Rob.. Here's the first prototype from the newest board design.. works perfectly! Thanks!

John

Charlie... Here's the link for efficiency ratings between toroid and e-core transformers.. You may be in disbelief but take a look... good reading with charts etc..

John

http://www.cse-transformers.com/toroidal.html

Hi John,
you have been busy.....incase i missed something can you tell me what you have done to sweeten this circuit up ?

i did read voltage is up is that correct ?

does it triger from the pic opto setup and can the rest of the timing board stay the say

did you change the circuit in your early post 750 as it show the same or not ?

cheers
brett

Thanks John,

Thats good info. I think I will try to wind one. The first time I tried to wind one I had a shuttle that was a foot long. I think this time I will use a shorter one.]

The first picture is from a small (25x25mm) E-core transformer with pulses.

Brett, It triggers just fine from the opto timer board.. The only real changes were parts placement on the new board. As you can see, I only use a toroid transformer, wound at 400-450 turns. That combination with the NPN transistor I use makes it efficient as well. Battery life is around 2 hours +.

John

The only thing that may require tuning is R-1 on the CDI board..initial tests it was set at 1.5k ohm..some transistors may require down to 1k ohm to oscillate consistently. Since Rob made the changes to the board design to accomodate either direct hall input or timerboard connections, I think it should be the final version of my board circuit.
Thanks to Rob and his efforts in board design, we now have many different choices that all work! Input and suggestions are always welcome since it is an open source project.
I'd like to hear from anyone building the latest board design. Please post your results so we can all benefit from the results.

Many thanks to all who have added to this great project!

John

Mmmmm.....this sounds like a farewell []

I have no idea what was wrong. I did a fresh install of microC and used your project file, and it just worked..

Here is a short video of the first test on an engine:
http://www.youtube.com/watch?v=PbuHh0UleUU

This is using the stock magneto coil, and the IGBT connected to the stop-button leads on the coil. However, polarity is wrong, so I now engine ground is not on the same potential as control board ground.. I want to use the receiver battery for processor power supply, and that might be a problem now.
Will se how I solve that, now I'm just happy that I can control the ignition timing

No it is not a farewell by any means...I have to work sometime to raise money for my hobby and I just got a part-time job.
I wish I would still have the 20+ hours to devolp things here but time is money.. and we all need that for sure.
I am processing about 6 of these boards a day. Testing each one before it leaves is time consuming as you well know Rob. There has been alot of interest in this forum since we have a completed functioning and tested project..I hope the support and fresh ideas keep coming.
Many new members to the forum as well as those who just read the info are interested in a do-it-yourself type of project and this one with it's many variations and support for each design should be around for a long time to come.
I for one, would like to look to the future of R/C engines with the knowledge gained from all of the members. This is what a hobby is all about!! I have offerred my circuit to the users as well as you Rob and JMJ with the beginning of the forum here.
Let's hope it continues to improve!

Thanks again..


John

Maybe I could get suggestions here on how to solve the ground potential issue? There must be a lot of electronics experience here based on the complexity of the CDI

I'm using a Fairchild ISL9V3040P3 ignition driver IGBT (n-channel). (datasheet: http://www.fairchildsemi.com/ds/IS/ISL9V3040P3.pdf )
For some reason there is no spark if i connect the grounded side of the primary coil to the emitter, and non grounded (marked +) to the collector. However, if I switch them, it works.
When I do this, the voltage spikes from the primary coil will be directly connected to the ground of the control board(since the IGBT needs the control board ground as a voltage reference for the gate) and I suspect this will cause problems. Everything (radio) I connect to this pulsating ground is likely to be disturbed.

-I tried to look for a P-channel ignition driver instead, but no luck..

-If I use a optical isolator, the control board will be separated, but the secondary side of the optocoupler still needs a supply voltage and IGBT emitter/ground, and again, that ground will be pulsating.
I really want to use the radio battery pack for the control board too, to keep everything as simple as possible (not needing to charge multiple battery packs before driving, adding weight etc).


Suggestions are welcome!

How was the ignition original, meganical switch or electronic switch ?
Is the coil not switched by power plus ?
Do you have a schematic of the eriginal ignition ?

@John, thx for sharing your schematic with us, we have all learn of it.
Within the 5 years of the topics are on the forum, there are many people who start to build, but never let us know if there projects positive finish.
Most project are dead after / within a (few) year(s), but with enthusiastic people who start to build and let us know it works it for us nice to help.
You are one of thoose guys who was enthusiastic of the project let us know it works and thats wat we need........positive feedback to keep the topic alive !
Thx again and I hope some people you have served a CDI will tell us here there experience with the CDI.

