6 pin CDI problem under temp.

[ischmitz]

Wow, that is a lot of mechanics for triggering the boxes. I needed something that runs indefinitely for long-term testing so I came up with a little 555 timer based trigger circuit on a bread board.

Originally Posted by HKZ Bob

I use a new Bosch Hall distributor. It is the std. hall device Bosch used for Hall distributors. I drive it with my Makita

It works for both 3-pin and 6-pin boxes. With this design you can choose between a high speed around 6000 RPM and a low speed setting around 1000 RPM. It shouldn't take more than half an hour to put this together. If you want to get more sophisticated you could put another voltage devider and RLC network for pulse shaping onto the 6-pin output to closer simulate the real trigger signal of the 6-pin units.

Any news on your testing?

Ingo

[HKZ Bob]

Dear All
I do see multiple trigger on the Gate.
Gate of the SCR is brigded to Kathode

Is that normal.

Regards Bob.

[HKZ Bob]

Originally Posted by ischmitz

Wow, that is a lot of mechanics for triggering the boxes. I needed something that runs indefinitely for long-term testing so I came up with a little 555 timer based trigger circuit on a bread board.





It works for both 3-pin and 6-pin boxes. With this design you can choose between a high speed around 6000 RPM and a low speed setting around 1000 RPM. It shouldn't take more than half an hour to put this together. If you want to get more sophisticated you could put another voltage devider and RLC network for pulse shaping onto the 6-pin output to closer simulate the real trigger signal of the 6-pin units.

Any news on your testing?

Ingo

That trigger you see I used in the past perfect. I use now a 2 Volt pp sinus signal. I also bought a function generator. It is still in the mail. Will have it soon.


Regards bob

bob

[ischmitz]

Originally Posted by HKZ Bob

Loren I did it as you said, and the SCR is not firing. Not even at 20V. I did it with & without trigger. Does it means the failure is been initiated through the gate?

Originally Posted by HKZ Bob

Dear all & Ingo, Loren, the second test with gate shorted to kathode without the capacitor showed no spark as well. I also measured a little 0.35V sawtooth signal at Gate with 15V supply just before the SCR ignited. Bob.

Originally Posted by HKZ Bob

I do see multiple trigger on the Gate.
Gate of the SCR is bridged to Kathode. Is that normal

Bob, your above statements are contradicting each other. First, you reported a couple of days back there was no spark whatsoever with the and without a capacitor over the SCR while the gate was tied to GND. Second, you keep saying that you measure something on the gate while at the same time it is tied to GND. This becomes very confusing.

Let me re-phrase what you are saying: You have the gate shorted to the cathode, in other words C6 is shorted. If that is the case you will not be able to see anything on the gate because it is tied to GND. I assume the latest scope traces shows the voltage at the anode of the SCR. When you raise the supply voltage to 16V and beyond you start to see spikes on the scope, correct?

If my above description is correct there are only two possibilities at this point. Either the SCR breaks down electrically or you have an isolation problem.

Inspect the mounting of the SCR closely. I don't remember what the last SCR is that you have swapped into your box. Was it the 2N6509? TO220 cases have the anode present on the cooling fin. The heat sink in the box is at GND. If you don't use the proper washers and isolators around the metal screw there will be arcing going on. It most likley happenes between the metal screw and the inside of the hole of the cooling fin. Look for the tell-tale black marks.

Ingo

[HKZ Bob]

Originally Posted by ischmitz

Bob, your above statements are contradicting each other. First, you reported a couple of days back there was no spark whatsoever with the and without a capacitor over the SCR while the gate was tied to GND. Second, you keep saying that you measure something on the gate while at the same time it is tied to GND. This becomes very confusing.

The gate conection was open. SCR Gate & Kathode shortend. I meaured between R14 & R15.

Let me re-phrase what you are saying: You have the gate shorted to the cathode, in other words C6 is shorted. If that is the case you will not be able to see anything on the gate because it is tied to GND. I assume the latest scope traces shows the voltage at the anode of the SCR. When you raise the supply voltage to 16V and beyond you start to see spikes on the scope, correct?

