I agree totally that turbochargers are fantastic and most all cars I have built have been turbo however its just not a good form of power for a mid-engined well balanced car designed for turning.Chris,
From my experience with forced induction and the K-series trans I'd advise you to use a supercharger rather than a turbocharger in your Exige. Couple of reasons: Its well known that the gears Honda builds aren't that strong (especially 4th in the K-trans), and while cryo-treating can extend their lifespan the bottom line is you really need to look at the kind of torque curve different methods of FI produce. Worst case scenario for drivibility and time to distance is a centrifugal compressor, because boost and therefore torque production are directly linked to engine RPM. Conversely this rather weenie torque curve, where the torque peak is generally right at the redline, is good for trans longevity since drive-train parts only feel torque and not HP. Torque production in a turbocharged engine is weak at low RPM simply because turbine housing sizing is always a compromize, and to get great low-end torque the rising backpressure kills the top end. Using a bigger turbine housing is great for top end power but makes the engine lazy until spool-up and then watch out because the torque curve is going to climb like a Saturn rocket. Most all of this is negated with a VT (variable turbine) turbocharger, but the turbo manufacturers have been thrashing on making this setup work reliably for about 35 years without a lot of success. Its the spool-up part of the torque curve that's troublesome for the trans because that torque spike is always somewhere in the mid-range, and then the torque falls off as engine RPM increases. (This can be negated to a degree by using very agressive boost control, but its expensive and take a lot of dyno-time to get it right.)So in most cases your peak HP production happens at an RPM where the torque is far lower than at its peak, so the trans is taking a lot more abuse than it needs to for the HP generated. Best case scenario from a drivibility standpoint is a positive displacement supercharger. Bottom end and midrange torque production is higher than with a turbo without the spike, and since it keeps making more boost as the RPM climbs the torque curve stays relatively flat all the way to the redline. Secondly, its a whole lot easier to drive a supercharged car fast because there are never any surprising spikes in the torque curve so the HP production is very linear relative to RPM. An example you'll be able to relate to: The 2,308cc "K23A" (M90 supercharged) engine I have in my Elise makes 454 [email protected],000 RPM with 12 psi on Arco 91, and is only producing a maximum of 288 lb/ft of torque. If that engine were turbocharged and making the same HP the torque peak would be more like 340 lb/ft and the trans would last nowhere near as long. I started turbocharging cars in 1975, and I LOVE Turbochargers, but I've never even considered using one in an Elise or Exige and I've been converting them to Honda Power since 1989.
Torque curve manipulation is really what engine building is all about, and if its done properly you always get the most HP from the least Torque.
Take care and have fun,
Here's a dyno chart from back in July '09 when I finally got around to putting my "K23" on the dyno. I machined 4 differently sized 8-rib pulleys, but only 3 of them are on this chart. 12 psi came from running the M90 at 58% overdrive, 14 psi at 69% overdrive, and 16.5 psi at 81% overdrive. This was done more for personal data gathering than anything else, so we didn't even rev the engine to its redline which is 9,200. Basically I wanted to see what shape the torque curve would be with the little IPS supercharger cams since they're small enough to give really sweet mileage at a cruise. Its my street car so that's important, and with the gearing I'm using I'm hoping to see somewhere close to 40 MPG at an 80 mph cruise. From other S/C dyno testing I've done bigger cams would flatten the torque curve out above 6,500 and make substantially more top-end HP, but considering the car only weighs 1,820 lbs I think this setup will be OK as a starting point. The 12 psi run was done using Arco 91, so that's the street setup. The other runs were done using MS-109 since I had some and that's our standard dyno fuel for N/A and low boost testing. The highest post intercooler temp in the manifold was 138 F. but since this was done on the dyno the airflow through the heat exchanger was piss poor, so it'll be substantially lower with 80 mph airflow.
On the subject of centrifugal compressors, you'd need a drive system that would be variable relative to crank rpm to achieve what you were talking about, and as far as I know nothing like that exists. I saw engineering drawings of one that used a planetary drive to get the wheel up to a usable speed quicker, but there was no consideration given to maintaining that speed as the engine rpm increased. Generally with a crank driven centrifugal compressor the torque and HP peaks are at the same rpm, right at the redline, so the torque curve is often curved in the wrong direction, concave rather than seriously convex like with a positive displacement blower. Centrifugal compressors need to be driven by something other than the crank to make a decent torque curve. Something like really hot exhaust gas.
