RT rescued from graveyard

[megagame]

Next step is to finish tube subframe to hold both oil and cooler radiator. It's partly finished but still needs some work to do to hold plastic parts and then zinc plating.

Final thing will be hand brake lever. Braking with feet is so strange. I have another smaller three-wheeler with hand and foot braking and I never use foot because it's too slow and unpredictable for sudden situations and without nanny everything is sudden situation. I still need to get used to driving this vehicle. It is very strange and unstable. Not sure about newer versions, but mine has very long wishbones which is poor design in my opinion, and after reading about struggles of other people with bump steer, I understand why there are so many safety systems present.

 

[Mikey]

:clap::bowdown: Wow, what do you do in your spare time!! My hat comes off to you sir!! This just shows that if there's a will there's a way around things that drive us crazy. Good job!!!

 

[megagame]

:clap::bowdown: Wow, what do you do in your spare time!! My hat comes off to you sir!! This just shows that if there's a will there's a way around things that drive us crazy. Good job!!!

Thanks . I hope this kind of information will help people with their struggle with fixing things, and maybe encourage new to analyze how things work. There is so much electronics everywhere and quite often no way of fixing it - this drives me crazy. So I'm trying to prove to myself, and a bit to others, that things can be fixed and it makes sense to do it.

 

[BertRemington]

megagame -- thank you very much for this information.

I had searched for CANBUS information but never discovered your project. Thank you for decoding those messages. I never would have decoded the XOR checksum calculation although I'm only receiving messages not transmitting any.

Here's some message information I've decoded:

// Can-Am Spyder uses Keyword Protocol 2000 version of CANbus 2.0a
// CANbus 2.0a is Standard 11-bit Message Identifiers
// CANbus 2.0b with Extended 29-bit Identifiers was pre-released 18 Apr 2012
//
// ECM = 102h 103h 120h 122h 300h 308h 310h [342h] 343h
// TCM = 220h [221h 222h 223h]
// VCM = 020h [021h] 230h 430h
// WPM/EPB = 460h
// Cluster = 400h [401h 402h]
// MSL-MSR/SWM = 340

// Multi-Function Switch (Left) functions via CANBUS [0x340 message identifier]
// Reverse: Free-Reverse
// PTT: Free-Push-to-Talk [Byte 2: 40]
// Shift: Free-UpShift-DownShift
// Windshield: Free-UpWindShield-DownWindShield
// RECC: Free-Up-Down-Left-Right
// Mode: Free-Mode
// Set: Free-Set
// Turn: Free-Left-Right-Cancel [Byte 3: 01=LT, 02=RT, 04=CT]
//
// Multi-Function Switch (Left) functions via direct wire
// Horn: Free-Horn
// Headlight: Low-High-Pass
//
// Multi-Function Switch (Right) functions via MSL via CANBUS
// Cruise: Off-On-Set-Resume
//
// Multi-Function Switch (Right) functions via direct wire
// Start: Free-Start
// Run: Stop-Run
// Hazard: Free-Hazard

I'm using an Arduino Uno with KS0411 CANBUS Shield. The GitHub CANBUS library was kinda old and patched together so rewrote most of it. I had to learn another language (C++). I'm only using round-robin scheduling -- no events whatsoever. The Duo is a bit cramped for memory but then I'm not using much either.

To capture the CANBUS messages I used a reader ("sniffer") but it's filter function doesn't work so I'm overwhelmed in messages and picking out the ones of interest is difficult. So again thank you for sharing your decoding work.

I have two other projects keeping me from completing this project and will create a new topic when I get a chance to finish it.

PS I have a black cat too.

 

[BertRemington]

CANBUS messages as Word .doc

I reformatted megagame's CANBUS message decoding HTML table as Word table.

Thanks again megagame for this valuable information. You're my inspiration to get back to work on my Spyder CANBUS project!

 

[BertRemington]

megagame -- concur with "long wishbones" leading to steering issues. Rubber bushings contribute. I'm thinking of asking a friend to mill some Delrin bushings. I've identified the material but haven't pushed the Buy button yet.

