As some of you know, I have been working on an ignition upgrade called the "Transistor Trick". It appears that the next development is going to require adding some complexity to the design (microcontroller). However, having a microcontroller can potentially add a bunch of features. I will start on the new design after we get some things wrapped up with the current design. I won't be able to start prototyping, coding and testing until I return to the US (around October or so). What I am looking for is a list of what people want in an ignition system. All of the features will probably not be available in the beginning, but I would like to choose the micocontroller and design the board so that all the features can be added. Here are some things I have thought about:

- rev limiter
- possibly have a 2nd gen CAS as an optional input
- 1st gen ignitor as the other input option
- 2nd gen ignition on trailing if CAS is used
- boost retard for the turbo guys running a dizzy
- ability to create your own ignition advance curves (have to figure the software part out)
- output for a digital tach?

Some features would only be available with the Crank Angle Sensor (CAS) input.

I am not sure if there is something else people would want. I will do my best to add all of the features possible. One thing that I am thinking is to choose the parts and design the board based on all the features wanted. If people wanted a cheaper/simpler option, the extra parts could be left out on the board. The parts could be added later as an upgrade. Maybe do 2 boards: 1 with the main hardware, the 2nd for additional features.

The goal is to get a nice upgrade on the ignition system and keep the price as low as possible. Let me know what you guys think. Your input will go into the final design.

Thanks,

Kent

 

sounds pretty awesome you can count me in! the 2nd gen CAS is something i would diffently like!

 

Wow. I think this is definately something I'd be interested in. The features you have listed would be good for an initial system. Making it too complex with too many bells and whistles to begin with can cause problems.
I think those features are all that is really necessary though.

 

I think that is a great idea.

The 2 boards is a great idea, that way for us you just want it simple for now can have that and then, If or when we need it we can add the features that we want.

 

very nice idea not somthing the avarge rotary guru can do

 

Glad you guys like the idea. Basically, I will try to figure out all the hardware and layout of the board to have these features. The rest will be coming up with the code to do these features. Using the CAS and/or having boost retard would require a MAP sensor (manifold pressure). I would have to see what the maximum boost you guys want to run. Perhaps a 3bar sensor (be good for 2bar boost (~29psi)) would be good. The sensor that is used in the MegaSquirt ECU is a 2.5bar (good for 21psi boost). The 2.5bar is probably good enough. People running higher will probably have a standalone setup.

For the input from the J-109 ignitor, the circuit would be just like the TT (leading only). Perhaps could do some sort of fixed boost retard. Could have an output for a digital tach. The other features would be available with the CAS input. For this, we need a map of values (advance values given RPM and manifold pressure). Perhaps I could load some general map for NA cars and another for boosted. If I could figure out the software part, I could add a way to connect to a laptop and you could put in your own values.

So, I will see about the hardware and layout. I will include the harware needed to make it all work in the circuit, but all may not be active at first. I know the trailing ignition can be tricky. Also having an interface where you guys can put in your own values would come later. I will have to see the best way for connecting as well. The MegaSquirt uses a serial connection. That is maybe the easiest to do. The problem is many laptops don't have a serial connection anymore. I guess they do have USB-to-serial converters you guys could get if it is a problem.

I will think about it more and see what I can come up with.

Kent

 

Sound great cant wait to her more about this set up. I would be definatley interested.

 

WOW! Great idea!!

I would love to be able to move to the 2nd gen CAS and the 2nd gen trailing system. It just looks cleaner and I can get some junk out of the engine bay.

I have a cas and trailing coils if you would like to borrow them for dev work. Will the stock tach work with the 2nd gen cas?

 

Thanks guys. I'll start laying things out shortly. I am currently still on dial-up since I moved. Makes it hard to download datasheets, diagrams, and so on. The DSL should come soon. Basically I need a rough idea on what it will take to do all this and then select a microcontroller based on that.

I guess I will separate the two boards having what would function like the TT on one board. The second board would contain the parts to use the CAS, boost retard, and so on. This way the basic board should be around the cost of the current TT depending on the cost of the box, connectors, microcontroller, and so on. The add on board may be around the same cost as the main board depending on the cost of the MAP sensor. I think that the whole setup should be pretty reasonably priced. I think the most difficult part is making everything easy to install/use and to make sure the circuit is protected in case people connect things backwards.

