locating pins bolted to the table I can take a cut, check the center, put it back and make the corrections, and if I'm not well under .005" by the third shot I'll be very much surprised. I'll be shooting for less than .0005" TIR. But worst case, I could mark the high spot and note the runout figure and match that on the AA80E case and be right on the money. Since it's a stationary part all it really needs is clearance.

But before I can do any of that I have to finish the boring bar holder.

Now the converter adapter may be a different matter. The one I made that fits the 5 speed was cut from the drop and since that is MIC-6 plate I'm considering a complete re-do on that. Billet would be considerably stronger.

JB

The center hole is machined. Woo Hoo!! Almost had to eat my words though, on the second pass I began to doubt I could hit the mark with just one more. But, on the 3rd go around I had .0025" TIR, well under the .005" I had predicted. Part of the problem was that the locating pins were spreading the ring slightly causing the hole to go oval. That meant I had to jack around with the pins on the mill table at the same time I was fighting the x and y axis offsets, and to complicate it further, due to the bind one of the pins was moving just a smidgen every time I took the ring off the mill.



Eventually though I got it all sorted out and on the last cut indicated .0002" TIR, which I was quite excited about. But then I thought I'd better put the other three bolts in and see if that made any difference and sure enough, as I tightened them I could feel the plate pull in ever so slightly. Just enough to add another half thou to my reading, putting it at .0007" TIR. RATS! Missed my target. But it wasn't by enough to make it worth another cut so I let it go. To put that in perspective, it means a center-to-center misalignment of .00035" which is about one tenth the thickness of a human hair. Any but the most high precision bearings will tolerate that much misalignment easily so I'm not going to worry about it.



You may think this is overkill and that the flexplate will handle it and that is true. But if there is no misalignment to handle to start with everything works better. So next is drilling the holes for the transmission locating pins and bolts and then external clean-up. For that step I will clamp the ring to the bellhousing, dial it in to center, and stress it side to side to bring it to round like it is when mounted to the engine and then transfer the holes.

JB

Jim, Looks like a "lot" of work.Nice progress. I look foward to having "some" time for my progects. I have a10" vert./horiz., rotary table and for larger pcs., I install a "cheater" plate on top,(approx"16" dia. X.5" alum plate). This still clears the crank handle. Not heavy duty, but enough to squeak by. roverman.

Jim, I see TCI is advertising a tranny control with rotary dial or touch pad. I don't know the details or cost.[www.tciauto.com] I clicked on pricing,Ouch! $793.



Edited 3 time(s). Last edit at 07/26/2010 01:23AM by mgb260.

We rented a Nissan Altima on vacation last week so I got to drive one of the newer 6 speed automatics. This one has two modes. Full auto, or you can slot the shifter over to an up-down shift slot for manual. Now we were in flat Florida and the family was with me so I never had any real chance to wring it out, but my general impressions just driving around were that this new technology is very very good.

Fellas, I think the manual transmission is on it's way out. Seriously.
The brakes otoh, were way touchy.

JB

Jim said: "I think the manual transmission is on the way out". Wasn't that the general consensus, when automatics were first introduced, oh so long ago ? Everything is on the way out, but what time frame ? Perhaps now is a good time to discuss merits of a "pre-selector" manual box ? Should have less parasitic losses than auto's. Computer controlled shifting sequences? Dual inputs allow "preselecting", and reduced shift times/near seamless shifting. Stick it to the pavement. roverman.



Edited 1 time(s). Last edit at 08/09/2010 07:18PM by roverman.

Yeah, but this time...
I dunno Art, with all your pre-select, computer controlled clutch and the like I think your manual transmission is morphing into a (gasp!) automatic.

And they said the non-synchro transmission would never die either...

JB

Jeep XJ (Cherokee) AW4 2WD 4 speed Auto with GM TH400 tranny mount. A pic of the TCU(controller) and link to manual shifter page. Jim B., maybe you could use 2 of these in series.

[radesignsproducts.com]



Edited 2 time(s). Last edit at 09/06/2010 01:41AM by mgb260.