 

[BertRemington]

What I see is that within timespan of 12s steering is moved to right, left, right left and so on.
Data which looks like torque (0x150) changes in first 4s and later is mostly the same.
There is one bit in 0x150 which changes after 0,5s when steering is sensed, and after 2s it is back to old state.

megagame, Idaho -- I'm thinking the Cluster uses this information, along with road speed, to determine turn signal cancellation. Having the DPS monitor and perform steering angle movement greatly simplifies Cluster programming requirements. Having tried to replicate the Spyder's turn signal cancellation, this 150h message is a tremendous simplification.

I'm checking to see if my CANBUS sniffer software has been updated to provide filtering...

 

[safecracker]

WOW, what a great job you are doing. Bruce

 

[megagame]

I reformatted megagame's CANBUS message decoding HTML table as Word table.

Thanks again megagame for this valuable information. You're my inspiration to get back to work on my Spyder CANBUS project!

Seeing how this could help someone makes me very happpy . If you have any questions feel free to ask.

I'm not C++ programmer either . Normally I work on Microsoft erp software - something completely different than microcontrollers, and there is working debugger there. With arduino/teensy it's hard and slow process but once it works, it works forever.
My personal choice was teensy because of speed. With so many messages it's already almost at it's limit when you need to do some small additional stuff and TX. RX only should be fine though with arduino. The problem with TX is that delay of 1ms will in some cases be too much and ECM will show error with missing device.
But again, RX only and arduino should be fine.

 

[megagame]

megagame -- concur with "long wishbones" leading to steering issues. Rubber bushings contribute. I'm thinking of asking a friend to mill some Delrin bushings. I've identified the material but haven't pushed the Buy button yet.

If Delrin is POM than it's a very good material, way better than PA. There are some variations of POM, some with better greasing capabilities, but overall, it's very good. Very nice to work with (lathe/milling machine) and very tough. Looking at shore parameters it's similar to harder polyurethane bushings (90 shore) so it should be fine. Also, wishbones are made of steel tubes on Spyder so the whole suspension has some elasticity if bushings were to be too hard.

 

[megagame]

megagame, Idaho -- I'm thinking the Cluster uses this information, along with road speed, to determine turn signal cancellation. Having the DPS monitor and perform steering angle movement greatly simplifies Cluster programming requirements. Having tried to replicate the Spyder's turn signal cancellation, this 150h message is a tremendous simplification.

I'm checking to see if my CANBUS sniffer software has been updated to provide filtering...

Don't forget there is separate sensor for turn angle below DPS - its message id is 0x0C3. I would assume this is used to sense when steering is straight again.
You could always disconnect DPS for a moment and check if cancellation still happpens. Unfortunately I will be away for a week because of vacation so I'm not able to check how it works on mine without DPS.

 

[megagame]

After couple of rides, I've decided to stiffen the front end a little bit. Fox suspension is great, and I wouldn't like to change how it behaves on uneven roads, but there is too much swinging from left to right during normal riding, and especially on long turns. My wife hates it when the Spyder leans so much on turns.

I've decided to buy a BajaRon Sway Bar, but first I would like to try doing my own. I've bought 42CrMo4 (4140) bar with min. 31HRC. It's already in +QT state so it should be fine as a sway bar with limited movement. It's not spring steel, but it is used in sports where a sway bar has limited movement compared to the OEM. I will probably do some stress tests to check how far it can go without doing permanent damage.

I will bend it cold, no welding of course. If that will still be too stiff, the center part will be brought down to a lower diameter with carbide tools on a lathe. 4140 is quite cheap ($10USD for a 1m length of 14mm bar) so testing it will not harm my budget.

 

[KX5062]

Amazing work! Keep us all in the loop as you progress. :congrats:

 

[megagame]

I've made some tests on a standard sway bar. I've checked how much it rotates on its own axis when 10kg is applied at the end. Red square indicates the part which is held by the vice. Green indicates steel plates which are holding the sway bar where the blue lines are. Thicker green line indicates the steel bar which is going under the sway bar and acts as a lever. With that sway bar, it is rotating without any movements away from its axis. When 10kg is applied 50cm from the center of the sway bar, it rotates such that the end of the plate is lowered by 6.5cm. This kind of test is not perfect because not only the sway bar is bending, but the green plate as well. It's not a problem though, because the same test with the same plate will be performed on the new sway bar, thus the error will be the same on both of them and I'm interested in the difference in bending. Based on test 2, I'm expecting around 4cm rotation with the new bar - but we will see .

