[PestaCX]

Hello CXers,



I am just registered now, however, having a CX650E for 10 years now.



I started a new project on my other bike, and that received the transistorized ignition system like of the CX's.



The only thing I need at the moment is some kind of schematics, relevant to the ignition electronics itself.

I would like to couple it to a PIC (programmable microcontroller), and need to know the circuit inside the transistorized unit.



Any usable help is highly appreciated.



Kind regards,

István

(from Hungary)

 

[DaveF]

I am not aware of any schematics for the transistorized ignition unit. The design is most likely proprietary and not released to the public.



Several weeks ago I took one apart. It has a ceramic substrate module with several semiconductors mounted directly on it, and a handful of passive chip components. If one knew what the semiconductor die components were the design would be fairly easy to reverse engineer.



The ceramic substrate is mounted to the inside front of the metal case, presumably for heat sinking.

 

[markmakeitso]

Hi Istvan,

I've been toying around with a project like this for awhile, mostly with Microchip PICs due to their free sample program.

I would guess that an engine position pickup is required, and you could use an internal timer to determine RPM. Another timer could control the delay before coil activation, and the same timer could be re-used to turn off the coil. Delay time would be based on RPM (decreasing with engine speed), while coil saturation time could be static or modified by RPM. If you have interrupts on your hardware this could be relatively simple.



Have heard, but not verified, that almost any modern transistor or MOSFET will open fast enough to cause the spark on the secondary side. Match current capacity, maximum voltage difference, and NPN and PNP and control with a digital output pin and you've got a good beginning.



Starting from scratch give you some interesting options, like multiple spark events during cranking, timing advance based on both constant and delta RPM. If it were me I'd like serial output (or possibly even CAN for OBDII compatibility) and some rudimentary troubleshooting capability during operation. Maybe an display output also, possibly a replacement for the traditional gauges using a character LCD for a couple bucks.



Okay, somewhere back there I gave you all I know right now and went on a tangent. Sorry, keep us posted, or even PM me if you're digging into it.

Good luck,

Mark

 

[DaveF]

I've also been kicking around the same basic idea for a couple of reasons.



There have been several instances of failed ignition timing pickup coils documented on the forum over the past few months. Replacing them involves dropping the engine or moving it forward enough to remove the timing pickup cover. Although this is not a hard job to do in the garage, it would be a real pain to deal with out in the middle of nowhere.



The pickup coil assembly actually has two coils, one for each cylinder. But only one is needed to provide a timing reference if the ignition system is designed appropriately. Why not design an ignition system that could use either pickup coil? It could sense if one failed and automatically use the other one in that case.



Another potential goal would be to automatically adjust the timing during starting to reduce/prevent the kickback that plagues the 650's. This should greatly reduce strain on the starter clutch.

 

[PestaCX]

Hello DaveF, markmakeitso,



I am really impressed for receiveing very meaningful answers such soon.

Thank you very much for this!



Please be aware that electronics is only my hobby (and I am quite novice), and playing with this ignition issue is going to be done on my MZ ETZ250. (This type is probably unknown for you, they were produced in former East-Germany.)

But I really like this bike, and think that I can made some improvement on its ignition system.



I already installed the NEC transistorized igniton on it (is was quite simple), keeping the old breaker-point, as a trigger for ignition.



The improvement was really tremendous, I have never experienced such steady idling before on my MZ. But the MZ's ignition-timing is fixed, no advance-regulation!



I would like to use the PIC to control the NEC ignition (timing), but for this I need some knowledge of this ignition unit (the Honda's NEC I mean). I am more or less familiar with the wiring of the unit (from outside), but still debating what the method of operation with the pickup-coil.



I measured the voltage betveen the wires which would connected on the CX pickup-coil, and found that there is 2,6V.

I suppose (but not sure yet), the pickup-coil gives a reverse-voltage on the circuit, which causes the voltage to drop (cannot measure, but maybe down to 0V?). Or: does it superpose additional voltage on the system?



But if I consider the mechanical breaker-point operation: if closed (that is, grounded, however, practically zero current!), voltage drops to zero, and if breaker opens, voltage gets back to 2,6V for a short time, and ignition happens.



The things you mentioned are also interesting for me, like preventing kick-back while kick-start or cranking (yes, my CX650E also suffers that issue! :-( ).



But first (as being a novice) I need to know, how to wire the PIC to the ignition. I already scrapped 3 PICs, probably killed by high-voltage. If you have any hints on this, also welomed.



I think the basic-software is already made (tested it in MPLAB's simulator), but the hardware is still not ready.



If I would succeed in this issue, definitely would like to go on for improving it with additional functions, and probably would make some development on the CX also.

But nowadays wife is worrying for me, and sometimes angry for being awake till late night, playing around with my stupid idea.



Sorry for the long post, I hope you did not loose your interest.



Better wheather!



István

 

[PestaCX]

I am not aware of any schematics for the transistorized ignition unit. The design is most likely proprietary and not released to the public.



