Tired torsen? 4 channel wheel speed logging
#1
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
Tired torsen? 4 channel wheel speed logging
I had my first sprint (for both myself and the car) yesterday and it was my first time launching the car. The car did well, but one issue stuck out in particular, it sounded like it was doing some severe wheel hop off the line. This was evidenced by me having one of the slowest 64ft times of everyone there, 2.75s was my best!
A friend filmed the start and you can see there is what appears to be a torsional mode of the diff with the right rear (and presumably the left rear in antiphase, since the engine speed seemed smooth) visibly speeding up and slowing down.
Pertinent car info
FM2 with GTX2860 ~280whp here
Supermiata big grip kit 700/400 and race alignment
15x8 with 205 Federal RS-R (yes I know but the tyres on my 9Js weren't legal)
6 speed w/3.6 torsen type 2
BBFW low NVH mounts and diff fillers
I have datalogging with wheel speed sensors at each corner and this is what it showed (note, may not be same run as the video, but it did it every time)
Frustratingly, the data doesn't look to be logged at a high enough rate to determine whether it was in phase or anti-phase (it says 25Hz, but that's just the polling frequency, it only updates at 12.5Hz) but this isn't what I would have expected. Has anyone taken similar data that could show to compare?
The odd behaviour doesn't end there, the first turn (visible on right of above plot where MAP goes positive again) is a 3rd gear left hander and it looks like the diff is having the same torsional mode but at times goes open. It seems to me that having a torsional vibration of that kind of amplitude is just going to break the rear tyres loose all the time like an impact gun.
I've only ever driven this car on the Supermiata Big Grip Kit and I've never been fully comfortable with how it's driven. I've changed the rake, the alignment, the ARB settings, swept front to rear damper clicks, tried 225/9, 205/8, 195/7 road tyre (wet), removed and checked (measured stiction) at every bush, changed all ball joints and never been happy with how the car transitions from off power to on-power mid corner and has zero traction on exit. Every time I rake tune, I end up with the rear so much lower than the front I tell myself that's ridiculous, can't be right and raise it back up. I even tried rear ARB off and it was only a tad pushy. Whereas I tried this set of dampers (but stock ARBs) on an otherwise stock NB with supermiata race alignment and the same 205/8 Federal RS-R and it was by far the nicest handling Miata I've driven. So I'm hoping the diff is the cause of all of this!
Anyway, here's a video of the sprint. I managed to win the class (A5 - 2WD >2301cc road cars) despite just running the car I had, rather than building one to the rules. The driver of the GT3 RS in second place didn't seem best pleased with a rookie in a 25 year old Mazda beating him!
A friend filmed the start and you can see there is what appears to be a torsional mode of the diff with the right rear (and presumably the left rear in antiphase, since the engine speed seemed smooth) visibly speeding up and slowing down.
Pertinent car info
FM2 with GTX2860 ~280whp here
Supermiata big grip kit 700/400 and race alignment
15x8 with 205 Federal RS-R (yes I know but the tyres on my 9Js weren't legal)
6 speed w/3.6 torsen type 2
BBFW low NVH mounts and diff fillers
I have datalogging with wheel speed sensors at each corner and this is what it showed (note, may not be same run as the video, but it did it every time)
Frustratingly, the data doesn't look to be logged at a high enough rate to determine whether it was in phase or anti-phase (it says 25Hz, but that's just the polling frequency, it only updates at 12.5Hz) but this isn't what I would have expected. Has anyone taken similar data that could show to compare?
The odd behaviour doesn't end there, the first turn (visible on right of above plot where MAP goes positive again) is a 3rd gear left hander and it looks like the diff is having the same torsional mode but at times goes open. It seems to me that having a torsional vibration of that kind of amplitude is just going to break the rear tyres loose all the time like an impact gun.
I've only ever driven this car on the Supermiata Big Grip Kit and I've never been fully comfortable with how it's driven. I've changed the rake, the alignment, the ARB settings, swept front to rear damper clicks, tried 225/9, 205/8, 195/7 road tyre (wet), removed and checked (measured stiction) at every bush, changed all ball joints and never been happy with how the car transitions from off power to on-power mid corner and has zero traction on exit. Every time I rake tune, I end up with the rear so much lower than the front I tell myself that's ridiculous, can't be right and raise it back up. I even tried rear ARB off and it was only a tad pushy. Whereas I tried this set of dampers (but stock ARBs) on an otherwise stock NB with supermiata race alignment and the same 205/8 Federal RS-R and it was by far the nicest handling Miata I've driven. So I'm hoping the diff is the cause of all of this!
Anyway, here's a video of the sprint. I managed to win the class (A5 - 2WD >2301cc road cars) despite just running the car I had, rather than building one to the rules. The driver of the GT3 RS in second place didn't seem best pleased with a rookie in a 25 year old Mazda beating him!
