Pulse generator

Does anyone know what the cam sensor does?

Maybe help the spark box identify which pulses to use and which to ignore? The crank basically rotates twice for each cycle while the cam rotates once per cycle.

Then they found a cheaper way to do the same thing for the 87-89 bikes? Just a guess.

 

I can confirm that for you. The '86 has four connectors - three red multi-pin and a single bullet.

Just out of curiosity, I checked what the '86 connectors do:
The largest connector is a six pin. It has the four coil outputs, a grounding wire and gets power via the kill switch.
There's a four pin too. It gets the signal input from the pulse generators.
Then a two pin. It gets the signal input from the cam sensor.
The single bullet outputs the tach & fuel cut relay signals.

 

The '86 spark box is wired to all four coils independently, I'd bet that so they can fire individually. You'd need a cam sensor to fire sequentially like that. The earlier 360deg V4s I've owned all ran wasted spark with dual output coils, no cam sensor required for that. As for the later bikes, maybe they picked the cam signal somewhere else? I'd suspect Honda reverted to the wasted spark setup despite the independent coils.

 

It's impossible to run a (2 coil) wasted spark design on the 2nd gen 180* crankshaft bikes because the other cylinder would be on the intake or compression stroke and the engine would not run. Or else explode.

You can only run (single or dual coil) wasted spark designs where the other cylinder is on the exhaust stroke. That's the only time a spark would not cause any problems.

The 90-93 used a mechanical speedo drive on the front sprocket. The 94+ changed it to an electrical drive in the same location, shape, etc... to drive the electric speedo.

 

That sounds right to me.

Here are the connectors for the 88-89 spark/ignition box. It has a red 6 pin connector and a white 6 pin connector.

The red connector has:
the 4 outputs to the spark plug coils
power to the coils from the kill switch
and a ground wire

The white connector has:
3 wires going to the pulse generators. (each generator shares a common wire)
1 wire to the tach. (the fuel cut relay gets it's signal from the #4 coil wire)
1 wire to the neutral switch
1 wire to the clutch and side stand safety switch

So even though the 86 and 88/89 boxes are different, I bet you could interchange them with a few wiring mods plus you'd need to swap the starter clutch due to the different crank triggers. The tach signal may need an adapter as I think the signal speed is either 50% faster or slower. An easy fix. The ignition timing differences seem minimal as well.

I don't really know about the 87 ignition/spark box. I guess we need details on it as well.

I've looked into new, aftermarket, programable ignition boxes for the 2nd gen motors and have not found any yet. I did find them for the 1st gen motors, though.

 

Have you seen these
http://www.ignitech.cz/en/info

Roger

 

I'm not confident they can make one for the 2nd gen VFR's as they are more complicated than the VF, V twin, parallel twin and inline 4 bikes.

In fact, I wish someone could explain to me how the VFR ignition boxes work. To me, it seems like each coil should have it's own pickup coil and they should be mounted on the cams. That makes sense to me. Then it's just a matter of each channel advancing the timing based on rpm. I'm leaving out all the extra junk for starting, etc... The Honda engineers must be using a few tricks to use only 2 pickups and have them on the crank. It's a little beyond my understanding.

 

I wouldn't think they're that complicated, but reverse engineering them would not be a simple task. It's just a CPU with an integrated timing lookup table and four hefty Darlington transistors to switch the coil grounds. It might even be analog control...but given that this is late-70's/mid-80s tech, I'd figure the processor is digital. Might be an 8-bit 20-ish MHz unit. Given a clock rate like that, it should be able to time ignition events with very high precision. I'd figure the slow part of the circuit is actually the coils; they can take a bit of time to charge up. Most likely that's why Honda went with four coils, and possibly why they're fired sequentially. At 12K RPM you don't get a lot of time for an ignition event, but the transistors and CPU can easily handle that sort of speed. I'd also bet that much of the spark box is filtering for the crank sensor signal.