Take care and have fun,
138 at 16.5# is even more impressive. I can only imagine what kind of core/heat exchanger that was. I'm surprised you were able to package the core in the S2 Elise. Were you running the heat exchanger up front?WhatsADSM,
It was 138 degrees at 16.5 psi.
Now I see where you're going with the Rotrex thing. I'm basing what I've said on the experience we had with Hondata's Bonneville RSX. Oscar supplied the biggest thing Rotrex made and it simply wouldn't flow enough air to get the job done, so in the end they ended up using 2 of them. Even with 2 Rotrex's the torque curve was concave! Max torque was something like 390 lb.ft. at 9400 rpm.
Now if you were only interested in an extremely mildly boosted engine I'm sure what you talked about would work OK, but you'd still get a sweeter torque curve with a positive displacement blower.
I've taken broken Rotrex's apart and its obvious what their problem is, they're appealing to people with limited available space so their compressor housings are small in diameter. Good marketing but poor engineering. The compressor wheels they're buying are designed to be efficient in MUCH larger housings, so overall its a very inefficient package.
Quite a few of them went into self-destruct mode at as little as 5% over what they say is the safe RPM. The shaft snaps off where the diameter is reduced to accept the compressor wheel, and the dummies in England who make them say its just fine to produce that part with a zero root radius even though there's plenty of room for a very generous root radius. Not an impressively designed or built unit in my opinion.
Take care and have fun,
Christ, yea that's a nice setup. You were about | | <-- this close to just making it from scratch.Murray,
My water/air intercooler core is 7" long, 5.3" wide, and 2.25" thick. I started with a JR manifold, put it on an angle plate in my milling machine and cut the bottom half of it completely off, everything below the lower manifold to head bolt bosses. Then I built a new manifold using the M90 and that I/C core as guidelines. Space above the core, below the core, as well as smooth flow into the ports were also optimized, especially the #1 port which is absolutely horrible on both JR and Comptech manifolds. I machined all of these parts and then welded it all together, so the core itself has zero leakage. 100% of the discharge air HAS to flow through it. I machined the end tanks so the water flows through the I/C core twice, and sealed the bottom plate to water tubes using O-rings. All of this took about 60 hours. It fits with the stock firewall in both the S1 and S2 Elise/Exige, just barely but it fits.
Interesting... I still have the infrastructure for the stock oil coolers in the front and could put them in the loop but for now I think I am just going to use a heat exchanger setup developed by BOE which runs in the back. Has some ducting off the scoop and a few nice fans as well... Not ideal, but then again my little drop in IC core in the JR access panel is likely the weakest link anyways.In the Elise I use the stock up-front oil coolers as my heat exchangers. Derek followed my lead and did the same thing. Its not ideal because they're essentially cheap English oil coolers and rather small, but Lotus was having trouble cooling the 2ZZ motor with just water so they added the oil coolers to help with that problem rather than making a bigger radiator. At some point I'll have Ron Davis make a pair of bigger heat exchangers that are actually made for water rather than oil, but when I was using the M62 S/C K20A setup in the Elise they worked pretty well. Only 325 HP (stock motor with 9 psi), but the air into the engine never got over 106 F and that was on a LONG WFO run out in the desert at 90 F ambient.
Yea what is interesting is that I don't expect near 16psi. I actually think it will be in the mid/high 13s based on what I saw with my last pulley (3.5"). Not sure what to make about that. I do have more displacement, but your K23 I'm sure outflows my stock K24A2 based on VE (your built head + cams I'm sure outweighs my ~50cc more displacement). I do have an aftermarket M90 which may trade off blower VE for more overall flow at high blower speeds?!With your 5.98/3.25 setup you're at 84% overdrive, so just a little higher than the 81% I ran to get 16.5 psi. The E-85 helps in your situation. You just don't have the rpm potential I have because of the 'tractor motor' 99mm stroke. That's one of the biggest reasons I built my "K23". Its 90mm x 90.7mm (2,308cc) and has a 1.75:1 rod/stroke ratio. No matter what it'll always make more power than a K24 even though its 46cc smaller. The only reason Honda used a 99mm stroke in the K24 was to get the displacment over 2,350cc so they could call it a 2.4 liter motor. Pure advertising, not engineering.
