Raze wrote:
Um, simple question, but did the radiator in stock location have any ducting between the body and the radiator that was removed to fit the IC? .... . . . .
That and get a wideband, if you can datalog (usually the expensive part) the sensor itself is cheap $50...
My FC RX7 absolutely required that the under tray was on or temps would be through the roof.
Re the O2 sensor: The sensor alone is useless with out the controller. Some controllers have data longing built in and a whole lot more. I just put together the SLC EVK unit form 14Point7 and bench tested it last night. It is dumb as a stump simple little unit with no on board data-logging. Just wide-band and narrow band outputs. I went this rout since I am going to use the MJLJr and eventually the MS to data-log.
If you already have something in place that can data-log form a 0-5 volt output and you want to keep the car running on an ecu that uses a narrow band signal this may work for you. The unit I got I think is $125 with the Bosch Wide band sensor included. However assembly is not included!!! If MS does not intimidate you then this may be a cheep way to get where you want to go.
This would give you a much better look at what is really going on with your car.
There should be ducting in front of the radiator to direct air flow. I was trying to get a screenshot of the parts catalog but my work computer doesn't have paint or anything to save it as an image that I can host.
Sometimes it's not even an undertray that is essential. 914's with highway speed overheating problems in high ambient temps could have a pair of small spoilers added just ahead of the cylinder heads on each side. That created a low pressure area under the cylinders at high speed, helping the engine driven fan push enough air through to keep the engine temps where they needed to be. I've seen similar 'lips' under the radiator on various cars that serve the same purpose.
You'd need to first make sure that there is no way air can escape going through the radiator; fill any gaps with foam or use that stick on foam weatherstrip. Don't forget the underside of the hood where it goes over the radiator core support or the openings for the intercooler piping, etc. Then make sure the whole area between the lower part of the air intake and the bottom of the radiator is closed off.
EvanB wrote:
There should be ducting in front of the radiator to direct air flow. I was trying to get a screenshot of the parts catalog but my work computer doesn't have paint or anything to save it as an image that I can host.
Right, that might be the first problem - instead of ducting, I have a big intercooler there. I guess improvising some ducting might be a good starting point.
PhilStubbs wrote:
as a "tuner" (i dyno tune cars) i see a few issues.
one being your IAT sensor. if its showing cooler than actual intake temp, you will get more timing than you should.
That to a certain extent was a brain fart at my end. The IAT sensor is part of the MAF on a 1st gen DSM and in this case sits right behind the air filter. So of course that'll give me the outside temperature more or less (which it did) but doesn't make any allowance for the way the air is heated up by the turbo and the intercooler. I'm still not sure that it's working as well as it should be as I would expect further changes in IAT when the car is sitting at idle at a stoplight, for example.
All that said, I do need to research what the ECU in this car expects to see IAT-wise. It might make some educated guesses based on IAT at the air filter.
PhilStubbs wrote:
also, you dont need to hear knocking to have too much timing for building heat. knocking is near the maximum power the engine will make. 99% of turbo cars are tuned/built/driven for short bursts of acceleration. coolant temp will climb during a 1/4 mile run, but not fast enough to cause a problem.
Unfortunately it's beginning to look like this sort of drag-tune is part of the problem here.
PhilStubbs wrote:
i am shaking my head at you playing with your tune with out a wideband. that narrowband is 110% useless for this. the all mighty 14.7:1 is for idle and cruise only. any acceleration should be richer than that. your narrowband should show rich only during acceleration, but switch back and forth during cruise/idle.
Before I get handed the Dunce cap (hey, if it fits 
) I would like to point out that I'm not playing with the tune - I did get the car like this after the PO tuned it old-school with a narrowband and a pyrometer. It did run fine in cooler temperatures after the waterpump was changed. What I am currently trying to do is to figure out what is going on with the car by using the datalogging facility of the ECU, not tweak the ECU settings. And yes, I'm aware that the car should only be running stoich or potentially slightly leaner under cruising conditions or idle. My original comment re the narrowband was to see if the ECU was thinking the engine was either running permanently lean or permanently reich. Narrowband shows rich under acceleration, lean under deceleration and is switching during normal cruising.
My original plan was to see if I can figure out what's wrong (if anything) and then make an educated guess if it's safe-ish to drive the car to the nearest place that is supposed to be able to dyno-tune it. I don't feel safe taking a street tuned car onto a road race track, but I needed understand if it was safe to take it over the mountains as is - the company I was thinking of giving my money is down in the Bay Area and there are a couple of 7000'+ passes between them and me.
So far it looks like it's not safe taking the car down there as is unfortunately so I need to track down a few of the issues I'm aware of and install a wideband to have a better look at the tune. I do actually own a wideband (LC-1) but given the underhood temperatures of this car I'm currently researching alternatives that would allow me to keep the electronics out of the engine bay. Nevertheless I pretty much have to install one the next few weeks.
PhilStubbs wrote:
another afterthought on timing. spark plug/piston top readings are your only real gauge as to whats going on. knock sensors are garbage, plain and simple.it is a microphone tuned to pick up one frequency. im not trusting my engine to that. if that engine has ever been opened, you mill the head, change pistons etc, the tone of the knock changes and that sensor is useless to you.
Part of the issue is that the ECU insists of using the knock sensor as a way of figuring out if it needs to retard the ignition, though. Plus the darn exhaust is loud enough to drown out any knock noises.
PhilStubbs wrote:
post up some data log screen shots starting before it gets hot to when its hot and also when it gets hot, pullover, pull a plug or two and take good clear pics. that should give me a good idea of whats going on
I'll do that next weekend when I have a little time to play with the car again. Thanks for offering to help!
triumph5 wrote:
It sounds like your knock sensor with excessive advance is the root of all evil. Since the fittings are there already, you might want to look into what it would take to get the coolant lines to the turbo.
