Suzuki GT750 EFI Special [Archive]

Kettlekev

30th December 2020, 22:04

I am new to the forum so bear with me on posting issues.

Over two years ago, having just restored an old GT750, I decided it might be fun to take one of those bikes and push the design as far as I could.

Two years down the line and pending paint and final EFI work I have this.....

348201

I wanted a suspended and tucked engine so started with a VFR 800 frame and designed new suspension/engine mounts to graft in the SS swing arm.

I hated the front radiator on the GT so found a nice underseat Benelli Tornado option which meant adding the Benelli subframe and fans.

Front end is Suzuki GSXR 750.

Clocks are suzuki Bandit ....and there are various other bits from around 13 bikes now.

348202

The main bike is Arduino powered (including the clocks) and includes;
RFID ignition lock
GPS
Digital speedo and tacho
Electric water pump and fan management
Proximity sensor alarm
SD card logging
Real time clock

The engine mods are broadly speaking;

Ported near to TR750 specs
Reed valves
Dry clutch
Close ratio gear box
Custom exhausts
Replaceable (tuneable) squish heads
Zero deck O ring head

For its first dyno run I have carbs fitted but have started an EFI project with a speeduino.

I have also managed to get a set of Triumph Tiger throttle bodies working with fly by wire throttles so plan to use them in the EFI.

I've got to the point with the speeduino where the basic ignition is kinda working except the spark modes I can find dont allow me to run 3 independent ignition channels which is my first hurdle.

I've done lots of reading especially on TZ350 excellent thread here (and still have lots to read) and am just getting into the fuel planning for the EFI. I believe the summary of what I have read includes;

1. Although my engine is relatively low revving (8.5k peak power), throttle body injectors probably wont work so I need transfer port injecting
2. The Kettle has two ports so do I need one injector per port? I assume so?
3. Best position is probably to squirt under the piston
4. For smooth running I may need to consider TZ350's pseudo MAP approach, so a MAP sensor in one cylinder under the inlet reed is probably best?
5. From a injection timing perspective, is it better to start injecting on the down stroke or wait until BDC and then inject until the transfer closes to get best mixing?
6. I probably need an injector flow of around 300cc
7. Given my rev range I shouldnt need staged injectors (small/large) although bottom end response might benefit from it.
8. With my Spark and VE tables on my Alpha N system, keep the steps leading to peak power fairly short

As I get into the injector choice/setups I am sure there will be more observations/questions, the first one that springs to mind is how much fuel am I trying to inject at each squirt and what calculates it?

I will post some more detail of the build thus far and the EFI work as it goes on.....Hope this helps...

Bonez

31st December 2020, 05:15

Welcome to the crazy zone.

Nice work you've done there.

TZ350

31st December 2020, 08:19

.
Hi Kettlekev. Looks a very interesting project. I am impressed with the quality of your work.
.

From a injection timing perspective, is it better to start injecting on the down stroke or wait until BDC and then inject until the transfer closes to get best mixing?

Think in terms of where to end the injection cycle. Setting the "Ending" somewhere between BDC bottom dead center and TPC transfer port closing seems to work best.

The engine will run smother and smother as the time (measured in crankshaft degrees of rotation) increases to the point where the injection cycle covers the time the transfer ports are open. Use the smallest injector possible to get the longest injection cycle that will fit inside the maximum RPM time limit. The injection cycle could eventually stretch out to nearly 360 deg of crankshaft rotation at say 9k RPM. An injection cycle includes opening, injecting, closing.

348207

These are two injectors in a 350cc Kawasaki singles crankcase "C" port wall. They are fired together as one logical injector. This injector position worked very well in the big single.

348206

The drill indicates the firing position of my transfer port injection in the air cooled 125. This combined with under piston injection for staged injection also worked very well for a high revving 125. And is my favorite but hard to replicate on a water cooled cylinder.

348208

This is my "B" port and "C" port under piston injection on a water cooled cylinder.

The "C" has the small injector and the "B",s are fired together as one big logical injector for staged injection.

KTM's does their staged injection by alternating the "B"s and then firing them both together when they need a bigger logical injector.

This works because there is always some residual fuel floating around in the crankcase. It seems that it takes something like 6 to 10 crank cycles to completely flush the crankcase. So injecting just maintains a working average of air/fuel ratio in the crankcase.

Personally, for the GT750 engine I would use a single injector a cylinder and look to copy the 350cc Kawasaki or fit the injector in the reed block between the reeds and crankcase.

husaberg

31st December 2020, 16:08

Cool bike.......
I just boiled the kettle
It reminds me of a modern silk/scott

Grumph

31st December 2020, 18:52

Very nice work indeed.

Does your engine still have the oil draining crosslinkage between the crankchambers ?

I'd be tempted to block that off on the grounds that it could produce conflicting or unreliable signals on your sensors.
It may not of course. I'm not qualified to judge. But that feature is I believe unique to those engines.

Kettlekev

31st December 2020, 21:42

Thanks for the feedback guys...

So I guess throttle body injection even with reeds just isnt going to get the mixing right even though I am lower revving but squirting just the other side of the reeds and still into the inlet port might work (if I am understanding TZ350's post)?

I have a spare barrel so will start looking at places to inject. I think there may be some space to do it more like the Kawa 350 so will post some photos later. I will also try to squeeze a MAP sensor feed in to one cylinder just in case I need it. I guess the sensor can be remote (end of tube)?

My current issue is with Speeduino/Tuner studio in that there are settings that allow the injectors to fire in sequence (1,2,3) at 120 degree intervals but the spark options (for crank only sensing) either pair the ignition circuits (1&3, 2&4) or put them all on cylinder 1, so I cant get 1,2,3 in sequence. I need to get this one sorted first.

