Hi Bernard,

Thanks for the circuit.
It sort of makes a little sense to me (i'm slow).

Would there be any benefit to going with one circuit over the other (from the ones suggested on the thread by others)

In the end, I would like to look for the most reliable, with the least amount of components/simplest.

Thanks!

I think you should ask Bernard for a plot of comparator input voltage vs RPM (or frequency) and ripple voltage (and equivalent ripple RPM). I ran some sims. The sensitivity is low and the ripple is high. It doesn't look too good, but I could be wrong.
Also, I have not tested or simulated the circuit in the LM2907 datasheet, so I can't comment on its performance.

 

Napoleon,
I know this is a bit late in the game, but there are a few things that should be addressed before you actually try to implement some of these ideas.

Clutch-less shifting at high engine RPM (even without the engine firing) will place tremendous stresses on your entire drivetrain, even if the gear ratios are relatively close together. Expect all sorts of nasty things to happen rather quickly, like broken teeth in the gearbox, worn-out synchros, bent sprockets, sheared flywheel keys, even a broken crankshaft and broken gearbox. Additionally, slamming the gearbox from one ratio to another suddenly will most likely cause your rear wheel to break traction with the road surface.

The clutch is there to smoothly match speeds and re-connect the gearbox with the crankshaft after the gear ratio changes. If you bypass using the clutch, you'll rapidly have severe (and expensive) problems.

Secondly, interrupting the ignition will result in your exhaust system rapidly filling with a fuel/air mix. When the ignition source is restored, it is extremely likely that the unburnt fuel/air mix will be ignited simultaneously, resulting in an explosion. If your exhaust system is stock, the muffler may be split wide open. If it has been modified to a low-restriction exhaust, the noise will be deafening.

If your engine is fuel injected, it would be most preferable to interrupt the fuel injector signal rather than the ignition. If it is carbureted, closing the throttle would be the preferred option.

 

Napoleon,
I know this is a bit late in the game, but there are a few things that should be addressed before you actually try to implement some of these ideas.

Clutch-less shifting at high engine RPM (even without the engine firing) will place tremendous stresses on your entire drivetrain, even if the gear ratios are relatively close together. Expect all sorts of nasty things to happen rather quickly, like broken teeth in the gearbox, worn-out synchros, bent sprockets, sheared flywheel keys, even a broken crankshaft and broken gearbox. Additionally, slamming the gearbox from one ratio to another suddenly will most likely cause your rear wheel to break traction with the road surface.

The clutch is there to smoothly match speeds and re-connect the gearbox with the crankshaft after the gear ratio changes. If you bypass using the clutch, you'll rapidly have severe (and expensive) problems.

Secondly, interrupting the ignition will result in your exhaust system rapidly filling with a fuel/air mix. When the ignition source is restored, it is extremely likely that the unburnt fuel/air mix will be ignited simultaneously, resulting in an explosion. If your exhaust system is stock, the muffler may be split wide open. If it has been modified to a low-restriction exhaust, the noise will be deafening.

If your engine is fuel injected, it would be most preferable to interrupt the fuel injector signal rather than the ignition. If it is carbureted, closing the throttle would be the preferred option.

SgtWookie,

Thank you very much or your concern on this.


This is a race bike, and clutchless upshifts are already in use.
Watch any bike race on TV and they are using clutchless upshifting.

Its sort of a necessity.

And, the bikes are designed to allow for that.


There are no synchros on bike transmissions, and they are already designed to be able to shift without using the clutch.

The key is in smooth clutchless upshifting, with is harder to do without some electronic aid.

Re: ignition
The kick stand switch is connected to the ECU, who then properly turns off ignition and fuel injectors. So this is already handled properly there.

The exhaust already has some back firing by design.
Fresh air is injected into the exhaust when the engine is under low load to burn unburned fuel.

But, the point of this project is to precisely control the chain of events in the upshift so that it is smooth and does not upset the bikes geometry or dynamics. It simply feels like one longer gear, constant acceleration.

