[*sean*]

I want to add exhaust gas temperature sensing to my Insight as part of my efforts to add "missing meters" to my car, to keep it alive in an era when parts from Honda are no longer available and third party parts might dry up.

The EGT numbers might provide a warning of fueling issues that are missed by the existing systems. I am wondering if a misreporting EGR valve sensor can cause oscillation between under- and over-fueling that the fuel trim system is not fast enough to compensate for and causes the car to alternately lean out and dump fuel into the catalytic converter, leading to damage. Wondering if this is happening during "herky-jerky". I want to monitor EGT before and after the primary cat.

Anyone (since some of you guys play with performance cars) have ideas on how to accomplish this?

 

[Balto]

Buy two EGT gauges.

 

[*sean*]

Buy two EGT gauges.

Already visited Amazon, saw plenty of gauges and sensors. I don't need a gauge; the sensor will be read directly by a microcontroller. I'm looking for ideas for what sensors I should look into and ways of physically connecting.

Perhaps I should examine measuring the voltage drop across the O2 sensor heaters. Certainly this is accessed easily enough and without changing the sensor operation. It will lag a little due to some thermal mass but it is in the exhaust stream and requires no exhaust mods. It assumes some low and constant resistance in the wires and FET etc. But it might not give me enough resolution - MAF sensors work this way but use a long skinny wire to increase sensitivity. Also, any change in the 12V supply voltage will also need to be measured and factored in. But even 10-20C per bit should be plenty.

Thoughts?

 

[eq1]

Just a random idea: Do you have an OBDIIC&C? I know you used to be able to connect a 'melexis' temp sensor to one of those. Probably hard to find a sensor to withstand the actual exhaust temps, but seems like you could stick one somewhere that heats up with the exhaust, just not to the same high temp, and then calibrate/scale the data...

 

[Balto]

Already visited Amazon, saw plenty of gauges and sensors. I don't need a gauge; the sensor will be read directly by a microcontroller. I'm looking for ideas for what sensors I should look into and ways of physically connecting.

Perhaps I should examine measuring the voltage drop across the O2 sensor heaters. Certainly this is accessed easily enough and without changing the sensor operation. It will lag a little due to some thermal mass but it is in the exhaust stream and requires no exhaust mods. It assumes some low and constant resistance in the wires and FET etc. But it might not give me enough resolution - MAF sensors work this way but use a long skinny wire to increase sensitivity. Also, any change in the 12V supply voltage will also need to be measured and factored in. But even 10-20C per bit should be plenty.

Thoughts?

Most EGT sensors should have the voltage+ temp curve data to be used with aftermarket ECU's, which you could just use to program it to a Arduino.

Something like this? EGT Probe Exhaust Gas Temperature Sensor - 1/8" Diameter

 

[CoreyS]

Why not just use an aftermarket lambda sensor like an LSU4.9? It’s response time will be much faster as well.

 

[DW_CT9A]

The EGT metric used to be ubiquitous in the 1980s/90s performance/ tuning world when folks were using them in conjunction with narrowband o2 sensors and maybe reading plugs.
Widebands were exotic at that time.

With the advent of the affordable wideband o2 aka lambda sensor, EGT sensors lost popularity due to their slower response time as mentioned above.

We still use the crap out of them in aviation. I'm a mechanic not an engineer but i would imagine that is likely a durability/ reliability thing mostly. The good ol K type thermocouple is hard to kill in spite of its (relatively) lethargic nature.

As it pertains to the ZE1 I'm aware these cars have a LAF sensor, and I pay some attention to its values on my OBDIIC&C. The resolution there is not great- as voltage changes, the correlating AFR value jumps by several tenths at a time rather than a single tenth as is more common.

 

[DW_CT9A]

Cont.

A decent aftermarket wideband should be able to spot the sort of Air/Fuel Ratio oscillation you are suspicious of. Here's a couple snapshots of a single datalog I have from a few years ago in a different car... I think I was on the freeway at the time and just slightly moved my right foot. Over the 774 milliseconds notice the fluctuation in TPS and the AFR oscillation. Also kind of interesting is the ECU commanded evap purge (note the "PCS" annunciator in the bottom right) when it saw a decrease in TPS. The wideband picked that up too.

The wideband I was using here was an AEM 30-0310. Over the course of a few months I kept trying to optimize cruising fuel economy in that car. The wideband fed a 0-5V signal directly to my aftermarket ECU which translated that into a relatively high resolution AFR value. On my daily commute (80mi round trip), I would capture a bunch of data and then make small changes to the fuel table often twice per day, rinse and repeat. It was fun and increased cruise/closed loop fuel economy by about 16%. The software made it easy.

In our case here we need to capture the data from your car. Having the wideband is great but how do we analyze its output? Off the top of my head, capture and plot RPM, TPS and AFR at a minimum. Ideally you could also see EGR commanded position as well as actual position. (I doubt these have the latter capability?? I'm not familiar enough with the ZE1). And any other conditions/value that the ECU looks at for EGR operation...maybe ECT, LAF, IAT...

 

[*sean*]

@DW_CT9A Looks exactly like what I want to measure. I think that between good lambda readings, EGT readings, and making sure the sensor signals are clean (like, bad EGR sensor which car won't detect) and within expected range - as measured against a baseline) - I'll know whether or not my car is running properly.

I am replacing both catalytic converters soon on one of my cars. Perhaps this is a good time to have a shop weld on another O2 sensor bung - this would allow me to finally have a way to test an oxygen sensor's performance against a known good sensor, and weld bungs for EGT sensors at the top and bottom (to detect too hot or too cold conditions or fuel being burned in the cat? (Reason: I am wondering if an intermittent EGR valve causes fueling to quickly cycle between too rich and too lean, perhaps dumping unburned fuel into the cat? I mean, there is a reason they can melt down - I'm looking at one that has - so why not instrument it to detect it before it happens again?)

So three bungs added in total to the top cat: two to the top (one more O2 and one EGT bung) and one to the bottom (EGT bung).

Thoughts?

(again, I'm not looking to change the stock behavior, just to instrument the car so that I can verify that all components are working together to make it run like stock, and have a way to detect problems well before they would be noticed.)

 

[DW_CT9A]

Sounds great. Would be even better if we could log a known good car first to have as a baseline. In any event, should be interesting. Any bigger defects should be easy to spot.