A new turbo 
. The OEM turbo and several of the aftermarket ones do have the coolant line fitting but it looks like the one on the car doesn't.
triumph5 wrote:
Add an oil cooler when you can, and, the ducting to the radiator can surprisingly effect how well the radiator works. I'd make sure all the ducting/shields are in place.
Part of the problem with the ducting or lack thereof is that both the intercooler and the radiator are non-standard. So of course there is no ducting...
Oil cooler seems to be a better and better idea the longer I think about it; I'm also planning to fit an oil temp gauge.
triumph5 wrote:
Lastly, as a last resort. Take the windshield wiper hose and reroute so it sprays the front of the radiator. It's a temp fix, but it just might get you over that hump. A couple of squirts or more when things really start to get hot. Pinched brass tube in front of the radiator to spray the fluid. Yeah, sounds wacky, but, coolant is coolant.
Actually the Evo II, the DSM's close cousin, has exactly that sort of facility. Come to think of it, the Evo II I owned also turned out to be a disaster area. Slow learner, me.
dean1484 wrote:
Re the O2 sensor: The sensor alone is useless with out the controller. Some controllers have data longing built in and a whole lot more. I just put together the SLC EVK unit form 14Point7 and bench tested it last night. It is dumb as a stump simple little unit with no on board data-logging. Just wide-band and narrow band outputs. I went this rout since I am going to use the MJLJr and eventually the MS to data-log.
If you already have something in place that can data-log form a 0-5 volt output and you want to keep the car running on an ecu that uses a narrow band signal this may work for you. The unit I got I think is $125 with the Bosch Wide band sensor included. However assembly is not included!!! If MS does not intimidate you then this may be a cheep way to get where you want to go.
This would give you a much better look at what is really going on with your car.
Well, the part that can datalog the 0-5V (actually, it'll translate it for you to the O2 reading) is called the 'ECU' in this car (it's got the ECMtuning ECMLink modded ECU + software). They do support a bunch of Widebands out of the box which is probably the easiest way to go. I either need to find an extension cable to go between the wideband I have and it's controller box or find one with a long cable so I can move the controller box out of the engine bay.
Did you have to actually solder components to the EVK board (because I can solder, but not good enough to do SMT) or was it just the soldering work necessary to hook up a readily assembled board? I think that particular kit might well be the reasonable alternative to my LC-1 that I'm looking for.
If you don't have ducting or someway to make the air go thru the radiator then it's going around it and that's your problem.
You can't disable the knock sensor with the tuning software? I'm not overly familiar with DSM's on a tuning level, most of my experience is with Honda and gm stuff. I'm not saying the knock sensor is causing your overheating problem, but it can kill your power.
I also ASSumed you were tuning, since you aren't I take back my head shaking. Lol. I am shaking my head at the iat sensor before the turbo on a factory turbo car. Thats just another nail inthe coffin for the DSM engineers in my book. Lol
If the car came with ducting around the radiator then it will cause problems if it isn't there. Most cars don't have anything like that so I'm not expecting lack of ducting to be your issue.
I'm really leaning towards an inexperienced tuner causing your problems here. I can't wait to see the logs and plugs.
I doesn't look like it's possible to to disable the knock sensor for this ECU, nor am I sure it's a good idea as the ECU expects it to be there. However after posting an idle datalog over on a DSM-specific forum, several people came back and said "check your knock sensor, I've had similar logs out of a duff one".
Someone also solved the IAT "mystery" by making a passing remark - the car appears to have a GM MAF (common mod on these for better flow) and guess what, that doesn't have an IAT. If you don't add a separate IAT, you'll get the sort of readings I do - ie, intake air always being in the 76/77F range. Doh. Guess that goes on the shopping list really quickly...
Well Damn!! There is a good chance that's your problem. Timing for 70° on a 90° day will sure as hell put more timing in than you want.
So many times retards do a wire tuck and bring me their car. While I'm hooking up to the ecu I see the IAT sensor under the dash and shake my head. I tell them how Important it is to have it in the intake manifold and they look at me like a confused puppy dog. I tune it very conservative and when the weather changes they always call me complaining.
Use deduction and logic on this. You've got some very substantial information already.
Recently replaced waterpump, now it overheats.
Wouldn't cool down while idling.
Not all waterpumps flow well. Sounds to me like this replacement one may not. Yes, you've got to get air through the radiator to transfer heat. As well water through the radiator. And the heat has to transfer through the radiator walls (which if it's new, likely isn't slimed or scaled).
Were I you, far more than worrying about possible lean mixtures, I'd be chasing a cooling problem instead.
Make sure that fan belt is good, and properly tensioned btw. If it's not spinning the water pump, it won't move water well.
No fan belt involved here - on a Mitsu 4G63, the timing belt drives the pump, too. And yes, that's nice and tight.
The 'new waterpump, now it overheats' might be a red herring as the car was driving fine with the new waterpump until someone set the thermostat from "late winter" to "summer" out here.
Cooling system is something I'm looking at, it did need an overflow tank which I've fitted and I'm planning to fit a new OEM thermostat this weekend.
It looks like I'm dealing with a bunch of potentially unrelated issues here that all seem to conspire to make the car into a winter beater.
Plastic scoop and cut a hole. 
That would go well with the 3" "arrest me" exhaust and the "rice, rice baby" fake carbon fibre A-pillar gauge pod.
Neither of which I put on the car - it was the previous owner, honest, occifer.
BoxheadTim wrote:
It looks like I'm dealing with a bunch of potentially unrelated issues here that all seem to conspire to make the car into a winter beater.
It is a common thing to have multiple issues contributing to the problem. Seems like you have a good list to start checking off
