Motorcycle CDI Pulse output

[Les Jones]

Hi Max,
The new pictures of the scope trace are much better. I would not trust normal divide by 10 probes with such high voltage signals. The resistors in them could flash over. I think you may need to make your own voltage divider. I was assuming the information I found on the web would be similar to your CDI unit but as the voltage pulse on yours is much higher it must be different. We will need to do some more research to get a better idea of what is inside your CDI unit. I would not try such a large spark gap for testing as the voltage may get so high it might cause it to flash over inside your CDI unit and damage it.
By the way I use the printscreen with 64 bit Windows 7

Les.

 

[spec]

This probably represents the innards of the Hus CD ignition - pretty standard stuff.
(there may be some some additional electronics to alter the spark timing in relation to revs)

spec

​

Reference about 2 stroke ignition timing:
https://www.dragonfly75.com/motorbike/timingcurves.html

 

[spec]

POST ISSUE 3 of 2016_11_10

Below is the schematic of post #22 above, turned into English.
(there may be some some additional electronics to alter the spark timing in relation to revs)

If this is in fact what the Hus WR300 CDI comprises, a charging and trigger circuit will be reasonably straight forward.

spec

​

ERRATA
(1) 240V should read 230V

 

[spec]

Hi Max,

I am trying to figure the details of the Hus WR300 CDI unit. You mentioned that it may be mapped to suit the 300cc, single cylinder, two-stroke engine (motor).

But from what I have read, trail bike two stroke engines often do not have mapping. Instead they just fire at top dead center (TDC) throughout the rev range, typically 1.3K RPM to 8K RPM.

Do you know if the CD unit does in fact have mapping?

This will probably not affect the CDI trigger and power supply design, but I would just like to get an overall picture of what is going on.

Maybe you could check for any mapping by looking for any timing change between the trigger pulse and the spark, for the simulated rev range for your engine, on your lathe.

spec

 

[Les Jones]

Hi spec,
Like me you have re drawn the schematic to make it easier to understand. I think you have put one diode in the wrong place. I thing the bottom end of r5 should go straight to the chassis and the diode should go between chassis (+) and the end of the capacotor connected to the coil. I have not worked out what D3 does. If there was a capacitor in series with the AC input the circuit would be a voltage doubler.

Les.

 

[spec]

Like me you have re drawn the schematic to make it easier to understand.

Yes, I noticed. How some people can work/understand the convoluted schematics you sometimes see around is beyond me. They must be extremely clever. I have studied my automobile (car to us) wiring diagrams and, unless automobiles use a special type of electricity, they are beyond me.

I think you have put one diode in the wrong place. I thing the bottom end of r5 should go straight to the chassis and the diode should go between chassis (+) and the end of the capacitor connected to the coil.

Oh yes. It was just a test. I wanted to see how long you would take to spot it! (I don't think) I will correct the schematic- thanks.

I have not worked out what D3 does. If there was a capacitor in series with the AC input the circuit would be a voltage doubler.

UPDATE: No not sure about D3 either. As shown the diode will short the 230V for negative excursions which, in conjunction with the voltage generator coil. may have a voltage multiplying effect.

spec

 

[spec]

2016_11_16 PLEASE SEE POST #33

INVERTER LINK: **broken link removed**

spec

 

[Les Jones]

Hi spec,
That looks like a nice inverter for the price. I think it is using the KA3525A driver chip
This is **broken link removed** of it used in a similar inverter to the one in your link.
It has the facility to adjust the dead time so we could get short breaks in the output that might be long enough for the SCR to unlatch. I have not been able to find out how long they take to unlatch. I think we need to get Max to run the original trigger generator as the lowest speed the engine runs at to see how low a voltage we can get away with for the high voltage pulse. (I find it hard to believe that the originl unit puts out 2KV pulses.) The fact that he wants to power it from 22 volts also creates a problem. I don't see why he can't power it from the 12 volts sytem used for lighting. I don't know if he has considered the charging system for the 22 volt LiPo starting battery. It will be a lot more complex than charging a lead acid battery.

Les.

