K-Jetronic - Measuring Duty Cycle, Adjusting A
K-Jetronic - Measuring Duty Cycle, Adjusting A
K-Jetronic - Measuring Duty Cycle, Adjusting A
K-Jetronic: Measuring Duty Cycle, Adjusting Air-Fuel Mixture, & Setting Idle
Overview
Before doing anything, let's discuss what CIS, K-Jetronic, lambda, and duty cycle are so that you get a very basic understanding of
how the system works and what it is you'll be testing.
K-Jetronic is a mechanical fuel injection system developed by Bosch in the 1970s and is used in a wide variety of vehicles around the
world. K-Jetronic is a continuous injection system (CIS), which means that the fuel injectors receive and produce a constant flow of
fuel, provided there is proper pressure, whenever the engine is running. CIS is a mechanical injection system because fuel is
metered by the mechanical link between the fuel distributor and airflow sensor plate, which are linked together in one housing. In
other words, fuel is metered in proportion to the incoming air flow.
In order to adhere to stricter emissions requirements in North America, Bosch enhanced K-Jetronic by adding lambda control. With
the addition of an oxygen sensor (aka lambda probe) feedback system, K-Jetronic's air-fuel mixture can be more precisely controlled
while the engine is running under various conditions (cold, warm, full-throttle). K-lambda (as it's often referred to) has better
emissions, in theory, than its basic predecessor. It is called lambda because Bosch uses the Greek letter lambda (λ) to represent the
stoichiometric ratio (14.7:1) of the air-fuel mixture the system strives to achieve for peak operating performance.
While it is often called an electronic control unit, the Jetronic controller is not a computer
and should never be referred to as such. Because this control unit is analog and merely
processes and reroutes voltage signals, it is very wrong to refer to K-lambda as an
electronic fuel system… it is not an electronic fuel system! While the frequency valve
operates based on an electronic voltage signal, the air-fuel mixture still takes place
mechanically inside the fuel distributor as described above. The K-Jetronic lambda
controller installed in Volkswagen Cabriolets (and Rabbits, Golfs, Sciroccos, and a host of
other vehicles) consists of nothing more than a circuit board with a bunch of capacitors,
inductors, resistors, transistors, and diodes. As you can see from the photo at left, the
Photo by Myndex from benzworld.org forum.
controller is really nothing more than a relay (of sorts) on steroids.
1. Open loop - cold: Engine is cold-started, thus providing cold-running enrichment during warm-up period.
2. Open loop - warm: Cold enrichment has ended; system is at 50% duty cycle, waiting for oxygen sensor activity; also known as
limp-home mode.
3. Closed loop: Oxygen sensor is sufficiently heated and is now sending voltage signals to control unit during the engine's warm
running period.
4. Open loop - WOT: Full-throttle (aka wide-open throttle) switch has been activated, providing fuel enrichment override.
1. If you start the car when the engine temperature is cold, the oxygen sensor thermoswitch closes and sends a signal to the
control unit.
2. The control unit, in turn, sends a fixed 80% duty cycle (65% for 1988+) signal to the frequency valve. That fixed cold-
running enrichment signal (80% or 65%) remains in place until the coolant reaches a specific temperature (25°C or 45°C,
depending on the thermoswitch installed in your car).
3. The thermoswitch then opens, cutting its signal to the control unit. If the oxygen sensor is not at its operating temperature
when the thermoswitch opens, the system reverts to warm open loop, or limp-home mode (50%).
4. Once the oxygen sensor is heated to its proper temperature, it sends a signal to the control unit and the system enters
closed loop, or its normal regulating function (50% ±8% duty cycle). Should the full-throttle switch be activated, closed-loop
is overridden for forced full-throttle enrichment (65% duty cycle).
Up through and including the 1987 model year, cold-running enrichment (80%) overrides full-throttle enrichment (65%). From the
1988 model year onward, cold-running enrichment does not override full-throttle enrichment because the two values are the same
(65%).
"Open loop will also happen when the car has been parked briefly, the oxygen sensor has cooled off, and the thermoswitch
hasn't closed; upon restarting, it takes a minute or more before the oxygen sensor heats enough to produce a signal and send
the system into closed loop. That's why oxygen sensors are now heated; they start working faster, systems go into closed-loop
sooner.