Dan: I willl let you know. Thanks for the offer. I may pick them up locally because it may be easier that way. For the tach, you would need to move the stock tach connection over to that black wire that comes out of the 2nd gen coil (it connects to the leading coil -). I think that there is a spot on the 2nd gen trailing that you could connect to as well.

Kent

 

Kent, if you could find a way to have me put a 2nd gen CAS in and run 2nd gen leading and trailing coils, I'd pay a bunch for that. Not only would it be a great ignition upgrade, but it would also clean up my engine bay.

Why exactly does it take a MAP to use the CAS? (if you don't mind me asking, my understanding of the CAS is very limited)


Personally, not running boost and not caring too much about using a computer to map my ignition curves, I'd opt for just the simplest system that can take a 2nd gen CAS and run 2nd gen coils.... though I guess it would be nice to re-create the vac and centrifical advance of the stock dizzy... I guess that's where the MAP comes in eh? for the vac advance?

Jon

 

i have no understanding on the 2gen cas either please do explain

 

http://www.teamfc3s.org/info/articles/demystifying.html

 

Gah, my head hurts... but that's very informative!

Jon

 

Kent,

Sounds like you've got a pile of work ahead of you. Have you considered simply using the existing Megasquirt in spark-only mode? That would be way easier than starting from scratch. Granted there are some components on the board that aren't needed for spark contol, but a builder could always leave those out to save some cash. If after that experiment you decided a new rotary ignition-only solution was really warranted, you could make use of much of the open-source code and circuit designs to make up your own boards. In the end, I doubt you'll end up significantly cheaper than the existing Megasquirt, but you could reduce the form factor a bit.

Jon,

In the dizzy world, there are two advance mechanisms, mechanical (aka centrifugal), and vacuum. Pop the lid off the dizzy, and you'll see evidence of both of them. Mechanical advance adds advance with increasing rpm, firing the plugs sooner. A simple computer-controlled ignition system can easily emulate that advance scheme, as it's already counting teeth on a wheel, and knows the rpm from that input.

Vacuum advance is provided via a dashpot on the dizzy connected to a vacuum source on the intake. This provides added advance under high-vacuum conditions to improve efficiency. Under heavy load (low vacuum), this extra advance is removed to prevent detonation. An electronic ignition controller commonly uses a MAP sensor, as you suggested, to measure engine load and apply appropriate ignition timing corrections.

 

If I get this new job I'll have some extra funds and would definately want to help you with this to get in on the ground floor of it. It sounds like an awesome idea and I want to be one of the guys running it with how good it sounds like it will be. Especially if I do the 13b TII swap that I want.

Oh boy do I hope I land this new job.

 

Yup, I already know how Mechanical and Vac advance work quite well, I was just curious about how the MAP was used. I figured the MAP would provide the vacuum signal you could base vac advance off of, but it doesn't provide information on how fast the engine's spinning, so how can you replicate the mechanical advance? I guess your ignition controller would be able to figure out from the signals it gets from the CAS how fast the engine's spinning and compensate, which explains the ignition maps....


Jon

 

Jon: Basically as you have said. The MAP sensor is there to take care of vacuum advance. It is a measure of engine load. The RPM for the mechanical advance can be calculated from the teeth inside the CAS. For the MegaSquirt, there is typically an 8x8 table. You have 8 bins for RPM and 8 bins for manifold pressure. At each point you have a number that gives advance at that pressure/RPM. If boosted, it is the same thing, only you put some bins beyond atmospheric pressure (greater than 1 bar). So if you run NA, you have a more refined curve as the points are spaced out within a smaller range. I could preprogram with a curve that matches the stock setup or some other curve that works well. You could run without the MAP sensor, but it would be like running without vacuum advance connected. The MAP sensors are not very expensive, and the connection is easy (just take the vacuum hose that connected to the vacuum advance on the dizzy and connect it into the MAP sensor).

Roger,

I know what you mean. No one ever said it would be easy. Basically, I have found that we need a micocontroller or similar to control the pulsewidths well on the TT circuit. I figure it is worth a little more development and a few extra dollars to make it expandable. I figure I can have one board that would be the next development of the TT and have a daugtherboard to add the other features. This way people can upgrade in steps and not have to shell out a bunch of money at one time. I know how the budget is for most of us 1st geners.