Several weeks ago I took one apart. It has a ceramic substrate module with several semiconductors mounted directly on it, and a handful of passive chip components. If one knew what the semiconductor die components were the design would be fairly easy to reverse engineer.



The ceramic substrate is mounted to the inside front of the metal case, presumably for heat sinking.

Dave,



didn't you make any pics of the inner part of the NEC unit?



Maybe interesting to see...



István

 

[marshallf3]

I already scrapped 3 PICs, probably killed by high-voltage. If you have any hints on this, also welomed.

Could be a number of things but I'd look for the most obvious first:



1) Is the PIC driving anything inductive that could induce a reverse EMF when the current stops flowing? If so, a snubber diode is mandatory across the offending device. If it's an AC signal then you'll have to use two zeners back to back to form it.



2) Are there any input signals coming into the PIC that might exceed their input voltage capability or possibly generate a negative spike? If so, snubber time again.



3) Is the PIC being called upon to drive something that takes more current than that particular output pin is able to source or sink? If so, a logic-level MOSFET would be in order as a buffer.



Simple stuff I'm sure you already know, but it never hurts to double check for situations like this.

 

[MotorcycleCommuter]

There are several people using General Motors HEI ignition modules on motorcycles in place of the NEC boxes. Here's an example that would work for the CX/GL TI engines: http://home.comcast.net/~loudgpz/GPZweb/Ignition/CBheiMod.html

 

[marshallf3]

Our big jump in advance doesn't seem to occur until around 5,000 rpm?

 

[DaveF]

Here are a couple of photos of the one I took apart. Since I had no idea what was inside there was some damage done to the substrate as it was pried apart.



The ceramic substrate is bonded to the front cover section shown.

The bulk of the device is simply potting compound in which the wires are secured.



The circuit does not look very complicated. Note the power transistor in the lower left corner - it's probably an NPN silicon, with the collector bonded to the substrate. And there are chip capacitors and screened resistors that can be measured. The biggest question is what is the IC in the middle? [Edit: It's probably similar to the MC3334.]

[Edit again:



According to Don in Oz, this is the schematic for the TI modules, either the Oki MPS200 or the NEC 5194.

It is my understanding the diagram was taken from a GL1100 manual, hence the two plugs external to the module.



Given this new information, the IC at the center of the module is a transistor array.]

 

[PestaCX]

Our big jump in advance doesn't seem to occur until around 5,000 rpm?

I am not aware of 500s, but as far as I remember from the Honda workshop manual, in case of 650s from 3000(!)rpm full advance is applied (which is close to 45 degrees, or something like that).



Unfortunately, cranking is a bad situation. Really most of the CX650s (of not all) suffering it, having difficulties (better) with starting, or resulting in mechanical damage (worst case) after a longer time.



Too pity for them.

 

[PestaCX]

Here are a couple of photos of the one I took apart. Since I had no idea what was inside there was some damage done to the substrate as it was pried apart.



The ceramic substrate is bonded to the front cover section shown.

The bulk of the device is simply potting compound in which the wires are secured.



The circuit does not look very complicated. Note the power transistor in the lower left corner - it's probably an NPN silicon, with the collector bonded to the substrate. And there are chip capacitors and screened resistors that can be measured. The biggest question is what is the IC in the middle? [Edit: It's probably similar to the MC3334.]

Dave,



thank you very much for the pictures, however, I expected some more old-fashioned components (bigger ones).

It seems that Japanese producers were already very improved that time too...



To be honest, these parts do not tell me anything... :-(



(But will not give up! ).

 

[PestaCX]

Could be a number of things but I'd look for the most obvious first:



1) Is the PIC driving anything inductive that could induce a reverse EMF when the current stops flowing? If so, a snubber diode is mandatory across the offending device. If it's an AC signal then you'll have to use two zeners back to back to form it.



2) Are there any input signals coming into the PIC that might exceed their input voltage capability or possibly generate a negative spike? If so, snubber time again.



3) Is the PIC being called upon to drive something that takes more current than that particular output pin is able to source or sink? If so, a logic-level MOSFET would be in order as a buffer.



Simple stuff I'm sure you already know, but it never hurts to double check for situations like this.

Thanks a lot for the hints!



I am considering to install protective diodes to everywhere, and also opto-couplers.



After 3 PICs dead, I must be more careful, as ordering new ones take time (any money too...).

 

[markmakeitso]

Hi Istvan,

Like I mentioned, Microchip has a pretty generous sample program, and they list shipping to Europe. You might want to check whether you can get free ones shipped your way. I've been getting 6 or 8 a month for awhile and they haven't charged me anything yet.



http://www.microchip.com/samples/



Hopefully that links, otherwise it's not too hard to find.



Good luck

Mark

 

[PestaCX]

Hi Istvan,

Like I mentioned, Microchip has a pretty generous sample program, and they list shipping to Europe. You might want to check whether you can get free ones shipped your way. I've been getting 6 or 8 a month for awhile and they haven't charged me anything yet.



http://www.microchip.com/samples/



Hopefully that links, otherwise it's not too hard to find.