 

[spec]

That looks like a nice inverter for the price. I think it is using the KA3525A driver chip
This is **broken link removed** of it used in a similar inverter to the one in your link.
It has the facility to adjust the dead time so we could get short breaks in the output that might be long enough for the SCR to unlatch. I have not been able to find out how long they take to unlatch. I think we need to get Max to run the original trigger generator as the lowest speed the engine runs at to see how low a voltage we can get away with for the high voltage pulse. (I find it hard to believe that the originl unit puts out 2KV pulses.) The fact that he wants to power it from 22 volts also creates a problem. I don't see why he can't power it from the 12 volts system used for lighting. I don't know if he has considered the charging system for the 22 volt LiPo starting battery. It will be a lot more complex than charging a lead acid battery.

Hi Les,

glad you like the inverter- very useful (and cheap) device.

I have a detailed design for the CDI controller in my head now, but was just going to put it into EAGLE, when missus forced me to go for lunch in town and then go shopping.

The plan is to generate two periods starting from T0 (trigger point)
(1) The TRIAC will be triggered for period Pt
(2) A high-voltage switch will be turned off for period Pin which will be longer than Pt. How much longer is the question. The high-voltage switch will disconnect the 200V supply from the CDI while the inverter continues running non-interrupted.

You have pin-pointed the critical area: preventing SCR latch- up. The trade off is to inhibit the high voltage just long enough to prevent SCR latching, as the inverter needs maximum time to charge the capacitor before the next SCR trigger. This is critical for high RPM. The best approach, I have found, is to inhibit the inverter whenever any SCR current is present. I used this approach on the CDIs I designed by putting a resistor in parallel with a diode between the SCR cathode and 0V but, unfortunately, in this case there is no access to the SCR, so another approach will be required.

The high voltage switch approach gives the fastest cut off response and fastest charging of the capacitor, I think. The exact value of inhibit will need to be determined though. The trigger frequency at 10K RPM, 10000/60s = 167 Hz, which only gives 6ms between triggers, so it is all a bit tight.

About the power supply, I think that Max said that the ignition will always be operated from 12V. If the CDI Controller is in fact required to operate from 12V and 22V, I have got an off-the-shelf DC/DC SMPS in mind to produce a constant 12V or so.

In terms of charging the 22V LiPo battery, I see that as a standard buck converter/battery charger board, but have not investigated a suitable board. Charging the LiPo battery is really a separate exercise.

An area that I haven't considered in any detail is the sensor. I know that hall-effect is one contender, but there are others, magneto-resistance being the most reliable and rugged, as far as I know, bearing in mind the environment: a motorbike, especially a highly-tuned off-road two-stroke, is going to put the electronics through some extremes of temperature, mechanical shock, and vibration.

Apologies for the rambling post.

spec

 

[spec]

I asked if the CDI does, in fact, have built in advance: I have now confirmed that the CDI does have built in advance, so there will be more electronics associated with the trigger input than shown in post #23. But, hopefully, this should not affect the design of the CDI Controller.

spec

 

[MaxD]

G'day Spec, Les, just back from a work tour, and checked the timing tonight, with a timing light connected. Yes, correct Spec, at the minimum rpm it would idle at, it fires at 18 degrees (plus or minus one) before TDC, and revving to as high as I am game to take it to in neutral, ie blipping the throttle hard, it moves to about 8 degrees (give or take one) So yes, it has decreasing advance with rpm, as many modern two stroke have.

Now, I think I might be wise to ask you guys for advice on using an oscilliscope / probes correctly, to make sure I am not giving false info.
The setting you see on that screen grab are what was actually set on the probes etc.

Les, I take on board what you mentioned about high voltages and scope probes.
I assume to use a voltage divider, I would use two resistors in series, one very high value, one lower, connected between the point I want to measure and ground. The resistor values should be in a ratio to drop the expected voltage at the junction between them to perhaps a hundred or so, then connect the scope probe to that junction. The only problem I see, is that the trigger coil is not connected to ground, at least outside the CDI unti itself.
When I took the screen shot I provided, I did connect the scope probe grounds to the engine case.

