OMG, my EGT went over 1900 degrees!

[mock_glh]

Having tuned my Omni with the Zeitronix wideband to run at 24 psi I had felt pretty comfortable with an 11.9 A/F ratio. It seemed to be running pretty well like that. Then I installed the EGT sensor and took it out for some more testing. I jumped on the freeway, found some space for a 4th gear blast, and got on it. With my laptop on the passenger seat I could catch what was going on. As soon as I planted the throttle the screen flashed the 1600* warning but I stayed in it until about 6000 rpm. Nothing ugly happened but when I looked at the readout I saw the yellow EGT line climb to the top of the graph and slide straight across until I shut it down. It went over 1900 degrees! I added more fuel and got the A/F into the 10's but the EGT didn't improve any. What's up? Can I maybe add some more advance? Or could it just be the quick response probe I'm using?

[johnl]

Did you see this thread?

http://www.turbo-mopar.com/forums/sh...ad.php?t=20292


I had the same problem (also with the Zeitronix quick response sensor) which was placed in the throat of the outlet of the TU welded el stainless header.

I attributed my high temps to the location of the sensor - continuous exposure to heat - zero pulsing of heat/no heat as you might get if the sensor was isolated on single cyllinder's heat output. I tried adjusting fuel as well as cam and ignition timing to get it down - never did - then the cast pistons in that stock block cracked but by then I was pretty much ignoring the EGTs when they flashed above 1600*. Saw 1700s regularly. Never saw anything as high as 1900 though.

In the TU cast header, I put the Zeitronix sensor in the throat of #2. We'll see in a few days . . . .

[overlordsshadow]

I think if you have it in the collector then that would be a normal temp? I have mine in #4 and it gets to 1600 alone on a hard run. I can't see why the turbo ro collector would be the same temp.

[moparzrule]

If the probe is in the collector than it could be up to 200 degree's more than what it really is. Something about all 4 cylinder pulses hitting it messing up the reading. I have mine in the number 4 runner.
I was hitting 1700 EGT's until I started burning my own cal's. Having the CEL flash on knock is a nice tool when you don't have a scanner, just turn the timing up until the light flashes and then back it off a little. My fuel was set pretty rich so it was all timing. Then again I'm running a G head so I need a lot more timing.

[mock_glh]

I suspected that since there have apparently been no ill effects from this (yet). I have my probe in #4 runner (stock manifold). Perhaps there's some kind of velocity/friction thing going on with the probe. I'm going to try some timing tweaks to see what comes of it.

I hooked the "user data 1" input to the knock sensor wire and it gives a nice sound wave output on the Zeitronix graph. This will be helpful.

[johnl]

I suspected that since there have apparently been no ill effects from this (yet). I have my probe in #4 runner (stock manifold). Perhaps there's some kind of velocity/friction thing going on with the probe. I'm going to try some timing tweaks to see what comes of it.

I hooked the "user data 1" input to the knock sensor wire and it gives a nice sound wave output on the Zeitronix graph. This will be helpful.

Could be a swirl deal where pulses from 123 are pushing up onto the sensor in the 4th runner.

Wow, good idea for "user data 1." Better get out there into that 40* garage and wire that up.

[Aries_Turbo]

needs more timing. is the CEl off... ie no knock?

Brian

[Ondonti]

and the plugs look good and happy?

Does it do much good in that location if your tuning changes cant even be detected in that location?

I thought it was a good idea to hook up the user data to a knock sensor but since I dont have a factory knock sensor I need one of these devices to lengthen the frequency to a 0-5v range the user imput likes.

The graphs I have seen online look really helpful when it comes to knock datalogging. If you have the ability to get very creative in your tuning (standalone with a lot of cells to fill in or well done cals) then that can help you be real aggressive with your tune

[mock_glh]

The plugs look great. Kind of brownish, no glazing or detonation freckles.

I tried adding some timing above 5000 rpm but the knock light would come on. Then I tried a cal with the swirl head "Advance from MAP" tables and G-head "Advance from RPM" table. The power response was awesome but the knock light was going crazy. Maybe I should go with 3:1 on the 91 octane/C16 ratio instead of 5:1 like I've been running.

Doesn't the stock knock sensor put out 5 volts? The D-cal files all show a 5 volt scale on the "KnockSensorVoltageFromKnock" tables. I did however have to set "User Data 1" output to "Calibrate to Plot Range" because the "Default Display" would only generate a near-flat line at the bottom of the graph.

[johnl]

Man, be careful with adding timing to push EGTs down when the "high EGTs" are only deemed high because of somebody else's rule of thumb and when several others also have "high EGTs" AND they share a similar sensor location - throat of #4/throat of the manifold outlet. I suspect the 1600* ceiling is derived from sensors on #1, or maybe #2.

When I was dealing with the same issue, I had audible detonation at the transition between 3/4 and 4/4 throttle. I was trying to avoid it by quick pedaling the throttle to WOT - table jumping but then 'puter is too slow.

You're onto it - it's the gas - more C16. Or, maybe alcohol injection but then that's on top of the RR regulator/pinch valve. Getting complicated,

[moparzrule]

The plugs look great. Kind of brownish, no glazing or detonation freckles.

I tried adding some timing above 5000 rpm but the knock light would come on. Then I tried a cal with the swirl head "Advance from MAP" tables and G-head "Advance from RPM" table. The power response was awesome but the knock light was going crazy. Maybe I should go with 3:1 on the 91 octane/C16 ratio instead of 5:1 like I've been running.

Doesn't the stock knock sensor put out 5 volts? The D-cal files all show a 5 volt scale on the "KnockSensorVoltageFromKnock" tables. I did however have to set "User Data 1" output to "Calibrate to Plot Range" because the "Default Display" would only generate a near-flat line at the bottom of the graph.