Good luck

Mark

Hi Mark,



thank you for the link. I tried, and found that Microchip doesn't supply samples to Hungary... :-(

However, I must be able to get samples from Austria, there is availability of samples.



On the other hand, we do not need to be afraid as PICs are also well-available here, the only concern is that a genereal order delivery-time is a few days.

Itself it is not a problem, but these days I feel some disappointment after facing with the problem of interfacing the PIC to the transistorized ignition.



I am really considering to build my own transistorized ignition, which I am aware of how to connect/drive with the PIC.



I feel a little bit tired, so need to have some tome to regain devotion.



István

 

[msleboda]

Hi Istvan,



I was surfing the net last night and found this fantastic information site. It shows the CDI ignition schematic, and I remembered this post. Here is the site:



http://drop.io/CX500_GL500#



If this link doesn't take you directly, then you want page 2 for the schematic files.



Hope it is what you were looking for.

 

[PestaCX]

Hi Istvan,



I was surfing the net last night and found this fantastic information site. It shows the CDI ignition schematic, and I remembered this post. Here is the site:



http://drop.io/CX500_GL500#



If this link doesn't take you directly, then you want page 2 for the schematic files.



Hope it is what you were looking for.

Thank you, msleboda,



this site is really wonderfull, full of information.

However, the CDI and transistorized ignition are quite different.

I do not know why Honda made such modifications within the same family of bikes, but one thing is for sure: I never heard any TI units to be broken, meanwhile CDIs are going to be dead after a long period of time.

As I read, CDI has its advantage in case of higher RPMs, but TI also can fulfill the task in case of the max. speed of CXs (like 10 000 rpm for the 500, and 9 000 rpm for the 650s).



Once again, I must say thank you for the link, though that is not I did search for, but really-really useful.



István

 

[markmakeitso]

I should have asked earlier, what PIC are you using Istvan? I assume you have the free MPLAB IDE, are you using C or do you know assembly? I've been using CC5X and CC8X compilers and really like them, especially for a free program. In my opinion life is too short to use assembly, but some (read: many) are more dedicated and talented than myself.



I also use a MPLAB ICD2 for programming, but had access to a PICStart Plus earlier, that one was really handy. I'd like to pick up one of the cheap ICSP programmers, but don't know how well they work with MPLAB.

 

[PestaCX]

I should have asked earlier, what PIC are you using Istvan? I assume you have the free MPLAB IDE, are you using C or do you know assembly? I've been using CC5X and CC8X compilers and really like them, especially for a free program. In my opinion life is too short to use assembly, but some (read: many) are more dedicated and talented than myself.



I also use a MPLAB ICD2 for programming, but had access to a PICStart Plus earlier, that one was really handy. I'd like to pick up one of the cheap ICSP programmers, but don't know how well they work with MPLAB.

Hi Mark,



I tried (and killed... :-s): 12F508, 12F509, 12F629, 16F819 (just because they were available for me, you know, if a project is really exciting, nothing is enough to stop me in creation...).

Now I have another 2pcs of 12F629, and make some "dry-test", before installing it to a real circuit connected to the bike.



Of course, I use MPLAB IDE, and use assembly. A long time ago I was interested to program C16 in assembly, so have some feeling about it. And, assembly is the language which is preferred in Microchip's manuals too (I mean the datasheets of PICs).

The code I wrote is not big (because being simple), after coding it is 196 words only, and just enough for adjusting the ignition timing from below 1000 to over 7000 rpms in 1000rpm steps (that is, 8 possible setpoints).



Other features would be nice too, but that definitely would require a much sophisticated program, and I am not a professional, though...



But, you are definitely right: if any of yours would know, how many hours (mean: nights) I spent on this project, probably would call me insane.



However, I am still eager to make this modification on my MZ ETZ250, and -if succesfull-, would make some similar on the CX too.



For programming the hardware I have a Velleman kit, (8048, if I remember correctly), and a JDM2 programmer also.

In fact I am a chemical engineer, but definitely interested in electronics, and this is a way for me to learn...

 

[markmakeitso]

I've heard of the JDM family of serial programmers before, is yours like these?



http://shop.ebay.com/i.html?_nkw=jd...grammer&_osacat=0&_trksid=p3286.c0.m270.l1313



Is it selectable under Programmer in MPLAB? The ZIF socket is very nice, much faster than me seating and unseating a DIP in a breadboard, and also needing to match up pins for ICSP.



If you have mastered, or at least achieved functional competence in assembly, I salute you. Although, to be honest, I would not be interested in switching PICs if all my code was assembly. You could probably learn CC5X or Hi-Tech very quickly and reduce development time, not to mention re-use more code if you switch part families. Just my knee-jerk reaction, but bear in mind that you can always jump into direct assembly language from within most C compilers.