Yes I am still running the SRIS oil scavenging system and the mechanical pump, it is way better than premix and I dont think it will mess with any sensing but we will see.

I clearly need to get my head around some injector volume/timing maths but will do that in another post......

Kettlekev

1st January 2021, 04:24

Having played with Tunerstudio/Speeduino settings a bit I think this is the only way to fool it to be a multi (3) cylinder 2 stroke?

Any thoughts anyone?

348217

TZ350

1st January 2021, 09:26

Having played with Tunerstudio/Speeduino settings a bit I think this is the only way to fool it to be a multi (3) cylinder 2 stroke?

Any thoughts anyone?

348217

I have only worked with single cylinders but 6 cylinders and wasted spark looks good for (a single trigger and) 120 deg firing order for a 3 cylinder two stroke.

TZ350

1st January 2021, 09:48

.

If you do decide to try to work with crankcase pressure as a pseudo MAP reading.

For a crankcase pressure sensor you could use the same type (2.5 bar MPX4250) pressure sensor that the Speeduino uses.

I use a 2.5 bar sensor left over from my Ecotrons days.

348223 Mount with a short tube 2 to 3 cm long. Angle it down so that it drains.

My sensor puts out a reading every ms for 1 ms. I expect the MPX is the same. You will probably get a reading like this.

348222 1,800 RPM.

348221 10,000 RPM.

TZ350

1st January 2021, 13:00

So I guess throttle body injection even with reeds just isnt going to get the mixing right even though I am lower revving but squirting just the other side of the reeds and still into the inlet port might work (if I am understanding TZ350's post)?

Yes ......

Flettner

1st January 2021, 21:29

Dont waste your time and money with TBI.
If want to introduce the fuel at the TB, use a carburetor

Kettlekev

1st January 2021, 22:42

husaberg

1st January 2021, 22:49

Thanks guys. Port injection it is then.

I will hack my development barrels and post some images.

I am trying to get my head around fuel load/injection timing now and will post a spread sheet later.

BTW does anyone know a link to a Keihin injector spec sheet website? I cant see anything on line. I have some GG12 (pink dot) injectors and would love to know the specs.

https://www.injectorrx.com/fuel-injector-cleaning-and-flow-testing-service/fuel-injectors/fuel-injector-data/keihin-fuel-injectors/

gg12 is showing up much? any other numbers?

Kettlekev

2nd January 2021, 00:01

Thanks, I did see that one but I think it is for a different injector type. Mine come from a Tiger 1200 and the only number on it is GG12. Will keep digging.

Just for fun I have been trying to understand injector maths....I am sure there is a simpler way to do this and I am probably missing something major but with the end goal to define when to stop injection at the various rev numbers, I came up with a auto calculating excel sheet. I cant seem to attach it? but here is a screen shot.

Caveats are;
1. I dont have definitive injector specs yet
2. I am not sure of the required fuel number (see comment)

Anyway I am sure someone will correct or it may actually be a useful exercise??

348226

If anyone is interested I can email it unless there is another way to attach?

husaberg

2nd January 2021, 10:46

https://injector-rehab.com/knowledge-base/flow-rates/

*Disclaimer about Keihin numbers. These numbers do not represent a Part Number, but yet a manufacture date and batch codes. There is no way to tell the flow rate of a Keihin injector by the number alone!

email these guys with a few pics of your bike and i am sure they can help they have most others listed.
http://www.factorypro.com/about/indexabout.html
the other option is just measure them yourself.

https://www.youtube.com/watch?v=bgUAxgQcvis

https://www.youtube.com/watch?v=jvfyVgqHWpk

https://www.youtube.com/watch?v=e_g9K0VEsfE

or just get someone to test them..

Flettner

2nd January 2021, 12:17

The TPI YZ 250 ran two 510 injectors. Or perhaps they were 520, NZ EFI has them on the shelf, many sizes infact.
My E85 fueled 350 Kawasaki runs two 700's as an example.

TZ350

2nd January 2021, 20:03

https://www.youtube.com/watch?v=e_g9K0VEsfE

or just get someone to test them..

If you are testing injector flow, best to use kerosene, much less flammable than petrol. Also a respirator is a good idea as the injector also creates a fine mist that is pretty hard on the lungs. Been there done this home made testing thing.

husaberg

2nd January 2021, 20:27

If you are testing injector flow, best to use kerosene, much less flammable than petrol. Also a respirator is a good idea as the injector also creates a fine mist that is pretty hard on the lungs. Been there done this home made testing thing.

using mineral spirits is on i think the third one as an added bonus it cleans out the varnish.
When they do the proper testing they use a reference solvent N heptane or something.
what was the sizes you ended up using TZ i see they do 25cc ones now

Kettlekev

2nd January 2021, 21:24

That's great thank!

The main thing I am trying to sort out is dead time (open/close). Keihin have a decent UK office so I will ping them and share what I get.

Any though on the inject end timing spreadsheet? Is there an easier way to do. Is there any way to attach and Excel sheet to the forum?

New scope probs on the way now so I can confirm the Tunerstudio/Speeduino hack for triple ignition and injector timing :yes:

Cheers

TZ350

2nd January 2021, 21:33

.
My 125 running Av Gas and 20:1 oil uses two 128g/min injectors and one 60g/min. 60g/min for low speed and all three for high speed.

The Speduino software allows you to select the injection end point in crank degrees.

The actual amount of fuel injected remains pretty constant for any engine Load with changes in RPM.