#2
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
In an attempt to understand what is actually going on, I've exported the data to Excel and had a play around with it. Having the lines on top of each other let me see a few things
1. When I've almost released the clutch, the revs reach their lowest and I'm starting to add TPS, (31.2s on plot) the rear left and right wheels (red and light blue) lines start to oscillate in anti-phase
2. This oscillation gets worse and worse and by 32.3s on plot the oscillation in phase. I'm not even sure what could be causing an oscillation this big at this kind of frequency. It doesn't appear to be clutch because calculating the speed of the gearbox output/diff input shaft from the rpm channel gives a value that appears to be close to a time averaged value of the left and right wheel speeds. I have to back out to <25% TPS for over half a second to bring it back under control. Could this be a severe case of traditional wheel hop and the entire drivetrain/PPF is bouncing about? The mode is ~9Hz. This is with BBFW low NVH engine mounts and diff fillers in the rear. Gearstick vibration is visible in the onboard, but I've not yet done any calculations as to whether that kind of vibration could cause this kind of speed fluctuation.
3. When I shift into second, as soon as I tip in, the same behaviour resumes
4. The GPS speed lags the front wheel speeds by 0.2-0.3s which seems high!
It turns out the wheelspeed data was logged at a true 25Hz, but I can turn it up the logging frequency 500Hz. There probably aren't enough teeth to get usable data at this frequency and speed but I'll get some measured, though I'm away for a while so can't do it soon.
I'm struggling to take conclusions out of this data though.
1. Is my torsen broken? If yes, maybe another torsen would be enough for me for the moment, if not, I'd better get eating ramen and saving for an OS Giken.
2. Maybe I need stiffer diff and engine mounts, but the oscillation may be exacerbated by the torsen being broken?
1. When I've almost released the clutch, the revs reach their lowest and I'm starting to add TPS, (31.2s on plot) the rear left and right wheels (red and light blue) lines start to oscillate in anti-phase
2. This oscillation gets worse and worse and by 32.3s on plot the oscillation in phase. I'm not even sure what could be causing an oscillation this big at this kind of frequency. It doesn't appear to be clutch because calculating the speed of the gearbox output/diff input shaft from the rpm channel gives a value that appears to be close to a time averaged value of the left and right wheel speeds. I have to back out to <25% TPS for over half a second to bring it back under control. Could this be a severe case of traditional wheel hop and the entire drivetrain/PPF is bouncing about? The mode is ~9Hz. This is with BBFW low NVH engine mounts and diff fillers in the rear. Gearstick vibration is visible in the onboard, but I've not yet done any calculations as to whether that kind of vibration could cause this kind of speed fluctuation.
3. When I shift into second, as soon as I tip in, the same behaviour resumes
4. The GPS speed lags the front wheel speeds by 0.2-0.3s which seems high!
It turns out the wheelspeed data was logged at a true 25Hz, but I can turn it up the logging frequency 500Hz. There probably aren't enough teeth to get usable data at this frequency and speed but I'll get some measured, though I'm away for a while so can't do it soon.
I'm struggling to take conclusions out of this data though.
1. Is my torsen broken? If yes, maybe another torsen would be enough for me for the moment, if not, I'd better get eating ramen and saving for an OS Giken.
2. Maybe I need stiffer diff and engine mounts, but the oscillation may be exacerbated by the torsen being broken?
#4
The reason why you like the stock powered car so much is because you can drive the throttle like a lightswitch. Go from 0% to 100% throttle coming out of a turn and nothing weird happens. Part of it is just realizing that you're not going to be able to put the pedal straight to the mat coming out of lower speed corners in autocross environments.
Wheel hop can be helped with better suspension bushings, better diff bushings, better engine mounts (the good kind, not MazdaComp or stock).
Your torsen could also be tired but if you're looking for outright performance then maybe consider just going straight to the OS Giken. One of the downsides of the torsen is that if one wheel gets really loose, it can't recover from it and suddenly "lock up", it'll just let it spin, so if one wheel has much less weight and you throw all 300WHP at it at once, well, game over. Throttle control is key.
Wheel hop can be helped with better suspension bushings, better diff bushings, better engine mounts (the good kind, not MazdaComp or stock).
Your torsen could also be tired but if you're looking for outright performance then maybe consider just going straight to the OS Giken. One of the downsides of the torsen is that if one wheel gets really loose, it can't recover from it and suddenly "lock up", it'll just let it spin, so if one wheel has much less weight and you throw all 300WHP at it at once, well, game over. Throttle control is key.
#5
OS Giken will be far easier to drive than Torsen with 300whp. As Marcus pointed out, the Torsen has a more abrupt nature when one wheel begins to slip. The OSG masks that. Customers often ask us, "what will the Giken feel like?". Our answer is that when a diff is working correctly, you don't feel anything. The car just accelerates and turns according to throttle inputs and is more stable overall.
Torsens do wear out. Only two sources in the US I know of that service them. Both are Spec Miata shops in the midwest. Search spec miata group on facebook for more info on that.
Torsens do wear out. Only two sources in the US I know of that service them. Both are Spec Miata shops in the midwest. Search spec miata group on facebook for more info on that.
__________________
#8
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
Minor update,
I upped the logging frequency of the driven wheels to 250Hz, just took a quick bit of data on the road here, 1st gear right turn and the inside wheel spins up and oscillates at 10Hz but the outside one isn't affected while this is happening. Seems very odd to me! I'll be able to get some proper data tomorrow during a track evening at Brands Hatch.