Honda uses an unusually wide single crank trigger...I'd estimate it's ~20 degrees. I'm not sure why, but their newer crank wheels are far more complex. My humble '80s tech Bosch Motronic 1.7 (installed in a '91 BMW 318i) can switch four coils with a crank and cam setup using only a single pickup on a 60-2 crank wheel, as well as managing fueling, idle control and even providing error code storage & retrieval.


IIRC the units were even designed to fail gracefully, locking in a fixed ignition timing to keep the motor running. They'll also safely shut down the motor in case of a catastrophic system failure...unfortunately that "feature" causes us aggravation!

Now that I'm thinking on it...I'd suppose that a Megasquirt in timing-only configuration would be the best answer to fixing these systems permanently. They're cheap and open-source with good community support. Once someone makes it work, the ignition tables can easily be posted & shared. If one even wanted to adapt later FI throttle bodies to the earlier Vif, I'd presume it wouldn't be all that difficult. I'd be tempted o try if I was a little better with fabrication.

 

From what I've read online, it was the 87+ ignition boxes that went fully digital, hence why they did not need the camshaft sensor any longer. I dug up 2 pictures of an 87 circuit board. They are attachements. EDIT: These are both 86 ignition boards.

I found this in the back of the Factory Service Manual for 1986.




As to why they used 4 spark plug coils, they had to. The firing order was 1-180*, 3-270*, 2-180* and 4-90*. (Interestingly, the newer Yamaha R1 with the crossplane crank has the same firing order and firing intervals. That's why it sounds just like the 86+ VFR bikes.)



Check out this video to better understand the firing order of the 180* crank V4's.
http://youtu.be/MwEbwKBic6w

 

Here is a picture of a 1986 Ignition box (with a broken off top left corner). What's strange is that it looks just like the 87 box (wrong, my earlier picture is of an 86 box as well). Can the 86 be partially digital while the 87 is fully digital and they look so similar?

 

I'm sure they're all digital. Most digital control units accept analog data (like sensor inputs), but they use a DAC to convert it into digital form before processing it. That all has nothing to do with the cam angle/position sensor. From my limited understanding, any timing computer requires the cam angle to figure out if the piston is at TDC on the compression stroke or the exhaust stroke. Crank sensors can only determine piston position. Any 4-stroke multi-cyl. engine without a cam sensor must fire coils in some sort of a batch.

I've heard some guys with newer Vifs talking about COP upgrades, IIRC they wired them in wasted spark mode. I took a look at the '95 wiring diagram to see if it had any differences, but the '95 has four coils wired independently just like the '86. In fact the computer wiring looks very similar to the '88-'89. No mention of a cam sensor on the '95 diagram either, I'm afraid.

Sorry for the rambling post. The short answer is...I'm baffled. Nuthin' new there! layful:

 

Im sure ingitech do a box for our bikes. I did email them when my 87s box stopped working, they did respond with prices etc...
So that says to me that they can make one.

Roger

 

Dave, on the crank trigger ring (not sure of the name) there should be something unique about the number 1 trigger (87 and up because the 86 has the cam sensor). Either a double or longer trigger that the ignition box looks for only on the cylinder 1/3 pulse generator. (it actually has 11 triggers all equally spaced except where a 12 trigger tooth should be. That 1 tooth is missing and the gap is double wide. That's how the 87+ boxes get their reference point. It sees steady pulses then a missed pulse)

It's possible the 87 and up bikes DO use a wasted spark design, but only for each of the 4 coils. No sharing of coils with other cylinders due to the unique firing order. It would mean the 86 bike fires it's number 1 coil once every cycle while the 87 and up bikes coils fire twice every cycle. Meaning the 87 and up ignition boxes operate twice as fast as the 86 box. (its a theory) Maybe Honda could not affordably build a fast enough box in 86 so they needed the cam sensor to tell the slower box which of the 2, 360* apart, number 1 cylinder trigger pulses to fire on.