This is especially important in turns, where one is accelerating heavily while still leaning in a turn. It is dangerous to upset the bike as traction can be lost due to the jerking of upshifting too harshly.

When it is done correctly, the transmission shifts incredibly smooth.
No grinding or sudden jolts are felt. Just smooth acceleration throughout the gears.

Again, I appreciate the concerns, and they are valid ones.

Thanks!

 

I think you should ask Bernard for a plot of comparator input voltage vs RPM (or frequency) and ripple voltage (and equivalent ripple RPM). I ran some sims. The sensitivity is low and the ripple is high. It doesn't look too good, but I could be wrong.
Also, I have not tested or simulated the circuit in the LM2907 datasheet, so I can't comment on its performance.

Hmm, thanks for running sim on that.

I will probably try the LM2907 idea first.
Not only becuase of the info you posted here, but also because they have a specific circuit example pretty much for what I want to use it for.
Seems to make sense to me a little better.


Once the circuit is built...

Anybody have any recommendations/ideas on a good way to get this on a PC board so it can be mounted on the bike somewhere ?
Ideally something waterproof, or splashproof, and that can be opened for fixing/modifying.

I've seen some proto boards, but they all seem too big.
Are there really small ones just big enought to fit this circuit ?
2 ics and a handful of res/caps etc...

 

Napoleon, if I understand correctly, you intend to set the pulse width of the oneshot so that the RPMs will drop to exactly the value needed for the next gear during the time the ignition is interrupted, thus allowing for smooth shifting. is this correct?

 

Napoleon, if I understand correctly, you intend to set the pulse width of the oneshot so that the RPMs will drop to exactly the value needed for the next gear during the time the ignition is interrupted, thus allowing for smooth shifting. is this correct?

Well, not exactly.

I'm not trying to lower the rpm.
I just need to cut engine power for a few ms so the transmission is unloaded just long enough so the ned gear falls into place almost on it's own with minimal pressure on the shift lever.

But that part of the circuit I already have solved with the 555 monostable.

Now I need to try the triggering part based on RPM.
(that lm2907 seems like it is made for this)

And then, I still need some suggestions on how to put this in a small enclosure to mount it on the bike.

I've gotten so much help here!
Thank you!

 

Well, not exactly.

I'm not trying to lower the rpm.
I just need to cut engine power for a few ms so the transmission is unloaded just long enough so the ned gear falls into place almost on it's own with minimal pressure on the shift lever.

But that part of the circuit I already have solved with the 555 monostable.

Now I need to try the triggering part based on RPM.
(that lm2907 seems like it is made for this)

And then, I still need some suggestions on how to put this in a small enclosure to mount it on the bike.

I've gotten so much help here!
Thank you!

If RPM does not drop during the shift, won't you get a jerk when the transmission re-engages? That is one of the things you said you wanted to avoid.

 

If RPM does not drop during the shift, won't you get a jerk when the transmission re-engages? That is one of the things you said you wanted to avoid.

once the timing on the trigger and delay period are properly set, there will be no jerkiness on the shifts. Specially once I expand the circuit ( with help of you guys here) so each shift point gets it's own specific delay.

The transmission has close ratio gears, flywheel is lightweight, and engine had high compression. The gears are also always spinning(even the ones not currently engaged).

This all combined makes this work remarkably well.
I already do it, but manually, by lowering throttle briefly with pressure on the shift lever. The next gear falls I'm smoothly if I time it right.

 

Ok, I have a circuit I made that integrates all the suggestions here.

The 555 monostable with adjustable delay, which can be triggered either by a push button, or by an LM2907 setup to receive a pulse train, and trigger at a specific pulse train frequency, and and above that frequency.


Please have a look and let me know if you spot any issues so far.
(there are a few values that are not specified in the drawing. I either have not calculated those, or dont know what they should be)

Thanks for the help!