"There isn't always an open-loop stage during warm-up. If the engine is cold (real cold, like winter in Montana cold), and the
oxygen sensor is in great shape, the oxygen sensor's heater and the exhaust heat may get the oxygen sensor up to operating
temperature before the thermoswitch warms enough from from the coolant to open. In that sort of circumstance, the system
will go straight from cold-running enrichment to closed-loop; no open-loop transition at all. Same car, same cold weather: The
oxygen sensor will cool quickly when the car is parked, while the coolant will retain heat much longer; on restart, there will be
no cold-running enrichment -- the system will be in open-loop until the oxygen sensor warms up enough to put the system into
closed-loop." ~tolusina of VWvortex.com
Likewise, if the car is in a warm climate (or during warm seasonal months), the system may revert to open loop upon initial start-up,
skipping cold-running enrichment due to the coolant already being at or above the thermoswitch's pre-set temperature.
The preceding is only a very brief overview of how K-lambda operates. For much more detail on the theory, application, component
descriptions, performance, modifications, and troubleshooting, please obtain Charles Probst's book, Bosch Fuel Injection & Engine
Management.
In order to read your lambda system's duty cycle, you need a digital or analog meter that can read duty cycle (%) and/or dwell. A
simple volt-ohm meter will not do the job. This is also a case of "you get what you pay for". While you can visit Harbor Freight with
your 20% off coupon, the quality of meters they sell is substandard and there are numerous reports of inaccuracy. You don't need
the most expensive meter on the market, but you want quality. Additionally, while you can use a dwell meter, it has been reported
that using a dwell meter does not provide quite as accurate of a reading as a duty cycle meter, but it gets close enough. If you
already have a dwell meter, go ahead and use it; if you are new to the game, try to find a duty cycle meter (or better, a meter that
has both for additional automotive troubleshooting).
An oscilloscope, like those above, is actually the best and most accurate device for this job, but it is not a multifunction tool for your
toolbox. There is no need to buy this type of instrument just to read your car's duty cycle, unless you wish to have a neat toy to play
with.
So, what do you look for? Example meters are on the following page. Remember, these are merely examples highlighting what
setting your perspective meter should have in order to read your lambda system's duty cycle. I.E. these meters are not necessarily
being recommended (for the record I, personally, use a Craftsman duty cycle meter that is, ironically, not shown in the examples).
If buying a duty cycle meter, the meter MUST have the ability to switch to %/duty. Otherwise, you will be measuring Hertz (Hz),
which is not included in these instructions.
Requirements:
Car should be otherwise well-tuned (see below for further information)
No, and I mean zero, vacuum leaks
Distributor advance mechanism(s) should be operating correctly
Timing should be correctly set with a closed throttle plate
Air-fuel mixture plug removed and mixture screw free of gunk
Frequency valve should be buzzing/vibrating whenever the fuel pumps are running; if it is not buzzing, please test it.
Do NOT, under any circumstances, disassemble the fuel distributor and/or airflow sensor unless you know for
certain what you're doing! Taking this piece of precision equipment apart on a whim because your CIS fuel injection
system is not running right is a disaster waiting to happen, and usually ends with you having to buy a new (actually
used/rebuilt) fuel distributor!
If you feel that you and your car are prepared for testing, proceed to the next section.
Tools needed:
Volt-ohm meter that measures duty cycle (dwell meter can also be used, but duty cycle is a bit more accurate)
3mm Allen wrench (T-handle version advised)
7mm crescent wrench
Connect the red positive (+) meter wire to the blue/white *If you have installed spade terminal disconnects in place of the
wire's terminal in the two-pin test connector by the cold-start OEM connector, simply attach the red (+) probe’s alligator clip to
valve. Connect the black negative (-) DVOM wire to any the blue/white wire terminal and the black (-) probe’s alligator
convenient ground*. clip to the brown wires’ terminal.
If using the dwell setting, set meter to "4 cyl";
If using the duty cycle setting, set meter to "frequency %" ** **If using duty cycle, be aware that Cabriolets produce a
negative slope reading. Many meters read only in positive slope.
Read your meter's instructions; some meters require an Easy test: Connect the test leads as described. With the engine
additional step, such as pressing a button such as “range”. running, actuate the full-throttle switch. If the reading is 35%,
reverse the test leads: red (+) to brown wire, black (-) to
blue/white wire. Actuate the full-throttle switch once more; the
reading should be 65%.