I measured the output of T4.
The gate conection was open. SCR Gate & Kathode shortend. I meaured between R14 & R15.


If my above description is correct there are only two possibilities at this point. Either the SCR breaks down electrically or you have an isolation problem.

Inspect the mounting of the SCR closely. I don't remember what the last SCR is that you have swapped into your box. Was it the 2N6509? TO220 cases have the anode present on the cooling fin. The heat sink in the box is at GND. If you don't use the proper washers and isolators around the metal screw there will be arcing going on. It most likley happenes between the metal screw and the inside of the hole of the cooling fin. Look for the tell-tale black marks.

Ingo

I measured the output of T4.
The gate conection was open. SCR Gate & Kathode shortend. I meaured between R14 & R15.

[HKZ Bob]

Originally Posted by HKZ Bob

That trigger you see I used in the past perfect. I use now a 2 Volt pp sinus signal. I also bought a function generator. It is still in the mail. Will have it soon.


Regards bob

bob

Ingo what output do you use? 2Vpp.

When I use 2VPP sinus and I increase the volume from 8 to 14V I also increase the noise signal ratio and the schmitt trigger gets false pulses.

When I use a higher input signal the for exmple 6Vpp the box works fine.

So what is the values to test the box. I read in the bosch bible that the output on the magnetic pickup varies. so what is the low & what is the highest values.

bob

[HKZ Bob]

Originally Posted by HKZ Bob

That trigger you see I used in the past perfect. I use now a 2 Volt pp sinus signal. I also bought a function generator. It is still in the mail. Will have it soon.


Regards bob

bob

Lets talk more about you trigger cuircut?

bob

[HKZ Bob]

Originally Posted by Lorenfb

The signal amplitude varies with the RPM, i.e. it increases as the RPM increases.
So use a sine wave with a 1.0 volt peak to peak. Since a six cylinder at 6000
RPMs produces a spark every 3.3ms, the corresponding freq is 300 HZ at 6000
and 400 HZ at 8000.

"the negative slope of the signal wasn't always quick enough to trigger the SCR"

Nope! Remember, the input circuit is a Schmitt Trigger which is INDEPENDENT
of the input slope, i.e. That's the reason for using it.

With all respect Loren with a 1.0V peak to peak I'll get so much noise overlaying from the cuircut. Are you sure this is the value to go with.

Bob

[jmchrist]

I have recently be given to repair a 6 pin CDI with a similar problem under temp.
A lock up occurred only at high temp and at high rpm.

At first try on my bench I was unable to reproduce the phenomenon, even after half an hour of 8000 rpm at 14.4 V.

Finally I was able to reproduce the lock up by switching just for a second my generator to the equivalent of 80000 rpm !

The SCR was an original AEG T10N 600.

The phenomenon was identical with either a 16RIA80 or with a 2N692 (extracted of my drawer).
As the frequency of the converter increased to about 10 kHz when in lock up condition, I suspected a Tq problem, so I got rid of the problem with a BTW30-800 from ST Microelectronics, which is with a TO 48 case and has a Tq specified at only 12 µs max !
Unfortunately it is rather expensive since I got it from my electronics parts distributor for 20 euros (about 30 $) each !

I do not know if it is easily available in the USA since ST Microelectronics is an European semiconductor company.
I can provide the full specs of this SCR if somebody here is interested.

[HKZ Bob]

JMChrist,
now Warren is missing.

Do you have the chance to monitor the voltage on the load capacitor

I assume what could be your problem.
Can you monitor the voltage and you will see that your signal shows a lock up of swichting.






I figured out that the 2N6905 performes better that the OEM Thyristor.

See my plots



and now the noises and low level caused by the power consumtion.





Good luck would change the THYristor to 2N6905 and redo your bench test.

bob

[HKZ Bob]

That what I assume what could happen at high temp.

[rick-l]

HKZ Bob

What are the traces in the scope photos? I think the first one is the high voltage cap but what are the others?

[jmchrist]

HKZ Bob

You wrote :

> now Warren is missing.<

Unfortunately. Yes I know.

