Honda CX500 Motorcycle CDI ignition unit

[roadrash]

Originally Posted by redwire

The tricky part will be the thermistors. You would have measure their resistance at a few different temperatures to find something that matches. As the SCR's heat up, their sensitivity goes up and the thermistor lowers how much gate drive they get.
I did see a remake of the box called Finnbox G47, the thermistors are on the SCR's tabs.

I have seen the finnbox and others like the Ignitech one too. The finn box is no longer available as far as I can find out as their website is now dead. The Ignitech is very expensive in the hundreds of British pounds.
The thermistors you mentioned, are they maybe what those resistors wired to each other are at each end of the original unit
they just look like a standard pair of resistors wire wound together with copper wire are sit high up at each end of the PCB.

[roadrash]

Originally Posted by stj

maybe re-designing it to use mosfets is the answer.
or darlingtons

Gone over my motor-mechanics head stj. Quite happy to experiment with this old unit now we have access to it as I still have my working unit still fitted to the bike.
Apparently there are extra winding's in the stator of the bike that generate 12v to power the unit but in a aftermarket cdi unit called a "Ignitech" (click here for info) they dont use the 12v supply from the stator and instead just take the 12v supply from a switched 12v output from the battery. By doing this it means if those 12v winding's in the stator pack up it doesn't matter anymore. I don't see why this original unit cant be modified to take its 12v supply from the battery too as a sensible modification. We can try all these improvements once I get it working again.

[roadrash]

Originally Posted by SMDFlea

(off-topic) The old CX500 flying maggot , a friend of mine had one of those

yes thats it except we called it the plastic maggot after the little plastic cowl over the headlamp. Most bikes in that era were what we call naked and never had plastic cowls or fairings so hence "plastic"

[redwire]

Originally Posted by roadrash

I have seen the finnbox and others like the Ignitech one too. The finn box is no longer available as far as I can find out as their website is now dead. The Ignitech is very expensive in the hundreds of British pounds.
The thermistors you mentioned, are they maybe what those resistors wired to each other are at each end of the original unit
they just look like a standard pair of resistors wire wound together with copper wire are sit high up at each end of the PCB.

Those are just calibration resistors to precisely set the circuit for the two particular SCR's sensitivity. One is 330R, the other 470R+560R parallel = 255R (SCR gate-drive).
The thermistors would be in the tall black towers, it looks like heatshrink tubing or something with parts inside. They are likely 470-500R at 25°C.

I started drawing the schematic while a friend was boring me to death on the phone with their life's drama. It's a way to reverse engineer these modules, mirror the board and then doodle in the parts you know. I did not finish it because I don't know many of the parts, and well, people have already done it.
The other pics are more of the Finnbox, should be able to clone it. Finnbox builder's pics: https://www.flickr.com/photos/lotjak...7632938667888/
The TIA02-14 part number I'm not sure, usually the SCR's are special CDI parts which can take a massive current surge, like S6992 or S6A13.

Some guy's CX500 repair photos https://www.flickr.com/photos/7849153@N05/

P.S. your bike looks really nice

[roadrash]

Originally Posted by redwire

Those are just calibration resistors to precisely set the circuit for the two particular SCR's sensitivity. One is 330R, the other 470R+560R parallel = 255R (SCR gate-drive).
The thermistors would be in the tall black towers, it looks like heatshrink tubing or something with parts inside. They are likely 470-500R at 25°C.

I started drawing the schematic while a friend was boring me to death on the phone with their life's drama. It's a way to reverse engineer these modules, mirror the board and then doodle in the parts you know. I did not finish it because I don't know many of the parts, and well, people have already done it.
The other pics are more of the Finnbox, should be able to clone it. Finnbox builder's pics: https://www.flickr.com/photos/lotjak...7632938667888/
The TIA02-14 part number I'm not sure, usually the SCR's are special CDI parts which can take a massive current surge, like S6992 or S6A13.