Step 2 ~ Warm the engine Step 3 ~ Read your meter
Target reading:
50% duty cycle (±8%)
Run the engine up to operating temperature (oil at 80°C); the 45° dwell (±7°)
cooling fan should cycle on/off at least once. Keep all
accessories off. If it is not within this range, continue to step 4. If it is within this
range, skip to Step 5.
Make note of where the air-fuel adjustment screw is on the
airflow sensor (do not put the Allen wrench in the hole until you Note: If the needle bounce/swing is higher than a range of 10
adjust the mixture!!). Note: The factory anti-tamper plug must (ex. 55-74), there are issues that need resolving before
be removed first! Please click here for details on removal. continuing.
Your air-fuel mixture setting is too lean and the control unit is Your air-fuel mixture setting is too rich and the control unit is
making a rich correction. making a lean correction.
Check for vacuum/air leaks; fix any that are found. Check cold-start valve to verify that it is not leaking.
Insert your 3mm Allen wrench into the adjustment hole and turn Insert your 3mm Allen wrench into the adjustment hole and turn
the wrench clockwise. You are richening up a lean condition. the wrench counterclockwise. You are leaning out a rich
condition.
NOTE: The weight of the wrench will affect the mixture! Turn the wrench a little, just enough to feel it move, then lift the wrench
out of the adjustment screw hole. Repeat this process until 45° ±7° dwell / 50% ±8% duty cycle is achieved.
Have your 7mm open-end wrench handy; you may have to tweak the idle bypass screw (backside of the throttle body) if the idle
changes too much after adjusting the mixture. You need to stay below about 900 RPM while adjusting the idle. Use the tach on the
dash and your ears, you know what sounds right and normal. If you go over 900 RPM, the centrifugal advance starts to kick in,
changing the timing and confusing all your readings. Set the idle RPM at 900±, but above the threshold of the idle boost valve.
If the idle boost valve kicks in while you are adjusting, you've dropped too low on idle RPM. You can unplug the idle boost valve
electrical connector, or better, pinch off either of the hoses going to it. When the idle boost valve (or hose) is reconnected, idle
should not change if you've adjusted the idle correctly.
Switch on the A/C (if installed) to make sure that the A/C idle boost valve works; idle should stay about the same with the A/C
compressor running.
Upon replacing these 2 parts the engine would run great and start fine cold. So well in-fact that I equipped the car with a Remote
Starter unit. The problem was that when the car was hot (Oil temperature between 80ºC - 100ºC) and had been sitting for 10, 15 or
20 minutes the car would be almost impossible to start or would start and studder like crazy and require you to step on the
accelerator to clear it up.
My dad (mechanic of 35 years) and I spent days trying to remedy this problem and drove ourselves nuts. We checked fuel pressures,
we checked the warm up regulator, the cold start valve (also used in hot-restart). The relays and control modules, the thermo time
switch and the list goes on and on.
We found out that the Dwell of 45º that is listed on this site and others my not apply to all models. What we did was rev the engine
to 2,000rpm and set the idle-mixture screw until the dwell fell to 45º at this RPM. Then it would naturally enrichen the mixture when
the idle was allowed to fall to the 925 ±25 rpm as indicated on the hood sticker. It was only then that the car would start/restart
properly under hot conditions.
I wanted to share this information with you to help anyone else banging their head against the wall like we did trying to get it to
start properly when hot with this Hot-Pulse Start system.
Please note that we checked all cold/hot fuel pressures and residual pressures on shut-down both cold/hot and they were within
spec. The car also has brand new injectors, thermo-time switch, fuel pump, accumulator, filter, air-filter, cap/rotor, spark-plugs,
spark-plug wires, vacuum enrichment switch, Oxygen Sensor, idle-stabilizer relay, fuel pump relay, Hot-Pulse Start Relay. As we
replaced all this trying to remedy this problem. So we know that this is the setting that it should be at."
~ Steve C.
Notes ~ Questionable readings
Should you find questionable meter operation while measuring frequency valve duty cycle (ex., the reading never changes):
Connect your test lead to the Hall sender green/white wire and test your meter. It should read a very steady 36-37° on the 4
cylinder dwell scale, 40.5-40.6% duty cycle.
