Transient Fueling Control on Port Throttled Bosch K Jetronic 1

[Marquis]

First of all let me outline the imperative requirements of my system which can not be altered. My subject project engine must run on Bosch K Jetronic/CIS fuel injection or I will/may conceed to a purely Bosch mechanical system (reluctantly) as say used on old early 1970s Porsches and Mercedes. The ignition system on the car will be breakerless but of a conventional distributor design, with centrifugal advance and a vacuum capsule for part load advance. I am fully aware of the benefits of a modern solid state mapped Hot wire mapped system and don’t need some after market supplier harping on about the benefits of these to make a quick profit-I work with OEM systems like this everyday thank you!I am NOT interested in any kind of modern mapped system (either ignition or fueling) for this project.

Now the subject engine is a BMW M20 belt driven 2 valve small six cylinder bored and stroked to the limit to 3 litre capacity. This will be mounted in a light weight 1970’s 3 series coupe. The camshaft profiles and a lot of the intake manifold runners will be designed and optimised by myself. But this is the twist: I intend to retro fit the port throttled inlet manfiold out of the E36 M3 engine (euro spec)-S50-with it’s 14 litre plenum volume to get full performance benefit. The single K jetronic air mass metering system and fuel distributor will be mounted upstream of the plenum. The port throttles of course will greatly enhance throttle response. The big problem I forsee is that during a throttle position change, it will take evacuation of a lot of the 14 litre plenum for the K jetronic Air mass metering disc to realise the new requierement for fueling. This will probably provide poor transient fueling response and most likely cause lean-flame-burn out misfire. Does anyone have any suggestions of how to avoid this phenomenon while keeping to the above constraints?
Now the warm up regulator does have a vacuum line attatchment for manifold pressure changes for some transient fueling compensation but I’d imagine this is limted. Does anyone know for sure? Can the port hole here be bored out to play with this fueling?
Does anyone have any background with high performance K Jetronic applications?
Does anyone know of any cars that were produced that had the Bosch K jetronic and port throttles? Early Porsches such as the RS’ had port throttles but had a Sequential Bosch mechanical system I believe- similar in operation to the Kugelfischer system.
Any relevant help would be appreciated.
(again no talks of multiple carburetters or Solid state mapped systems- thank you!)

 

[Chumley]

Every Bosch K-Jet system I've seen, including my own
K-Jet/Lambda in the NA version keeps the throttle body
closer to the mid volume point between the intake valves
and the air flow meter. The turbos are different.

While moving individual throttle bodies closer to the
intake ports might get you better individual air control,
what you gain there you will most likely lose in mixture
stability, one reason for using a K-Jet system over a
carburetor in the first place.

I'm unclear about what your intake runner length would be,
but without sufficient length/volume you might experience
some mixing/vaporization issues at lower throttle positions
because of the relatively lower injection pressures and the
CIS injector design compared to the later Motronic systems.

Having a large plenum in between the TB and the AFM
would not only give you a temporary lean out at rapidly
increasing throttle settings, it would also give you a
major rich bump at throttle closure.

Realizing that the that the K-Jet system is at least 25
years old, it did have fairly good throttle response in
its stock form. Why do you want to change it? AFAIK
Bosch never did. They just gave up on it. That tells me
something!

This probably doesn't help you much, but I hope it does,

Chumley

 

[andyv8]

Looking at my Bosch book, it says that the warm up regulator can incorporate a full-throttle enrichment function and it does directly affect the control plunger in the fuel distributor. Hence I think that you can probably adapt it for transient fuelling.... Maybe by adding restrictors into the vacuum lines?
Apparently a seperate add-on is also available for closed loop control (probably v. basic) which you might be able to use if you could borrow a UHEGO sensor from work for a few days.......

 

[Marquis]

Thanks for the prompt response Chumley.

It's no secret I'm a fan of port throttles.
You mentioned the K Jetronic system being quite good in terms of throttle response out of a box, and this is probably partly due to that the fact the systems inherently do have relatively small throttled volumes as you outlined also. They also tend to have fuel "sitting around" the ports, as the injectors inject continuously, so compared to say, L-Jetronic (pseudo sequential-pointless really!), for a giving throttle angle change there is always ample fuel around!

In terms of why I am a fan of port throttles and the benefits- I'll copy and paste what I have written before

"For one they do not give you extra power when bolted on to an EXISTING SET UP infact they can lose you power due to the extra restriction they present to the air flow so near to the ports. The main and primary advantage of using port throttles much sharper throttle response- it’s easy to see why: consider a conventional single throttle engine operating at a particular load-speed site. When the throttle is pushed down further, before an airflow change is realised by the engine, the big volume of the plenum and runners and anything else between where the throttle is and the intake valves are, must be filled –this usually occurs over a few engine cycles. With a port throttled engine the throttled volume between the throttles and the intake valves is so much less there is a more instantaneous response.*quicker air control as we have agreed upon-bascially*
Now this is how port throttles CAN increase power if done correctly: Because of the above response issue on a conventional engine the throttle volume, consisting of plenums and runners etc etc is constrained , with a port throttled engine this constraint is now effectively removed, and the plenum volume can be increased. This is shown on for example on the S52 BMW engine which has a plenum volume of 16 litres, which along with appropriate cam duration and timing no doubt helps it to achieve it’s great power AND torque figures.
The other effect it has –is that it allows you to use longer duration cams with much more overlap without the corresponding degrading of idle speed quality and high low-speed residual exhaust gas flow back into the intake- so common with other high overlap applications. So therefore you can effectively have better low speed Vol.Eff and therefore torque with a longer duration cam with more overlap then you’d normally be able to use. This in itself is another removal of a design constraint which again allows higher power and torque. "