I've made an additional test (test 2) to measure how much both bars bend when the force is applied perpendicular to the bar, just to compare later if the bending difference is the same on both tests. The new bar is 14mm in diameter, the old is 13.5mm (it is a tube not a bar/rod). If the same difference is visible during test 1 on the new bar as on test 2, then the new bar will be 50% stronger (which is not that much). TBH, I was expecting 100% but forgot to measure the diameter of the old bar - I was sure it's less than 13mm. I'm quite curious how much stronger BajaRon's sway bar might be. I'm sure it's tougher and it will be hard to buy rod with the same strength without additional heat treatment.
If I decide to buy from him, I will ask how big of a difference (%) in strength it has compared to standard sway bar. Maybe he will allow me to share that information.



Seeing how much the standard sway bar rotates with just 10kg, it's amazing that BRP allowed that on our Spyders. 6.5cm is more than 2 inches. It's like 1 inch up on one side, 1 inch down on another. 1 inch movement where the sway bar is connected is at least 3 inches where wheel is. Off course, this is 50cm from the center of the sway bar, and the actual value is probably 1/5th of that (from the center of the sway bar to the bar ends), still, it would be more than 0.5of an inch with just 10kg at 50cm from the sway bar. Quite a lot. With a strong wind you could expect to lean that much between left and right and change direction on the road as a result!

 

[megagame]

I've made measurement with new bar with test 1 and it's not as good as I was expecting originally but in line with test 2. Basically it will be around 30% stiffer than original, but still far from 100%. I will try to look for 15mm and 16mm bars online.
For bending I've used hydraulic tube bender which can be bought for around 100 USD. I've had one already because of other projects so no additional cost there fortunately.

 

[davev1pa]

Just a quick "kudos" to megagame and BertRemmington.....two members that have enriched the last two years of SPYDERLOVERS with their persistence and knowledge in repairing what seemingly would appear to most of us as "lost causes" in the revival of wrecked Spyders. I would love to see the two of these guys get together in a shop environment. I am amazed at the tenacity and effort that these two guys put into these projects....I promised to never bitch about an oil change or rear tire removal I do myself or any expense I incur at a shop. Thank you guys you have shown us all the endurance of the human spirit.

 

[Rednaxs60]

Just a quick "kudos" to megagame and BertRemmington.....two members that have enriched the last two years of SPYDERLOVERS with their persistence and knowledge in repairing what seemingly would appear to most of us as "lost causes" in the revival of wrecked Spyders. I would love to see the two of these guys get together in a shop environment. I am amazed at the tenacity and effort that these two guys put into these projects....I promised to never bitch about an oil change or rear tire removal I do myself or any expense I incur at a shop. Thank you guys you have shown us all the endurance of the human spirit.

Agree wholeheartedly. These gents and the other contributors have gone over an beyond what us mere mortals would do. I also hate throwing things away. My 1985 Honda Gold Wing FI model ECU replacement project pales in comparison. Do like the Arduino family though. Have liked following and will stay to the end.

 

[megagame]

It is frustrating quite often when thing go wrong, many times multiple things in a row.
But it makes me happy when I read such kind words . When I sometimes question this whole project, small things like that allow me to keep going.

So, back to the topic .

I've made some calls to spring producers and steel suppliers and found one company which sells high quality steel bars and tubes. It is around 2km from my hope, and they've had 16mm 40HM+QT in their warehouse. I've had to buy one whole bar (6m), but it was around 30 USD.
Again, I've made same tests and I think I'm finally happy with the results. I've updated the image in my previous post with tests. Test 1 indicated around 3.5-3.7cm which is almost twice as strong as original. But there is an error for sure (explained also in previous post), and the bigger the difference in results there is, the more significant error is when calculating change 100%. For example, if 1cm of bending is due to error, then actual results are 5.5, 4.1 and 2,5. 2,5 is more than twice as strong as 5.5. Fortunately I was able to observe how everything is bending during tests and I don't thing there is so much additional bending due to material as test tool.