After a few months of sicknes I've also back to work for some houres a day and can't spent all my time to the forum.
This isn't mean I'm not at all on the forum, but it takes a little bit more time to the forumguest to get awnsers from me.
But Charlie, Brett, and a lot of others know even more then I do and are my continuators [8D]

Rallyfinnen why dont you try the recommended component ( BT151 ) to see if that works first, that n channel is a Mosfet is it not and we use a SCR


Hi John i made your power board up last night but used my bobbin coil, i get a good spark across the cap ,

but i dont here the bobbin oscillate.......is it the low voltage i tested at......or maybe my battery was nearly dead


Brett

Rallyfinnen use the TCI-board and this board is specialy for using a low Voltage bobine / sparkcoil.
But, the sparkcoil can be fired to ground but also to power plus batterie.
I don't know if he use the enginepower for the sparkcoil.
If he is only grounded the TCI to the engine there can't be a potential differends......there can't be a potential differends on a one wire powersource.

Sad to hear you've been sick, glad to hear you're getting better

The original ignition is just a standalone coil, like on many chainsaws. Standard coil on Zenoah (and clone) car engines.
Here is a picture of it:


I moved it a little bit to have it at approx 40deg BTDC, it was abt 35deg to start with.

It only has a spark plug wire out, and two wires for the kill switch. I have no schematic, all I did was to connect the IGBT in parallell with the kill switch to 'hold the spark'.
There must be an internal transistor connected to a primary winding, and somehow triggered by the magnetic field. The resistance between the wires to the stop switch is abt 0,8 ohm, with one connected directly to GND and marked -, and the other one marked +. I beleive this is the primary winding controlled by the internal transistor.
Don't ask me how, but it works

The beauty of it, in my opinion, is that it's almost fool proof. If the controller dies (and hopefully leaves the IGBT open) the engine will still run, but with full advance. It consumes next to none in power, since the ignition charge comes from the magnet.

kermet: I will look in to that BT151. I think it needs to have quite low on-resistance to work in this application though.

You have a magneto ignition, no transistors or something like that.
The flywheel makes the magneticfeeld for some time and when the flywheel has past the coil it ignite.....thats all.
Some Zenoah (like PUM260) have two coils, a red one (timerboard an HV-coil) and a grey one (HV-sparkcoil).
Into the redcoil there will be some electronics and and the pickupcoil, see the picture.
The graycoil is the same coil a CDI use and is NOT a 6 Volt sparkcoil.....this coil can't be used with a TCI !
It's also impossible to put the TCI between this coil and the killswitch, it will never work.

I wasn't see the picture of your ignitioncoil befor, now I can see and it's realy a magneto ignition.
On the flywheel there are 3 magnets, This magnetes makes the magneticfeeld., first two magnets positive, middle and last magnet negative.
This negative Voltage is also the start to unload the power from the ignition to the sparkplug.
Into the ignitioncoil is 1 sinuswaveform made by the 3 magnets the hole ignitionpower to get a spark.

Wat you need is a coil showing in picture two, a normal 6Volt ignitioncoil.

So what you are saying is that it should not work?
It does.. I can easily see it by disconnecting the power to the control board. Then the ignition advance jumps to another position, and the ignition advance is (as good as) constant over the rev-range. With the board connected it varies with RPM.


There is a comercially availible ECU connected the same way, just connected in parallel to the stop switch to adjust ignition timing. Oddified ECU: http://www.oddified.com/services.php
The Oddified ECU does not let the user make his own curves, but it has a nice feature that there is no need for batteries, they are using some of the energy from the ignition coil to power the ECU. This is due to racing class restrictions I think.

I believe there is only one magnet on the flywheel, north and south is all i could get on the hall-sensor. By using a 'latched' hall sensor I get only one pulse/rev.

I don't know this system and have never see it
If you connect only the ground to the timer/powerboard and place the other wire direct on the IGBT it can be a problem.
Don't connect any other parts to the engine, this will blow your electronics, the coil produce a high Voltage, >15,000 Volt.
I 'll draw how I think you have to connect the boards.