But the start of injection automatically happens earlier and earlier in terms of crank degrees as the RPM increases and the time for a crank revolution decreases.

TZ350

2nd January 2021, 21:47

TThe main thing I am trying to sort out is dead time (open/close time)

Easy to get a good handle on the dead time. Arduino is your friend here.

Run an injector at 100% for a time then pulse it at 50% for twice the amount of time. The difference between the two volumes is the dead time.

Kettlekev

3rd January 2021, 05:13

TZ350, did you take a look at the excel?

I have tried to make it so you input certain things such as port timing, injector characteristics then it spits out end inject degrees based on RPM.

I understand allows end time input in degrees and then adjust inject time but how does it know the injector characteristics or does it just assume?

Just trying to get a feel if the excel is a useful tool or not. Problem is the forum wont allow the upload as far as I can see?

TZ350

3rd January 2021, 07:33

348232348233

I use injectors left over from my old Ecotrons kit. Dead time was 1.6ms.

TZ350

3rd January 2021, 07:38

I understand allows end time input in degrees and then adjust inject time but how does it know the injector characteristics or does it just assume?

Sorry, no I have not looked at the spreed sheet.

Speeduino has a place for entering the injector characteristics. I will see if I can grab some screen shots.

TZ350

3rd January 2021, 08:01

.
You are using a latter version of Speeduino firmware to me but I expect the basics are the same.

Injector characteristics:-

348235348236348237

This is 125cc single two stroke, single stage with a 330cc 1.2ms dead time injector. (330cc could be two 165cc/min Physical "B" port injectors fired together)

Speeduino assumes 12V and high impedance, ie normal injector not peak and hold.

TZ350

3rd January 2021, 08:35

.
Speeduino has a BlueTooth connection option so you can log data to a smart phone and review it later on your laptop. This is handy because you can have the phone in your pocket and data log the bike under real riding conditions.

You need a Speeduino BlueTooth card and the EFI Analytics ShadowLogger for TuningStudio http://www.tunerstudio.com/index.php/shadowloggerms

348241 TunerStudio for tuning the Speeduino EFI firmware and logging the basic results.

348239 MegaLogger for logging and analyzing the results from TunerStudio changes to the Speeduino.

348240 ShadowLogger for logging on your smart phone.

348238 BlueTooth connectivity to a smart phone or laptop.

I use Bluetooth connectivity to my laptop for tuning with TunerStudio and data logging.

And a USB cable for any Speeduino firmware upgrades/changes ie if I have fiddled with the Speeduino source code.

TZ350

3rd January 2021, 09:31

.
For anyone that is interested in how the Tuner Studio and Speeduino Software connects together:-

There are three parts.

348242 (1) The Speeduino firmware. Source code written for the Arduino hardware.

348243 (2) The Speeduino Configuration file that tells TunerStudio what the TunerStudio screen and IO input/output should look like.

348244 (3) TunerStudio paints its screen the way Speeduino Configuration wants it to look and then TunerStudio works in the background and handles the Human Interface with the Speeduino Firmware. This saves fuel injection software developers from having to do all the Human Interface development too. Human Interface is often the biggest part of a software project. Tuner Studio can be adapted by developers for their own EFI projects just by writing a Configuration File.

Kettlekev

3rd January 2021, 21:15

Cheers TZ350.

I've got registered Tunerstudio and Megalog viewer. Will look at the bluetooth link.

I saw the injector characteristics stuff. So that deals with deadtime and flow.

Also in the constants/required fuel tab I assume the two numbers in the top are in relation to injector opening times/volume of fuel delivered and are basically calculated from the engine displacement etc. entered in the Required Fuel sub menu, so to define the volume of fuel delivered you dont need to know HP/RPM? Or is this all just a starting point? Why are two numbers calculated?

348255

The thing that seems to be missing from all this setup is the injector end times calculator in that you have to manually enter an end time with the graph? That is what I was trying to work out with the spreadsheet. I must be missing something here? Does Tunerstudio help with that as well?

Sorry, too many questions. Once I get my head around fuel calculations. I will shut up until I have some more mechanics done and some test data.

I wish I could find a way to attach the excel as I think it might be useful in calculating injection end times.....

TZ350

4th January 2021, 08:10

Also in the constants/required fuel tab I assume the two numbers in the top are in relation to injector opening times/volume of fuel delivered and are basically calculated from the engine displacement etc. entered in the Required Fuel sub menu, so to define the volume of fuel delivered you don't need to know HP/RPM? Or is this all just a starting point? Why are two numbers calculated?

348256

I am not sure why there are two numbers there. You are right, they are calculated by the system. I have not paid them much attention other than to know that they are the base fuel for the VE and Alpha-N fuel mapping.

I also once tried to figure out if they included the dead time or not by comparing them to a log file of injector pulse width. But I have forgotten the answer. It does not matter much.

348257

By specifying in the "Engine Constants" that the "Control Algorithm" is "TPS" you change the "VE" table to Alpha-N.

I think this is how it works.

The numbers entered in the table multiply the "Required Fuel" i.e. 100 = 100% of the Required Fuel amount. 60 = 60% of Required Fuel amount and 220 = 220% of the Required Fuel Amount.

Me, I just throw numbers at the map. But if its important to you. You should maybe check this tech detail with people on the Speduino user projects forum. The map numbers can be anywhere between 0 to 256.

The numbers on your map are determined by trial and error.

Alpha-N = TPS vis RPM .... VE = (the engine's volumetric efficiency for) MAP vis RPM

TZ350

4th January 2021, 08:24

The thing that seems to be missing from all this setup is the injector end times calculator in that you have to manually enter an end time with the graph?

