Programmable Datsun Dizzy

[brodster]

After reading quite a few books on fuel injection systems, I have been kicking around the idea of designing my parts for my Datsuns. There are quite a few things on the list (and I would appreciate other ideas) but here is the one I really want to do first:

 

A fully-programmable ignition module for L-Series electronic dizzies. I realize there are already a lot of options out there. However, I am looking for a unit that fits somewhere in between using the stock dizzy with very limited control and a distributorless ignition system. I have kicked around ideas of stripping out a dizzy completely and doing things like optical triggers, etc. but that defeats the purpose of this project. Plus they already exist (except not for the Datsun that I can find). I want as little modification as possible and something cheap.

 

Basically, the unit would replace the matchbox module that is mounted to the dizzy. I would also remove the vacuum advance mechanism and replace it with a MAP sensor, which would feed to the device. The coil inside the dizzy would provide a base trigger point for the new module. Then, just like most EFI controllers, the device would compare the inputs against a user-defined advance table. Then it would trigger the coil accordingly. Of course, it would be a cinch at this point to let the user control dwell time and set a rev limit.

 

I have spent the last day thinking about problems with this design and if they are relevant. The most important issue is that the unit will have to “guess” when to trigger the coil. As you all know, the stock setup uses a mechanical advance and vacuum to adjust the trigger points. However, the electronic version needs to measure RPM and try to determine where the appropriate angular position for triggering resides. This shouldn’t be a problem if the RPM stays constant, however quick changes in RPM (e.g. a hard downshift) would cause slight error. (Please correct me if any of the following is wrong, as I am a bit new to engines).

 

Your typical 4-cylinder dizzy, unfortunately, will only trigger the coil four times per 2 revolutions of the crankshaft. So this means the crankshaft moves 180 degrees between trigger points. From the point of view of the simple programmable unit, this means that the current speed of the engine is only measured four times per 4-stroke cycle. This only accounts to a few milliseconds, but it MAY cause a problem when driving a car hard. (I guess this also explains why a tooth-triggered crank pulley has MANY teeth in comparison). I also figured that a mechanical advance also has some “lag time” before the weights inside the distributor can reach a stable position when RPM changes rapidly. Haven’t done the calcs on that yet. Any experience here would be much appreciated. I also considered using the 6-cylinder dizzy as a base unit. This would improve things a bit.

 

The only other major issue is the necessary mods to the dizzy. I would of course have to lock down the mechanical advance and vacuum advance plates. I would assume the easiest thing to do for the mechanical advance is to find a spare cam and JB-weld the slots so the weights can’t move. Then it would also probably be a good idea to lock the plates down that move from the vacuum advance. I have disassembled a few dizzies for overhaul, but some advice on how to mod it without destroying it would be much appreciated.

 

My final question to you guys is whether you think it is worth the effort? I have plenty of other projects to work on, but this seems like it could be really useful and fun. Plus if it worked out, I could put up design info on Ratsun or sell some to your guys. Like I said I am trying to keep the price tag cheap, like under $100. I think anymore than that, people would be better off just buying a distributorless ignition kit.

 

One request: I am looking for spare “matchbox” modules to experiment with. My first step is to reverse engineer the thing . I am hoping it will be easy to overhaul. If I figure that out I will post that up in the how-to section. From there, I will use it as a base for my design. So if you have any broken ones laying around, let me know! I really don’t want to cut any working modules up. These things are getting much harder to find.

 

Thanks for reading all that!

 

 

[datzenmike]

Plus they already exist (except not for the Datsun that I can find).

 

The late 280zx uses a Crank Angle Sensor instead of an EI dizzy. I know it's a 6 cylinder but it's the ECU that figutes out the 6 firing positions.

 

 

I have spent the last day thinking about problems with this design and if they are relevant. The most important issue is that the unit will have to “guess” when to trigger the coil. As you all know, the stock setup uses a mechanical advance and vacuum to adjust the trigger points. However, the electronic version needs to measure RPM and try to determine where the appropriate angular position for triggering resides. This shouldn’t be a problem if the RPM stays constant, however quick changes in RPM (e.g. a hard downshift) would cause slight error. (Please correct me if any of the following is wrong, as I am a bit new to engines).