Spec, "About the power supply, I think that Max said that the ignition will always be operated from 12V. If the CDI Controller is in fact required to operate from 12V and 22V, I have got an off-the-shelf DC/DC SMPS in mind to produce a constant 12V or so". The battery voltage is 22 volts. I imagine it is easier to choose a single voltage to run the charge/trigger circuit from, whether that is 12v or 22v I am open to advice, probably guided by standard off the shelf components if possible.
The cdi will need to be driven from the battery at least until the engine is actually running, as the generator output will not be available until the start cycle is complete. The generator is now in fact a starter/generator, and the two functions can not run concurrently.
I would not anticipate a problem dropping from 22v to 12v if prudence suggests it to drive the charge/trigger circuit, but this will always be powered from the battery.
"Charging the LiPo battery is really a separate exercise". Yep, I do not see a problem using off the shelf chargers for this, reliability might be another issue later though.

Les, "I think we need to get Max to run the original trigger generator as the lowest speed the engine runs at to see how low a voltage we can get away with for the high voltage pulse. I hope to do this tomorrow.(I find it hard to believe that the originl unit puts out 2KV pulses.) I also struggle to believe it, given that those measurements were taken at maybe 3,000 rpm, what is the voltage at 9,000 ? One would expect considerably higher. The fact that he wants to power it from 22 volts also creates a problem. I don't see why he can't power it from the 12 volts sytem used for lighting.
The above answer probably covers that questions about running it from 12 volts. I will be regulating the generator output to feed the original lighting circuit.
I have chosen a LiPo for the size/power/weight ratio. Lead acid is not viable for this mod. I have tried a 12v, 200 CCA Gel Cell LA. But that battery in series with my 12v van battery works

My lathe will only spin to just over two thousand rpm, so getting higher rpm readings has to be done on the bike itself.

Tomorrow I hope to re-do the coil readings with the engine running and using a voltage divider. Both minimum and highish rpm.

Thank you for the time and effort so far guys, shame work gets in the way of more interesting pursuits huh

 

[Les Jones]

Hi Max,
For the voltage divider I would suggest the bottom resistor being 1.1 meg and the top resistor being 9 megs. Make the 9 megs up using 9 x 1 meg resistors, (This so the voltage across each individual resister is not enough for it to break down. ) It is probably be easier to make the bottom resistor from 1 meg in series with a 100K. The reason it is not just 1 meg is because it will pe in parallel with the 10 meg input resistance of a divide by 10 scope probe. Looking at your scope traces it looks like the duration of the high voltage pulse is about 1/3 of the full cycle. As you are using the wasted spark system you would need to cut two notches out of the timming disk for the high voltge pulse each with an angle of 60 degrees. This is assuming you are using hall sensors that behave like the ones that I suggested. I have never tried them like this as I have only been interested in counting pulses for tachometers. The data sheet seems to suggest that they will work like this. It will be nice if 250 volts is high enough for your CDI to work as then the inverter that spec has found should be ideal. When the starter/generator is in generator mode does it output DC directly or does it output AC which is rectified ? Does it use permanent magnets or can it's output be controlled like a car alternator ? Will you also have a 12 volt battery on the lighting circuit ?

Les.

 

[spec]

POST ISSUE 2 of 2016_11_16

Hi Max & Les,

Thanks for clarifying the points above in post #31 Max.

In view of what Max says, below is how I now see the overall block diagram of the Husky WR300 power and ignition system. It does look complex but most of the items are off-the-shelf, relatively compact, and low cost.

spec

​

LINKS
(1) Battery Charger (6 cell, LiIon or LiPo): TBD
(2) Balance (6 cell, LiIon or LiPo): TBD
(3) Buck SMPS (12V OP): **broken link removed**
(4) Inverter (12V to 200V DC): **broken link removed**

​

 

[spec]

