[alfaparticle]

A few years ago I did some experimenting with different triggers connected to the V6 Bosch module. I triggered it from points, VR sensors and a 123 distibutor. I used a dual trace oscilloscope to monitor the trigger signal and the coil current. I could make good sparks with all of them but none of them controlled the dwell and phased the spark correctly.

The V6 uses a Hall sensor distributor which produces a 12 volt square wave signal. The sensor is horseshoe shaped with a magnet on the inside and the Hall device on the outside. The rotor is a can with six slots that passes though the horseshoe. When the can is in the sensor the output from the Hall is 12V and when the slot is in the sensor the output is pulled down to 0 volts. The ignition is timed so that the slot is just entering the sensor at the correct ignition point.
The output signal goes then to the ignition advance computer which applies phase shift according to rpm and also INVERTS the square wave output signal. This signal then goes to pin 5 of the Bosch ignition module which triggers the coil on the leading edge of the signal. You should take this into account when you are driving the Bosch module from any device. The Bosch module triggers on the leading edge and it computes the dwell angle from the coil current signal.

 

[Graham Hilder]

Thanks Ed. Would I be right in proposing that your essential findings with the V6 module, the -111, would be that it can control dwell and coil current, but you only got
that result if it were triggered by the signal from the V6's computer, which is a 12V sq wave connected to module pin 5 in the correct phase.
I wonder why you didn't see dwell-control happening when you triggered from points, as surely that signal is is also a 12V sq wave, with rising edge at spark time? (Perhaps it was a matter of debouncing the signal? Or lack of a pull-up resistor? But I'm sure you would've taken care of these in the test rig.)
And would I be right to say that if a Pertronix Hall trigger also gives a 12V sq wave, with rising edge at spark time (using a pullup resistor if needed), then this too should give the good result if connected to pin 5 on the -111 module.
If all the above is right, then the only puzzle is the VR trigger applications. We know they do work in the standard factory design for various Alfa models. All the ones I've seen use a balanced VR connection to module pins 5 and 6, in the correct polarity (which is set by the factory connector keying). Perhaps this is the thing that matters - the VR application works but uses pins 5 and 6, and the Hall application also works but uses only pin 5 input. Wonder if this is how you connected up your VR test rig Ed?
Thanks again,
Graham H

P.S. I've just noticed that in my first post, nr 34, I wrote -011 all over the place when I meant -111 !!!!
Sorry about confusion this may be causing to anyone. To put it straight, I do confirm that std OE for Alfa boxer engines (and early Alfa 75 twincam) is -111 with reluctor dizzie and correct Bosch coil.

 

[alfaparticle]

Hi Graham,
I gave the folder containing all of my V6 ignition information to the guy who bought my GTV6 last year and I am now posting from memory. I think that the mark-space ratio of the square wave has got something to do with it.
There is a good description of the operation of these modules in "Automotive Electric/Electronic Systems" published by Bosch. ISBN 3-18-419110-9

 

[alfaparticle]

I was thinking about this on my morning bike ride today. The dwell angle range of the Bosch module is about 10 to 70 degrees. It is possible that it can shorten the dwell angle of the input pulse but not elongate it. My points simulator was a reed relay driven by a signal generator and I could control the frequency up to 100 Hz - the limit of the relay, which corresponds to 3000 rpm on a 4 cylinder engine but I had no control over the dwell angle which was most likely less than 70 degrees. Who knows what is the effective dwell angle of an inductive pulse? It is most likely pretty short.

 

[Graham Hilder]

Ah, thanks again Ed. Too bad about that folder - it must be a gold-mine of info! And that Bosch book - I'd love to see that too, but have no idea where I would find it.

In the absence of those, I have found an old photocopy that I was given decades ago, of three pages from some Bosch handbook, with a cct diag and detailed description of a Bosch TCI module (no type number given). This type has an unbalanced reluctor input, and controls dwell by an internal RC timing cct that's influenced by the input pulse duration, i.e. by the RPM. It could be quite an early system - it doesn't have any coil current monitoring; it's just a simple open-loop control.

And on the net I found some technical article on ign systems in general, and this describes two dwell-control methods. One is just like the above. The other is for a balanced reluctor input and this one controls dwell by developing an RPM-based DC voltage from some property of the input waveform (maybe from its pulse-width, but I don't know); anyway this varying DC is applied to one leg of the reluctor input-pair to offset that input signal so its waveform crosses the module's input threshold at a different point of the cycle, to vary the start of dwell. This same strategy is used by the Motorola reluctor ignition IC that was popular for electronics magazines' ignition projects in the 1980s and 90s. I wouldn't be surprised if the Bosch -111 uses this strategy in its Alfa boxer reluctor applications. But I can't know for sure.