Hi Brett.. on all the CDI's I run here, the oscillations are either at such a high frequency that you can't hear it or it's the fact that it is a happy circuit without audible hum or noise. In other words it is efficiently using the current. If you have an actual noise coming from it, there may be excess current draw. I use only CDI style coils. I tried a home made coil but couldn't get it to keep working because of the high voltages (15-20,000 volts). I wasn't interested in making excessive voltages because of using them for Radio Control planes..it causes alot of unwanted RF energy (noise to the radio control reciever). I never run shielded wiring either.. nothing on the sparkplug wire..just a normal seperation between the systems of 6 inches. That seemed to work for the purposes here. Charlie can tell you first hand about RF signals and radio recievers..
If you need that high of a voltage, add more turns to the toroid secondary..I think you are looking for the voltage to gain reliability at higher rpm's than we need with our planes..
Hope this info helps..all things considered, you have an exceptional system with higher voltage requirements. All I can say is experiment with the windings..

John

Rob.. with all the people looking at this project ( looked at the number of hits for the files and pictures), we are sure to start getting feedback on the project.. We all started the ball rolling..now you will see many people input ideas. Some with more of a background in Pic programming I hope..
As I see it now, electronics are fully developed, with the exception of a few bugs in a certain part of someone changing the design to match their needs. I have recieved emails from people wondering how and why we give this information out for free...You and I know the joy of helping others learn and be happy when it all works! I do it every day.. I even learn from beginners with their questions..
Wish I had more time to dig into the programming.. I'm itching to learn it but no time.. There has to be someone doing this secretly at home and not sharing.. Oh well..Hopefully someone will point us in the right direction with some good info.

John

thanks John........ have you tested the max volts the circuit likes to run at



brett

I've tested mine at 6 volts input from a 4 cell alkaline battery pack and it worked fine...might be able to go up to 9 without too much heating of the transistor...but if you go that high, you might need a voltage regulator...not sure. I am in the process of using a LM317T for that purpose.
My bench supply uses one and I turned up the voltage to 8 volts for a few seconds and got alot of heat from the transistor. You have to consider that the 1n4007 on the timer board (original design) takes about .5 volts away from the input voltage..so at a fully charged nicad pack you'd have 5.3 - 5.5 volts ending up with about 5 volts.. and at that voltage, I still got almost 2 hours run time. Spark was great though either way.

John

If you increase R1 will that muck up the RC in that end of the circuit ??? and change the oscillator

wonder if you could fit another resister between negative side if C1 and ground to current limit T1 if more volts were added ?

Im just thinking alot of people (and me) would look at a 2s lipo pack as they are so easy to field recharge less weight and more power so 7.4 volts
sounds a nice voltage to set it up for and try not using the voltage reg as there quite power hungry unles its switch mode

What do you think

brett

Brett, lowering the value of R-1 in the circuit makes the transistor fire a little harder keeping a constand rate of trigger on it for the oscillations in that part of the circuit.. Going higher in resistance till it stops functioning and then lowering it till it's consistant by using a variable resistor is how I find the sweet spot!

Using a Li-Po 2 cell pack would probably be an ideal setup.. just be aware of the heating it may cause..if it's not excessive, you have a reliable circuit.
Give it a try..shouldn't be too much voltage for it.

John

Hello!

I have now done some more testing and scoping on the ignition.

I have noticed that the ignition timing is lagging with higher revs even if the angle is set to a constant angle.
To test this I made an ignition curve with a constant 40deg advance. I can then see that the angle decrease almost 10deg at approx 17k.
When looking at the scope, it sems there is a delay of abt 60us from the pulse on the input arrives until the output reacts. This is with the 40deg curve (all constants in the code is 00).
When calculating it, a delay of 60us would give a spark retard of 3,6deg @ 10k rpm, so this seems logical (since I only have an approximate delay value).
I think all processor controlled ignitions have some lag, and that is why the trigger always needs to be set in advance, so there is time to compensate for the lag.

The question here (to the one who made the code I guess) is if the 60us delay I measured seems reasonable based on processor speed/number of operations?

I think there could be a possible solution to the problem as well:

-If I'm not mistaken, the C-code contains delay-values for diffrent RPM's, and if we know the format of these, we could remove 60us from all delay values to compensate for the delay in the processor.
I think ths could easily be incorporated in the Excel sheet?

-The trigger would also need to be a few degrees ahead of the maximum ignition angle to give the processor some time to do the math(higher reving engine would need trigger to be further ahead).
I think the 'trigger advance angle' could also be included as a constant in the Excel sheet?

What do you think guys? Maybe it's not so critical for you? I guess the airplane engines operate below 10k?

I have been running the engine quite a lot on a stand now, and the ignition seems to work flawlessly exept from the delay-issue.