Paging the old ignition gurus (long).

BIL is trying some magneto - CDI ignition / 12V lighting on an old
Lambretta scooter he's restoring.

Parts were supplied in a kit form a couple of years ago and he's just
got round to fitting them (new crank, stator, flywheel, reg and CDI
unit).

A bit like this: http://preview.tinyurl.com/lac4pj

He's got no spark, bought another CDI unit but still no spark.

I've been though the wiring with him and *I* think the supplied
diagram is wrong.

On the stator there are 4 heavy duty coils, wired in series and
earthed at one end. This I think is the lighting / battery charging
cct. It reads single digit ohms and produces some AC when the flywheel
is turned by hand (brown wire).

Another coil appears to have both ends brought out in two wires (green
& white). This I think is the exciter coil for the CDI, reads about
400 ohms and produces quite a voltage when the flywheel is turned by
hand.

The last coil (yellow wire/ 90 ohms) doesn't produce a measurable
output but it's resistance reading changes when the flywheel is
rotated (suggesting it's doing *something* and not open cct / short
etc).

FWIW The CDI unit is possibly similar to one used on a Ducati (I've
seen mention) and is a combined CDI unit and HT coil. It has 4 spade
terminals (marked white, red, green and green), a captive earth wire
and ht lead output.

I think the white wire from the stator goes on the white terminal of
the CDI (earth side of the exciter coil). The yellow wire goes on the
red terminal (trigger) and the green wire on one of the green
terminals (other side of exciter coil and the wire to the kill /
ignition switch (also 'a kill' function)).

So, assuming all the coils, wires, the flywheel and spark plug are all
ok, does anyone know how you could test the CDI unit on the bench
please? If I had a schematic of the inside of the CDI I could probably
work something out myself but can't find anything as yet. The CDI
could be substituted again (assuming he's killed the first two by
inappropriate connection) or test these on another scooter but he only
has one such system at the moment.

So, if we knew what the charge - input voltage is to the CDI unit,
could I use something like a mains transformer to simulate the exciter
coil (240V > 100 acv or whatever) and then a suitable battery (1.5v?)
to reproduce the trigger input pulse (assuming the input device is a
generator coil not a hall effect switch etc. This 'trigger coil' is
activated by two iron pole pieces extending from two of the flywheel
magnets).

Cheers, T i m

 

If he has access to a known good CDI unit, simply cross-check terminal
to terminal for resistance values and make an X-Y chart, then do the
same for the suspect one. It won't tell you what the fault is, but it
will show you the dud one isn't good.

Btw, that coil that apparently doesn't do owt, I'd suspect that's the
trigger for the CDI.

 

Did the bike originally have CDI? I couldn't figure out from your
description.

 

He doesn't have a known good one but could do the same with the two
(now) unknown units he's got. At least to see if there is any symmetry
(it could even show up easier against a good one as / when he gets
one).

That's what I think. On the few CDI unit's I have found the schematics
for it seems the trigger input could be the gate of an SCR so it's
quite possible we wouldn't see signal much on a shite old analogue
multimeter (it showed the output of the other coils easily). I nearly
took the scope round there but CBA to dig it out.

I was just thinking it might be nice to have a simple bench go / no-go
type test jig.

I have asked elsewhere about what would be the easiest way to test the
rotor / stator and a I think the pillar drill could be the simplest
solution. However I think in this case those bits are ok.

Cheers, T i m

 

I think the Lambrettas made in India started getting 12V lighting /
electronic ignition so Co's started importing that stuff in kit form
as retrofits for 6V / points machines.

I might well do that now the seller is nearly back from his hols. ;-)

T i m

 

It happens that T i m formulated :

That will be the trigger circuit to fire the CDI.

To see the output from the trigger you would need an oscilloscope, it
will be much to fast for your meter.

 

Result. ;-)

Right and what I thought (thanks again).

Sooo, assuming the trigger coil is doing what it aught (as it has
resistance and seems alive), how could I make up a non mechanical jig
to test the CDI unit on the bench (I was talking elsewhere of putting
the flywheel in a bench drill and holding the stator on the drill
table etc).

I was thinking that if the exciter inputs of the CDI unit were looking
for (say) 100V AC [1] then that could be replicated via suitable mains
transformer (with 50 Hz being a real-world equiv of 750 rpm [2]). Then
I would just need to rig up a means of triggering the SCR (or whatever
is in there), HT lead to spark-gap and we should have sparks aplenty?

T i m

[1] Could be measured on a working bike with the same setup. ;-(

[2] I think because ... 50 (Hz) x 60 (sec/min) = 3000 (Hz). In the
rotor there are 4 magnets / rev so = 3000 / 4 = 750 equiv 'rpm'?

 

T i m formulated the question :

As you suggest - Measure the voltage output from a working bike and
find a transformer which will recreate it. I'm not at all familiar with
how the trigger pulse works or how you would recreate it. Is there a
separate magnet for the trigger? I would guess there must be, or there
would be four trigger points per revolution.

 

Indeed. I *think* the trigger coil is set at 90 degrees to all the
other coils so not directly affected by the rotor magnets. However,
between one pair of magnets there are two soft iron fingers that reach
from adjacent magnets, creating a magnetic field, parallel to the gap
between said two magnets (so also at 90 degrees to them and now
parallel to the trigger coil).

Hence this coil only activates once per rev.

I should have a couple of CDI / HT units on their way, one to replace
his (potentially) two faulty (killed?) units and one for me to play
with.

Cheers, T i m

 

T i m was thinking very hard :

I would hazard a guess that the range of what is needed as a trigger
pulse would have quite large margins. Probably something as simple as a
12v battery with a 100uF cap in series, then fed to the trigger input
would be enough to cause it to fire, when connected briefly. Add a 10k
resistor across the cap to discharge it and it would then be able to
repeatedly fire it. Maybe a home made impulse signal generator would
provide a means of firing it at regular intervals -All very much guess
work, but a starting point.

Getting back to the heavy duty coils....

Are we assuming these feed the ignition unit directly, or that it is
fed from the 12v rectified DC? If the latter obviously you cannot
simply feed this from a transformer.

 

I may have a similar trigger coil from an old stator I could try when
I get the spare CDI unit.

Indeed. I believe the *heavy* heavy duty coils feed the rectifier >
battery / lights etc

I think I understand the other single coil feeds the capacitor
(probably through a single diode or bridge). This was the one that
produced the most voltage from my hand rotation of the flywheel. This
would be the cct I would simulate with a transformer.

Cheers, T i m