1987 VW Vanagon

[bertbeck]

Once I have the mandatory items:

Will I have to do something special for the Hall sensor, and the other various input sensors ?

Will I have to do anything special (hardware interface wise) to control spark, fuel injection, fuel pump etc ?
If so - is it mostly verify signal level, type, characteristic (and massage) or are these mostly similar and ready to be connected ?

It is a Digifant system - 1987 VW Vanagon.

I'll look for a used/dead ECU to grab a connector from.

Bert

[AndreyB]

Define 'something special'?

The intention is to use the OEM hardware as is, so that you can always go back to the stock setup. The Frankenstein board actually has two options when it comes to Hall sensors - you can either use some of the op-amp analog channels, or you can use the special hall/VR input chip which is also on the board. I can solder a jumper wire for one setup and worst case scenario you would need to re-mount it to the other input circuit.

The injectors - if they are "high impedance injectors" - and 97% of them are - are expected to work right away. Ignition - http://en.wikipedia.org/wiki/Digifant_engine_management_system - that's a bit confusing Hall control unit ? maybe something like the EDIS where you control only the advance? Hopefully would be drive my the 5v high-side schematics. If not, you have an options to move the jumper wire and make it +12 high side if that's the way it's controlled.

Bottom line - you would not be alone, we will figure it out

[AndreyB]

http://car-part.com search for 1987 VW Vanagon "Engine Computer" shows "Ignition Cont" and "Fuel Inj Cont" as separate items, the first is cheap - the other one is not.

Maybe the connector is shared between this vehicle and something more popular?

[kb1gtt]

Please tell me if I remember the digifant correctly.

Qty 4 of standard highz injectors one for each cyl, distributor with one coil. The coil has the "power" stage integrated, such that the ECU sends a 0-5V or 0-12V signal to create a spark. One CAM hall under the distributor cap, and the other was some where else, perhaps fly wheel, but I seem to recall it might have also been embedded in the oil pan looking at a crank wheel or something like that. However that may have been digifant II. I seem to recall the mechanical injection was called centronics or something like that. Do you have a fuel rail? Or do you have the kind of mechanical injector thing with a bunch of little hoses that runt o each injector? Kind of like a distributor cap, but with fuel.

The odd 3% cases are typically when you have one injector in a throttle body or if you have very high performance 1kHP with port injection. That's generally when you will be low Z. Most standard flow and standard OEM port injectors are high Z. That centronics thing was .05% of the time, and really odd.

Do you know your crank wheel pattern? I think that's going to be the biggest "real" issue. You'll either have to sniff it, or get details about the angles. Basically russian will need to know the pattern such that he can make a wheel decoder. Other than that, the hardware should be good, it's all a software issue after that There will of course be some other less "real" issues, like migrating wires and tuning.

[abecedarian]

Digifant, specifically Digifant II has: 4 tooth + Hall in distributor, volumetric air flow sensor (may be movable vane or hot wire depending on year / model) + intake air temperature sensor, coolant temp, idle and WOT switches on the throttle, O2 and maybe knock (depending on year and or model vehicle). Some models may have throttle position sensor instead of idle/WOT switches. It is a common rail electronic injection system with all injectors wired and driven in parallel, so no alternating or sequential injection. Idle speed was controlled by an electronically controlled bypass valve that opened and closed rapidly in response to signals from the ECU as well as modifying ignition timing at idle.

The mechanical system used by VW and Audi (and Porsche and others) was known as Jetronic, primarily K; D and L were electronic systems. And it was hardly 0.05% use- nearly every FI VW and Audi and Porsche during the 70's through mid-80's used it. Type 1 through 4 air cooled engines (Beetle, Transporters, 911/912, 914, et cetera), Rabbit, Scirocco, Jetta, Dasher / Quantum / Audi 4000/5000, 924, 944 are all among the list of users of this. I seem to recall some Mercedes and BMW's used it also.

It was (at least) one fuel distributor that air flowed through. Air flow would displace a plunger centered in the housing which in turn varied the fuel pressure sent to the mechanical fuel injectors. This system also had a warm-up regulator that increased main fuel pressure to the fuel distributor in order to enrich the mixture with a cold engine. It may or may not have a bi-metallic air bypass valve to add additional air to the intake to bring idle speed up during warm up. Later models incorporated lambda control by using a 'frequency valve' mounted to the side of the fuel distributor. When the O2 sensor signal was stable, the ECU would alter the duty cycle of the frequency valve. This valve (essentially an electronic fuel injector) would allow primary fuel pressure to bypass the regulator and return to the tank thus allowing for finer control over the fuel mixture than the fuel distributor plunger and controls could accomplish by itself.