I know that using the MegaSquirt would be easier, but it would be nice to make something as small and as simple as possible. I still recommend the MS to people who think they may want EFI in the future. It also gives me something to keep me busy and I know I will learn a lot in the process. I think the hardest part for me would be a software interface for people to make their own ignition curves.


Good luck on the job. I am sure I could use some help. It will be a big project, as renns was saying. I will do what I can on the design here so that we can get a good jump on the project when I get back. Once the circuit is figured out, I can start testing the hardware portion. Then start adding code. The next advancement of the TT would come 1st. Then probably the CAS, then trailing control, then rev limiter, and finally the software interface.

I do have a MegaSquirt back at home, so I can compare what it does for spark control to what this circuit will do for spark control. I know that it may not have everything the MS has going for it. I do hope to come up with something that is at least comparable and is smaller, simpler, and cheaper if possible.

Kent

 

Thanks.
I think my desire to throw money at things involving my car is the fact that I will finally have the money to throw money at my car. It's another sophisticated hi-tech LED computerized blinking light I can insert in my dash.
And it's not a ricer mod but will be a functional mod and be very helpful in tuning my car.

I've dealt with Megasquirt in building and tuning Volkswagen and it'd be awesome to run a megasquirt program on a 7.

 

I downloaded the EAGLE layout software. Took 2hr+ on this dial-up. I am starting to layout everything. I am following much of the V3 MegaSquirt. The V3 has a lot of protection features (overvoltage, surge protection, reverse polarity, ...). I am going to incorporate these features as well. One thing to think about is temperature rating of the parts. Normally the microcontrollers are rated 0C to 70C or from -40C to 85C. The second is called Industrial rated. It may be fine for us. However, there are a couple Automotive rated microcontrollers. They have a range of -40C to 125C rating. This may be what we want since this circuit will live in the engine bay. If I can find the microcontroller with that temperature range and that will do what we need, I can also get the remaining parts to handle those temperatures as well.

I am trying to select the connectors for the input/output. We want something that is common and supports the number of wires needed. I am thinking probably D-sub connectors (similar to like a computer serial connection).

Kent

 

I just got my powdercoated, uber-pullies from trochoid

I'll take pics later, but I FINALLY know without a doubt where my timing marks are, or at least I will once I get the pulleys mounted.

That means that I'll finally be able to time the car and install TTv1.

Honestly, add me to the list now, I want one of these CAS-run Ignition Microcontrollers with MAP sensor.

We just need a good name for it now. Transistor Trick was just too tricky to write all the time and TT got confused with "Twin Turbo", so when we name this damn thing, we should keep that in mind

Or you could just give it a person's name.. like Bob... or "Iggy the Ignition Controller"

Jon

 

Jon: Cool. Now you can time your car and have some more bling-bling too. Yeah, mount up the TT and see what you think. If we can pick a good ballast resistor, you can get the performance and MPG gain with the TT and not have to worry about the coil pack failure.

We do need a good name. Something simple would be good. Maybe something with rotary in the name since it will be a rotary only design. The persons name thing could be funny: "I just put Bob under the hood and now my car runs great" People would be like: : WTF?

Anyway, I think it would be nice to have two separate harnesses that could be plugged in (one for like the current TT operation, the other for CAS). I could try to set it up so the controller would detect which one is connected and switch between the CAS mode or the TT mode accordingly. Shouldn't be too difficult to do.

One thing that I am thinking about is the MAP sensor. If we have the MAP on the 2nd board and use one of the ADC in the microcontroller (1st board) to convert the signal, I worry that a poor connection, corrosion, etc. at the connector could degrade the readings over time. There could be a similar problem with the CAS signals. The MAP thing could be corrected by putting a separate ADC on the 2nd board and then send the digital output to the main borad. Not sure on the CAS signals yet.

Having two boards would be the easiest for you guys if we can make sure we keep a good signal quality between the boards. The other option would be to everything on one board. This would reduce the cost for the people getting the full setup. The people that don't get the full setup, I can leave out the extra parts. You could get the other parts later and solder them in or you could send the board to me and I could install for you. The disadvantages of using a single board are the more difficult upgrades (in terms of installing things and if I came up with additional features in the future) and the size would be bigger (in terms of width/length). I will see what I can come up with.