 

 

Electronics work in the billionths of a second so don't worry it can more than keep up. At 7,000 RPMs the dizzy is only firing 233.3 times a second or fires every 4.25 thousandths of a second. How many billionths of a second pass between each plug firing ? Well it's 4.25 million. Lots of time to adjust.

 

The only other major issue is the necessary mods to the dizzy. I would of course have to lock down the mechanical advance and vacuum advance plates. I would assume the easiest thing to do for the mechanical advance is to find a spare cam and JB-weld the slots so the weights can’t move. Then it would also probably be a good idea to lock the plates down that move from the vacuum advance. I have disassembled a few dizzies for overhaul, but some advice on how to mod it without destroying it would be much appreciated.

 

Just remove the weights and springs on the mechanical advance and disconnect the vacuum line or remove the vacuum advance mechanism itself.

 

My final question to you guys is whether you think it is worth the effort? I have plenty of other projects to work on, but this seems like it could be really useful and fun. Plus if it worked out, I could put up design info on Ratsun or sell some to your guys. Like I said I am trying to keep the price tag cheap, like under $100. I think anymore than that, people would be better off just buying a distributorless ignition kit.

 

The drive spindle is driven off the end of the crankshaft and is connected to the dizzy somewhat like a screwdriver slot. This is fine when designed back in '68 but isn't accurate enough when splitting a circle into 360 parts. (or more) It is going to 'wander' several degrees from the slack. The later CAS are splined.

 

L series slot drive spindle.

 

 

New style CAS drive spindle

CAS dizzy

 

 

One request: I am looking for spare “matchbox” modules to experiment with. My first step is to reverse engineer the thing . I am hoping it will be easy to overhaul. If I figure that out I will post that up in the how-to section. From there, I will use it as a base for my design. So if you have any broken ones laying around, let me know! I really don’t want to cut any working modules up. These things are getting much harder to find.

 

Well here is some info that may help speed this up. The matchbox used on the 280zx and the Maxima have a modified dizzy matchbox on them. I know of someone who used one of these boxes on his L20B dizzy and as he revved up the timing would retard 20 degrees. So while the mechanical was advancing the overall timing was retarding. Eventually someone noticed the wrong matchbox on it. Anyway, the 280zx box has an extra two wire plug on the side and perhaps the 280zx ECU uses this to adjust the timing??? Anyway here's a picture of the matchbox..

 

E12 93 with extra plug lower right side

[DanielC]

Honestly, this is where you should go. Get a controller that will talk to the Ford EDIS system, that will meet your budget concerns.

Even in the 1970's Nissan, then Datsun saw the advantages of using a crank fire ignition system.

No oil pump interference, no slop in the gear drive between the crankshaft, no slop in the mating of the distributer to the oil pump drive spindle.

 

I believe the Megajolt Jr lite will control the timing. It is out of the price range you mentioned, but it is an option. It does have an input for a MAP sensor.

[brodster]

Thanks Mike. I appreciate your help. Very good info!

 

The late 280zx uses a Crank Angle Sensor instead of an EI dizzy. I know it's a 6 cylinder but it's the ECU that figutes out the 6 firing positions.

 

I didn't know about this one. I will check it out. Do you know which years? I will look for the manual.

 

Electronics work in the billionths of a second so don't worry it can more than keep up. At 7,000 RPMs the dizzy is only firing 233.3 times a second or fires every 4.25 thousandths of a second. How many billionths of a second pass between each plug firing ? Well it's 4.25 million. Lots of time to adjust.

 

On the "elapsed time" between firings, I probably wasn't clear with my thoughts. I am not worried about the microprocessor or electronics keeping up. In fact, I will probably design it so the microprocessor is just barely overpowered. That way it will be cheap and small.

 

What I am concerned about is how much input data is available for the microprocessor. In the case of the 4-cyl dizzy, that thing will only provide a new data point every few milliseconds ( which is of course an eternity for microprocessors). Given that the microcontroller only has input from the coil and the vacuum, this would mean the microcontroller will have to determine the correct firing position based on time and engine speed. The speed would be calculated from measuring the elapsed time between each pulse from the coil (unless impulses from the MAP sensor can be implemented) and knowing that the crank has travelled 180 degrees. This is fine easy driving, but I was wondering how quickly an engine will change RPM during a hard downshift or upshift. I did some research and found out that rpm per second of Indy cars can exceed 25000 (not sure if this accurate). I am not sure what happens in your typical car but if a sudden downshift causes the engine rpm to reach those numbers, this would mean a change of about 100 rpm per few milli seconds. Basically, the microprocessor will be calculating a firing angle based of old data. Or in other words, the microprocessor is lagging behind actual speed because the input data isn't provided fast enough. I don't think it would be a huge problem though. But I am not 100% sure.

