Been busy on the final prepartions for the emerald m3d ECU installation. I've gone the whole hog and opted for a wasted spark setup as it allows me to get everything under the control of the ECU where I can keep an eye on it. I'm using a mondeo coil pack which has necessitated some custom HT leads. The crank trigger wheel is a custom unit that I have had laser cut from 5mm steel.
Currently the loom is now pretty much complete and all the sensors test OK. Can you spot the difference between the old and new looms though?
Old setup:
New setup:
As you can see the original 2-part cossie loom is about 6 feet longer than is necessary. Whilst I went a bit OTT on the new loom in some respects there's an awful lot less of it to go wrong, in spite of me pulling all the fan wiring onto it. The one thing I didn't do was bring the chargecooler pump control up to the emerald, in hindsight I perhaps should have as it would have enabled me to leave the pump off until things warmed up a bit.
The techedge is a wideband lambda sensor system with some built in datalogging facilities. It is connected via a little sub-loom and a 7-pin connector so that it's removeable in about a minute if I wanted to use it elsewhere. It's hooked up so that the emerald can log the output of the wideband but it can't use it directly other than via a simulated narrowband which is a limitation of the emerald.
The little green connector at the bottom of the image is for a Honda resistor pack which is a very cute little box which means that I'll be able to swap between high and low impedence injectors very easily.
I've still got to wrap a few odd bits of wire, most notably the ones from the relay but it's largely a done job. All the sensors seem to read OK, the only big gotchas being the emeralds stupid use of the permanant feed and the slightly confusing/missing documentation on earthing strategies and IACV control. The only outstanding issue now is to get my crank pulley to paragon for final centering and then I can try firing it up.
After a few false starts I finally got the car started and idling happily, as usually turns out the frost hasn't melted all day and I'm sitting in the car trying to use the laptop to map the engine.
I've had to stop for a bit because even with gloves on I couldn't feel the keyboard on the laptop any more. Driving around with the bonnet and nose off, a stupid hat on and a laptop screen glowing in the passenger seat certainly raised a few eyebrows though. Better wipe the grin off my face and get back to mapping :)
Oh and for anyone who was in any doubt at all a wideband lambda sensor makes mapping *really* easy. You just set drive gently around for a mile or two park up, review the logs, add/remove fuel as appropriate and go off for another drive. Whilst it took me an age to get the car idling happily getting it to the point where you can drive it happilly enough has only taken a couple of hours. Given another couple of hours I should have most of the rev range mapped out acceptably well and I can start adding in the turbo. So far it's working up as far as 4500rpm on light throttle, tomorrow I'll be opening the throttle some more.
My day at emerald went fairly well. There were a couple of niggles, one being a firware bug which Karl spent an hour tracking down and then documenting, the other being an incorrectly documented pin (you'd think that PWM output 1 configured to use the Aux 1 output would use the aux 1 pin wouldn't you :) but we got there in the end.
The car looked pretty comical on the rollers, we strapped 4 25KG bags of sand to the boot and I sat atop the roll bar and equipeed as such cured the wheelspin problems that plagued us early on.
Going into the session I was worried about running out of injector duration as they're only 380cc/min and also that the turbo maps say that the turbo shouldn't be able deliver enough power for the power that the tuners quote for stage 3 cossies. Well at 260bhp we were running at 93% duration which answers one question and the turbo had run out of puff beyond 4500rpm which answers the second. All this means that tuners who are putting out 330bhp stage 3 kits are being pretty damn generous in their measurements, it's possible that a bigger hybrid would get them the right side of 300bhp but they'll be running right on the ragged edge.
So that's the bad bits, onto the good. 260bhp is 5bhp more than we got back in March and we obtained that using less boost. We're down a shade on power at 3500rpm because we're not overboosting but what we have got is a much, much smoother power delivery as a result of mapping the wategate for each speed site individually. Peak torque is up 17lbft to 270lbft and is 5-10lbft in the places you care about it.