Some guy's CX500 repair photos https://www.flickr.com/photos/7849153@N05/

P.S. your bike looks really nice

That's some very interesting info redwire. I did see a supposed circuit diagram online but people who tried to make it say it doesn't work.
I recently had to send a Kawasaki ER5 CDI box to a company in the Netherlands called Carno who did manage to repair it at a cost nearing 200 UK pounds.
Do you want me to remove those thermistor stacks and unwrap them to see what's inside? I am removing some parts carefully and cleaning up the pcb of that resin as I go and resoldering all other parts. Hopefully when finished I should have a working CDI box and all the info that can be used to repair or recreate clones of the original. The pcb looks pretty straight forward and not overly complex. As long as there are not any custom components it should be possible.
Thanks for the remarks about my bike. It's not as nice as it could be but if you saw the condition it was in when it was dragged out of a leaky old shed where it sat for over 30 years you would have thought it was scrap only and for that reason I never did a full restore as I was not sure it was possible to make it usable again with even the frame rotted out in some places. Now it's done it could be stripped again and a full as new restore done. But like it is now it's used a lot and not just for show.

[roadrash]

This capacitor C7 (see pic) is showing up as a resistor on my component tester with a value of .55 ohms. Is that normal? unfortunately it has no value marked on it, just what looked like a "N" symbol. Ive also just removed what my tester says is a thyristor.
also another up tp date picture of pcb now cleaned up a lot better so that component markings can now be seen.

[stj]

first pic is a mylar film capacitor!!

[redwire]

That film cap C7 is shorted. It should be 0.047uF (473) 47nF low voltage 100V, it filters the Left TDC pulse with R15 330R. It doesn't see high stress so I'm not sure why it died edit: the module might have been wired up wrongly to the HV coil and that killed it. C7 is across SCR P anode-cathode. The larger black film caps are 47nF 400V.

There's only two real mysteries in the CX500 CDI module.
The thermistor's value and temperature curve. If you measure the ohms at room temp and maybe the fridge we can find the part. The black towers are weird, I'm not sure what's in them. One thermistor connects to R13, other to R14, if that is the bead in the black heatshrink. I don't think scraping them out will tell us anything. Maybe the Finnbox guy is willing to share what he used?
The other mystery is what the factor calibration resistors are about, Finnbox guy had a prototype with trimpots and then he got rid of them.
I understand the internals of the module so no prob to make one if they are not available at a fair price.

This bike seems very advanced for it's time - V-twin, water cooled and Honda does go crazy with complexity. I think there's six crank pickup coils lol.

[roadrash]

Originally Posted by redwire

That film cap C7 is shorted. It should be 0.047uF (473) 47nF low voltage 100V, it filters the Left TDC pulse with R15 330R. It doesn't see high stress so I'm not sure why it died edit: the module might have been wired up wrongly to the HV coil and that killed it. C7 is across SCR P anode-cathode. The larger black film caps are 47nF 400V.

There's only two real mysteries in the CX500 CDI module.
The thermistor's value and temperature curve. If you measure the ohms at room temp and maybe the fridge we can find the part. The black towers are weird, I'm not sure what's in them. One thermistor connects to R13, other to R14, if that is the bead in the black heatshrink. I don't think scraping them out will tell us anything. Maybe the Finnbox guy is willing to share what he used?
The other mystery is what the factor calibration resistors are about, Finnbox guy had a prototype with trimpots and then he got rid of them.
I understand the internals of the module so no prob to make one if they are not available at a fair price.

This bike seems very advanced for it's time - V-twin, water cooled and Honda does go crazy with complexity. I think there's six crank pickup coils lol.

Thats incredibly useful info Redwire. I will get those measurements done so we can identify what these unusual components are. Re the cap C7, I couldn't have wired anything wrong at any time as its all connected by the original connectors so it was probably another component failure that caused a over voltage. I did some more work yesterday and when testing one of the trannies my tester said it was not recognised or faulty. It looks the same as the other green TO220 one I removed earlier. Do you happen to know what type these are as this one is going to need replaced.