I'd imagine Derek has told you how much he dislikes the K24 in his Elise because it can't rev. Eventually he'll get around to building his 'real' motor, which will be 89 x 90.7 and R1900 supercharged. Then he'll be a really happy camper!
Take care and have fun,
Yea ultimately I'm not sure how it will work, but they do seem to be somewhat effective and in general Phil does a good job with his products so we will have to see I guess.Joe McCarthy said:I'd advise against using Phil's heat exchanger setup, there's nearly zero airflow in that location and a couple of little fans don't so shit as far as airflow goes. It may be easy, but its NOT the right way to do it. You need a good high pressure area like the front of the car, and since the coolers are already there its just a gimmie to use them, AND no extra weight involved. Cut off the stock 1/2" ID English fittings and weld on fittings for 3/4" hose. Firmly attach lengths of 3/4" heater hose to the stock oil cooler hoses and pull them through. It's not always easy because there are foam bulkheads inside the door sills that get bunched up as you're pulling, but on both Derek's and my car only one side got stuck to any degree. Also don't go the Mickey Mouse route with the I/C water supply. Put a header tank in the system so you'll never have any concern with air bubbles.
Agreed boost in a PD setup is basically a function of engine restriction to flow. Which is why I am surprised still. Sure I didn't spin to 9k, but as you know once the blower is spinning reasonably quickly the boost is pretty flat overall. I know while my boost does increase from say 5000 to 7600 its not by a lot. It increases in part due to engine VE falling and also in part to blower VE increasing slightly.Joe McCarthy said:Yeah, you won't make 16.5 psi at 84% overdrive because your engine can't rev to 9000 like mine. If you keep things safe from a piston speed standpoint and never go above 7,600 you'll be spinning your M90 to 14,000. I spun mine to 16,300 to get 16.5 psi after the I/C. Your setup will be just about like mine with my 12 psi setup, only you'll be making more boost but not pumping as much dense air because it'll be hotter. A boost number by itself is meaningless, you have to know the temperature of the air the engine's actually breathing as well.
Pretty much. The supercharger originally (i.e. the lobes) came from the early gen 1 or gen 2 M90s. However the case is actually pretty new and efficiency-wise is up there with the newer gen stuff. It's not ported per say but rather an entire new casting. The stuff I know he changed is that it does away with any noise suppression stuff and opens up the inlet a lot to help flow/IATs at high blower speeds. Granted it is a different engine but lightening guys have seen over 500whp (550ish rings a bell) with it, so although it is still a "old" roots blower the ability to move large amounts of air is still there.Joe McCarthy said:From the photos I've seen of the case and bolting arrangement your M90 originally came off a S/C Thunderbird that was new 20+ years ago. I can't be certain because I haven't seen one up close but yours could also have been ported to make it a little more efficient, and certainly rebuilt at the price they're charging for them. I've thought about doing this too because it's a really cheap alternative to a new R1320. I've found lots of those old T-Bird M90's on ebay for under $200. Mine was given to me by Moss Motors, free. Its in a "Scott" case, same one TRD used on their S/C pickup trucks. No integral bypass valve, which if its done wrong (like on the Eaton MP-series) generates a pumping loss at the inlet. That's been cured on all the TVS models I've seen.
Makes me feel a little better, just wish I would have went OEM from the get go. Oh well I guess you live and learn.Joe McCarthy said:Don't feel like the 'Lone Ranger' on bad engine building. At least 90% of the guys screwing K20's together don't have a clue what they're doing, and that includes a LOT of the drag race guys who don't live in the real world. 10 seconds isn't reality as far as I'm concerned. The monkeys from Monkey Wrench Racing found this out at Bonneville when they blew their 1.5L 2ZZ up trying to make sucessive 3 mile runs. The real course is 5 miles!