> Good luck would change the THYristor to 2N6905 and redo your bench test.<

I agree that a 2N6905 would also solve the problem since it is specified with a Tq of 15 µs typical. But the TO 220 case of the 2N6905 is much less practical to install than a TO 48 one.
Since my lock up problem has been solved with a BTW 30-800 which has the same TO 48 case as the original SCR from AEG and has a Tq of 12 µs max, I have no intention to put instead a 2N6905 even though it is cheaper...

[ischmitz]

Guys - the issue is most likely NOT the SCR. I agree that putting in an SCR with lower tq will help but you are band-aiding the real problem.

The main capacitor becomes part-ohmic with elevated temperatures. This increases the load on the DC-DC converter producing even more heat from the 2n3055. Heat degrades the SCR performance (tq). And it leads to the capactor to become even more ohmic. It's a typical run-away situation.

Eventually the SCR is unable switch off and the box locks up. Even devices with tq=15 usec will lock up eventually. Ask me how I know.

Change the capactor C8 and your problems will magically disappear, even at 8k RPM on a hot summer day.

Ingo

[jmchrist]

Thanks Ingo,

changing C8 was one of the actions I done when trying to solve the lock up at extreme conditions. It was unsuccessful.

My feeling is that the transformer of the converter in this particular box could be at the lower limit (I do not know what it is) in term of inductance This could explain that its operating frequency seems to be higher than usual. I measured 3.72 kHz when idle at 14 V. In another (operating) box I found 3.35 kHz.

I presume that, mpore or less, higher is the frequency, shorter is the time allowed to the SCR to return to the open state.

[ischmitz]

I don't think the frequency of the DC/DC converter affects the SCR's ability to comutate. You can see each swing of the half-wave rectifier on the rising voltage at the anode of the SCR as a little riple. I have seen anything between 3 and 4 kHz in different boxes and it did not seem to affect the stability of a given box.

If C8 wasn't the issue how about the snubber across the SCR. It is to limit dV/dt which is the other factor that can make an SCR to turn on. If dV/dt becomes too large the junction capacitance can trigger the SCR with no signal to the gate.

Ingo

[jmchrist]

Originally Posted by ischmitz

I don't think the frequency of the DC/DC converter affects the SCR's ability to comutate. You can see each swing of the half-wave rectifier on the rising voltage at the anode of the SCR as a little riple. I have seen anything between 3 and 4 kHz in different boxes and it did not seem to affect the stability of a given box.

In lock up situation, I observed that the frequency of the converter climbed up to about 10 kHz. This left less than 100 µs to the SCR to go off.

If C8 wasn't the issue how about the snubber across the SCR. It is to limit dV/dt which is the other factor that can make an SCR to turn on. If dV/dt becomes too large the junction capacitance can trigger the SCR with no signal to the gate.

Ingo

Of course I had also checked that C7 and R16 were OK.
The only major component that I was not able to check or to change is the transformer.
May be it is the real culprit, with an inductance that I found a little too low by comparison with another box (operating correctly).
Now that the box is correctly operating with a BTX30-800 I am not really motivated to try to change the transformer.

jmc

[HKZ Bob]

Originally Posted by rick-l

HKZ Bob

What are the traces in the scope photos? I think the first one is the high voltage cap but what are the others?

The other shots are on the Gate of the SCR.

bob

[HKZ Bob]

Originally Posted by jmchrist

In lock up situation, I observed that the frequency of the converter climbed up to about 10 kHz. This left less than 100 µs to the SCR to go off.


Of course I had also checked that C7 and R16 were OK.
The only major component that I was not able to check or to change is the transformer.
May be it is the real culprit, with an inductance that I found a little too low by comparison with another box (operating correctly).
Now that the box is correctly operating with a BTX30-800 I am not really motivated to try to change the transformer.

jmc

JMC What are you paying for an BTW30 -800

Bob

[jmchrist]

HKZ Bob wrote :

> JMC What are you paying for an BTW30 -800 <

As I said earlier, I paid 20 euros (about 30 $) a piece for the BTW30-800.
jmc.