Just wondered if you had seen this article about the CX500 CDI as its has some useful info too Here and and here

[roadrash]

Ive looked at the tower thingy's and they are not tunable pots of any kind, just resistors inside a plastic tube filled up with that same rubbery resin. Ive now finished checking everything on the PCB and most is ok with the exception of one transistor and the cap you mentioned Redwire. What kind of caps are these?

[redwire]

Oh, the heatshrink just has resistors inside. Now I see the nearby two green thermistors TH1, TH2 but I figured out their values, they should be 500 ohms at room temp. and around 1.2k in the fridge.

I have to look at the SCR's because they come in maybe 5 different sensitivities, i.e. IGT 6mA, 10mA, 15mA, 20mA, 30mA and the old green Toshibas I have to dig for. Later I can give a parts list, I have them all figured out.

I would check the 14 diodes (!) the board has too.
I can't see a way to zap the capacitor C7 so either it got stabbed scraping the potting compound off, but maybe it did fail short. The Left ign timing coil connects to it and is supposed to be low voltage.

The cap pictured is a 47nF 400V film cap, best is polypropylene like Epcos B32621A4473K000 or Panasonic ECW-F4473HL. I'm not sure who your parts disti's are in the UK.

[stj]

howdoes it work in general, i'm not a bike person.
is it fed a fixed dc voltage?
does it get a clean trigger pulse to trip the output??

i have a schematic for a german kit for CDI unit to add to a car - maybe it could be doubled and modded for higher rpm limit.

[roadrash]

Originally Posted by stj

howdoes it work in general, i'm not a bike person.
is it fed a fixed dc voltage?
does it get a clean trigger pulse to trip the output??

i have a schematic for a german kit for CDI unit to add to a car - maybe it could be doubled and modded for higher rpm limit.

Not a expert on this but I do know the stator supplies a isolated 12v just to power the CDI which is not powered from the ignition voltage as you would think so the bike can be started with a totally flat battery if you bump start it in gear.
I did read somewhere there are two outputs so the cdi, one does low speed and he other high speed.
You would need to Google the Honda CX500 ignition to get all the details. I am still using a 40+ year old unit now and I have a feeling it's on its way out to because there are occasional pops and bangs out of the exhausts that I cannot cure and it frightens old ladies and horses too.

[redwire]

This Honda CX500 CDI module gets high voltage from the (G47) stator winding "source coils", to charge the 2.2uF caps up to around 300VDC. Other CDI's use an inverter powered off the 12V to make the high voltage DC (i.e. IgniTech), but this factory ignition module is entirely powered off the stator coil pulses.

For each cylinder, that SCR is triggered ON by one of two pickup coils depending on engine speed. Either Idle-pulse (low-speed) or Advance Pulse (high-speed) coil, and there is a third coil Neg Pulse that seems to mute the SCR from firing. Suzuki uses a system where the trigger coil's signals are summed up to make the advance curve (pic related). I'm not 100% sure how advance is done in the CX500.

It's a complicated setup, and there seems to be no camshaft sensor so the engine has odd timing where one cylinder fires shortly after the other?
The small SCR seems to prevent cross firing, so when the Right is firing the Left ignition SCR is muted?
One guy thought this little SCR was instead making the timing advance, but it's (small SCR) only (muting) the Left cylinder's SCR from triggering and making a spark.

It's possible to clone the module, just have to lay out a PC board. New parts are stronger and more reliable. IgniTech seems way too expensive, and shipping from Czech on top of that. Ouch.

I don't own the bike or know it, I can just read the electronics from pictures.

TIA02-14, Honda part number: 30400-415-005, 30400-415-015.

[stj]

damn - typical japanese crap.
doing complex stuff the hard way.

if it was mine i'd run the primary supply through a schotky rectifier into a cap bank, then switch it with a pair of fets.
i would control it with a microcontroller that times the pulse-train and then divides it by 2 to get the split and then add advance-delay based on rpm.

it would of course be better if it had a vac sensor for that so you can create a timing table/curve.
i wonder if it would missfire if you went wasted-spark and just fired both cylinders together??