In any case, it's still a mystery how this single 111 type can work so well in both the balanced-input reluctor application and the unbal square wave GTV6 app. But there seems little doubt that it does!

Best regards,
Graham H

 

[alfaparticle]

Too bad about that folder - it must be a gold-mine of info!

The guy who got it has been banned from the ABB so we are unlikely to see any of it posted.

 

[Graham Hilder]

That's too bad - looks like we can kiss that info goodbye (apart from your memory, perhaps!).
Rgds, GH

 

[Graham Hilder]

G'day, it's me again! I've started some bench-tests on my -111 module, the units that almost all Alfa boxer cars use together with their reluctor dizzies. I'm wanting to see if I can make it work with a Hall Effect trigger as I plan to fit to my old 1750 GTV dizzy. In the factory GTV6 Hall application, I think for GTV6 the square-wave Hall-type signal was connected to pin 5 of the -111, with pin 6 (the second pin of the -111's balanced reluctor input) left unconnected. So I made that connection in my test setup. To begin with, I've simply toggled pin 5 by hand, connecting it alternately to +12 or 0v. Sure enough, this makes a spark, on the falling edge of the input wave. So then I tried connecting the signal to pin 6, with pin 5 unconnected, and oddly enough, that worked in just the same way.

So it does basically work.Next thing is to find some way of observing whether or not the dwell control function works, which might be difficult as I've only got a very primitive scope, one that can't handle inductive spike voltages on its input, and is only single-beam too.

Regards all,
GH

 

[Graham Hilder]

Back again. I've done some tests with my -111 module, using a square-wave multivibrator trigger, my basic scope, and a low-ohms ballast resistor to vaguely simulate a coil-primary load without causing inductive spikes to wreck my scope.
With the unbal square wave trigger input, I found the coil drive simply followed the waveform of the trigger. This applied over the whole RPM range of about 1,000-9,000 rpm on a 4cyl, and also with a couple of different Mark-space ratios.
This agrees with a lot of stuff I found on the 'net about using Bosch modules in various aftermarket ECU projects. Many of them indeed said that if you use a reluctor-type smart module with a Hall-type input, then the coil simply follows the trigger, and for dwell control to operate with Hall trigger you do have to use a Hall-type smart module. This seems reasonable, as Bosch presumably makes two different types of modules for a good reason!

Given the above conclusion, what I've been worried about is the contradiction that a single smart module, the -111 type, is used both by the Alfa GTV6 Hall-type cars and by all Alfa boxer reluctor-type cars. The important point was that Ed's post Nr 1 seemed to indicate that in the GTV6 it was triggered directly from the Hall pickup (and I know that it's also triggered directly from the reluctor in the Alfa boxers).
However, on re-reading Ed's later post, Nr 41, I see that in the GTV6 the Hall signal actually goes through “the ignition advance computer which applies phase shift according to rpm.....(and) then goes to pin 5 of the Bosch ignition module which triggers the coil.”
I suspect that if indeed there is any smart dwell-control in the GTV6 application, then it's done by the spark computer and not by the -111 module - does this seem reasonable do you think?

If that's the case, then the whole thing makes sense, in that yes, the -111 is a smart module when used directly with a reluctor balanced-input trigger, but is not smart when used directly from an unbal sq wave Hall-type trigger. And this situation would be consistent across my tests, Ed's tests, and the two Alfa applications, with no contradictions.

That's the conclusion I'm tending towards now. If so, I should go to the dismantling yards and find myself a Hall-type Bosch smart module for my 1750 GTV Hall-converted distributor. BIM-137 is often suggested, and this is/was used on many VW models I believe.

Best regards all,
Graham H

 

[alfaparticle]

I suspect that if indeed there is any smart dwell-control in the GTV6 application, then it's done by the spark computer and not by the -111 module - does this seem reasonable do you think?

I believe that is incorrect. the spark advance computer controls advance and has an rpm limit, nothing else. The behaviour of the Bosch module is described in the Bosch Book that I referenced.
When I ran Megasquirt ignition I first used the stock GTV6 Bosch module and I could not get the car to run properly above 6000 rpm. It turned out that both Megasquirt and the Bosch module were trying to control dwell and the result was a screwed up charge/spark cycle. I posted Megasquirt plots of this in a Megasquirt thread. I replaced the -111 module with a "dumb" module from a Porsche - the one recommended by the Megasquirt designers and which I had overlooked, and the car ran great to 7000 rpm.
I ran the _111 module on the bench with a square wave input and a coil and I monitored the coil primary current with a DC probe that was connected to the second input of the oscilloscope. I was able to measure the dwell angle of the current and determine that it varied between 10 degrees at low rpm to 70 degrees at high rpm. I think that I posted photographs of some of these current plots.

