Electronic Shifting

[boss451t]

What electronic shfiting systems does everyone use? I had selected a manufacturer but they are being very unresponsive so I am thinking I need to go elsewhere. So what manufacturer's and how do they shifters work? Thanks.

[boss451t]

What electronic shfiting systems does everyone use? I had selected a manufacturer but they are being very unresponsive so I am thinking I need to go elsewhere. So what manufacturer's and how do they shifters work? Thanks.

[Charlie]

Make your own. They are not terribly complicated systems, I would guess that there are more custom systems out there now than anything prepackaged.

-Charlie Ping
Auburn University FSAE (http://eng.auburn.edu/organizations/SAE/AUFSAE)
5th Overall Detroit 2003
? Overall Aussie 2003. http://fsae.com/groupee_common/emoticons/icon_smile.gif

[Sgt Slaughter]

We tried using click-tronics system, they're an aussie based company but had some power consumption problems causing the selenoid not to shift sometimes, but then again, we never tested it with the engine running.

"Lets kick the tires and light the fires..."

[karter]

One of the shifter kart guys here claims he built an electronic shifter from a Ford door lock "solonoid", I think in reality it is a DC motor with a recirculating ball screw, don't know how they got the shift lever to return to center but, sounds cheep to do if it has enough force.

[Griff]

Charlie's right, they're really not that complicated and you can build them relatively cheap. We tried going the doorlock route, but it was heavy and couldn't supply enough force to shift. We switched to an air solenoid system

[boss451t]

how do you control the ignition cut when you are shifting? getting a solenoid that will provide the power to shift wouldnt be too hard, but I am not sure how to controll the MSD.

[Ryan Schoffer]

just cut the signal from the crank angle and cam position sensors with a dual pole relay

Vehicle electronics leader

www.ucalgary.ca/fsae (http://www.ucalgary.ca/fsae)

[Chris Boyden]

It might be better to cut the ignition in another way boss451. cutting the supercritical crank angle and cam sensors with a mechanical relay might not be the most reliable or fastest system. But hey, if it works, more power to you. Try using a power mosfet to switch the ground of the MSD. Connect the drain of the mosfet to the ground of the MSD box. Connect the source of the mosfet to the car ground. Then switch the voltage at the gate of the mosfet. Make sure that the mosfet is Normally on (Gate is high)during normal driving. During the shift, take the gate low and turn the mosfet off for a certain amount of time. This should cut power to the MSD box.

This is alot easier if the ECU you are using has an input that can be configured to cut the ignition. The timing can be controlled in several ways, the easiest way is probably with a small microcontroller if you are familiar with them.

The biggest problem that we had with our system was noise and spontaneous shifting until we isolated the power to the shift control box. That seemed to solve the problem.

Another problem with solenoids is sideloading as mentioned by others in previous posts. Finding a way to fix that is fun. Anyway, good luck.

[Andres]

Hey Boyden, I am looking for a depletion power mosfet, I havent been able to find one, do you know a model and a provider who sells those??

Equipo F-SAE USB
Universidad Simón BolÃ*var
VENEZUELA
www.formulasae.grupos.usb.ve (http://www.formulasae.grupos.usb.ve)

[Igor]

Regarding the power mosfet in the ground lead of the MSD:
I'm not familiar with MSD stuff, but there might be a problem if the coil current also passes through this ground lead.
If you cut the ground when there's significant current running, the flyback voltage may shoot through the parasitic drain-source capacitance and fry the mosfet by exceeding the max gate-source voltage.
If the design doesn't take this in consideration, you might get stuck with a car that will suddenly stop shifting.

Igor
Delft

[Akos]

Simplest way I found of cutting the engine is by turning off the power to the TPS sensor.

Regardless of throttle position, the ecu will read this as 0V on the TPS. You can set the engine map at that point to either retard the ignition timing or cut the fuel.

Similar setup might work with MAP sensor as well, I haven't tried that.

This saves you swithching high currents, and you can use a relay. It is also relatively fail safe, as long as the relay is off, the car runs as normal (easy to disable in a hurry).

Cheers,

Akos

[Ryan Schoffer]

thats the same theory behind cutting the crank angle position sensor - you can use a very small relay, or a transistor based setup so it switches faster - and when the ECU reads 0 RPM it wont meter any spark or fuel (obviously)

Vehicle electronics leader

www.ucalgary.ca/fsae (http://www.ucalgary.ca/fsae)

[karter]

if you cut the crank position the ecu may have to reinitalize, or may interpit the cut as the "skip" tooth, just dosen't seem like the way to do this.