 

Looks good,' would add a diode ,anode at junction of LEDs ,cathode to relay.a recent thread explains this problem.

 

SEE, charlieboy pg. 16 , post by jpanhalt.

 

SEE, charlieboy pg. 16 , post by jpanhalt.

This one ?
http://forum.allaboutcircuits.com/showpost.php?p=107422&postcount=16

It also mentions the 1000pf cap from output to ground.
Would this apply here as well ?

I added those two on the circuit.
Here is the revised one.

Please let me know if anyone has any observations or suggestions.

Thanks!

 

Sorry,but one more thought: I do not see any V supply filtering; something as simple as a .1μF or a zener diode, resietor and cap. The relay probably pulls as much I as rest of ckt.; if it were low side driven it could be left outside the regulator, Or use an adjustable IC regulator set for 9 V.

 

Sorry,but one more thought: I do not see any V supply filtering; something as simple as a .1μF or a zener diode, resietor and cap. The relay probably pulls as much I as rest of ckt.; if it were low side driven it could be left outside the regulator, Or use an adjustable IC regulator set for 9 V.

This will be used in a motor bike, so the 12v supplied by the electrical system can vary +/- a few volts from that.

Would just adding a 0.1uf cap between + and - do the trick ?
Will the voltage range affect the timing/operation of the circuit ?"

Thanks!

 

Ok, here is an updated one.
I added a 10uf to the + source after the on-off switch.

Would that work ?

Thanks!

 

The added cap. will help to reduce noise picked up in leads and switching within ckts.the timing variations in 555 in 5to 15V variation is suposed to be around 1%.The LM 2917 has regulation built in otherwise I think it is the same as the LM 2907. An outboard regulator can be added later if necessary. It has been a pleasure working with you,you answer questions promptly and keep us updated on progress.

 

The added cap. will help to reduce noise picked up in leads and switching within ckts.the timing variations in 555 in 5to 15V variation is suposed to be around 1%.The LM 2917 has regulation built in otherwise I think it is the same as the LM 2907. An outboard regulator can be added later if necessary. It has been a pleasure working with you,you answer questions promptly and keep us updated on progress.

Cool!
Thanks for all your help!

I'm sure I'll be having more questions as this project progresses.

Actually, I have one that I just thought about...

For the LM2907...

Currently, the plan is to splice pin 1 to the existing wire on the bike that carries the pulse train for the RPM.

How could I use some sort of inductive pickup to get enough of that frequency signal from that wire without actually splicing into it ?

Would this be practical/possible in this application ?

Thanks!

 

In a system where Vcc can vary, I would use LM2917, and power the threshold voltage divider (5k/5k) from the zener (vcc).
Since the -input of the input comparator is at ground, and your input signal doesn't go below ground, you should AC couple the tach pulses to the input. See the amended schematic attached.

Alternately, you could use the 14-pin part, direct-couple the tach pulses, and bias the -input to ≈2.5V.

 

Cool!
Thanks for all your help!

I'm sure I'll be having more questions as this project progresses.

Actually, I have one that I just thought about...

For the LM2907...

Currently, the plan is to splice pin 1 to the existing wire on the bike that carries the pulse train for the RPM.

How could I use some sort of inductive pickup to get enough of that frequency signal from that wire without actually splicing into it ?

Would this be practical/possible in this application ?

Thanks!

I read this after my last post. My post assumes a 5V p-p input.

 

In a system where Vcc can vary, I would use LM2917, and power the threshold voltage divider (5k/5k) from the zener (vcc).
Since the -input of the input comparator is at ground, and your input signal doesn't go below ground, you should use a AC couple the tach pulses to the input. See the amended schematic attached.

Alternately, you could use the 14-pin part, direct-couple the tach pulses, and bias the -input to ≈2.5V.

Ok, I think I understand this.
I'll go add those on my drawing...that should help me understand your changes better...

Thanks!

PS - Any ideas on an inductive pickup for this ? So I don't have to splice into the wire with the pulse train ?