 

 

Just remove the weights and springs on the mechanical advance and disconnect the vacuum line or remove the vacuum advance mechanism itself.

 

Do you think the little plate piece will float around without everything secured? This was my only concern.

 

The drive spindle is driven off the end of the crankshaft and is connected to the dizzy somewhat like a screwdriver slot. This is fine when designed back in '68 but isn't accurate enough when splitting a circle into 360 parts. (or more) It is going to 'wander' several degrees from the slack. The later CAS are splined.

 

 

This is good info. Probably posted somewhere, but are these easily mated to an L-Series??

 

 

Well here is some info that may help speed this up. The matchbox used on the 280zx and the Maxima have a modified dizzy matchbox on them. I know of someone who used one of these boxes on his L20B dizzy and as he revved up the timing would retard 20 degrees. So while the mechanical was advancing the overall timing was retarding. Eventually someone noticed the wrong matchbox on it. Anyway, the 280zx box has an extra two wire plug on the side and perhaps the 280zx ECU uses this to adjust the timing??? Anyway here's a picture of the matchbox..

 

E12 93 with extra plug lower right side

 

I have seen these modules and thought I read somewhere about that 20 deg retard issue. Anyways, in seing the pic, I may not even have to tap into the whole box. After figuring out how it works I could probably just tap right into that port...maybe. That would be a slick install. The worst case is that I could post up an article about how to use these boxes in place of the normal ones. I will have to get some of these to play with. THANKS!

[brodster]

Honestly, this is where you should go. Get a controller that will talk to the Ford EDIS system, that will meet your budget concerns.

Even in the 1970's Nissan, then Datsun saw the advantages of using a crank fire ignition system.

No oil pump interference, no slop in the gear drive between the crankshaft, no slop in the mating of the distributer to the oil pump drive spindle.

 

I believe the Megajolt Jr lite will control the timing. It is out of the price range you mentioned, but it is an option. It does have an input for a MAP sensor.

 

I agree with you. The crank fire system is by far superior. This is more of a experiment and a cheap alternative to whats out there.

[brodster]

I guess on that note, would anyone want something like this? I know lots of people use the electronic dizzies. Figured it would be a cheap way to upgrade without going all the way to crank triggered units etc.

[hacked521]

as long as the design is as good as crank trigger, can take a boost reference and is cheaper or same cost as crank trigger i think i would be interested. :cool:

 

how does the cas type spindle mate up to l series type stuff?

[datzenmike]

Good question. This spindle is from the Z24.

 

Two are from KAs and one from a Z24, the rest are Z20/Z22. The long ones...I forget. L20B???

 

 

Looks like the end could be welded onto an L spindle. The spline part on the dizzy has a set screw to hold it on

 

 

 

 

Hell use a KA spindle and a KA CAS. This gives a signal every degree instead of every 180.

.

[Sealik]

 

 

 

 

 

.

 

From the left....3 KAs......3 of Z22/Z24 or Z20....2 LBs...?

Not sure if there is a Z20 in the mix?...Maybe No. 5 then

[datzenmike]

Pretty sure the Z20 is the same as the Z22 and I have 2 Z20 spindles for sure. The Z dizzys are longer than the L20B ones I think.

 

Doug If you know that a Z24 dizzy will fit a Z22 then all the short ones are KA and various Z ones.

[Sealik]

Pretty sure the Z20 is the same as the Z22 and I have 2 Z20 spindles for sure. The Z dizzys are longer than the L20B ones I think.

 

Doug If you know that a Z24 dizzy will fit a Z22 then all the short ones are KA and various Z ones.

 

Ya.... the Z24 dizzy will fit the Z22....spindles are also the same.

[datzenmike]

Good I was fairly sure. The three short non spline ones are Z20 (2) and either a Z24 or a Z22. The long ones must be L20B.