[Ryan Schoffer]

it works just fine - the ECU just reads it as if the engine is running at 0 RPM - even if it reads it as the missing tooth that will only last for a tiny fraction of a second

Vehicle electronics leader

www.ucalgary.ca/fsae (http://www.ucalgary.ca/fsae)

[Chris Boyden]

Andres,

I could not find a depletion mode mosfet either.

But, we found that we needed a mosfet driver to drive the gate. They come in inverting and non-inverting configurations.

go to
www.microchip.com (http://www.microchip.com)
TL4420(non-inverting) is the part number for the mosfet driver. We also used IRL1404 mosfets from International rectifier.
all of these are cheap.

there is an application note that makes it real easy to select the proper driver. The main consideration is the Gate capacitance (~5000pF for the FET we used). in the application note, there is a table that tells you what mosfet driver to use based on the gate capacitance.

I guessing that you want a "normally on" FET that you switch off when you cut the ignition.


Anyway, good luck with your circuit.

[Chris Boyden]

Igor,

Yes, flyback currents can be nasty. Protection diodes should do the trick.

[Buhweet]

Hi Guys,

[Buhweet]

Hi Guys,
Just wondering does anyone blip the throttle on the down-change to match engine RPM with an electronic shift system? Is there time for the driver to do it or can it be done automatically.

[js10coastr]

currently the driver is in charge of blipping the throttle. Makes braking much more difficult.

www.calpolysae.org (http://www.calpolysae.org)

[Mick_P]

I am not directly involved with the shifter mechanism but if i remember correctly we dont use a 'blip' between shifts and the motec controls the gear changes and our car runs awesome with a pneumatic-solenoid shifter.

Mick Partridge
Swinburne University of Technology

[Chris Boyden]

During the initial testing on the shifter, I used to the clutch to downshift. Eventually, we tried to downshift without the clutch. Sometimes, the rear tires chirped, but a quick blip of the throttled during the button push seemed to work nicely. We have only tested in straight line runs so far. The downshifts were done after acceleration runs at a pretty slow speed. The first chance I get, I want to run an autocross and practice shifting in and out of corners.

Does anybody run a fully automatic shifting setup?

I'm considering implementing automatic upshifting triggered by a shift light signal, but I feel uncomfortable about downshifting at the wrong time and having the rear end break loose in a turn.

Mick, care to elaborate on the MOTEC setup?

[Daves]

I believe Ohio State used automatic upshifting for the acceleration event, but I am not sure about the other events.

http://www.letu.edu/formula

Formula Photos & Links (http://what3542.5u.com/dave/fsae/formula_photos.html)

[woollymoof]

Boyden,

Mick is right about the driver doing the blipping on our down shifts.

We have the option of running full auto upshifts that uses the shift light output from MoTeC as the trigger. We are working (thinking) on doing full auto downshifts but this requires some sort of actuator on the throttle and the rules basically outlaw this, you would need approval from the rules commitee before the event.

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR> 2003 Formula SAE® Rules 40
3.5.4.2 Throttle Actuation
The throttle must be actuated mechanically, i.e. via a cable or a
rod system. The use of electronic throttle control (ETC) or
"drive-by-wire"� is not permitted.
The throttle cable or rod must have smooth operation, and must
not have the possibility of binding or sticking. The throttle
actuation system must use at least two (2) return springs, so that
the failure of any component of the throttle system will not
prevent the throttle returning to the closed position. Throttle
cables must be at least 50.8 mm (2 inches) from any exhaust
system component and out of the exhaust stream. The use of a
push-pull type throttle cable with a throttle pedal that is capable of
forcing the throttle closed (e.g. toe strap) is recommended.
A positive pedal stop must be incorporated on the throttle pedal to
prevent over stressing the throttle cable or actuation system. <HR></BLOCKQUOTE>

Cheers,

Kirk Veitch
Swinburne University of Technology 2004

[Chris Boyden]

That's the same conclusion we came up with...
However, it would be an interesting control problem to tackle for race-cars in general.

[Igor]

But you could rig something using an idle speed control valve :-)

Igor

[Chris Boyden]

Igor,

I think that would be neat. Either a solenoid or stepper motor style should work. I think the trick would be getting enough of an rpm increase out of the engine to match rpms on the downshift when the valve is wide open...

Another idea, although is probably violates the throttle by wire rule, is to take a throttle quadrant from a passenger car with cruise control which has 2 throttle levers with one way bearings. Normally, the throttle is controlled by the foot pedal, but the second lever is connected to a small solenoid or linear actuator. When the downshift takes place, the second lever "blips" the throttle.
I'm not sure if the primary lever goes along for the ride or not.