HI Max,

Sorry to be a pest, but I would just like to confirm the battery & starter/generator arrangements on the WR300.
(1) There is a combined starter/generator rather than a separate starter and generator?
(2) The same connections on the starter/generator are used for both starting and generating functions?
(3) There is one LiPo battery pack, comprising six LiPo cells connected in series to give 6 * 3.6V = 21.6V nominal? The battery pack voltage may be 6 * 3.4V for Lipo, as opposed to standard LiIon batteries, but that does not matter.
(4) When starting the WR300, the LiPo battery pack is connected to the starter/generator?
(5) Can you give details of the battery pack:
(5.1) Is the battery pack off-the-shelf or did you make it from standard cells
(5.2) Is the battery pack protected
(5.3) Does the battery pack have a charging balancer built-in?
(5.4) What is the nominal voltage of the battery pack? It will vary from 6 * 3.4V to 6 * 3.6V depending on the exact battery chemistry.
(6) What maximum current does the starter/generator take when starting?

spec

 

[spec]

Here is a DC to high voltage converter which may do the job and elliminate the 12V buck inverter, but I am not sure it will charge the CDI capacitor fast enough. **broken link removed**

And here is a power supply that may be useful for charging the battery. **broken link removed**

spec

 

[MaxD]

Les, "When the starter/generator is in generator mode does it output DC directly or does it output AC which is rectified ? Does it use permanent magnets or can it's output be controlled like a car alternator ? Will you also have a 12 volt battery on the lighting circuit ?"
It is a permanent magnet, star connected, 3 phase generator, so no field control. The measurements I took of it running in the lathe gave an output of 12.18 v across 2 phases at 1053 rpm.
The single phase will be enough as it has reasonably heavy windings, which will be rectified and shunt regulated to maintain around 12 volts. No other battery will be used, the original bike does not use a battery.


Spec; "I would just like to confirm the battery & starter/generator arrangements on the WR300".
(1) There is a combined starter/generator rather than a separate starter and generator? Yes, new unit is a combined starter/generator, my own construction.
(2) The same connections on the starter/generator are used for both starting and generating functions? If you mean connections to the windings, yes. However, during either of the start or gen functions, the windings will need to be switched independently to either the starter controller, or the rectifier/regulator.
This is how I see the process:
To start, battery will connect to the windings via the start controller, at/after engine start, dis-connected.
As the start circuit dis-connects, the windings will connect to the rectifier/regulator for engine running, at the same time, the battery charger will connect to the battery.
(3) There is one LiPo battery pack, comprising six LiPo cells connected in series to give 6 * 3.6V = 21.6V nominal? The battery pack voltage may be 6 * 3.4V for Lipo, as opposed to standard LiIon batteries, but that does not matter.
(4) When starting the WR300, the LiPo battery pack is connected to the starter/generator? Yes, via a controller.
(5) Can you give details of the battery pack: 1 pack with 6 cells in series (6S) 5000 mAh. The nominal voltage of this battery pack is 22.2 volts, ie; 6 x 3.7 v
(5.1) Is the battery pack off-the-shelf or did you make it from standard cells ? Off the shelf, HRB-XF, 5000 mAh
(5.2) Is the battery pack protected. Not that I can establish. I will monitor it very carefully in the early stages of operation to ascertain how it performs. The charger will have individual cell management and the battery has individual cell monitoring facility.
This provokes a question of how will the CDI be supplied in the event the battery has to be switched off for any reason ?
I guess the SMPS could be manually connected directly to the gen output, before the engine is started, if the starter circuit is not activated, then kick start the engine as normal.

(5.3) Does the battery pack have a charging balancer built-in? Not a balancer, but connections for the charger to keep it balanced.
(5.4) What is the nominal voltage of the battery pack? It will vary from 6 * 3.4V to 6 * 3.6V depending on the exact battery chemistry. These type of LiPo are recommended to be operated between 3.0v minimum under load, (3.7v static) to a maximum charge of 4.2v.
(6) What maximum current does the starter/generator take when starting? I have not been able to get an accurate reading of this yet, but I know it is just north of about 30 ish amps. My 30A benchtop SMPS will sometimes turn over the engine correctly, and sometimes not, with the red overload light on.