So keep the piston speed down and your TSX motor will last a long time.
Actually the Elise chassis will handle a LOT of HP. With a proper rollcage its stiff enough to handle 8-900 HP without a whimper.
That is interesting, I wouldn't have guessed. In terms of the header I was just worried I would run into physical fitment issues with the K20, since the deck height, or lack there of, would basically put my existing header on the rear subframe... maybe I am missing something?Joe McCarthy said:During the gazillions of hours Derek and I spent on the dyno testing all the K-series headers I've built we found that a supercharged motor simply doesn't care very much what header is bolted onto it. Changing the primaries from 15" to 24" didn't make more than a 2 HP difference anywhere on the curve. So stop worrying about that part, OK?
Yea the TSX rations actually look really good even with the 4.3 FD... although I wouldn't mind an even smaller FD.Joe McCarthy said:On the subject of transmissions, what I've done is to build one very much like what we built for Bonneville. TSX 1st-5th with an RSX 6th and a 4.105 final drive with a Quaife LSD. This puts my Elise at 218 MPH at 9,000 in 6th, and only 3,300 at an 80 MPH cruise. With my original JDM Type R gearbox my S/C K20A was spinning 4,100 at 80 MPH. And it still got like 31 MPG. The best I EVER got from the N/A 2ZZ was 26 MPG, making a weenie 180 HP. Damn, that's sad!
Huh, that is pretty darn cool. Yea the K23 sounds like a fantastic engine overall, love the bore/stroke and rod/stroke ratios. And given it is done with an OEM crank and block makes it that much more appealing.Joe McCarthy said:The 90.7mm F22C crank needs a little work near the snout to make it drive the K-series cam chain and little work at the ass-end to pilot the input shaft. This isn't stuff you can do yourself, it needs to be done by a real crankshaft guy. Otherwise it fits into any K-series block. You also get the added benefit of wider rod journals and its beefier than any of the other K or F-series cranks. Honda went a little crazy when they made the K20 rods so skinny. I use nothing but custom made rods anyway, so all the engines I build get the wide rod journals, no matter if its a stock crank or a custom built SAE 4340 billet crank. None of this using weenie B-series rod bearings either, that's for morons who don't understand what makes a crankshaft strong.
Take care and have fun,
Yea up front is ideal, and FWIW for his personal 400whp 2zz car he has the same exchanger I do in back and also another centrally located exchanger up front as well. So... I know he knows it's no perfect back there.Joe McCarthy said:Murray,
I agree with you about Phil doing a good job, he tends to think outside the lines rather than following what everyone else is doing. Except in the case of the heat exchanger. Ask him what he'd do if it being a bolt-on wasn't a consideration and see what he says. That makes a big difference when you're designing something as a user installed kit. Even Lotus went to front-mounted water/air heat exchangers on their GT3 cars once they realized they'd made a mistake mounting an air to air setup in the hot-ass engine bay.
Thanks for the heads up on cleaning them I may be able to do that without too much issue. Unfortunately changing to -12 (or 3/4" standard hose) may be just more trouble than it is worth for me. It was a pain getting the -10AN (pushloc btw) through there, as well as having the -10AN fittings welded on. I honestly think it will be either run them the way they are now with the -10AN pushloc stuff or just skip them all together and save it for an upgrade down the road. Ultimately it wouldn't be much more difficult to add them down the road as it would be to add them now. That way I could always get some data on everything now and see how much it actually helps with better exchangers up front. Hey might be good data for the community in the end.Joe McCarthy said:Making the heat exchangers efficient is all about the volume of water you can flow through them, so no part of the entire system should be smaller than the I.D. of the pump's discharge port, and ideally a little bigger since the lines are so damn long. Its just unpressurized warm water in this case, so there's no need to use AN hose. AN hose is heavy too! Measure the I.D. of a -10 AN fitting, its a LOT smaller than you'd expect considering each 'dash number' is supposed to represent 1/16". On long-run systems I always use at least -12 AN if the system in question calls for it.