*Edit to add-
Had a brain fart and I have to take back my claims about K-Jetronic on air-cooled's: most of them used D or L Jetronic, not K.

[abecedarian]

Once I have the mandatory items:

Will I have to do something special for the Hall sensor, and the other various input sensors ?

Will I have to do anything special (hardware interface wise) to control spark, fuel injection, fuel pump etc ?
If so - is it mostly verify signal level, type, characteristic (and massage) or are these mostly similar and ready to be connected ?

It is a Digifant system - 1987 VW Vanagon.

I'll look for a used/dead ECU to grab a connector from.

Bert

Most likely you will not be able to use your air flow meter, and by extension the intake air temperature sensor since it's mounted in the meter; you will need a MAP sensor and intake air temp sensor.
Since your Hall effect is in the distributor, and only 4 tooth, you will need an alternate triggering system.
Again, most likely, you have the throttle body with idle and wide-open switches so you will need some sort of throttle position sensor system... perhaps a throttle body from a later model or other model (Golf, Jetta, et cetera) that used the system and had the TPS could work(?).
You will probably have to re-wire the fuel injectors too, so they can fire independently, unless this was handled in the existing ECU. I don't have a schematic and it's been a long time since I've seen a in the back end of Vanagon.

[AndreyB]

Most likely you will not be able to use your air flow meter

Why?

You will probably have to re-wire the fuel injectors too, so they can fire independently, unless this was handled in the existing ECU. I don't have a schematic and it's been a long time since I've seen a in the back end of Vanagon.

Why? There is simultaneous injection mode. Are you concerned about the driver IC? In simultaneous mode you can wire up all the injector drivers in parallel so that the load is distributed between them.

[abecedarian]

Most likely you will not be able to use your air flow meter

Why?

It's most likely the 'volumetric air flow' type meter where air flows through and physically displaces a 'door' which in turn is sensed by a variable potentiometer. VW didn't start using hot-wire MAF until well into the 90's.

You will probably have to re-wire the fuel injectors too, so they can fire independently, unless this was handled in the existing ECU. I don't have a schematic and it's been a long time since I've seen a in the back end of Vanagon.

Why? There is simultaneous injection mode. Are you concerned about the driver IC? In simultaneous mode you can wire up all the injector drivers in parallel so that the load is distributed between them.

So there's another option.
Didn't realize you could parallel injectors on the one end, and parallel the drivers on the other end.
If the drivers are wired in parallel output, how do you turn them all on at the same time to ensure one driver isn't handling the current spike by itself?

I pay attention, but not as much as I used to.

[AndreyB]

If the drivers are wired in parallel output, how do you turn them all on at the same time to ensure one driver isn't handling the current spike by itself?

Hm... Good point, there probably would be a some uS (?) delay between the turn on time between the ICs. Maybe these should be simply soldered in parallel in this case

[abecedarian]

If the drivers are wired in parallel output, how do you turn them all on at the same time to ensure one driver isn't handling the current spike by itself?

Hm... Good point, there probably would be a some uS (?) delay between the turn on time between the ICs. Maybe these should be simply soldered in parallel in this case

That was, if I remember, the way the stock Digifant system was wired.

[kb1gtt]

The VNS14NV04 is rated for 12A, if you can keep the chip cool enough and if you support 4A on the connectors and wires and such, you can drive multiple injectors with one channel.

I don't see a problem with the flappy paddle style of MAF sensor. In the end its a MAF sensor, so voltage correlates to O2 in the cyl. The issue is that it might not be linear, so you may need a software calibration table to predict the O2's in the cyl.

I only ever saw the distributor signal used for TDC verification. Many of the distro-cap's had pins installed such that you could not rotate the cap to change the spark advance. As well you could disconnect the distributor wire and the engine would run just fine. MIL would go bright, but it would run fine. I'm sure there is a more accurate crank sensor some where else, other than under the cap. However bertbeck will be the best suited to know what is actually installed on that engine.