 

[MaxD]

Hmm, it uploaded before I was ready. Here are the latest measurements, taken with the engine running, a Voltage divider of 7 x 1.2M ohm + 1 x 1.2Mohm + 1x100K resistors in series. This was connected to each of the two leads in turn from the charge coil to the CDI, ( one each end of the coil), the other end to earth.
So like this....CDI Input lead, 1.2M, 1.2M, 1.2M, 1.2M, 1.2M, 1.2M, 1.2M, Scope probe, 1.2M, 100K earth.
Hope that makes sense.
I used both the USB and stand alone Tektronix scopes to hopefully remove any doubt about voltages produced.
All readings were with the probes set to 1:10.
Pic 1, eg; Lead A , at idle, USB scope
Pic 2 Lead A, with revs, USB scope. (Sorry about pic qual. but no change in voltage, only frequency.)
Pic 3 Lead A, idle, Tek scope
Pic 4 Lead A with revs, Tek scope

Pic 5 Lead B, at idle, USB scope
Pic 6 Lead B revs, USB scope
Pic 7 Lead B at idle, Tek scope
Pic 8 Lead B with revs, Tek scope


Pic 9 Lead B at idle, but using a different brand probe at 1:10 also
Pic 10 Lead B at idle, probe of #9 above, but at 1:1 (note I forgot to change the scope setting to 1:1 )
The settings on both scopes were as you see them in the pics.

After uploading all those, I realized I'm probably labouring the point here somewhat. But I also wanted to give as clear a picture of what is going on as I could.

I was quite surprised to see that the voltages do not increase with revs, only the frequency as expected.

 

[spec]

Les, "When the starter/generator is in generator mode does it output DC directly or does it output AC which is rectified ? Does it use permanent magnets or can it's output be controlled like a car alternator ? Will you also have a 12 volt battery on the lighting circuit ?"
It is a permanent magnet, star connected, 3 phase generator, so no field control. The measurements I took of it running in the lathe gave an output of 12.18 v across 2 phases at 1053 rpm.
The single phase will be enough as it has reasonably heavy windings, which will be rectified and shunt regulated to maintain around 12 volts. No other battery will be used, the original bike does not use a battery.


Spec; "I would just like to confirm the battery & starter/generator arrangements on the WR300".
(1) There is a combined starter/generator rather than a separate starter and generator? Yes, new unit is a combined starter/generator, my own construction.
(2) The same connections on the starter/generator are used for both starting and generating functions? If you mean connections to the windings, yes. However, during either of the start or gen functions, the windings will need to be switched independently to either the starter controller, or the rectifier/regulator.
This is how I see the process:
To start, battery will connect to the windings via the start controller, at/after engine start, dis-connected.
As the start circuit dis-connects, the windings will connect to the rectifier/regulator for engine running, at the same time, the battery charger will connect to the battery.
(3) There is one LiPo battery pack, comprising six LiPo cells connected in series to give 6 * 3.6V = 21.6V nominal? The battery pack voltage may be 6 * 3.4V for Lipo, as opposed to standard LiIon batteries, but that does not matter.
(4) When starting the WR300, the LiPo battery pack is connected to the starter/generator? Yes, via a controller.
(5) Can you give details of the battery pack: 1 pack with 6 cells in series (6S) 5000 mAh. The nominal voltage of this battery pack is 22.2 volts, ie; 6 x 3.7 v
(5.1) Is the battery pack off-the-shelf or did you make it from standard cells ? Off the shelf, HRB-XF, 5000 mAh
(5.2) Is the battery pack protected. Not that I can establish. I will monitor it very carefully in the early stages of operation to ascertain how it performs. The charger will have individual cell management and the battery has individual cell monitoring facility.
This provokes a question of how will the CDI be supplied in the event the battery has to be switched off for any reason ?
I guess the SMPS could be manually connected directly to the gen output, before the engine is started, if the starter circuit is not activated, then kick start the engine as normal.

(5.3) Does the battery pack have a charging balancer built-in? Not a balancer, but connections for the charger to keep it balanced.
(5.4) What is the nominal voltage of the battery pack? It will vary from 6 * 3.4V to 6 * 3.6V depending on the exact battery chemistry. These type of LiPo are recommended to be operated between 3.0v minimum under load, (3.7v static) to a maximum charge of 4.2v.
(6) What maximum current does the starter/generator take when starting? I have not been able to get an accurate reading of this yet, but I know it is just north of about 30 ish amps. My 30A benchtop SMPS will sometimes turn over the engine correctly, and sometimes not, with the red overload light on.