If the meter proves to be good:
Conduct the tests in the lambda system testing section starting on page 8.
The test port connector itself could be faulty. Refer to the Test Port section on page 12.
Tools needed:
Volt-ohm meter that measures duty cycle (dwell meter can also be used)
Hose pinchers
New D-cell battery
Jumper wire using two male spade connectors
Test tables on page 11 and a pen
Target reading:
80% duty cycle with little fluctuation
72° dwell with little fluctuation
Test 1 ~ Cold Running Enrichment 1988+ Test 2 ~ Limp-home Mode all
This Vref is pretty much the heart of this (or any, really) 'closed loop' mixture control system. The ECU sets a steady reference
voltage of 0.45 VDC to 0.50 VDC (this voltage is steady, varies slightly between different ECUs) on the green oxygen sensor wire, the
black O2 sensor lead connects to this wire.
The oxygen sensor, once at operating temperature, outputs a voltage between approximately 0.1 VDC (a lean signal), to
approximately 0.9 VDC (a rich signal).
When the ECU sees an oxygen sensor voltage higher than its Vref, it correctly interprets that to mean that the mixture is richer than
stoichiometric (i.e. 14.7:1 air-fuel ratio), the ECU then leans the mixture to compensate by lowering the duty cycle to the frequency
valve below 50%. Once below 50%, it is now too lean; oxygen sensor voltage is now below Vref, so duty cycle goes back up to
correct.
On and on goes this continuous correction, the ECU constantly trying and failing to get the oxygen sensor voltage to match Vref.
The system is actually designed to strive and fail. There are two reasons it strives and fails. One is built in response delays, the other
is that those response delays are built in so that the catalytic converter will always have a slightly fluctuating mixture, something the
cat converter requires for its chemical reaction to work properly.
above information provided by tolusina of VWvortex
Notes
A sluggish oxygen sensor may cause a failed smog inspection while exhibiting absolutely no other drivability issues.
Should you see an operating range at the oxygen sensor ranging from -0.5 DCV to 0 DCV (instead of the normal +0.1 DCV to +0.9
DCV), your sensor has been permanently damaged by chemical contamination and needs replacing. You may also experience a
constant, very high duty cycle reading, like 96% during duty cycle check.
Cabriolets built from July 1987 through 1989 have heated (3-wire) oxygen sensors. Should you see 12 DCV (or charging voltage)
at the oxygen sensor wire, replace the oxygen sensor immediately -- the heater has shorted to the sensor.
Should the oxygen sensor control unit happen to be faulty, it will fail to compensate for the simulated lean/rich conditions
described above. The oxygen sensor control unit rarely goes bad, but is does happen. Before condemning the control unit as
being faulty, verify that the control unit is receiving power (pin-out diagram on page 15) and that all ground wires/connections
are good. Additionally, if the duty cycle stays at 65% or 80%, disconnect the cold running enrichment switch (if installed) and the
full-throttle switch, one at a time, then both together if need be.
If your meter is giving you odd readings and you have verified that the meter is working properly, the test port may be faulty as
evidenced by my own experience:
The first two tables below contain readings on my 1986 CIS-lambda Cabriolet. The initial test readings in attempts 1 and 2 point to a
possible control unit issue and/or meter issue (one reason a new meter was bought). However, the tests prove that while the
readings are bizarre, the idle changed as did the meter readings, which indicates the control unit is good as well as the meters
(meters were double-checked using the green/white Hall generator wire). The thermoswitch was proven good after conducting a
resistance test. The frequency valve buzzes every time the engine runs and the engine purrs like a kitten. The components as well
as the meters all indicated that everything was A-Okay. So, in the name of science...
Hypothesis: Test port is faulty, which was suspected at the very start.
Result & conclusion: As can be seen in the Attempt #3 table, the readings from both meters are within specifications (aside from #4,
which showed me the car had a too-lean mixture). The factory test port was indeed faulty by not providing good enough probe
contact, which can be seen in the photo below.
* * Remember, you are responsible for working on your car; Cabby-Info.com, KamzKreationz, VAG, VWoA, or anyone
else are not responsible if anything goes wrong while you are working on, in and under your car!
Use this information at your own risk!* *
Oxygen sensor
thermoswitch input