Duty cycle KE-Jetronic

Glad that you found the discussion (and technique) useful Vince. What model -- a late non-US KE3-Jet with Lambda 328 like Stefan's, a US KE-Jet with Lambda TR like in your avatar, something else?

 

Steve,
It is a 1988 US TR with KE-Jetronic (I guess that is KE3?)

So far I have only done the base setting (open-loop) and set the adjustment to get 700 - 800 mV. I am going to make an adapter today to be able to plug the O2 sensors back in and read the voltage closed-loop.

 

Steve,
When I plugged the O2 sensors back in both banks ran at ~ 500 mV when it first went into closed loop, then after a couple of minutes the left bank went to ~ 800 mV and the right bank went to ~ 150 mV. I assume that because the base setting was too rich it could not compensate?

I think it is OK to run open loop because I removed the cats and basically this would make the car a Euro spec.

 

Not the greatest IMO to always run open-loop on a US TR because during hard engine deceleration, the system uses the closed-loop system to prevent the system from leaning out too much (the euro TR have another gizmo, the vacuum limiting valve to deal with this condition that is not present on the US TR). But not sure that I follow your description -- when running closed-loop, the values on each should be constantly varying/wandering from 0.1~0.2V (lean) to 0.8~0.9V (rich) -- i.e., an average value of 0.5V DC, not just a constant 0.5V DC. However, the system only runs closed-loop when the water thermoswitch has opened (water temp > ~150 deg F) -- were you fully warmed up?

 

Steve,
Thanks for the reply.

My description was a bit vague: the voltage does vary open loop from 0.1 - 0.3 when accelerating and from 0.7 - 0.9 at idle or light throttle. The banks were always within 0.1 of each other. Closed loop though the left bank was way rich (not varying too much from ~ 0.8) and the right bank never went over 0.2.

If the right O2 was feeding info back to the left bank and visa versa, I think it would cause this condition, but I made sure that I plugged the connectors in the same way as they were before (i.e. not swapping right with left). But I have no guarantee that they were correct to begin with.

Yes, it was warmed up, but I did not go on a very long drive as I was afraid that this condition (esp the lean right bank) would screw something up, so I only drove about 5 miles. Also it was not running as well as open-loop.

I will repeat the test making sure I am over 150 F. Also I will try to check the right/left wiring.

 

Steve,
Found the problem: the O2 sensors were indeed reversed. Car runs great now in closed loop.

Thanks again for your help.

 

Stefan,
If you happen to have a sketch of the connector with some hints I would be interested in that.

 

That's a very nice (safe and foolproof) implementation Stefan (and glad that you were able to assist some of our KE-Jet brethren). The difficulties on a TR, like Vince's or mine, is that (without the 25-pin connector Bosch "break-out" box to use at the injection ECU connector): 1) there's no intermediate connnector between the injection ECU and the EHA so you have to get into the oddball bullet-socket connector built-into the EHA, and 2) they are really buried and hard to get to. I made a set-up using bare 4mm bullet and socket connectors from Radio Shack, but it doesn't give any polarity control, and there's nothing holding the connections together except friction -- so it's far more dangerous/difficult to use. I managed to blow up on of my injection ECUs so would even be reluctant to recommend this method to Vince

 

Steve, I might be able to repair the injection ECU for you. I've fixed a bunch of BMW Motronic ECUs. You probably just blew the injection driver, if you send me the ECU and the pin #s that you shorted I can take a look at it. Just let me know...

 

Stefan,
Thanks for the photos and description.

Steve,
Thanks for the warning. I think then I will stay with my O2 sensor readings and call it good.

By the way, as I was describing my method to someone today they told me that 0.8 - 0.9 V may be too rich and that 0.7 - 0.8 V would be a better goal. I think that may be generic info and not specific to the TR and also that due to the steep relationship of volts to Lambda that anyway I could not differentiate as it does jump. What are your thoughts on that? I will go with your expert recommendation.

 

See post #4 in this thread = 0.7~0.8ish average for open-loop operation is OK (when the system runs closed-loop it overshoots a little richer).

 

Thanks again for the advice. I think I will drive it for a couple of weekends, then take some more readings open loop and closed loop to get an average, then maybe tweak it 10 degrees CCW to get it 0.7 - 0.8 V open loop.