posted on April 03, 2017 14:01
External Combustion Rocket Anti Lag System + JDM Spec C Impreza STI
By Eric Hsu
The external combustion anti-lag "rocket" combustor was used on the factory Prodrive built WRC STi Subaru cars to improve the EJ20 engine's inherent mediocre torque and engine response. The rocket is so good that a version of it is still in use today in the 2012 Japanese Super GT GT300 STi BRZ which also uses an EJ20. Last time I wrote a teaser article, but this time we'll take a look at some of the details of the rocket combustor anti-lag system as it applies to my buddy Ben's JDM Impreza STi Spec C street car over in the UK.
If you missed the last article and video of the rocket at work, check it out here: External Combustion Rocket Zero-lag Teaser.
When I met Ben back in 2008, he had just left joined Cosworth Electronics and had recently left Prodrive due to the cancellation of the Subaru WRC program. It was almost perfect timing with Roland, our other buddy/colleague and the Pectel Program Manager at the time, needing somebody intimately familiar with production cars and Pectel ECUs and Cosworth USA wanting to order an entire product line of plug in Pectel ECUs. Being extremely good at engine calibration, problem solving, testing and all that good stuff, Ben was sent to Cosworth USA for the express purpose of developing the ECPro ECU [Cosworth's plug in version of the Pectel SQ6] for the USDM GR Subaru hatchback. During the work week we were working together most of the time. Ben sat at the desk directly across from me in the office and when we were out of the office, we were always either in a GR Impreza STi hatchback or EVO X developing, calibrating, testing what would become the ECPro. In fact, Ben got to experience most of Southern California in STis and EVO Xs which wasn't a bad deal for him. We got along pretty good because even though his roots were in motorsports, his passion was high end street cars. High end doesn't always mean Ferrari, Lambo, etc. In this particular case it means a properly built production car using high end parts and craftsmanship. At the time Ben's toy of choice was a built R33 Skyline GTS-t that he drifted all over the country roads near Banbury and beyond. The difference between Ben's R33 and the others was the SQ6M ECU [$7k by itself] with GPS, traction control, Omega dash, bang bang anti-lag, launch control, four switchable maps, sensors galore, and datalogging speed and capacity that probably exceeded Nismo's own LeMans R33 GT-Rs.
With Ben sitting across from me, he would sift though his images, data, notes, etc. looking for information and everytime he pulled up something really cool (which was quite often), he would be like, "hey check this out." And for me it was like car porn because how often did you get to see the behind the scenes WRC stuff? Whether pictures of failed WRC components (car or engine), Prodrive's facilities and dyno cells, WRC engine components, or pictures of test components that never went into production (e.g. high speed DC servo controlled wastegate), it was all ultra cool stuff. It was a good thing for me that Ben liked to take pictures of everything. The one thing that always caught my attention though was the combustor; especially after he told me about 5 bar of boost off the line being possible. Ben always affectionately called it the "rocket". Even back then he said, "I'm gonna sell the Datsun and build a demo car with the rocket. It'll be 'bad ass dude' [in a faux American accent]." The pictures in the article are the pictures of that demo car.
The Spec C's bone stock engine before the project start. The Spec C is the homologation special that Subaru built for FIA Group N rally. It comes complete with 12 liter intercooler spray bottle (even the factory knows the short comings of a top mount), transmission and engine oil coolers, hole in the roof with an air duct, high flow water pump and radiator, roller bearing turbo, STi strut tower bars front and rear, better flowing cylinder heads, and slightly more aggressive camshafts. What it doesn't come with is a stereo or any sound deadening. Air conditioning is an option.
