Steve,

I have a question regarding your DIY article "Evaluating Electronic Engine Controls". The KE-Jetronic duty cycle measurements taken at X11 pin 3 are shown together with scope displays of the waveforms. Figure 2 shows a square wave consistent with the 50% duty cycle reading. Figure 1 however, shows a waveform which is at a "high" level for a greater percentage of time than it is at a "low" level. I would normally interpret this as having a duty cycle greater than 50%. The duty cycle displayed is 30% however (rather than 70%). Figure 3 similarly shows 38.9% duty cycle which I would have expected to be more like 61.1%. Is this just a characteristic of your meter or is there a convention with measuring KE duty cycle where the inverted signal is used? Maybe I am missing something here.

I can use my multimeter to measure duty cycle on either my '90 190E-2.3 or my '90 300TE. With the meter negative connected to chassis (or X11 pin 2) and positive connected to X11 pin 3, I get "key on, engine off" readings of 30% for both cars. Obviously if I reverse my meter connections (ie. negative to pin 3, positive to chassis or pin 2) I get a reading of 70%. You suggest 70% is typical for US delivered cars and whilst mine are Australian delivered it appears the same might be the case (using your meter's convention). With the engines running I initially get 50% readings (prior to the O2 sensor reaching working temperature) followed by fluctuating readings slightly below 50% at idle rising to slightly above 50% at say 2000rpm (once system goes closed loop with the O2 sensor at working temperature). Using my meter, I read these figures as indicating the base air/fuel ratio being slightly lean at idle, going slightly rich at higher speeds (perhaps indicating slight vacuum leaks leaning the idle). Both cars incidently, are running fine. My reason for measuring them was purely out of curiosity and to check that they were operating somewhere around the desired 50% duty cycle. I have not yet verified these duty cycle measurements with EHA current measurements, as I have yet to make up a harness to do so.

I have checked my multimeter on the duty cycle range and and the readings are lower (<50%) with waveforms that have short "high" versus "low" states, and the readings are higher (>50%) with waveforms that have long "high" versus "low" states, as I would expect. These readings would be reversed following your meter's conventions.

Sorry to question you on this and I hope I have not "opened a can of worms". I appreciate the wealth of knowledge you bring to this forum and the time you devote to it. Many would not devote their valuable leisure time to doing what they spend much of their day doing at work.

Thanks,
Greg

__________________
107.023: 350SLC, 3-speed auto, icon gold, parchment MBtex (sold 2012 after 29 years ownership).
107.026: 500SLC, 4-speed auto, thistle green, green velour.
124.090: 300TE, 4-speed auto, arctic white, cream-beige MBtex.
201.028: 190E 2.3 Sportline, 5-speed manual, arctic white, blue leather.
201.028: 190E 2.3, 4-speed auto, blue-black, grey MBtex.
201.034: 190E 2.3-16, 5-speed manual, blue-black, black leather.

In automotive electronics polarity is a reference point. Many things are represented in ways the scientific community might express differently.

Manifold vacuum is such a concept. Even tyhe largest vacuum still represents some pressure absolute.

In the case of duty cycle my point of reference comes from that activity involved. Unfortunately by the time of KE the value on pin three had changed from the duty-cycle of the frequency valve to the mathmatical duty cycle called integrator. It now is just a representation not a true happening.

So let me explain a true duty-cycle such as used in the case of K-jet w/Lambda. Same pin. I will describe what an event cycle looks like on the scope. The picture i describe will be 30% duty cycle such as my scope pattern you reference.

The cycle starts high at battery voltage. The signal is read on the controller side of the frequency valve. It is battery voltage because the valve is constantly powered with the engine running. The other pin of the valve goes to the controller and is currently open circuited such that battery voltage appears on both sides of the frequency valve coil and nothing is happening. now the controller grounds the pin we are viewing, the voltage at the pin drops to ~0v. The duty period on the valve has started. The valve is being controlled when the signal is low since all these coil actuators are controlled on the ground side in a modern automobile (the only exception I can think of in MB are D-jet injectors they are controlled from the power side).

So when we look at a pattern on a ground control circuit and we look for the duty period it is the period when the signal is down not up.

In KE there is no direct duty cycle. It is a manufacturered signal called integrator. It still varies its signal electronically for interpretation like a true duty cycle of an activation circuit. It in this case is just a mechanism for a crude style blink code. All that matters when using it is that one gauges his perspective to that of the communications involved. KOEO = 70% worldwide except California where KOEO = 85%. Correct your point of reference to this and you won't wind up looking at the wrong fault code for 30% fixed, nor will you set the closed loop adaptation rich or lean improperly. When you turn the mixture screw in closed loop clockwise the duty cycle should go down (smaller). Thirty percent duty cycle in closed loop means that a rich mixture has been set basically and the fuel system has moved 20% (from the middle at 50%) to correct for it. And in subtracting the fuel it has kept the mixture at lambda out the exhaust.

__________________
Steve Brotherton
Continental Imports
Gainesville FL
Bosch Master, ASE Master, L1
33 years MB technician

Steve, by any chance do you also have an EE degree??
'cause you must be the most educated car mechanic I have ever encountered.
In most cases they only now how to plug the code reader and find the part that broke, or R&R parts.

You not only can tell people what is wrong with their cars and how to fix it, but you also can give a detailed lecture in electronics.

I admire you, and thank you for sharing your wealth of knowledge.

__________________
1999 Porsche 996 Carrera Convertible
1994 420E - SOLD
1986 300E - SOLD, what a car
609 Certified

Steve,

Thanks, all is clear now. I was stuck on the apparent disagreement between your scope diplay and duty cycle readings versus what I was measuring on my vehicles. Obviously if convention in modern automotive electronics is to control on the ground side rather than supply side, then duty cycle referenced to ground will be inverted. I was amused that you noted that control of D-Jet injectors (being on the supply side) was an exception. I guess I would hardly consider D-Jet "modern". As advanced as it was in its day, it seems prehistoric now with its use of germanium rather than silicon transistors in sections of its electronics. To give it credit though, I am amazed that after 31 years and some 350,000km, that same system in my old 350SLC soldiers on with only one transistor failure in the 21 years I have owned it.

Getting back to the KE system on my other two cars, the clue is as you suggest, to ensure that for KOEO I am reading 70% rather than 30%. I was unaware that 70% was used worldwide except California, thanks for this useful advice.

As Pesuazo suggests "you must be the most educated car mechanic". Steve, if more mechanics used your well considered diagnostic techniques rather than simply throwing expensive parts at the car, then I'm sure there would be fewer dissatisfied customers out there. Pesuazo, I think the only correction that should be made is to refer to Steve as a technician rather than mechanic as his knowledge is what makes the distinction.

Thanks for taking the time to reply and again, thanks for sharing your wealth of knowledge on this forum.

Greg

__________________
107.023: 350SLC, 3-speed auto, icon gold, parchment MBtex (sold 2012 after 29 years ownership).
107.026: 500SLC, 4-speed auto, thistle green, green velour.
124.090: 300TE, 4-speed auto, arctic white, cream-beige MBtex.
201.028: 190E 2.3 Sportline, 5-speed manual, arctic white, blue leather.
201.028: 190E 2.3, 4-speed auto, blue-black, grey MBtex.
201.034: 190E 2.3-16, 5-speed manual, blue-black, black leather.

Point well taken.
I appreciate candid statements.

__________________
1999 Porsche 996 Carrera Convertible
1994 420E - SOLD
1986 300E - SOLD, what a car
609 Certified