6 pin CDI problem under temp.

[HKZ Bob]

Dear All,

I have a problem with my 6 pin CDI. At room temp it works fine.
In the car it failes when the car gets hot.

I took the box on a test bench and saw with the scope that T3 was not swichting. I replaced it. The Box was not working. The T3 Base Voltage of the Schmitt Trigger was low . To check my theorie I repalced R10 by an 6.8K resistor. The box was working fine.

The next step was to check the box under temperature. I put a thermocouple in the box closed it and at about 60 degrees celsius the box was faling.

With an open box it was not able to reproduce the failure but I saw under temperatur that the resistor R7 was going down to 450 Ohm. So I replaced that one as well.

So I ll ended up to replace all the parts which are circled red.

At the bench at 60 C the failure did not occured again. I had it run for 30 minutes no problem but in the car after an hour the failure occured again.

But the was still a problem left which I thing has to do with the general problem of this box.

If I keep the resistor R10 at 12K at room Temperature the box won´t work.
With a 10K resistor it works. My Zener Diode is supplying the Schmitt Trigger with 6.5 til 6.8 Volt under load.
What failure could it be that causes this problem that with an 12K R10 resistor the CDI won´t work.
I checked R1 R2 & R3. What could be the driver that the Schmitt Trigger won´t work with 12K.




Thank´s Folks I apresciate your help. This Forum makes the tuff winter time enjoyable.

[HKZ Bob]

Is it nornal that the level of the voltage Zener 6.8 volt various from 6.5 V to 6.8V.
The 22uf storage capacior should make it constant ?

Bob

[jimmcc]

Hate to ask the most obvious question first but when it's not working is the unit producing the normal squeal or is it silent?

I would be more suspect of a solid state device rather than a passive component.

[HKZ Bob]

It makes noise and works at room temp. but I wonder why my schmitt trigger won't work with an 12K R10 resistor. and why at hot this box failes.

[HKZ Bob]

Do you mean if the if the schmitt trigger is get locked up C1 won't make noise ? Or could it be that I am not able under temp. to pull the Base Voltage of T3 down below 0,7V.

I already replaced all parts on the schitt trigger except the 3.3K.

[jimmcc]

Do you mean if the if the schmitt trigger is get locked up C1 won't make noise ?


No. The left side of the schematic (T1, C1 ...) is a free running oscillator driving Tr to produce the 460 volts that charges C8. It should be running and you should hear the squeal (as a warning to the HV present) anytime the CDI is powered.

I would suspect T3 is bad and intermittent with heat. Take a look at the data sheet


http://www.datasheet4u.com/download.php?id=127735

and you will note that it is a very high gain (avg 500) switching transistor. It wouldn't take much for that gain to change especially with heat and the transistor quit.

Do you have a can of cold spray? Let it fail and then spary T3 and see if it starts working again.

Notice on the schematic it you can also use a "selected" 2N2222A. I've never seen one with that level of gain.

[HKZ Bob]

Hello Jim,
thanks for your help.
I replaced T3 already by a new one. To test I took a heat blower and heated it up more than 70 degrees celsious. I had a thermocouple instaled in the box. I could get the box into failure.

Do you have an ID why the unit is not working with a 12K as (R10) resistor.

bob

[jimmcc]

Interesting problem. The majority of the failed units I have delt with were due to the failure of T1.

What did you replace T3 with? I doubt if you were able to obtain a real BCY58X.

Let me review the schematic again along with your notes.

On the bench, what are you using as a signal source and at what voltage level?

Have you attempted to replace T2 as well?

Some where here I have scope images of the raw sensor signal. I'll search for them and determine what the actual signal level is although the images I had was with the distributor being turned by a drill so the voltage/rpm level is unknown.

[Lorenfb]

"failure of T1."

That RARELY fails. But then again, take the DIY approach and replace ALL the parts.
When that fails, buy a good core, i.e. not touched by a DIY, and exchange it for
a Pelican Parts rebuilt unit.