Roland, Ben and I have all since moved on from Cosworth, but Ben's rocket demo car is an ultra clean, tastefully built, GD Impreza STi Spec C. It was purchased from Iain Litchfield at Litchfield Imports located in Gloucestershire, England. BTW, you Americans should try pronouncing Gloucestershire and not getting laughed at in the UK. Even if you can say it right, it doesn't sound right without the British accent. Anyhow, when Ben first got the car, he called me from his mobile and said, "HEY DUDE, I BOUGHT A NEW CAR!! IAIN CALLED ME AND SAID HE JUST GOT THIS LOW MILEAGE SPEC C IN!! WHEN I SAW IT, I HAD TO HAVE IT!! FUCK DUDE IT'S TOO LOUD!! LET ME PULL OVER!! THERE'S WAY TOO MUCH ROAD NOISE!!" As Ben slowed the car, you could literally hear the car come to a stop. The loud and sticky S compound tires coupled with the lack of sound deadening in the Spec C was pretty damn serious. You could even hear through the phone the road gravel being picked up from the soft tires hitting the sheet metal that make up the wheel wells. Some of the noise has been remedied with Pirelli P-Zero Neros for daily street driving, but it's still not exactly a Cadillac.
Enough of me reminiscing of the recent past. Here are some of the details of the Spec C and the rocket. Ben isn't afraid of sharing because he knows that it cannot be successfully duplicated from the pictures themselves. STi and Prodrive spent tons of time and money perfecting it. Lucky for Ben, he handled a lot of that development; especially from the controls standpoint. The controls are very complex and most commercially affordable standalone ECUs have no real chance of controlling the rocket either from a software or hardware standpoint.
Posted in: Magazine
, Beyond the Dyno
, Engines & ECUs
, Street Cars
, People I Work With
Tuesday, October 30, 2012 2:23 AM
Tuesday, October 30, 2012 3:11 AM
I have some very technical input to offer!
Tuesday, October 30, 2012 3:55 AM
This looks a bit similar to a gas turbine combustion chamber
An engineering master piece
Tuesday, October 30, 2012 5:18 AM
Whoa. Thanks a lot Eric and Ben; that's a lot more detail than I thought we'd ever see. The whole thing is damn cool.
In broad scope, the pics of the rocket internals fit one of the theories I had about how it would look, but with some added details. That's a long way from being able to duplicate any of this even on a test stand though.
Tuesday, October 30, 2012 6:22 AM
When running such high turbo speeds at low engine speeds, what does he do to prevent compressor surge? I suppose that is at least somewhat mitigated by pulling boost from the IC hot pipe to the rocket, and I suppose any other surge is avoided by calibrating the BOV to open enough so that there is enough flow rate to avoid compressor surge. Yes?
Tuesday, October 30, 2012 6:30 AM
This is basically a centrifugal jet engine. Freakin' awesome!
Tuesday, October 30, 2012 9:38 AM
Hey Eric, does Ben's rocket share any similarities with Prodrive's P2 anti-lag system? I bet he was involved in that project, back in the day.
Tuesday, October 30, 2012 9:39 AM
Good to see this is done on a road car , but couple of bits that are uber expensive like the ALS valve , Rocket housing and turbo.
Tuesday, October 30, 2012 9:40 AM
Also very rare bits as well
Tuesday, October 30, 2012 2:21 PM
This does seem in principle a "centrifugal jet engine" with the rocket being the "gas turbine combustion chamber." (csubaka and max_R35). Very sweet idea. The coolest things used to come out of the WRC. It's sad to see the series become so diminished in recent years.
I'd like to know what the exhaust gas pressure is when the rocket is turned on. I imagine that was taken into account when designing the "rocket" part. There's also going to be increased drag on the exhaust side when the engine is WOT and the rocket isn't necessary. But that's the price you pay for a full boost awd launch. The control strategy has to be overwhelmingly complicated. Essentially one is using a gas turbine to supply full boost at idle and then the turbine needs to disappear once the throttle cracks open (and airflow diverted from turbine to throttle). Holy cow. This is why I dreamed of working in the WRC. Dumb RSRs are NA.
Tuesday, October 30, 2012 4:57 PM
Eric, as usual thanks for dropping knowledge and sharing it with us common mortal. I truly enjoy your posts!
Tuesday, October 30, 2012 7:53 PM
Awww, no video of the car in action? How does he expect to sell it if nobody has seen and heard it perform?
Tuesday, October 30, 2012 8:01 PM
Hmm, I'm having a tough time with a couple of pictures but I'd imagine those pores as part of expansion to help create back pressure for the antilag, or to help spin that centrifugal piece? But hey, what the heck do I know!