It kinda makes sense. The 87 and up boxes have the same basic internal and electrical design, (no they dont) but the circuit chips just operate twice as fast, hence no need for the cam sensor.

I may be on to something here. What do you guys think?

 

That sounds great. I'm guessing they can also build one for the 86 bikes. If my above theory about the 87 and up boxes operating twice as fast as the 86 box is true, it's possible Ignitech's 87 box design would work on the 86. You'd just ignore the cam sensor wires. EDIT: (Nope, the crank sensor design is very different. They would need to build a unique box for the 86 bikes)

What were their prices if you don't mind telling?

 

Oh yeah, the original poster, a1gatorz, was able to buy a set of pulse generators on eBay and will get them soon. He may have forgot he has this thread here.

 

Yep, I think you're right there. I'd guess the 86 box can fire coils individually once per 720deg cycle. The later bikes still fire individually, but twice per 720deg cycle without any need for the cam trigger.

But there probably isn't anything special about the crank trigger ring - other than two sensors - that's a little odd in my experience. No doubt there's some cleverness going on there. Honda is pretty famous for trying out new ideas. Using two triggers on the crank ring gives a lot of data (even enough to run fuel injection). Each sensor can also pick up both sides of the teeth, so you could work out a lot of crank angles. It looks like the teeth are about 30 deg wide and 180 deg apart, in that case one crank sensor with two trigger teeth will give you 0, 30, 180, 210, 360, 390, 540, 570 & 720/0 degree positions. It'd be easy to time the motor, firing each coil once every 360 crank degrees, properly offset for their firing order. It's fully sequential ignition but fired twice per 720deg cycle. One hall sensor and one tooth is all you really need - but Honda kept them doubled up. I'll bet Honda still has two totally separate ignition computers set up in that single case to provide a fail-safe. RPM sensing would be near instantaneous too with a timer circuit, so calculating timing advance/retard in real-time isn't a big deal.

I'll also have to disagree on the upgrade. I'm thinking they're the exact same unit with a few detail differences, probably more for cost savings than any other reason. Costs are very high to develop special purpose electronics, but they come down rapidly in mass production. From what I've seen drop-in upgrades are rare, except for data burned on EPROMs. Besides, the processors and transistors were unbelievably quick, even in the 70's. If I understand this correctly, even a 70s CPU like the Z80 could run 4000 instructions per second. Those Z80 chips are so cheap & fast that they're still in wide use today. I'm not sure what CPUs Denso used, but I'd assume our unit can easily sample & resolve ignition events down to the millisecond range (100 times/sec). The transistors can be switched at insane speeds if they're well cooled, over 1000 times/sec. Small coils aren't slow either, and charge very rapidly. I also wouldn't be surprised if a set of four small coils could fire sequentially several hundred times per second if you could feed them enough amps. At 12K RPM, with the engine firing every other revolution, you'd only need to fire ten events/second/cyl. All four cylinders combined would only require 40 events/sec.. So if my math is correct, the spark box should be fast enough to fire a coil multiple times per ignition event.

 

I agree with you there, but only on the 86 bike since it had the cam sensor. The crank triggers were just 2 raised sections 180* apart and about 30* long in rotation.




The crank triggers on the 87 and up bikes are totally different. It has 11 raised sections each about 5* long in rotation, equally spaced with one missing. The missing trigger tooth is probably the computers reference point.

Below is the 87 and up VFR750 Crank Triggers



It's this difference that I think Ignitech would have to build a special box just for the 86 bikes. Unlike what I said earlier, which I've changed.

 

Just for reference purposes, here are pics of the 3 different ignition boxes. Notice that the 86 is slim and rectangular while the 87-89 is more square shaped and much thicker.

I now know that the picture I posted showing the circuit board of an 87 was in fact an 86. That's why they looked the same. I have yet to find a internal picture of an 87-89 ignition box.

 

Hard to say, I'm no expert. It could be both analog and digital - there are plenty of digital controllers that run analog I/O.

 

I had a parts bike that I've raided . Not two on one bike