Yea I didn't mean to say it was totally flat, I'm aware it is not. Just an observation of the difference in boost you saw versus what I saw with a similar blower/engine drive ratio. I can pull up an old log of mine and overlay them, it would be more clear.Joe McCarthy said:The boost curve is NOT flat, that's a pretty common misconception. It rises gradually all the way to the redline. On my 16.5 psi chart its at 10 psi at 3,250 rpm, 12 psi at 4,500 rpm, 14 psi at 6,200 rpm, and 16 psi at 8,400 rpm.
I agree the casting differences in the blower casing and throttle adapter castings are noticeable. HOWEVER I am like 90% sure it really is an aftermarket supercharger casting:Joe McCarthy said:1st gen. Eaton blowers had straight extruded rotors, so its gotta be later than that. If I remember correctly the S/C T-bird used a 2nd or 3rd gen. rotor group. Your case is identical to the T-bird M90, so I think the only new casting you got is the throttle body adapter. Look at the quality of the castings, they were obviously made by different people.
magnumpowers.com said:The MPx 90 cubic inch Supercharger case pictured is cast from 356-T6 aluminum that is stronger then the factory Eaton material. It has longer intake and exit port duration then the Eaton M90 allowing the supercharger to be spun much faster without loosing volumetric efficiency, which results in the generation of excessive heat.
In addition to the stronger 356-T6 material the MPx case has thicker wall thicknesses and a stiffening boss running the full length creating a very stiff case that will not flex while under load.
Since the case is a new casting there is no reason to send your M90 to Magnum Powers to have it ported. Simply order the case, which comes with a MP Intake Plenum, install your rotors, fill it with supercharger fluid and have fun.
Ahh yea I had a buddy of mine do a custom header for me. Collector, v-band all that is there just that the fitment is so-so. I just didn't want to risk the ~19mm of clearance. Maybe in the future?!Joe McCarthy said:The car dancing around is a function of the increased torque input, but its also due to the rubber suspension bushings and the absolutely horrible bump-steer Lotus puts into the front end geometry. That's the reason your steering wheel is dancing around like it does! I've already fixed both of those things in my car. Derek's car has the bump-steer fix, but is still using stock A-arm bushings.
With my K-series conversion kit there's the option of either K20 or K24 as far as the engine mounts and exhaust system goes, but one guy in Minnesota told me he was going to use a K20 and then after the header had already been built changed his mind to a K24. Fortunately I designed the header to sit low enough in the space that raising it up didn't cause any problem, and I use V-bands rather than bolted flanges so the 3" muffler still lived where it belonged. That one was done through a shop called RS Motors so I never spoke directly with the car's owner. Actually the entire kit was built before he changed his mind, so I had to build another engine-side mount as well. With your header I can't say what's going to happen...
Yea that's kind of what I figured. I'm thinking if, make that when, it does blow I will give a helical cut aftermarket set a shot. My working on car time has become much smaller recently and I really can't afford the time switching/rebuilding transmissions every season (or worse).Joe McCarthy said:Your concern about trans longevity is certainly valid, and I share it as well. That's the main reason I'll be sticking with just 12 psi most of the time. The gears have shown increased life with cryo-treating, but the bottom line is this: if you're going to be spending a lot of time at the track then you have to view the trans as a replacement item occasionally. 4th gear will die first because its narrower than the others. I don't have any first hand experience with the GearX setup, but at least the ratios are better than Honda's.
Too Gay or not, it was a fun video to watch and great for inspiration! If you ever put up any more videos of the thing, especially ones of it on track or even just fooling around I'd love to see themJoe McCarthy said:Yep, my Elise is a street car, pure street car. I'm not a talented enough road-race driver to make competition worth my time. Once I got back from Viet Nam most of the competitive fervor I felt as a kid had been completely satisfied, being hunted and killing people will do that for some of us, others never learn.
That American Touge (Derek and I call it Too Gay) thing I did was done on a whim because I know one of the producers. What's really funny is I didn't change anything before I drove it out to the track. That's the same setup I use for driving to the grocery store. Tsuchiya-san wasn't even driving the car all that hard, I don't think he ever took it above 7,000 rpm. Glad you enjoyed it!
Take care and have fun,