I really like the idle air idea, though.

[Buhweet]

Just a few quick ones.

Do you have to have slower downshifts than upshifts to give the driver time to blip it?

Does anyone know what the time difference is between the driver shifting with a lever and an electric/pneumatic shifting?

Do you need to use the clutch for gear changes with a motorbike gearbox?

[Charlie]

All this talk about powered shifting systems is overrated. I mean, when is the last time a pneumatic or electrically shifted car won a FSAE competition? Never!

So.. you should all just forget about it. http://fsae.com/groupee_common/emoticons/icon_wink.gif Nothing to see here. http://fsae.com/groupee_common/emoticons/icon_smile.gif

-Charlie Ping

I just need enough to tide me over until I need more.

[Lacy Lodmell]

In response to Buh-weet (nice!), both our 2001 and 2003 cars were manually shifted with good old-fashioned linkages and driven a fair amount by yours truly. The only time I used the clutch was on launch. I always thought that, if you had your drivers in the car enough, they should know the motor well enough by sound and the trans well enough by feel to upshift and downshift without the clutch. However, some of our drivers did downshift with the clutch, and I have no data (other than my calibrated ego) that shows that they were faster or SLOWER shifters than me. I'm sure my friend/rival Sam will contest that one. Oh, and we've never broken a trans [crosses fingers].

[Edited to clean up stupid grammar.]

Lacy Lodmell
Vandals Racing 2003
University of Idaho FSAE

[Sam Zimmerman]

I have to admit that after tearing apart the engine/tranny last summer, we found no significant wear on the gears. I will throw a few things in here though, just to make sure Lacy's predicition will come true:
<UL TYPE=SQUARE><LI>No doubt clutchless shifting helps when upshifting
<LI>Clutchless downshifting will get you in the right gear faster (providing it is done right)
<LI>I haven't seen a condition yet (in my own limited driving experience) where getting into the correct gear a fraction of a second faster helps me to get back into the throttle any faster. (Remember, they are timing your laps, not your shifts)
<LI>Never, ever, ever, ever, ever use a push-pull cable for shifting in a racing application. We learned that one the hard way. (Lacy was hard on them) http://fsae.com/groupee_common/emoticons/icon_wink.gif
<LI>Maybe it was me that was hard on the push-pull cable http://fsae.com/groupee_common/emoticons/icon_biggrin.gif [/list]

Sam Zimmerman
Vandals Racing (http://www.uidaho.edu/~racing)

[Dan Deussen @ Weber Motor]

Sam,

I don't think that push-pull cables are really bad for shifting. I do have negative experiences with them myself, but it was due to lack of experience and improper implementation. Reading Racecar Engineering magazine, I have seen amy hilclimbers with motorcycle engines using them and they seem to work just fine for one those cars. I've even seen them used in parallel on a car with two engines.

Daniel Deussen
www.walbro-italy.com (http://www.walbro-italy.com)

[Chris Boyden]

Charlie, wake up on the wrong side of the bed?
Or do you need a little more to tide you over? http://fsae.com/groupee_common/emoticons/icon_wink.gif

The pushbutton shifting isn't just about shift times. In the end, it's about driver ergonomics. It's really nice to not have to take your hands off of the wheel in order to shift.

Also, the time savings on the track can add up.
If you had ten upshifts in an autocross, and saved 4/10ths everyshift, that's 4 seconds of time spent shifting and not driving.

[Big Daddy]

Just go with a CVT and you won't have to worry about doing anything but driving http://fsae.com/groupee_common/emoticons/icon_wink.gif

"A woman is a lot like a beer, they look good, smell good, and you would run over your own mother to get one." Homer Simpson

[Charlie]

Who needs a CVT-- You know I never shifted during the FSAE-A endurance event? Not once! So much for our fancy shifting system. http://fsae.com/groupee_common/emoticons/icon_wink.gif

I was really just trying to be conservative and consistent though, as I knew if I made no mistakes, we'd likely come out on top. I have a tough time being conservative though.

-Charlie Ping

I just need enough to tide me over until I need more.

[Akos]

I guess I'll toss my 2 cents into the discussion.

Pneumatic shifting is not needed on FSAE cars. Shifting force on bike engines is very small, you can easaly do it by hand, you don't need any power assist.

Having run a pneumatic system I can tell you it wasn't any faster than a manual shifter with ignition cut, so why add on extra weight? I guess keeping your hands on the wheel is convenient, altough I didn't find much difference with driving with a paddle shifter vs a side shifter.