[abecedarian]

You might be correct regarding VAFM type sensor being usable. One thing I've seen though is that these often use 7v or 12v for supply, not 5v like many would think. Also, have to remember that it is a "Volumetric Air Flow Meter" not a "Mass Air Flow" meter. The 'volume' of air is measured, not the mass, which is why they include an air temperature sensor in the device so the ECU can derive mass from volume and temperature.

One significant drawback is that the vane door has mass and therefore inertia, and changes in the output signal lag changes in airflow because of this. Blip the throttle open and it will momentarily report less airflow than actual until it reaches it's appropriate open position, meaning you'll be leaner than expected; close the throttle quickly and it'll report more airflow than actual until it settles, resulting in a rich condition. Some engines with these often include Helmholtz resonance chambers in the intake tract for 'noise control' which further exacerbates the issue due to more volume in the intake tract itself.

Another issue with this type of sensor, again stemming from the vane door having mass, is operating the vehicle on less than ideal terrain can physically displace the door in spite of no change to airflow. Case in point would be the Toyota trucks of the mid-late 80's through mid 90's: if the vehicle was on an incline, the vane would have to overcome gravity as well as react to airflow and pitching from one side to the other, or simply being inclined to the left or right, would affect the fuel mixture.

And finally, it's a substantial restriction in the intake tract, one which prompted some people to develop VAFM-MAF conversion kits as well as VAFM-MAP conversions.

On the up side, it's the only type of airflow meter that can offer a physical way to affect the fuel mixture. They typically have (sealed): a bleed screw that allows air to bypass the vane which can enrich or lean the fuel mixture at idle if the ECU doesn't implement closed loop / O2 feedback control at idle; a spring that counter-balances the vane against air flow, which can usually be adjusted to increase or reduce the vane's ability to react to changes in air flow- relaxing the spring results in higher airflow being reported than actually is and tightening the spring results in lower airflow than actual being reported. Turbo Toyota Celicas, Supras and Pickups owners often resorted to these tweaks to help tune the engine.

My suggestion that the VAFM be replaced with something more modern still stands. Replacing it reduces restriction within the intake, provides better response to changing throttle conditions and wouldn't require complex voltage conversion (in the case the VAFM uses 7v or 12v for operation- no schematic so I can't verify this), nor require modifications to the RUSEFI code to support it.

[bertbeck]

1987+ Vanagon Digifant ECU links:

http://www.van-cafe.com/home/van/page_870/digifant-system-components.html
http://www.vanagonauts.com/files/Vanagon%20Digifant%20Fuel%20Injection%20System.pdf
http://www.vanagonparts.com/digifant.html
http://en.wikipedia.org/wiki/Digifant_engine_management_system
http://forums.generationdub.com/showthread.php?t=26511

The injectors all open at the same time.
Is there an advantage to controlling them individually ?

[kb1gtt]

Individual control allows individual cyl tims, which has pro's and con's. It's more complicated, but can make things run a bit better.

Use of the flappy paddle is all good, until some dust causes it to stick, then it's all messed up and wrong. I would agree, the flappy paddle is not optimal, but they also work for many engine setups. This is why there were on cars for a reasonably short period of time, and superseded by better technology.

[abecedarian]

...The injectors all open at the same time.
Is there an advantage to controlling them individually ?

Individual control allows individual cyl tims, which has pro's and con's. It's more complicated, but can make things run a bit better.

I agree. Being able to adjust each cylinder's fuel individually can let you 'balance' things. I.e. if one injector flows 1% more fuel than the others, you can adjust that injector's properties in the ECU and compensate for things, to an extent. So, it is more complicated and puts more demand on the ECU.

Use of the flappy paddle is all good, until some dust causes it to stick, then it's all messed up and wrong. I would agree, the flappy paddle is not optimal, but they also work for many engine setups. This is why there were on cars for a reasonably short period of time, and superseded by better technology.

If you call from ~1974 to well into the 1990's a reasonably short period of time, I'd have to agree.

Hundreds of millions of cars used them, from Toyotas (Celica / Supra, HiLux / Pickup, MR2 and Camry, Tercel, Paseo) to models from Nissan, VW, Audi and Porsche, Ferrari, Mazda, Subaru, Volvo and I lost track. To the best of my recollection, Chrysler, Ford (domestic produced, not Mazda sourced), GM (not Toyota or Isuzu sourced), Mitsubishi *(self or Diamond Star sourced), and a handful of European (predominately UK based) manufacturers did not use VAFM sensors.

Companies like AEM and HKS offered VAFM eliminator kits, way back then.