Many thanks for you helpful and comprehensive reply to my pesky questions Max.

If more posters were as forthcoming as you, it would be a great help and would save us a much sweat trying to find out what exactly the OP wants.

I do not see any insurmountable problems from what you say.

I do suspect, as anticipated, that the starting current will be high though.

spec

 

[Les Jones]

Hi Max,
There are a few points I would like to clarify. I was under impression that one end of the charge coil was connected to ground and the other end went to the CDI unit. Now you are talking about two wires from the charge coil. Do you mean that you have disconnected the one that went to ground or is there two wires from the charge coil going to the CDI unit and some point on the coil is connected to ground ? I do not know how your scopes display the volts per division. Is it the vots per division at the input socket on the scope or is it the actual votage befor the probe division ? I have never used a scope where you enter the probe division ratio. I have only ever mentaly multiplied by the division ratio. It is my impression that none of these reading show anything as high as the 2KV in your original tests. am I correct ? I second specs thanks for providing information. If you look at the thread on the broken RS232 cable you will see how difficult some people make it for us to try to help them. I don' think the OP in that thread has answered any question put to him.

Les.

 

[MaxD]

Les, "impression that one end of the charge coil was connected to ground and the other end went to the CDI unit. Now you are talking about two wires from the charge coil."
I must not have made my description clear. The original Charge/Trigger coil (one single coil) has both ends going into the CDI unit. This coil is not connected to ground, physically, nor would I think electrically, outside of the CDI anyway, as evidenced by the voltages being measured at the junction in the wiring loom from that coil to the CDI loom.

If you have a look back at my post #7, (2009 model) the schematic shows the alternator with three wires coming from it. One is the AC to power the lights etc. (one end of those coils are earthed internally at the stator end) The other two wires are straight into the CDI, as they are on the bike itself, with no internal or external earth connection. Only the High Tension coil, the one to the spark plug is earthed.
I did not dis-connect any leads to take the measurements, I simply tapped into the two wires from the Charge/Trigger coil to the CDI for the scope probe.
Unfortunately I do not have access at the moment, so can not check whether that front panel "Probe Attenuation" setting changes the trace value or not. I will check it after the 24th.

Please correct me if I am wrong, but the measured voltage will be:
Screen trace vertical value in divisions, x Front Panel Volts/Division setting, x probe ratio (1:1, 1:10, 1:100) as the case maybe, x any voltage divider value applied between the circuit point of interest and the scope probe.

If we refer to Post #37, Pics 5 and 9, the trace reaches 6 divisions, the V/div is 5.0, probe setting is 1:10, and the voltage divider is reducing the circuit point of interest to 13.4% of circuit voltage. (7x1.2MOhms) + (1.2MOhms+100KOhms) 6 x 5 x 10 = 300 being 13.4% of the circuit voltage gives approx 2,200 volts.

Referring to Post #37, Pic 1, we see the trace reaches 6 divisions, the V/Div setting is 0.2 V/Div, setting on the probe is 1:10, and the voltage divider is dropping the voltage at the circuit to 13.4% of circuit voltage. . My maths give about 90 volts for that wire. It is the other end of the same coil, so this voltage I think must be being held low by being loaded.

Hmm, this confuses the issue, Pic 10 shows what happens to the trace when the switch on the probe itself is set to 1:1. So say 15 Divs, x 10 x 1 (probe ratio)= 13.4% of circuit.....1,100 ish volts, half the previous readings. Can anyone see if/where I might be going wrong ?
These readings are taken as the engine itself is running.


"Is it the vots per division at the input socket on the scope or is it the actual votage befor the probe division ?" This Volts/Div is at the input to the scope, not counting the probe division.
I will have to wait until I'm home to re-confirm the Probe Attenuation setting and how/if that changes the trace on the screen.

I also just realized after seeing the pics I posted of the Tek scope, I had the Probe setting at 1:100 on the scope. I was however, using the probes themselves set at 1:10.

I hope that has cleared up a few things, and not created more confusion.

Thanks guys, regards
Max