The stock settings to detect knock are very conservative and can make the light flash too easily even from valvetrain noise. They can be bumped up. Stock they are 3.75v to detect knock above 6K RPM. But at 5K they are 2.15V.
Here's where I have mine set, which is very similar to what Rob had my first cal from him set at-
0-2500 RPM- .75v
2500-4500- .75-2.50v
4500-5800- 2.50-4.00v
5800+ -4V

Rob had mine at 4.25v to begin with at 5800+, but I backed it down a hair.

[mock_glh]

You're onto it - it's the gas - more C16. Or, maybe alcohol injection but then that's on top of the RR regulator/pinch valve. Getting complicated,

That Cartech regulator is gone. I'm all computer now, 3-bar and 750cc injectors.

[Aries_Turbo]

The plugs look great. Kind of brownish, no glazing or detonation freckles.

I tried adding some timing above 5000 rpm but the knock light would come on. Then I tried a cal with the swirl head "Advance from MAP" tables and G-head "Advance from RPM" table. The power response was awesome but the knock light was going crazy. Maybe I should go with 3:1 on the 91 octane/C16 ratio instead of 5:1 like I've been running.

Doesn't the stock knock sensor put out 5 volts? The D-cal files all show a 5 volt scale on the "KnockSensorVoltageFromKnock" tables. I did however have to set "User Data 1" output to "Calibrate to Plot Range" because the "Default Display" would only generate a near-flat line at the bottom of the graph.

the stock knock sensor doesnt directly drive the input of the A/D converter in the LM. it drives the base of a bipolar junction transistor. the collector of the transistor is connected to a +5v source and the emitter is connected to input of the a/d..... in laymans terms, the knock sensor just triggers a transistor which sends a 0-5V DC voltage dc voltage to the microprocessor so thats why the scale is only +5v max. the sensor outputs a sine wave which is AC so its difficult to accurately graph it without an oscilloscope.

The stock settings to detect knock are very conservative and can make the light flash too easily even from valvetrain noise. They can be bumped up. Stock they are 3.75v to detect knock above 6K RPM. But at 5K they are 2.15V.
Here's where I have mine set, which is very similar to what Rob had my first cal from him set at-
0-2500 RPM- .75v
2500-4500- .75-2.50v
4500-5800- 2.50-4.00v
5800+ -4V

Rob had mine at 4.25v to begin with at 5800+, but I backed it down a hair.

i wouldnt mess with that unless you make a detonation can/stethoscope setup to actually listen to the motor with your own ears before you start modifying that table.

Man, be careful with adding timing to push EGTs down when the "high EGTs" are only deemed high because of somebody else's rule of thumb and when several others also have "high EGTs" AND they share a similar sensor location - throat of #4/throat of the manifold outlet. I suspect the 1600* ceiling is derived from sensors on #1, or maybe #2.

When I was dealing with the same issue, I had audible detonation at the transition between 3/4 and 4/4 throttle. I was trying to avoid it by quick pedaling the throttle to WOT - table jumping but then 'puter is too slow.

You're onto it - it's the gas - more C16. Or, maybe alcohol injection but then that's on top of the RR regulator/pinch valve. Getting complicated,

if you make the timing and fuel tables the same over 5psi, you dont have to worry about the slow puter. and you only have to tune the table once.

Brian

[moparzrule]

Brian I have mine set more conservative than Rob had my other cal. 4.0v max instead of 4.25v that he had it set for. And I did have the CEL flash with his cal before at WOT 24 PSI down at cecil and nothing broke even with cast pistons.

[Aries_Turbo]

but you did bump up the midrange right? (i didnt put in your values into a cal to look at them....) i still have mine set to stock 87t2 values and i dont get knock till round 22psi with pump gas. those curves are set with the valvetrain noise and safety in mind. just be careful.

now that said, its probably ok given the fact that the S60 cal knock threshold is set at 3.95 at all rpms....

Brian

[moparzrule]

but you did bump up the midrange right? (i didnt put in your values into a cal to look at them....) i still have mine set to stock 87t2 values and i dont get knock till round 22psi with pump gas. those curves are set with the valvetrain noise and safety in mind. just be careful.

now that said, its probably ok given the fact that the S60 cal knock threshold is set at 3.95 at all rpms....

Brian

Yeah 4000 RPM is 2 volts vs 1.09 for stock settings.

[johnl]

OK, ole blue "Louise" (T2 glht) is up and runnin. Pretty much same set up as before - 2.2, ported head/2pc, +1mm valves, G head, TU cast header, 3", 120 cranking, stock 87CSX LM

Zeitronix' quick response egt sensor is in #2 of the header.

Car runs spooky strong. Was seeing high 1500*s and a couple 1600*s. Added fuel to 65lbs and then this am, the data grab/lock feature showed a peak of 1700* after a WOT, rolling start/clutchless "let's spin em up" blast in first gear - fuel was 10something, forget exactly. 10* static.

Before I was seeing 1600*+ with the sensor in the throat of the TU weld header outlet; now I'm seeing it again with the sensor in #2.

Obviously, timing/calibration is in order but I thought maybe this might add, maybe, to what's known about egt sensor locations?

[2.216VTurbo]

Third time is the charm for Louise I hope

[johnl]

Third time is the charm for Louise I hope

Thanks, but there's a knocking noise in the block I don't like and I changed the oil filter at 50 miles and 150 miles and both times came up with foreign matter - grit, flecks, filter fluff, slight brassy color in the oil, sludgy stuff - getting better with each change.