348258

The injector closing point is chosen by moving the (blue) points on the "Injector Timing" graph.

This is handy because you want to end the injection cycle about BDC to TPC. BDC because that is where the pipe should be sucking its hardest and you want all the injected fuel to be transferred to the cylinder.

I find BDC (180 deg) works best for me. Flettner with his 350cc Kawasaki single prefers TPC (240 deg).

Nothing is perfect, at some RPM some fuel will be swept out of the exhaust port and/or some will be blown back out of the inlet. And you have to allow for that in your fueling table. But with care you can minimize it. Do not worry about where the injection cycle starts, the system will calculate that for itself.

You can also extend the RPM points on the graph by changing the code in the Speeduino Configuration file and re loading it.

348259

Your job is to run/ride the bike and change the numbers on the map. A bit like changing carburetor jets but much easier.

Kettlekev

5th January 2021, 00:46

Cheers,

I totally get all that and thanks for confirming.

What I was trying to do with the excel which allowed for rpm, TPC position and fuel volume was to estimate where the end inject point needed to be based on how long it took to get the fuel load in at each RPM.

I can see its easy enough to do with trial and error but was just trying to add a bit of calculation to understand a start point.

If you are at all interested I can email the excel sometime?

Anyways, bits are at the machine shop. All good fun and I think I have found space in the barrels to squeeze the injectors in....:msn-wink:

TZ350

5th January 2021, 09:34

.

The Neil Hintz story and the birth of Two Stroke TPI injection that KTM later copied.

https://www.drd.co.nz/2-stroke-tpi-conspiracy-theory/?fbclid=IwAR3rXfB4T-3jDOMVR6MoVdQDZZAnAvxTvjqe2Z2lc_QkNyGjmrXAhLNFCxM

TZ350

5th January 2021, 09:53

What I was trying to do with the excel which allowed for rpm, TPC position and fuel volume was to estimate where the end inject point needed to be based on how long it took to get the fuel load in at each RPM.

The amount of air/fuel transferred to the cylinder is roughly the volume contained in the Transfer ducts. The crankcase transfer cutouts are also part of the transfer duct. Very little extra if any comes from the crankcase proper on each individual cycle.

348260 Engine Mod2T simulation of bulk flow in the transfer duct of my 125.

By finishing injection at BDC or shortly after. Pretty much that transfer duct's worth of air/fuel mixture that is drawn into the cylinder will be at the correct air/fuel ratio.

I would be very interested in your spread sheet. I will PM you my email address.

Kettlekev

9th January 2021, 07:08

Quick update.

Managed to work out how to kid Speeduino how it can be a 3 cylinder, 2 stroke controller...

348286

Signals look good on the scope (ignore the required fuel and mix ratio for now, corrected already).

I've found what looks like a good injector outfit who have helped with characteristics definition....https://www.fiveomotorsport.com/keihin-hayabusa-gsxr-husaberg-zx10/

I have made an auto calculating spreadsheet to calculate injector end inject times which hopefully works.

So crank wheel is back from the machine shop next week then its run the new ignition with carbs on the dyno to get a baseline, strip barrels and work out where I can squeeze injectors and a fuel rail and a MAP sensor in....

All good fun...:rolleyes:

Kettlekev

21st January 2021, 06:23

Managed to get a knackered set of barrels to play with, and I think my injectors will fit into the base of the transfers.....

348362

348363

348364

Just gotta make some fancy ports and a fuel rail but its looks like it will be neat enough :yes:

TZ350

21st January 2021, 21:21

.
It is looking good. I can see how the injectors will fire across the rearward transfers but how are you going to place the other injectors for the forward transfers?

Kettlekev

23rd January 2021, 23:55

I was kinda hoping I wouldnt need 2 injectors per cylinder but I guess from your post you are advising I do. Thinking about it makes sense to get better mixing.

What if one injector were more angled down into the crank, do you think that would work. If not I need 6 injectors in pairs I guess......

TZ350

24th January 2021, 10:00

.
I have tried one injector on one side. It did not work, I got rough running with a sort of hit and miss detonation. My guess is that the air/fuel ratio was not consistent.

6 injectors fired in pairs, an injector firing across the base of each transfer on a triple is the go.

KTM run an injector both sides but at low fuel demand they fire them alternately. I expect they get away with it because there will be wet fuel still evaporating of the duct walls on the alternate side to the firing injector and that balances up the air/fuel ratio a bit.

TZ350

24th January 2021, 10:17

This is my staged injection air cooled cylinder. The outer injectors fired across the base of the transfers and impacted on the far side of the transfer duct. Any wet fuel that made it that far was then evaporated of the hot metal. Evaporated the fuel and kept the transfer duct cooler. This arrangement worked well.

348389

Could you place just one injector in the middle that fires up under the piston. The injector could be in the front or back or even the underside of the crankcase between the crank wheels. I remember Suzuki twins of the GT750 era had drain plugs in the bottom of the crank cases. Maybe the 750 has them too. They could be modified for an injector. Worth a thought.

Kettlekev

24th January 2021, 10:32

Under the crank would definitely work. I can make space there.

I still maybe be able to squeeze 6 in but the fuel rail will be a challenge and it wont look too pretty.

I'm sure I can machine features under the crank. Good idea!

If its a no go I will look at double injectors on the barrels. Hopefully the speeduino has enough current capacity for dual injectors.

TZ350

24th January 2021, 14:12

.
Disclaimer, I have not done this myself but have thought about it and would like too.