[roadrash]

Originally Posted by stj

damn - typical japanese crap.
doing complex stuff the hard way.

if it was mine i'd run the primary supply through a schotky rectifier into a cap bank, then switch it with a pair of fets.
i would control it with a microcontroller that times the pulse-train and then divides it by 2 to get the split and then add advance-delay based on rpm.

it would of course be better if it had a vac sensor for that so you can create a timing table/curve.
i wonder if it would missfire if you went wasted-spark and just fired both cylinders together??

There is a repalcement CDI that uses thge ignition/battery voltage to power the CDI called a Igntec see here but its quite expensive.

[roadrash]

Originally Posted by redwire

Oh, the heatshrink just has resistors inside. Now I see the nearby two green thermistors TH1, TH2 but I figured out their values, they should be 500 ohms at room temp. and around 1.2k in the fridge.

I have to look at the SCR's because they come in maybe 5 different sensitivities, i.e. IGT 6mA, 10mA, 15mA, 20mA, 30mA and the old green Toshibas I have to dig for. Later I can give a parts list, I have them all figured out.

I would check the 14 diodes (!) the board has too.
I can't see a way to zap the capacitor C7 so either it got stabbed scraping the potting compound off, but maybe it did fail short. The Left ign timing coil connects to it and is supposed to be low voltage.

The cap pictured is a 47nF 400V film cap, best is polypropylene like Epcos B32621A4473K000 or Panasonic ECW-F4473HL. I'm not sure who your parts disti's are in the UK.

I have now finished cleaning and checking all the components and all diodes test as ok.
I think it may be a good idea too replace all the caps unless you think the tested ok film type caps are still good enough to use. A lot of the caps have no markings or they are faded. Are all the large film caps the same 47nf film type. Then its just the 2 other types used like the little rounds green ones marked with a value of 351 and the smaller oblong green caps that just have what seems to be a "N" symbol on them. Then just need to know what type these thermistors are? I can get components from all the usual sources here in the UK like RS, Mouser, Digkey, Farnells etc and of sometimes ebay.

[redwire]

These are the parts I would suggest as replacements:
Capacitors
2.2uF 400V PP Panasonic ECW-F4225 LS=25mm bend them+glue, all others almost 1" height w/short leads; RS has long lead Kemet R75MR42204040J. Or CBB22 225J400V.
47nF 400V PP Panasonic ECW-F4473 LS=10mm
47nF 250V mylar MKT Panasonic ECQ-E2473, Epcos B32529C3473, Kemet R82IC2470 LS=5mm

Diodes
Black - not sure what original part is.
I would use 1A 1000V 1N4007, FR107
White - original is either V08G 600V 1.1A controlled avalanche 1600V, V06E 400V 1.1A Hitachi 3usec
I would use 2A 1000V fast-recovery: MUR2100, HER208G, BYV38 etc. (replaces white ones)

SCR (two with tab)
TYN608RG 8A 600V 15mA, TN5015H 50A 600V but cheap, BT151-500 12A 500V, 6015LTP

SCR (one with no tab) - have to think about it.

The black bead part "351" is a thermistor. Can you measure the resistance, I thought around 500R at room temp. Bead thermistors 470-500R@25C Beta ~3450, not critical. Like MF52-501, TDK Epcos B57164K0471J000 (3560K), Vishay NTCLE100E3471JB0 (3977K).

If you want, it is possible to remake the PC board or maybe we should track down the Finnbox guy Andy in Finland? He has a few blogs and I just fired off an email to him.

[stj]

about the fast diodes - UF5408

[roadrash]

Originally Posted by stj

about the fast diodes - UF5408

Thanks a lot thats great! I have now finished testing all components that were on the pcb.

All Diodes test as OK including the white ones.
All resistors are the correct value.
All the original Caps except one were ok.
one SCR faulty.

is this really a thermistor?? It looks like a green cermic cap and it measures 347.4 ohms

I was also wondering what is the purpose of this little coller that you see on some resistors and diodes like the one on the resistor on the right in the photo and is it important?

Do you still want me to measure the one working SCR i have at room temp and in the fridge?