Anyway, I think it should be an interesting project. It will be cool once I can really start coding/testing. I will put up some board layouts as I get them made. It is a little slow since this is a new program for me.

Kent

 

Well, I have the power supply, input/output connection, and communication laid out. Right now, I am just going to use a serial connection for the communication. It is the easiest to do. USB would be cool, but probably harder to implement. It looks like I could fit everything on a reasonably sized board (3.25"x4") and not have to use a 2nd board. That would mean that when you upgrade, you would either have to solder the couple parts yourself (real easy) or you could send me the board and I could do it for you. The adavntage of this would save the cost of the extra connectors, extra board, and the possible signal degradation through the connector with age. If I add plenty of jumper pads for all the unused input/outputs, power, gnd, and so on, we could still add a 2nd board if needed in the future. I will go with the single board setup for now. I can then try a 2 board design if we think it may be better.

The design is iterative. Basically, I lay out the parts and see how big of a board I need. Then I need to find a box that will fit that board, then adjust the board size/layout depending on what is available. I am currently using a 15 pin connector on the input/output. This way it can't be confused with a normal serial port or something. I figure we will use about 5 pins for the regular TT operation, and maybe 8 or 9 for CAS operation. The remaining could be used for future upgrades and/or a setup to determine what harness you have plugged in (CAS or normal TT).

Kent

 

I have started looking at the code and working more on the layout. It looks like it may be a tight fit on the maximum board size that this software can do (free version). I will see if I can make everything fit, or separate between two boards, or see how much the pay version of the software is. I still like the idea of two boards, but I think signal integrity and total cost would be lbetter with a single board. It is up to you guys. If you don't mind soldering a couple parts or shipping the board to me for the upgrade, we can go the single board route. If you want plug-n-play for the upgrade without the extra hassle, we will go with the double board setup (this version would be more expensive because of the extra connectors and that two boards cost more to make than a single larger board).

Kent

 

Seems to me that a single board with sockets for the upgrade chip and wiring would be ideal.

Would there then be different PICs with different code versions for different options, or would a single PIC be able sense the options and run the appropriate code?

Regardless, I'd prefer the megaboard version, with all the bells and whistles for a bone stock 12A n/a car.

 

Thanks for the input, Dom. There would be 2 ICs (possibly a 3rd) that would be added for the CAS operation. Also, a pressure sensor to measure manifold pressure would need to be added. This would need to be soldered. Basically, if I go with a single board, it will require some soldering when people upgrade. They can either do it or could sent it to me to solder.

The two board setup would be more plug-n-play for the upgrades and would have a smaller footprint. The cost would be higher, though. It just depends on what people want. If most want the full setup from the start, then the whole 2 board/1 board issue does not matter.

I am basing the design around the Atmel AVR microcontrollers. These are powerful RISC based controllers. I will have a serial connection on the circuit so you will be able to download and install any new versions of the code. There will be just the one chip no matter the mode of operation. There are many input/output connections on the AVR that we won't be using. I plan to set a few of these lines up to detect the mode of operation. There is be a connector on one end that will either be wired for TT style operation or CAS style. Depending on the pins that are connected, the AVR will be able to detect which setup is plugged in and run that mode of operation. That is the plan, at least. I am trying to select the exact controller. I have found some that are full automotive grade (high quality, -40C to 125C working temperature range). The only problem is that they appear to be only available in a surface mount setup. This makes it harder to solder, may make things a bit difficult when I first program them, and difficult to replace is the controller were to get damaged. I would like to get the DIP design if I could find one. If we are stuck with the surface mount style, I could make/buy a small board to adapt the surface mount to DIP style, or just solder directly and if you ever have a problem with the AVR, you could just send the whole box to me so I could swap it out.

Anyway, just let me know what you guys think. I know it depends on price and all that, but let me know if you think you would like the basic board and then upgrade or the full board from the start. If you want the basic board and upgrade later, let me know if you are comfortable with soldering or not. I want to make it as simple and as functional as possible. If I can find the connectors that will mate up with the CAS and the 2nd gen leading/trailing pack, this could make it a very simple and quick install.