If you are firing directly up from the bottom of the crankcase between the crank wheels you should only need one injector per cylinder until you run into the rule of thumb engine performance > 10kRPM/10Bar point where staged injection could become necessary.

The ideal is one injector big enough that it can deliver the max amount of fuel required in the smallest amount of time available at high RPM and small enough that at idle/low RPM it does not drown the engine with its dribble when its opening in the dead time zone.

Injector sizing is a balance. When using just one injector, think smallest injector possible.

husaberg

24th January 2021, 14:21

.
It is looking good. I can see how the injectors will fire across the rearward transfers but how are you going to place the other injectors for the forward transfers?

I was kinda hoping I wouldnt need 2 injectors per cylinder but I guess from your post you are advising I do. Thinking about it makes sense to get better mixing.

What if one injector were more angled down into the crank, do you think that would work. If not I need 6 injectors in pairs I guess......

.
I have tried one injector on one side. It did not work, I got rough running with a sort of hit and miss detonation. My guess is that the air/fuel ratio was not consistent.

6 injectors fired in pairs, an injector firing across the base of each transfer on a triple is the go.

KTM run an injector both sides but at low fuel demand they fire them alternately. I expect they get away with it because there will be wet fuel still evaporating of the duct walls on the alternate side to the firing injector and that balances up the air/fuel ratio a bit.

I can't remember the rule of thumb, but modern engines have something like 3x the cylinder volume stored mostly in the transfers.

TZ350

24th January 2021, 14:39

348390 348391 348392

It might not be that hard. If the crankcase plugs were adapted for a injector. Some crankcase nuts could be replaced with pedestal nuts that also act as mounting points for a conventional style fuel rail. You will no doubt have to make the fuel rail and nuts to suit.

Kettlekev

24th January 2021, 21:36

I am liking this idea more. I have a suspended engine so no frame getting in the way and I can work around the exhausts I think.
Time to buy a knackered lower casing and have a tinker me thinks.....

I honestly think there isn't one component on the bike I haven't modified now.....:rolleyes:

Flettner

25th January 2021, 09:06

My Kawasaki, cant remember, I may have already post a picture of this.

TZ350

25th January 2021, 19:00

My Kawasaki, cant remember, I may have already post a picture of this.

Flettner hid two injectors in the back wall of of the "C" port duct that fire directly across the top of the flywheel's. He used two fired together to get the volume for E85. But it could have been one larger one or the two could be staged injection. Lots of possibilities for the Suzuki 750.

Kettlekev

26th January 2021, 00:29

I've got a new lower case coming for the weekend.

In my head I am envisioning squirting up between the crank lobes, right up the middle of each chamber.

Other than the benefit of not hitting a hot transfer wall to help mixing it should be in a good place as far as timing goes in that I can start injecting as early as transfer close right round to BDC (and beyond a bit if needed).

Will post pics when the casings arrive.....

Kettlekev

30th January 2021, 07:55

Ok now I am confused.

I am running a bosch ignition trigger module to feed my coils. The output of the speeduino feeds the trigger then the trigger feeds the coils.

My original CDI coils wont work because their primary coil is just a single connector, the other side of the coil being grounded through the case.

This is no good because the trigger needs to see a coil with 12v on one side and the trigger on the other which switches the earth.

The Speduino on the input then goes high and on the low pulse edge fires the trigger. All good so far.....but....

I bought what I thought was the right 2 wire primary coil (yamaha YZ) but the input resistance is 0.6 ohms (I know that's not the true impedance of the coil but I cant measure that ).

Long story short is all I am doing is blowing fuses now. My coil resistance must be too low? What resistance coils should I try to use on a set up like this and does anyone have any coil recommendations?

Kettlekev

30th January 2021, 23:14

I worked out the coil I had bought was for a CDI ignition, hence the low resistance.

I tried a simple test last night replacing the coils with high watt resistors to see the trigger firing but there is clearly an issue with the trigger now.

If I understand the trigger cct it is basically a FET and either a low or a high on the input forces the output low (shorting one end of the coil).

Initially I tried connecting the crank signal pulse and that didnt drive the ignition module then I resulted to just applying voltage to the input which should trigger the out put when its loaded (connected to the coil/coil simulator resistor). That didnt work either so I must have fried the trigger I guess.

EDIT: Worked it out. It is VERY easy to wire the module backwards. ...and that's exactly what I did. The upside is I have now mad a good module test system.... :D

Basically to test the trigger using a simulated coil, all you need is an 8 ohm (20 wattt or more) resistor which at 12v will draws about 1.5 amps (well inside the limit of the module). Connect one side to 12v and the other side to one output channel on the igniter. Now either connect the ignition channel from Speeduino to the same input channel on the igniter and either using a crank signal or speedysim's TPS control to generate a crank signel, fire it up and away it goes. Remember to ground the centre pin on the 5 pin (input side) of the module. To 'show' it is firing I also wired a LED and 540 ohm resistor across the 8ohm resistor so that I can see the fire.

The module prefers a 5v drive from speeduino and gives a much better square wave on the output of the trigger, the 12v option pulse doesnt pull low enough and the signal is a bit rounded on the trigger output which will affect the coil drive.

You can actually test it without Speeduino attached and just make and break (momentary) a 5v level to the trigger input

Ironically the original 3-4 ohm coils for the factory bike are probably OK but obviously not high performance

Kettlekev

30th January 2021, 23:22

and here's one I made earlier!!

348428

Kettlekev

1st February 2021, 00:20

I've been looking at the injector positions.

The X's marks the points where ideally I would put injectors so that they fire right up the middle of each cylinder.

348453

This would angle my injectors up at about 45 degrees and although they would be squirting towards the inlets, as I have reed valves I dont see a problem.