 

[Graham Hilder]

Cripes Ed, well, that puts the cat amongst the pigeons doesn't it !!! So then, your test was like mine, with unbalanced sq wave directly to one of the -111 module's two input-pins. The only difference was that your test used a real coil load and mine only a resistor. Yet we got different results. (I'm confident that I would have seen any change in dwell, even with my single-beam scope. Especially a change of the magnitude you mention.)

I wonder if the module does use coil current monitoring to control its dwell, and perhaps my load was too high a resistance so that the Bosch never reached its output current limit, and so never changed its dwell. I'll look into an amended test.

I wonder if you have any way of sending us a scan perhaps of the vital part of that Bosch Book that describes the behaviour of the Bosch -111 module?
I'd love to see that but I know of no way to obtain that book.

Also, if you could direct us to your Megasquirt postings on all this, with your plots, I'd love to read that too.

Well then, now we're still stuck with the contradiction that a single smart module, the -111 type, is used both by the Alfa GTV6 Hall-type cars and by all Alfa boxer reluctor-type cars, presumably works fully in both cases, even tho' one input's an unbal sq wave and other is a balanced sawtooth. The other contradiction is the matter of Bosch making separate modules for Hall and reluctor....

On Dtec.com.au there's a long detailed article on dwell-setting on ECU projects using Bosch modules. That writer reckons -
a. Do not use an inductive sensor triggered type with your ECU. Internally these modules use the slope of the reluctor sensor input signal in their dwell setting process and if triggered with a square wave it will not have dwell control available. Only use modules from Hall Effect systems (eg. BIM 137).
b. The Hall Effect triggered module type internally converts the square wave input signal (from your ECU in the case of an aftermarket project) into a saw tooth voltage and uses this new sloping signal in their dwell setting process.

All very confusing!
Best regards,
GH

 

[alfaparticle]

go to abebooks.com and search on 3184191109 in the ISBN field and you will find a number of copies of the Bosch book for sale

 

[Graham Hilder]

Rightio, thanks, will do! Best rgds, GH

 

[alfaparticle]

I did the testing in the second half of 2017 and I posted some details in post #309 of Installed Megasquirt today

I took some photos of scope traces but it was hard to get good ones so I took to sketching the waveforms along with pulse durations. Here is an example. You can see from my note at the bottom that beyond 5200 rpm the charge current was limited by the rise time and the spark time was 1.2 mSec. Also be aware that the time base is not the same for the different spark frequencies. I shortened it in order to see the pulse shape at higher rpm.

 

[alfaparticle]

Here is a photo of the current pulse at low rpm. Notice the long delay between the end of the charge and the start of the new one

 

[Graham Hilder]

Beauty, tks Ed! I'll visit those web pages and have a good look later today. But meanwhile, even just a quick glance at your images above do show what's happening.
I'll re-do my own test with a lower value load resistor, about 1 ohm, which should allow the Bosch to output enough current to require dwell restriction.
I haven't a current probe for my scope but monitoring the drive voltage to the coil -ve from the Bosch should show the same thing.
Best regards, GH

 

[alfaparticle]

I haven't a current probe for my scope but monitoring the drive voltage to the coil -ve from the Bosch should show the same thing.

The ramp rate is controlled as much by inductance as resistance. You need to measure the current to see what is going on.

 

[Graham Hilder]

Quite right, and if I had a current probe I certainly would, but I don't know that I need to see the current for my purposes......
All I want to do is to see some evidence that the module is applying some dwell control to the coil drive and not just following the sq wave trigger waveform. I suggest that if I observe the voltage waveform at the coil drive point (the Bosch output pin connected to load resistor in place of coil -ve), this'll tell me if anything's happening. Sure, if I do see that, I won't know just WHAT is happening in terms of coil current, but at least I'll know something is. It'll probably be the correct thing too, as I'd be using the module with its specified Bosch coil, in the actual car application itself,
That sound feasible do you think?
Regards again, GH

 

[alfaparticle]

Agree. The module limits at 7 amps so try to find a load that will let it get there - about 1.5 ohms.

 

[Graham Hilder]

Excellent, thanks Ed. I've a bunch of old ballast res in the parts box, which measure at 1.2, 1.8 and 5.5 ohms, so I'll play around with those.
Can't wait to try when I get a chance to play!
Regards, G