[brodster]

I love this forum! All sorts of new things I haven't even thought of.

 

My original plan was to mod an L-Series dizzy, but I am liking the possibility of using the KA's etc. The 280zx turbo dizzy from 82 and 83 is also looking promising. It appears that it is completely optical.

[brodster]

as long as the design is as good as crank trigger, can take a boost reference and is cheaper or same cost as crank trigger i think i would be interested. :cool:

 

how does the cas type spindle mate up to l series type stuff?

 

 

Well IF I used the old L-series dizzy as a base, it will only be as accurate and precise as the original dizzy. So given that there is a bit of slop, this will transfer right into the electronics: Garbage in, garbage out. In that regard it would never be as good as a crank trigger.

 

But, it would be a cheap way to customize the curves on a dizzy. I don't see any reason it couldn't sense boost at the manifold.

 

 

At the same time, I am starting to see lots of good ideas here that doesn't rely on the old L-series dizzy. I don't know if they would be as good as a crank trigger, but they will be pretty darn close.

 

 

[Unclejesse88]

search for ghetto efi and you will see a guy mate up a KA distributor to an L16 front cover with an adapter.

[brodster]

search for ghetto efi and you will see a guy mate up a KA distributor to an L16 front cover with an adapter.

 

Thanks for the heads up. I will check that out first thing in the morning.

[Pedro]

I just skimmed the thread so forgive me if this is not relevant.

 

But here is an SR CAS made to fit an fj20 block with a small aluminum spacer

 

 

 

Edit - just realised I said dizzy, meant CAS

[brodster]

I just skimmed the thread so forgive me if this is not relevant.

 

But here is an SR dizzy made to fit an fj20 block with a small aluminum spacer

 

 

At this point all is appreciated. My engine knowledge doesn't go much deeper than L-Series (and I am pretty new to these too). Before I started this thread, I wasn't thinking of any other dizzy but the matchbox style ones. However, these posts have got my mind thinking up new stuff.

[brodster]

 

Hell use a KA spindle and a KA CAS. This gives a signal every degree instead of every 180.

.

 

 

Okay, probably a dumb newbie question but does CAS stand for "Crank Angle Sensor" in this circumstance?

 

Mike, since I haven't seen inside a KA dizzy, what enables it to be triggered every degree? I am assuming it has a encoder/optical disk or something? Thanks

[brodster]

Okay looks like I answered my own question:

 

[brodster]

To go even further, I just found a megasquit article about a KA24e and how to trigger off the optical sensor. They already have a simple circuit: http://forums.nicoclub.com/megasquirt-write-up-for-ka24-series-engines-t221556.html

 

I think this will be the way I go for this project. It will provide WAY more accuracy and precision, require less modification and use dizzies that are more common. Looks like all that somebody would need to, in addition to the controller, is get an adapter plate built. Maybe I can get RacnJsn95 to build some. Thanks for all the help.

 

 

[hacked521]

sounds like a good plan, are you going to use this on an L20b? if so this will work out as a perfect guide for me :D .

[hacked521]

i found a couple links that might be of some help to you.

 

http://www.datfusion.com/forums/index.php?showtopic=86

 

http://ozdat.com/forum/viewtopic.php?f=32&t=19239&start=15

[brodster]

sounds like a good plan, are you going to use this on an L20b? if so this will work out as a perfect guide for me :D .

 

Initially, it will probably go on my L20b. But who knows. Could go on the L16 too.

 

Can't promise I will get started on this soon, but yeah I will post progress so others can try it too if they want. Also, when I get circuit boards built, you usually have to get multiple boards made. So I would be willing to part with these for pretty cheap.

 

Either way, I figure this will be mounted in my garage for awhile. I will probably attach the dizzy to a drill or something and build the whole system up my workbench until it all works.

 

 

 

 

i found a couple links that might be of some help to you.

 

http://www.datfusion...hp?showtopic=86

 

http://ozdat.com/for...=19239&start=15

 

Now I have more ideas. I didn't even think of taking another ECU and reprogramming the EEPROM. Using the 200sx ECU has always concerned me because it was designed for a specific motor. That would be awesome if you could use the computer and reprogram it at will. That would be way cheaper than megasquirt! Thanks for these links.

[510rob]

http://www.bgsoflex.com/mjl/mjl_edis_summary.html