Problem is the middle cylinder, X cant be in that place because of the water jacket, so it will be slightly to one side. The little red circles mark the injector positions
348455

Good news is the case looks thick enough to take a screw in, O ringed boss (not much room for welding one).

This means the fuel rail (which will obviously have to be custom should sit between the exhausts and the casing. Maybe!!

EDIT. Or, crazy thought? The oil feed SRIS valve positions (green circles) could be used as an inject point and I could make new SRIS ports which might be neater. One SRIS feed is a bit under a crank lobe though......

348456

TZ350

1st February 2021, 18:28

I've been looking at the injector positions. The X's marks the points where ideally I would put injectors so that they fire right up the middle of each cylinder.

348453

Hi Kettlekev. I am excited by the ideas, they look good but I do have to point out that I have not done this myself so you are leading the way here.

Because a real sucky pipe will make more power when there is more crank case volume to draw from. On my engines I have turned the inside face of the crank wheels back to the big end thrust washer surfaces to increase crankcase volume and skimmed an extra mm or so from the outside face and flywheel wheel rim to reduce hydra-static drag. If needs be you could maybe skim a bit of to un shroud the injectors.

Kettlekev

2nd February 2021, 00:30

The GT cranks are heavy but that helps the drive/feel.

I could lighten them but will probably look at that as a next step as there is a fair bit of space in there.

My only worry was the flow out of the injector getting inhibited but I guess as long as it all end up in the crank case somewhere and doesnt focus over one transfer then it should mix?

Looking more at the SRIS inject point option (green circles) the problem I see with that is the injection would flow horizontally first for about 50mm then vertical into the casing. Not only would that add some time but in might pool fluid in the track so that may not be as good as 'direct' injection near-vertically?

TZ350

2nd February 2021, 10:33

.
I think squirting the injected fuel up a narrow passage way will be an issue as the misted fuel will wet out on the walls of the passage.

If I was putting injectors in the bottom of the crankcase I would aim them so the injected stream of fuel impinged on the underside of the piston crown. That way you get good evaporation and piston crown cooling. At high speed and engine load the injector would be squirting for nearly a full crank cycle so plenty of time for the fuel stream to reach the piston and cool it. (piston cooling was the main reason I wanted to try crankcase injection.)

The only issue I see with this is that the injection stream will wash the bigend, maybe not a problem if you are squirting pre mix like I do but might be an issue with raw fuel.

348458

When I welded in the injector bungs I could only get to half the area needed. After they were securely tacked into place I sealed all around the bung with Devcon F. But any good metal glue would do.

F5 Dave

2nd February 2021, 12:29

I'm on my phone having chocolate fudge (very nice) so pictures hard to see, but I'd want to avoid bearing wash as TZ says but also the spray will cone out and wet the crank surfaces i would think.

Can you not move them near the front angled a bit up so they are nearer transfer bases and at the top of the crank?

There was an RZ350 on RD/RZ500 forum using R6 parts i recall. Into the throttle is looking more attractive if you have compromised positions. Peak power revs must be pretty low.

Disclaimer; I know jackshit about injection.

Kettlekev

3rd February 2021, 08:26

As you say raw fuel and bearing wash could be a bit of an issue potentially.

The 2 stroke oil inject point is on the inlet, post reeds (piston side) so kinda drops onto the conrod bottom bearings and if I am squirting up past them the other way it may be OK. For a good portion of the cycle the injector wont be obscured by the conrod (except the middle one is a bit of a challenge) because I want to stop injecting around BDC if possible.

Maybe I should try a more direct spray pattern but then I would reduce atomisation I guess.

Pro's and cons. Think I might just have to try it and hope I dont de-oil the bearing to much.

Still thinking......

Kettlekev

3rd February 2021, 11:02

OK, looking at the casings again....new plan.

To help with the oil mixing I am going to inject at the back of the crank chamber near the inlets and again straight up. It will mean loosing some casing ribs and some tricky welding but I think its dooable.

348463

The yellow circles mark the spots.

Hopefully the raw fuel wash will be minimized

F5 Dave

3rd February 2021, 12:11

Oil injection. Heck hadn't considered that properly. Do KTM mix first then injection both?

I wouldn't want to pass air over an oil dribble and hope it combines with the crank mixture to lubricate everything. I'd want the oil to be in direct path of the petrol spray which is how it occurs with carbs.

Are you sure you don't want to inject into throttle bodies?

What if you oil injected into the spray tube? Nah too much chance of blocking an injector with such a small volume and high pressure with 2 distinctly different viscocities and no mixing agitator.

My T125 has pressure feed to the crank, but you still need to mix gas with oil to do the bores.

Unless I misunderstood I think you need a revision.

husaberg

3rd February 2021, 16:41

Kettlekev

3rd February 2021, 22:08

Where I am planning will mean air and oil will come in together then will mix with fuel close to the inlet track, although the oil squirt might need to be a bit more directional now which is tricky......Will have to look at that.

If I put fuel back in the throttle bodies, I will likely loose efficiency although it would oil mix very well.

I guess that why they call it experimenting....Downside would be a seized crank though!!:eek5:

EDIT. I just reread what I posted and it is B******S......The SRIS system has 2 oiling methods. The pressurised feed from the pump is fed direct to the crank bearings and small ends. The remaining oil that doesn't stick falls into the casing and is pushed out via one way valves to appear again in the inlet ports. This then mixes with the fuel and lines the cylinder walls. SO the second part if my concern but I think there will be enough petrol wash to pick up the inlet oil and spread it around. I guess the temperature gauge and the blue smoke (or lack there of) will tell me!!