Wednesday, October 31, 2012 5:44 AM
@DrunkenMessiah: There is a link to a video of the car in action right after the first paragraph...???
Wednesday, October 31, 2012 6:23 AM
@Der Bruce: Think in terms of there being an outer shell that has air from the turbo introduced to it. I'm pretty sure the simpler holes by the collector and before it are just simply to introduce air to burn the over-rich exhaust, and the angled holes on the cylindrical part are to try to get everything swirling, maybe creating a bit of a curtain of air between the burning crap and the wall.
Wednesday, October 31, 2012 8:32 AM
Sorry guys, busy at SEMA.
sethulrich: there's no surge since the bypass is open and the turbo is cranking out such a high mass flow.
Marillionardo: yes very similar, Ben was involved in both the P1 and P2 I believe.
Kenku: sorry, I'll have to leave the theorizing alone at this point. The secret's in the sauce :)
Wednesday, October 31, 2012 8:53 AM
Wednesday, October 31, 2012 9:02 AM
Wednesday, October 31, 2012 9:26 AM
@Eric: Oh yeah, never figured you'd do otherwise - amongst other things, it would be a pretty crappy thing to give away a friend's secrets. I hadn't counted on seeing any more pics at all, so pretty excited you guys were nice enough to share - and if (big if; I'm under few illusions of what even a lashup development program would cost) I get around to building a whole testbed to try to duplicate it, that cuts out possibly a couple iterations of prototypes. Maybe. ;)
@logic: That's an S12, right? I always did like the look of the layout on those; always appreciate seeing what pics you share too.
Wednesday, October 31, 2012 9:39 AM
Yep it is an S12 , but with a 12B uograde
Wednesday, October 31, 2012 10:15 AM
Kenku: I had thought about a possible Venturi effect with the way things are shaped but I wasn't sure if those pores were for increased flow or expansion relief and utilization while combusting.
Eric - When SEMA people aren't swamping your booth(along with fans) maybe you could post a couple pics of what COBB's got cooking! Hey, saw recently that COBB's got an Op out of Portland now!?
Thursday, November 01, 2012 12:22 PM
That looks like my muffler did in high school when I was trying to free the horsepower in my muffler. They told me that there was some horses hiding in my exhaust setup. I tell ya what, no matter how much drilling, I couldn't shake a single pony out of that Honda.
Saturday, November 03, 2012 4:03 AM
I can understand the secrecy, but at the same time even if there was a exact understanding of the forces at work and purpose the chances of re-creation are minuscule at best.
That said, it is completely a jet engine combustion chamber. The turbo hot pipe feeds cool air over the housing. Taken in through the holes. Larger at the exhaust entry to help mix the overly rich exhaust and provide the pressure increase. Smaller holes throughout, which will actually help with cooling, and angled holes help encourage the vortex. I've read somewhere that as much as 20% of the total air through the chamber is just used for cooling.
The benefit far out weighs the detriments. Of which I really can only factor increased EGT and flow restriction. However, as long as at high RPM it flows better than the turbo itself I don't see it as an issue.
I can't really quantify how much I would love to play with things like this. I've got bottles of aerogel granules in my room with intentions of insulation. I can has job please?
Sunday, November 04, 2012 9:06 PM
There is actually a company that makes a "vortex" type muffler in the uk that has very similar tech behind it..they claim it increases engine efficiency by 30% improving both mpg and hp..Cherry bomb is even getting in on the act with there "vortex muffler" that even out performs there old glass packs.
Monday, November 05, 2012 7:33 PM
...and I'm STILL (over 4 months later) waiting on Cosworth's answer on expanding parameters of the EC Pro to some that I've asked if they can unlock. Really didn't think it was that hard of a question :/
Tuesday, November 13, 2012 12:34 PM
Please email me directly about the question you are looking for the answer on.
Private Message me and we will speak via email.
Tuesday, November 13, 2012 5:04 PM
Well apparently the site hates me. It won't let me message you and keeps up saying 'You must be a member of the site and signed in to perform this action' however I'm obviously logged in.