There is another thing I'm changing which is ride height. I don't want to stiffen the front shocks but I am sometimes hitting road with front plate under the trunk when there are big holes. So, I've decided to add 4cm of additional height to the front.
With FreeCad I've started to check where new lower mount for shock should be to add additional height. Having in mind that I don't want to destroy old mounting point, there are two possibilities - add mounting plate with hole lower and closed to frame, or move it up.
Depending on new location differrent forces might appear.



This is my overview for left lower arm. When shock mount is placed further from arm (up direction), then separation of mount from arm might occur. When it's close to the arm, and loser to the frame, there is higher probability that arm will bent because shock will act at start of lever.
Current original placement is almost perfect, because force from shock is directed to where force from wheel is applied.
So, i've decided that I will choose this option which guarantees lowest distance of extension of fforce fro shock to point which acts as force from wheel (end of lower arm) - option number 2. But, I will extend the mount towards the frame to minimize possibility of cracking near the arm.

Some calculations for this option in ca software:



Originally mounting hole is around 4.5cm from center of the arm. Now it should be 9cm.

If, because of shape of current mount, welding and shaping of additional plates will be difficult, it's possible i will just add some simple plates to have additional hole in the mounting. There holes will not hold the shock in that scenarion, instead they will hold separate mounting adapter.

This option would act as extension for the shock which would keep the forces in same place. And making new extension adapter would be fairly simple and would not require too much welding on the arm. I'm very much affraid of tearing and cracking and my welding skills are far from perfect. This is critical place so attention to details is really important here. There must be no doubts about rigidity of this modification.

 

[sledge]

I've moved photos to google. Don't know why but forum server probably rejected them. They were there at the beginning but disappeared. Now they are back.

i save everything on a stick.... and i always have them.

 

[megagame]

I've finished the subframe for holding the plastic side parts and most importantly, the oil cooler and radiator. I have a small shop nearby which powder coats rims and they've been able to fit my frame into their oven, so they powder coated some stuff for me.

I've resized images to 800x600 but size of 150kB is still too big for an attachment, so I'm adding images directly. Sorry for the inconvenience.
I can make it black and white and it will fit into attachment if anybody is interested.

I've made new brackets to hold the sway bar and instead of rubber I've used polyamide.

This is the press I've used to bend the new sway bar.

New sway bar is 16mm 40HM+QT bar. After some milage done on the Spyder, I must say that now it behaves as it should.

This is how I've been measuring the amount of force required to twist the old and new sway bars.

With the new sway bar, it still reacts as expected to strong, fast forces; for example, rocks on the road or speed bumps. In those situations, the force is too big for the sway bar to stop it and therefore the shock absorber is working when it should. But when it comes to small forces on uneven roads or on roundabouts, the shock is already static, and the sway bar helps stiffening the wheels together with the frame which causes less unexpected and unsteady behaviour during normal riding. Also, it helps very much during acceleration when the Spyder is leaning towards the back which has only one wheel, and it is very important to have wheels working together at the same time. Otherwise, it starts to wobble from left to right.
Since I don't have any safety mechanisms, I quite often see one of the front wheels off the ground, so it's very important for me to have front of the Spyder working as best as possible.

BTW. Considering amount of work I've put into the new sway bar, I must say that having the possibility of buying one at a cost of $300 USD is very much ok. Unfortunately, posting to Europe costs a lot now, and exchange rate EUR/USD is also not so good at the moment; so for me it would be more like $500 USD comparing to average income. But with my own work, counting the material costs (& I still have enough for another 6 sway bars! ) it was around $100 USD, so I'm happy.

Next to do is finishing the plastic parts, especially near my feet, because of the high temperature in traffic (probably from the oil cooler), then put an aluminum plate underneath the whole front, additional netting in front of the radiators to stop the bigger particles coming off the road, and finally a hand brake lever which is a must. Braking with just my foot is horrible. I have another 3-wheeler and I never foot brake there because it's too slow in sudden situations. Especially without ABS, brake modulation is a must and with my foot there is none.