F5 Dave

4th February 2021, 06:14

Oh well. Try blowing air through an open hole inlet size with a small dribble of oil on the floor (as it only dribbles tiny amounts of oil at low pressure into the inlet) - aimed at a piece of paper and see that you are happy it atomises and sprays.

I know where my money would go to the result.

Kettlekev

6th February 2021, 12:26

Yep I hear you. No one said this would be easy.

I'm going to get a baseline dyno with injectors in the throttle bodies first and see how that fares.

I may need to run two injectors (one small one in the throttles) just to get that return oil mixed and back in around the bores.

I t will be interstinmg to see what the crank pressure is and just how much squirt it generates. I am going to add a MAP sensor anyway so at least I can measure it.

My secondary worry with the crank case injectors is contamination......

F5 Dave

6th February 2021, 12:38

Heck yes, more things I didn't think of. Pull your engine down and it will likely contain a desert-spoon of oil at the lowest point of the crankcase (ok well usually not on small race engines) . The injector there would be soaking in oil.

TZ350

7th February 2021, 19:20

.
Cant say for sure as I have not tried it. But the fuel stream from an injector is pretty powerful. I expect it would blast up through a puddle of oil at the bottom of the crankcase.

F5 Dave

7th February 2021, 19:24

But the next time you started it, let's say 5 weeks, then what?

Kettlekev

8th February 2021, 12:09

The position I have them they wont be in the oil puddle at the bottom. They are quite a bit above it.

So the only real worry is the lack of oil on the inlets and just how much that return oil feed is needed to lubricate the barrels.

Anywhere I put the injectors will have the same oiling problem (other than on the throttle bodies).........

Kettlekev

17th February 2021, 05:01

This week has mostly been sorting out the Speeduino ignition system.

Lessons learned;

RF noise gets in everywhere and generates sync loss, especially when everything is closely packed.

Use screened cables on the ignition trigger module and earth both ends of the cable.

Use resistor (5k) plug caps and carbon resistive (not copper wire) core cables

My little VR crank sensor and the trigger wheel work great submerged in the oil bath. Use screened cable there as well.

RF ferrites help when power and triggers lines are passed through them (wrap cable round a couple of times).

At the moment I am running 3 ohm coils with a dwell time of 3ms (running). I am not sure f this is optimum but I have found a good dwell calculator (attached) to test my setup and work out best dwell and then speeduino can control dwell based on voltage. This should avoid any burnouts. I may try to switch to a high performance coil but will review after the dyno run with the new high compression heads.

348544

Kettlekev

20th February 2021, 21:09

Quick update

The standard 3 ohm coils wont cut it, I run into an Ohms law limitation and the dwell in effect limits out and cant really be controlled via Speeduino.

So I am going to switch to 0.5 ohm Dynatek DC3 coils.
348559

They will charge faster, give better spark and will allow the dwell control to work.

I am using a Bosch 211 ignition module which should handle the power.

New high compression heads back next week so we will see. :Punk:

F5 Dave

21st February 2021, 06:48

pete376403

21st February 2021, 08:22

I've seen a lot of ignition faults corrected by swapping the fancy looking car type HT lead back to copper wire. Car coils are usually much bigger than bike ones so have energy to burn, or at least waste.

Copper HT leads were one of the causes of electrical noise on this project and were replaced for this reason.

F5 Dave

21st February 2021, 08:54

Resistor plugs and caps are usually enough along with careful routing of wires so you are not inducing spikes from outputs to the ignition system with reasonable separation of the module from the angry end of the bike it should be fine.

Kettlekev

23rd February 2021, 05:15

Question about dwell time.

My ECU has programmable dwell and my low ohm coils (0.5 ohm) will allow me to run large dwells (current).

At the moment my start point is 3ms running (4.5 cranking). Clearly I can mess with the dwell until I get mis-fire then back it off but I wondered if anyone else had any 2 stroke dwell recommendations as a starter???

JanBros

25th February 2021, 07:26

ain't those coils for a CDI ignition ? CDI's do not use dwell.
such a low resistance can act like a short-circuit in a TCI system. they did in the AEPL-ignition I'm trying out.

Kettlekev

25th February 2021, 11:33

When I run the dwell/coil calculator in post #67 it all works out right.

Dynatek reference the coils as for use with dwell control ECU's.

They are low resistance but as far as I am aware its the inductance and coil layout that's key to making the dwell work with Speeduino.

I agree CDI coils are also low ohm normally but they rely on a positive charge from the ECU via a single wired primary (other side to earth). Dwell coils are two wire primary, one side to 12v and the other earthing/controlled from the ECU.

That's as far as I know. Testing this weekend with new heads so will let you know.:niceone:

Kettlekev

4th March 2021, 06:11

Quick update.

New 0.5 ohm coils and the Bosch 211 trigger all worked out OK with Speeduino

My current settings are below. My initial voltage compensation curve assumes centers around 13v's but when fully running I will tweak that.

348633

The dwell/spark settings are all from the calculator I posted earlier so will probably need tweaking on the dyno.

So far I am impressed with the Dynatek coils.

I've still got some high rev ignition noise causing the odd sync loss so that is this weekends project. :rolleyes:

Kettlekev

8th March 2021, 06:19

Clearly like all micro's, if there is noise around it will probably generate issues.

After a couple of weeks of trials, the items below, in order of importance, seem to be what helps the most for a small space, bike design.