"And finally the seldom published benefit is the fact that you can get a fuel economy benefit from using port throttles, because you are cutting your PV pumping loop during part load operation . Tests and SAE papers have shown typically this to be in the order of about a 3 % gain."

I've calculated- using cycle simulation code that a primary runner length of about 400mm will be needed of about 5.2 cm diameter at the plenum end tapering to the port size at the cylinder head.
I would say they (Bosch and other OEMs) gave up on it due to emission compliance (wall wetting) and it's associated inflexibility in terms of mapping, may be also the cost and fine tolerances associated with it.
I don't have the constraints of emissions in this way. Continuous injection is fine and would give me the sort of simplicity and reliability I dream of. I believe if this project were given to an OEM they would be fiddling with the vacuum line post port throttles and tieing this into the control pressure regulator -so transiently -when the throttle is opened-the control pressure is reduced to get more fuel injetced for a given air flow. I would imagine that this characteristic would have to be tuned into the Control pressure regulator-perhaps by altering the vacuum port size-may be the spring the diaphragm sits upon also.

 

[Marquis]

Andy I think we're thinking along the same lines!

I didn't think of the Uhego sensor aspect- that another last resort option-as I'm trying to steer clear of electronics if I can!

 

[Chumley]

Marquis: Another reason Bosch gave up on the K-Jet system
is that throughout the operating range, the "factory specs"
permit a difference of up to 10% in the fuel injection
quantities between cylinders. (10% is a lot!) Although
that is better that most carburetors, fully mechanical
injection can only get so good!

Not all the Control Pressure Regulators have a vacuum
enrichment option, however I have changed the CPR settings
to alter the overall mixture. Bosch used that method for
altitude compensation and WOT enrichment in some CIS
applications. It turned into a big mess for me so I reset
the factory CPR settings.

I read that you prefer not to go this route but have you
considered using a frequency valve/O2 sensor feedback loop
with the standard Bosch CIS/Lambda ECU? By slightly
altering the pressure in the lower chambers of the fuel
distributor you can make continuous minor A/F mixture
adjustments based on O2 content in the exhaust. That's how
Bosch initially got the K-Jet system work with catalytic
converters.

I didn't think about the fuel efficiency benefit of using
the port throttles but, I've got to side with Bosch on
that. With all the pressure for manufacturers to improve
fuel mileage in that era, if Bosch could have got it to
work while maintaining the basic CIS configuration and
gain 2-3% fuel economy, my bet is that it would all ready
be done.

Thanks for your previous detailed reply.

Good luck! This doesn't sound easy.


Chumley

 

[GregLocock]

While 10% injector to injector variation doesn't sound very good, I doubt that modern injectors are that much better. Had some figures once...

Cheers

Greg Locock

 

[Chumley]

Hi Greg, I'd like to see your 25 year old overall "figures." compared to the latest OBDII USA requirements.

The newer Bosch motronic fuel injection systems
compensate more quickly. Bosch CIS systems don't have
a chance against those, which were designed and promoted
by, . . . Bosch, 25 years ago. Who is running this
show?

With all due respect, I say, "Follow the money."

Chumley

 

[GregLocock]

You are right. Engine lab says 'down in the weeds'. Spec says around 3%. Experience says half that for a given engine set.

Cheers

Greg Locock

 

[Marquis]

Chumley thanks for your information.
I had totally forgotten about the K jetronic injector to injector flow rate variance. To be fair the later systems and especially the ir shrouded KE systems were a lot better then what’s fitted to my old tub. (I’m fitting KE inectors from a Mercedes 300 E).

I’ve found out that the injectors on our Euro stage 4 OBD2 compliant engines have about a 3 percent variance of fuel flow during the normal operating range (NOT really low loads and engine speeds).

Anyway, to get this thread back on track I don’t want to be talking about catalytic converters or OBD compliance. Yes, I know you can only control your mixture to your leanest cylinder and this may present a problem with a larger 10 Percent band, but not to my home grown vehical. I also don’t care to the obsessive anal degree of today, about idle stability ( to the point where good fuel economy is thrown away) or shunt/shuflfle, I think I’d like to leave those behind closed doors at work!

I don’t have the closed loop lambda control ECU on my K jet system. And short of rigging up some kind of wide band system, I’m glad too, as I don’t want to be throwing away good part load fuel economy to be forced to run stoichiometric on a car that doesn’t even have cats!