[HKZ Bob]

Hello Jim,
I replaced as well T2 and all in the schematic circled parts. T2& T3 with an BCX58x.
I use a 12V power supply with current limiter. I have an analog Hameg Scope 2 channel it is hard to post scope images.

I use the porsche setup as in the car with an original bosch hall distributor.

Bob

[HKZ Bob]

Hello Loren,
let me ask you this question?
Do you have an ID why the unit is not working with a 12K as (R10) but with an 10K resistor.

Maybe you have a ID.
By the way I used to live in Manhattan Beach for to years. You live in the right spot.

Bob

[Lorenfb]

Do you have an ID why the unit is not working with a 12K & not 10K?

Either should work as the "forced Beta" used is only ~ 100 & that device
or it's alternate has min Beta > than that. Otherwise, the actual device is bad.
The other resistor is 4.7K & not 47K.

[HKZ Bob]

Good point Loren.

Bob

[HKZ Bob]

Hello Jim,
what means on T3 selected.

The R6 is as well selected. How do I adjust these values?

"Notice on the schematic it you can also use a "selected" 2N2222A. I've never seen one with that level of gain"

Bob

[rick-l]

I don't have one of these I was just curious about how it works.

Does pin 7 go to a variable reluctance sensor, a coil of wire?

Is this the way to analyze the trigger circuit?


If you ground pin 7 do you get these voltages on the transistor? (Q1 = T2, Q2 = T3)

--- Operating Point ---
V(col1): 5.25113 voltage
V(base1): 1.70328 voltage
V(e1): 1.02285 voltage

V(col2): 1.09101 voltage
V(base2): 1.72703 voltage

Is TD a pulse that goes from 1 volt to 6 volts to the tach?

Select In Test parts (Select?) are selected durring final assembly/functional test to obtain a certain operating point or offset. A manufacturer will do almost anything to avoid this very expensive step.

[Lorenfb]

Q1 is biased "on" & Q2 is biased "off". The input signal turns Q1 "off",
and then Q2 turns "on". The hysteresis is determined by the common
emitter resistor. The devices DON'T need to be selected! R3 & R4 determine
the threshold.

"If you ground pin 7 do you get these voltages on the transistor?"

No! A simple calculation will indicate that pin 7 (distributor signal) must
go negative about 1.0 to switch the circuit.

[rick-l]

Here is how I was thinking this works.

With no signal, the diodes and resistors on the left side bias the base of Q1 at 1.7 volts and the emitter resistor has 1 volt across it or 6.7 mAmps through it (total through emitter of Q1 and Q2). At this bias point Q2 is saturated and 4 mA goes through the right side and 2 through the left.

With Q2 saturated C1 charges to about 5.5 volts through D5 (used to represent the emitter - base of the PNP transistor).

With a signal (here 1 cycle of a 1 volt sine wave) applied the voltage on the collector of Q1 falls on the rising edge of the input sine wave until V base2 < V base1 and Q2 turns off. The collector of Q2 goes to 6 volts and C1 discharges through D4 and R7. The collector of Q2 is the tach pulse.

When the input returns to the original bias point Q2 turns back on and C1 charges through D5 (the emitter - base of the PNP transistor) which triggers the SCR.

It looks like with this simulation (the sensor simulated with a single cycle of a 1 kHz. sine wave) the spark pulse leads the negative going zero crossing of the input by about 40 uSec

[HKZ Bob]

Nice Rick,

please use a 4.7 K for R8 and do it again and tell me what value you have on
Base Q2.


Bob

[Lorenfb]

" At this bias point Q2 is saturated and 4 mA goes through the right side and 2 through the left."

Wrong! It was clearly explained in the previous post.
Time to get another circuit simulator (maybe ICE).

Q2 MUST generate a negative pulse to trigger the SCR.

[jimmcc]

Rick,

Your simulation may be a bit off. The T2/T3 combination functions as a pulse shaper, schmitt trigger and zero-crossing detector noted by the C3 feed back.

Also, the sensor output is not a true sine wave. I'm enclosing an image of the reluctance sensor output showing the actual waveform. Sorry, but I forgot to record the scope settings but believe it to be about 2v p-p.