Apparently the rover tomcat injectors are a drop in replacement (oi DG, know of any lying around somewhere?) and since the rollers have confirmed what I'd initially thought from looking at the compressor flow maps it's also confirmed that the GT28RS has exactly the right profile for a genuine 330bhp with the potential to go up to 350. The GT30R is just a bit too big really for the boost profile that I want to run. If I could find a compressor map for the GT2871 then I could make an educated decision about that. Rather fortuitously there are people selling GT28s on ebay with a choice of T25 or T3 style flanges, less fortuitous is that they'll still need the downpipe tweaking.
Standard cosworth injectors flow as follows:
Most tuners go from the 803 greens up to to the greys but they're really oversized for most people and economy suffers. The blues are perfectly sized but aren't commonly available. A better bet may be a set of injectors from the Ford Mustang, they're usually sold in sets of 8 but that means that there will be 4 spares which has to be handy.
I've had a chance to consider what's happening and here's a quick summary: My turbo has asthma. At 3500rpm it has enough puff to feed the engine and get good boost, beyond that the engine is suffocating. You can see what the problem is by looking at the compressor map here: http://www.not2fast.com/turbo/maps/t3-40.gif
My 2 minute guide to compressor maps, on the X axis you have lbs/min of air being thrown towards the engine. As a rough example my engine consumes about 16lbs/min to produce 175bhp, 25lbs to produce 270bhp and 30lbs to produce 330bhp. The Y-axis is the boost pressure, note that it's based on 1 atmosphere being zero boost so my engine, currently peaking at 1.3 bar of boost is hitting 2.3 on that scale. The percentages on the graph are the compressor efficiency.
What that means is that at 3500rpm I get a little over 1.3 bar of boost and the engine is consuming 15lbs/min of air, if you look up 2.6 and 15 on the map you can see that the compressor is working at a reasonable 70% efficiency. As I accelerate to 4500 the turbo really comes on song, giving a shade more boost and we shuffle across the map to 20lbs/min and 65% efficiency. From here on in we hit a brick wall, the turbo can't really shift any more air but the engine is still slurping hungrily so boost drops away as does efficiency.
FWIW the dyno from the last session at emerald is here: http://www.g2.nu/chris/westie/pictures/20041127/med/dyno.jpg
OK, so that's the primer done. If you're still with me you should have worked out that I need a turbo with bigger lungs. If the turbo is too big though it won't do much until I'm hitting the redline. If I go too small I'll have a lovely bottom end but it will run out of puff even sooner.
So getting the compromise right is pretty important. Here are the maps for the leading candidates: http://www.not2fast.com/turbo/maps/gt28rcompress.gif http://www.not2fast.com/turbo/maps/gt28rscompress.gif There's also the gt2871 which is basically a 2871 with a bigger compressor wheel which gives more flow at the cost of lag, I can't find any maps for it at present though. I'm discounting anything GT30 based as IMO it's just too big for my application. http://www.not2fast.com/turbo/maps/gt30rcompress.gif
So the GT28RS looks pretty unstoppable, but there's the turbine maps to factor in, this is an area where I understand very little:
http://www.not2fast.com/turbo/maps/gt28rturbine.gif
http://www.not2fast.com/turbo/maps/gt28rsturbine.gif
http://www.not2fast.com/turbo/maps/gt30rturbine.gif
I think I understand how to read these and broadly speaking you want the line to start as low as possible as that's when you start seeing boost but not to top out too low as that would be choking the engine on the exhaust side. So this is where I'm at, aptturbo will sell me one of these http://shrunk.net/65f9eaf7-www.atpturbo.com that will bolt straight onto my T3 exhaust manifold with either a .63 or .84 A/R housing. The .63 will spool up quicker but might be a bit restrictive, having said that it should be miles better than my current one.
I'm tending towards the GT28RS with a .63 housing with internal wastegate does anyone have any comments? Most cosworth tuners seem to say that the T3 .55 housing is fine for 350bhp and the .63 good for an awful lot more than that. Frankly I can't see a good reason to use the larger housing.
| Previous | Next |
| G2 home | Chris' homepage | Westfield | Westie Rebuild |