Anyways, what really struck me about what you said was this

“Not all the Control Pressure Regulators have a vacuum
enrichment option, however I have changed the CPR settings
to alter the overall mixture. Bosch used that method for
altitude compensation and WOT enrichment in some CIS
applications. It turned into a big mess for me so I reset
the factory CPR settings”

I believe I have both a feed in the CPR to vent to either atmosphere or put a –pre throttle pressure feed for overrun and also a post throttle vacuum for limited transient enrichment. How did you fiddle about this and monitor the outcome-if I may ask?
What were the pit falls?

 

[ghamish]

This is probably not of much help, but would there be any way of tempoarily displacing the metering plate upwards with something like a small blast of compressed air from a small nozzle when the throttle is opened rapidly, sort of like an accelerator pump in a carburettor?

How well do these old CIS systems compare in power & drivability to a modern aftermarket efi? Would they get to within 10%, because they look like a very cheap way to put fuel injection on an older carbed engine.

 

[comnean]

Hello,
I have a lot of experience with all K injection sistems. The best way to modify (maybe a compromise) the feeding regime is to use one frequency valve mounted on the return line and one custom made controller piloted by one Lambda sonde. The controller can be made also in analog structure and can be full adjustable (amplitude and offset for the frequency signal). One adjustable Bosch IDLE Air Valve is great in aim to adjust the negative pressure from the anecoic chamber that is inside WUR. You can control also the IAV by one adustable cotroller. Factory made the controller operate at variable frequency and 9-12V but very best results i obtein with variable amplitude (6-12V) and minimum stable 40Hz generator. Conditions are:
1 To find one godd fitted FV (for flow and frequency reponse) or to modify one big injector (big flow)
2. Good shape and capacity for the cone from Air Mass Flowmeter (good fitted for engine capacity)
3. The WUR can be made adjustable (see Bob Tindel threads about Porsce CIS sistem-also similar)Thanks Bob!
Sorry for english

 

[ipm]

The newer the k jet, the harder it is to hotrod. Mix and match parts including the really big distributer from a mercedes for old mechanical systems. Once the electronic are added, the options drop fast and your m20 gonna not get enough fuel. Especially tough once ecu's and Eproms are utilized with the cis.

My m20 is motronic, not cis, and was gonna switch to gm speed density till I found enough maps on the eprom to tune.

 

[rgconway]

Marquis, I have been following your thoughts on ITB's on a BMW M20 with great interest. Do you have an update for us? THanks.

 

[Marquis]

rgconway: not yet mate, I haven't been coming to these forums much, but I STILL do intend to mate port throttles on my M20 BMW.
My priorities for this project are some sort of simple to work on mechanical fuel injection.

The lambda sensor comments above are the reason why I've decided to make myself scarce in this forum:the danger of thinking you know more then you think you do- Lambda sensors in the exhaust are for fueling changes- for closed loop-usually, and this has little to do with transient mixture control during TIP IN and the issues I speak of.
In fact relying on a lambda sond feedback control system would be even slower then my initial misgivings of using the existing port on the warm up regulator.

I'll go on using K Jetronic, and simply bore the vacuum port in the Warm up regulator-and place an adjustable apperture -perhaps with a orafice and screw and moderate it to suit-monitoring the K Jet control presure while I'm doing so.
If this leads to major problems, I'll use the old style Kugel fischer-purely mechanical system- as I know this IS compatible with port throttles as it was used in the BMW M1 and 1973 Porsche 911 Carerra RS 2.7.

Among the obvious throttle response and top end performance benefits I hope to reap (using a 14 litre plenum!)- according to work I've done and papers I've read (SAE-960580)-I look forward to gaining fuel ecconomy during part load due to faster pressure recovery in the port-leading to lower pumping losses in part load operation.
I'll keep you posted

 

[rgconway]

Marquis, thanks for the update. My interest is similar although I plan to use EFI. I am considering a megasquirt contoller referenced from MAP or alpha-N with a WB loop.

I appreciate that your challenge of getting good road drivability using mechanical FI are considerably greater than mine using EFI in a racecar.

Have you decided on a type/source for throttle bodies?

I'm not moving too fast on my M20 project, but would like someone to compare notes with on occasion.
Cheers, Bob

 

[Marquis]

I'll be trying to adapt ones from a euro spec S50 E36 M3 engine for it has the same bore centres-I have the engine already- and I'll try to make up a flange to adapt it to the 2 valve M20 heads.

If this doesn't work, I'll get the throttles specially made....to my specs, there are several companies here that could do so at a fraction of the price of the ones BMP design sell I imagine too!

 

[ghamish]

Marquis, what about the old Lucas timed mechanical system off a Triumph 2.5 PI? From what I read they ran ITB's as std.

 

[patprimmer]

I have seen individual runner, carby manifolds, with the carby bodies used purely as throttle bodies. The venturis etc were completely removed and fuel passages blocked with epoxy. The float bowls are optional, depending on how much effort is required to remove them.

Regards
pat pprimmer@acay.com.au
eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.