Proper clutchless shifting will not wear out a bike tranny. Improper clutchless shifting will destroy a tranny in very short time. The problem generally isn't dog wear, but one of the gear retaining snap rings popping out. This will normally cause two gears to engage at the same time, very expensive repair. Its a good idea to force newbies to clutch till they learn to blip properly.

I would say, unless you have to, say away from push pull cables. A solid linkage will weigh roughly 1/2 as much, cheaper, and way more reliable. On the FR2000 cars we have to replace the push-pull cable every 4 races, its a real pain in the #@%, I sometimes wish I could just put a solid linkage on it.

Cheers,

Akos

[Charlie]

Akos-

What did you measure as your shift dead-time and what method did you use, when you say you've compared pnuematic and manual shifting?

We've used peak-to-peak longitudinal G's.

-Charlie Ping

I just need enough to tide me over until I need more.

[Ben Beacock]

We used pneumatic shifting on our first car last year. We had 2 buttons on the wheel to activate the solonoids, and the reservoir was charged with an air compressor. The system was very heavy, due solely to the air compressor and would have been very easy to change over to nitrogen but we were short on development time and needed to fix other non-working bits first.

This year will be the same except that we will use nitrogen once the bugs are worked out of the system. The actual control of the solonoids will be in the ECU now, so that we can do automatic full throttle shifts(acceleration event) or ignition cuts in manual mode.

Ben Beacock
Co-Manager
2004 Gryphon Racing - University of Guelph (http://www.soe.uoguelph.ca/uogracing)

[Axle_Greece]

My only question would be "why?". I have never heard of an electronic shifter that was faster than a good driver except for some exotic setups with two clutches like the new bugatti (one clutch for each of two shafts, with every other gear mounted on an alternate shaft, instant shifts)
anyhow, the point is, you still have to have the reaction time to push the button, and the device has to pressurize a cylinder or jam current through a solenoid, and all those things add weight and complexity. I guarantee that my arm will not break during competition, and if it does, the rest of the car will not be drivable. I think it would seriously bite to blow off an air hose and lose the competition (or have an actuator seize up, or whatever) and be the guy who designed the system. I'm designing our traction control system, and i've put in failsafes EVERYWHERE so that if anything goes wrong, we typically won't even have to flip a switch to disable the TCS and still have a perfectly functioning car. I really don't know how to do that with a pneumatic shifter without having a shift knob bouncing back and forth in the cockpit under pneumatic control.

Anyhow, my whole bit is KISS : Keep It Simple Stupid. If it can break, break it, and see if the car still runs. If not, then either remove it or redesign it.

Take care all,
Orion

"They say time is of the essence... give me some and maybe i'll get the essence!"

[Charlie]

Now everyone's starting to get it! Pneumatics are terribly unreliable and can't be robustly designed. They can't be as quick and consistent as a good old human being, and on top of that they are terribly heavy!

2004 will be the year of the manual shifter for sure. http://fsae.com/groupee_common/emoticons/icon_wink.gif

-Charlie Ping

I just need enough to tide me over until I need more.

[Denny Trimble]

So Charlie, what's your peak-to-peak shift time? I posted ours a while ago, I think it was around .28 seconds (manual shifter, no ignition cut, no clutch).

University of Washington Formula SAE ('98, '99, '03, '04)

[Chris Boyden]

The solenoid shifter we designed gives us 100ms shift times, consistent and reliable everytime. That's the best thing about it. Granted, most of us (UNM FSAE) are a bunch of juniors when it comes to driving the car really really well. Shifting consistently can be a challenge for the team overall. The solenoid system really helps to insure quick and accurate shifting.

100ms seemed to be the limit. If we lowered the ignition cut time below a certain level, the system would miss. If we kept the ignition cut the same and lowered the solenoid actuation time, it would miss.

Yep, KISS is a good principle to live by. A solenoid system can be really simple and robust. You don't have to fuss with air tanks or compressors. The solenoid current is large but only for a very short duration. So, the load on the electric system is small overall.

At the very least, an ignition cut with a manual shifter would be pretty slick.

[Charlie]

How are you measuring this 100ms time??

-Charlie Ping

I just need enough to tide me over until I need more.

[Chris Boyden]

Right now, we're just adjusting the microcontroller pulse width time, which is verified with an Agilent o-scope. The total time after the button is pressed is 100ms (ignition cut plus solenoid w/ ignition cut). So there is still the time it takes your thumb to move an inch.

We installed the system on the 03 car, and just got the data logging going as well. I plan on recording the rpm for several runs, manual vs. electronic. Then we'll look at the total duration time of the rpm "dip" during shifts.