Can you try shooting me a PM when you get a moment?
Tuesday, November 13, 2012 5:10 PM
E-mail firstname.lastname@example.org and tell him your problem. Happened to me as well.
Wednesday, November 14, 2012 7:38 AM
Michah, you should be all set. Let me know if your not.
Wednesday, November 14, 2012 10:12 AM
HAHA this is comical. Mark only accepts PM's from friends and on my computer at work, I can't select send on the window to send a friend request!!!
Anyway, the last message I'd gotten on this was from Brad O. (VP Sales) on 10/18 and this was after I gave up with phone calls, as no one I spoke with could ever give an answer.
As for the flame tube/combustor here it seems quite straight forward in what is done....except for the vertical tube that comes down through the center. I can see that this would feed the 'feet' that come off down in the secondary'ish hole area, where the collector converges. It's kind of hard to tell where the primary, secondary and dilution/tertiary holes start and stop. The extended and angled tips to increase swirl and I imagine help keep temp and soot formation off of the inner wall.
As for the comment somwhere above, jets usually use 40-60% of compressed air is for cooling...or so I'd been taught.
As for this entire idea, look up 'hypercharging'. That is what it was called when done on diesel tanks back in the 60's or 70's. I think it was MTU that had developed it. From what I know of it, they had to throttle the turbo up by driving to start off the system, unlike what is done by Ben with modern electronics and controls.
Aaron, thanks again for helping straighten out my account.
Wednesday, November 14, 2012 10:20 AM
Micah, try messaging him one more time. I had the PM system locked down to friends only but we're a pretty respectful group here so I changed a setting. Just out of curiousity what browser are you using where the request button is blocked?
Wednesday, November 14, 2012 10:49 AM
IE 8. I'll try and shoot him a PM again.
Wednesday, February 06, 2013 7:37 AM
Was reading through my most recent copy of Race Engine Technology and spotted a familiar image. The AET Motorsports Zero Lag Technology ad is neatly tucked into the turbocharger article.
As for getting an answer from Cosworth on expanding the features in the ECPro (SQ6 repackaged)....still waiting LOL
Wednesday, April 05, 2017 7:15 AM
Anyone doing any experimentation with this kind of thing, or air bypass antilag in general? I'm hoping to get an engine dyno going in the next year or two, and may just (potentially) sacrifice a couple engines and turbos to learn how to make this sort of thing work.
Thursday, April 06, 2017 12:31 PM
We have been running an air bypass system on Ryan Tuerck's Formula D car for 3 years now. He even Runs it on his "street car". It doesn't use a rocket turbine but some thought has been put into the manifold design and we can achieve a high level of available ALS pressure. That has proven effective.
Friday, April 07, 2017 5:38 AM
Are there any details on it out in the wild, so to speak? I'm curious about control algorithms - whether the spark retard is still necessary, if it's targeting turbo speed or if open loop is workable, things like that. And no hard feelings if the answer is "no, we don't want to share the intellectual property we spent money developing".
Friday, April 07, 2017 1:15 PM
I'm not at liberty to divulge too much information, unfortunately. I can say Spark retard is needed to get the party started and it's more like we target pre-throttle plate pressure while monitoring turbo speed but it can be done many different ways. what management are you going to use for your research?
Friday, April 07, 2017 2:05 PM
Like I said, no hard feelings - we all have interesting stuff we can't talk about. I'm building an engine dyno around a Labview system, and my intent was to use "whatever" ECU (probably a megasquirt variant initially) setup to change spark and fuel based on analog inputs, and use a couple spare channels off the Labview system to feed inputs to the ECU and control anything else. Crude and not suited for a final installable version, but with the advantage of being quickly programmed to work on any control scheme I can think of.
Friday, April 07, 2017 2:16 PM
The Motec M1 ecu platform has packages available that are able to control air bypass systems.
Monday, April 10, 2017 6:04 AM
Perhaps misguidedly, I'm trying to come up with a lower cost solution, even if it does mean doing some electronics design and programming.