1. Resistor plugs (ones with a R in the code)
2. Resistor caps, initially I was worried about the additional 5k resistance but so far (pending dyno) they seem fine.
3. Silicone/carbon (not copper wire) ignition leads
4. The USB cable to run Tunerstudio live with must be high quality and not too long, also I found that using a couple of ferrites with at least one coil of USB inside, helped. Noise on the USB line will cause sync loss.
5. Wires between Speeduino and the trigger unit and from the trigger unit to the coils need to be screened cables. I use one for each coil feed with the earths connected at each end, which seemed better than a single ended earth.
6. The +ve to Speeduino benefits from a ferrite with at least 5 coils wound through it
7. The 0v to Speeduino also befits from a ferrite fitting

The latter two (6,7) need some more evaluation, as most of my problem was poor quality resistive plugs and caps and until I had those replaced, all of the above made improvements. I suspect that with quality plugs/caps, some of the other changes may not be required but again as a tight space build, they may still be useful.

Thanks everyone for the pointers. The real eye opener for me was the USB cable, I should have thought it would be an issue but didn't look there initially but it makes sense as its a direct line to the micro.

So all good. Final hurdle was to remember to zero the coolant and air temp sensor compensation curves (as I am not bothered about those sensors until I add the injectors) because they retard the ignition as an offset to the basic spark table and it turns out I was initially running at 15 degrees at startup (should be 24). I am amazed it started at all....the clue was running rich obviously.

Speeduino/Tunerstudio is pretty amazing but lots of ways to trip up!

Quick road run this week then back to the dyno then the fun starts with the injectors!

I am going to leave them in the throttle bodies as a starter then move to the crankcase later after another dyno run.

TZ350

8th March 2021, 21:45

1. Resistor plugs (ones with a R in the code)
2. Resistor caps, initially I was worried about the additional 5k resistance but so far (pending dyno) they seem fine.
3. Silicone/carbon (not copper wire) ignition leads
4. The USB cable to run Tunerstudio live with must be high quality and not too long, also I found that using a couple of ferrites with at least one coil of USB inside, helped. Noise on the USB line will cause sync loss.
5. Wires between Speeduino and the trigger unit and from the trigger unit to the coils need to be screened cables. I use one for each coil feed with the earths connected at each end, which seemed better than a single ended earth.
6. The +ve to Speeduino benefits from a ferrite with at least 5 coils wound through it
7. The 0v to Speeduino also befits from a ferrite fitting

Thanks for the tips. Very helpful.

TZ350

8th March 2021, 21:49

Quick road run this week then back to the dyno then the fun starts with the injectors!

I am going to leave them in the throttle bodies as a starter then move to the crankcase later after another dyno run.

My first step was to get it idling. Then I set the wide open throttle cells. I found wide open throttle was easy compared to trailing throttle.

Yours is a very interesting project. Good Luck.

Kettlekev

11th March 2021, 05:25

TZ350,

Did you ever post your spark table anywhere? It would be interesting to compare.

One more caution note on Speeduino and spark table.

Whatever you set on the spark table is affected by the cold start setting for IAT and CLT sensors.

There are offset temperature curves for both and if you dont check/set them you will have an offset to the standard spark table. If you are not using the sensors its best to run them to ground through say a 10k resistor and then calibrate them so that they read something sensible. If you dont, they become a floating input that can cause issues.

TZ350

11th March 2021, 09:13

TZ350,Did you ever post your spark table anywhere? It would be interesting to compare.

Hi Kettlekev. No I don't use the Speeduino spark side of things at all as I knew nothing about the car coil packs required and how to set them up.

I have continued to use the Ignitec digital 12Volt DC/CDI 2 cylinder Race system I had on the bike. I would love to know more about the details of the coil pack you use.

I will post a picture of my ignition curve if you like. But as the engine peaks at 12k rpm and the ignition goes into heavy retard to heat the pipe for additional over rev I am not sure it will be very helpful.

Kettlekev

17th March 2021, 08:24

TZ350, the coils are the 0.5 ohm dynatek ones. https://shop.dynatekuk.com/dc12-1-455-p.asp

I drive them from a BOSCH 211 ignition module (lots of clones out there)

Road test this weekend shows it is all working nicely with the new higher compression squish so dyno next week;)

Fuel next. I've started on the throttle bodies. They dont quite fit so need spacers/linkages designing but i've just finished that now.

Lots of moving bits to play with.....

348727

Misaligned by 12mmm 1 side and 8mm the other

348728

Tricky gearbox I managed to leave alone.

348729

Lots of bits to find a home for when it all goes back together!

SO I have the fly by wire all good with the fail safes built in so nearly there!!

TZ350

19th March 2021, 08:07

.
Thanks for the coil ignition module info.

The fly by wire thing is very interesting.

Kettlekev

3rd April 2021, 04:33

Got the throttle body bits back today.

Bit of a jig saw......

348818

All good and no screws or O rings left on the table!! Although Triumph dont make it easy with tamper proof Torx screws etc. :angry:

348817

The red bits are where I have had to make extension pieces. Some of them are O ring sealed to that had to be mimicked on the extensions. The lever arm on the RHS side also had to have a fancy extension and the LHS butterfly arm had to be extended as well.

Fired it up and it all works. GT750 fly by wire. :D

Just need to work out how to link to the oil pump now (although I am considering going premix after the strip down). If I extend the throttle position sensor linkage I can squeeze a lever arm in for the oil pump.

TZ350

3rd April 2021, 11:19

Got the throttle body bits back today. Bit of a jig saw......
Fired it up and it all works. GT750 fly by wire. :D

I think Flettner has his single on Fly by Wire but a Fly by wire big multi cylinder TwoStroke, that might be a first........ :clap:

Cie87k

15th June 2021, 13:38