I do like your idea of looking at g-sensor data to time the shift. What does the waveform look like?

[Denny Trimble]

Previous Thread w/ Plots (http://fsae.com/eve/forums?a=tpc&s=763607348&f=125607348&m=9546022104&r=3376024954#3376024954)

Our RPM drop was only about .18s (manual, no ignition kill, no clutch), but the G peak-to-peak was .28s.

University of Washington Formula SAE ('98, '99, '03, '04)

[Chris Boyden]

Thanks Denny,

that's interesting data. Are you guys working on an ignition kill?
I found the minimum ignition kill time to be about twice as long as our minimum solenoid on time. like I mentioned earlier, it was interesting to note that if we lowered the ignition kill time below what we've deemed the minimum and left the solenoid time at the "minumum" then the car would miss shift.
That tells me that the transmission hasn't unloaded enough and that it slows the solenoid down because it takes more force to shift.
This is with an 99 R6. Can't wait to try it with the F4i.

So, if you are shifting in .18s with no cut, an ignition cut should help speed up the process. And the dip in the TPS curve would disappear, because you don't need to lift. Although at first, it feels weird to not lift.

[woollymoof]

Has anyone looked into what these solenoid/pneumatic shifters do to the gearbox internals such as:
dogs
gears
selectors
cam/barrel
shifter mechanism
etc

Cheers,

Kirk Veitch
Swinburne University of Technology 2004

[Frank]

we use 100ms ignition cut,

and 40ms phase in

using MoTeC

so is that 100ms shift time or 140ms shift time?

http://fsae.com/groupee_common/emoticons/icon_wink.gif

[Ehsan]

woollymoof, are you concerned about overstressing the transmission components because the shift time is so short? Well I personally wouldn't be concerned. The one item that Japanese bike manufacturers are absolutely, positively amazing is in their transmissions. Now there might have been isolated cases of gear failures, but overall, the transmissions are about as bulletproof as they come. Especially for the low-duration life of an FSAE racer. Maybe if we're talking about doing this in a motorcycle for 35,000 miles, it could be another issue, but not in this application.

Ehsan
MIT FSAE

[Chris Boyden]

I'd call it 140ms. Does the ignition cut remain on during the shift?

We had to rebuild an R6 transmission. The dogs were worn and would pop out of 2nd gear. We also replaced the shift forks. The shift drum didn't show any wear. There was also a broken swivel on the shift lever that messed up the ratcheting action with the geneva mechanism. I think most of the problems were caused by the broken shift lever. But I think some bad shifting with the manual lever played a role as well.

Testing is really the only way to figure out what problems an automatic shifter might cause.

[Alan]

There is an interesting article in Racecar Engineering about a new transmission that has no delay between shifts. There were no details on how the system works but it did say that two gears are engaged simultaneously, albeit for only a little bit.

Kettering University FSAE

[Frank]

we actually damaged a box at the comp this year. (changed the whole engine overnight, to race the next day)

We had no "phase in" set at the time (brain fade)

we're not sure if the lack of phase in was the cause of the damage or not... but we're not going to experiment with it again

thinking no phase in = full HP switch

Frank

[Charlie]

I've heard that the R6 boxes were notoriously weak. The CBR box has been bulletproof for us, and we've subjected it to terrible things.

-Charlie Ping

I just need enough to tide me over until I need more.

[Chris Boyden]

Yea, we switched over to an F4i after rebuilding the R6 transmission.

[Robbo87]

Has anyone experienced any problems with solenoid shifters? We are currently running a Yamaha R6 engine with a Pro-Shift solenoid but it seems to be failing after a few minutes of use. It doesn't seem to be providing enough force to flat shift.

[tinkeringIdiot]

The R6 is notoriously difficult to shift. My personal feeling is that it's just a crappy shifter drum/linkage design over all. Solenoid-based shift systems do tend to be limited in the amount of force they can deliver, which could make them marginal on an R6. Pneumatic shifting allows you to easily increase the force of your shifting ram, but has its own host of issues to overcome.

[SamB]

We're on an R6 and have used the kliktronic unit for the past 2 years and it seems to be very tempromental. In the sprint at silverstone it worked, then in the endurance it just stopped after about 10 mins. Although on the other hand, the other driver didn't have as many problems. Saying that, in 2008 it worked ok but everyone who saw it looked very skeptical about it and we found out why last year. Hopefully switching to pneumatic this year though...

[Robbo87]

As I understand it, pressurised systems like pneumatic gear shifters need to have extra structural components to protect the driver. Do you find this adds a lot of weight? How have you got round this?