I'm trying to buy an OSG, but need it pretty quickly to have it in time for my next Sprint, since Supermiata are still out of stock, I'm wondering how much worse a European sourced Spec-S will be. I'm guessing it will still be night and day vs a Torsen.
I upped the logging frequency of the driven wheels to 250Hz, just took a quick bit of data on the road here, 1st gear right turn and the inside wheel spins up and oscillates at 10Hz but the outside one isn't affected while this is happening. Seems very odd to me! I'll be able to get some proper data tomorrow during a track evening at Brands Hatch.
I'm trying to buy an OSG, but need it pretty quickly to have it in time for my next Sprint, since Supermiata are still out of stock, I'm wondering how much worse a European sourced Spec-S will be. I'm guessing it will still be night and day vs a Torsen.
#10
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
https://www.ecumaster.com/products/adu/
I mainly wanted it to be able to have the brake locking lights at the side of the dash like a Porsche cup car. I got them to work but need to make a few presets for different tyres. You can see the lights working when trail braking in this video of a lap of Brands Hatch (it also shows the snappy nature of the diff!) I'm going back tonight so will capture some high resolution data of this.
The ADU is a cool piece of kit. I've got the GPS module with an inertial measurement unit and it can both receive and transmit data to the MS3 via CAN.
#11
Bit twitchy coming out of Paddock Hill bend there!!
If you aren't 100% on a Giken, it might be worth giving 3J a call - https://www.3jdriveline.com/
They have a plate type diff available in fair few setups, or can do a custom setup and it's a LOT cheaper:
https://www.3jdriveline.com/webshop/.../3j-mazda-mx5/
If you aren't 100% on a Giken, it might be worth giving 3J a call - https://www.3jdriveline.com/
They have a plate type diff available in fair few setups, or can do a custom setup and it's a LOT cheaper:
https://www.3jdriveline.com/webshop/.../3j-mazda-mx5/
#12
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
Bit twitchy coming out of Paddock Hill bend there!!
If you aren't 100% on a Giken, it might be worth giving 3J a call - https://www.3jdriveline.com/
They have a plate type diff available in fair few setups, or can do a custom setup and it's a LOT cheaper:
https://www.3jdriveline.com/webshop/.../3j-mazda-mx5/
If you aren't 100% on a Giken, it might be worth giving 3J a call - https://www.3jdriveline.com/
They have a plate type diff available in fair few setups, or can do a custom setup and it's a LOT cheaper:
https://www.3jdriveline.com/webshop/.../3j-mazda-mx5/
#13
Former Vendor
iTrader: (31)
Join Date: Nov 2006
Location: Sunnyvale, CA
Posts: 15,442
Total Cats: 2,103
Big mistake IMO. The car makes 3x stock power, you're putting an OSG tuned for stock power in. That means you will end up with a twitchy car as the diff goes to full lock at 1/3 to 1/2 throttle. The SPM tune smooths that out by allowing the diff to progressively lock. I would source it from them.
#14
Big mistake IMO. The car makes 3x stock power, you're putting an OSG tuned for stock power in. That means you will end up with a twitchy car as the diff goes to full lock at 1/3 to 1/2 throttle. The SPM tune smooths that out by allowing the diff to progressively lock. I would source it from them.
#15
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
Big mistake IMO. The car makes 3x stock power, you're putting an OSG tuned for stock power in. That means you will end up with a twitchy car as the diff goes to full lock at 1/3 to 1/2 throttle. The SPM tune smooths that out by allowing the diff to progressively lock. I would source it from them.
I did really want to get the SPM tune but with no firm ETA on them (and them still being the other side of the world when they do) I pulled the trigger on a European one to get it sooner as this car has spent way too much time in the garage, rather than on the track. Though it would have probably changed my mind if I'd read that before I'd ordered it....
#16
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
Time for an update, I've had the diff in the car for a month now. It arrived quickly, and I had the diff shimmed and installed into the old casing, giving me the Spec-S tune with the 3.6 ratio. I added poly diff mounts for good measure. It actually arrived a bit earlier than expected, so I was able to get a last minute entry to a sprint at the Lotus test track (Hethel).
Before I got there, I wanted to see whether the diff had fixed the inside wheel oscillation, so I went back to the same roundabout as before, and tried to replicate the manoeuvre I plotted earlier. The short answer; sort of, the long answer is that the inside wheel still starts to oscillate like the torsen, but the initial amplitude doesn't worsen like it does with the torsen. I don't think the OSG would be able to vibrate like that, so that led me to believe that it's either driveshaft twist or some form of compliance allowing the entire upright (and so the abs sensor) to have a relative angular acceleration.
I've no videos from Hethel (not allowed to film at a development facility) but there was still a lot of wheel hop at Coventry Motofest (this is a ridiculous event, they close the main ring road of a fairly large city in the UK, and have a Sprint around it!)
So, did the diff fix the wheel hop? No…. The Coventry data was in the damp with 205 AR1s, so it was never going to be a good launch, but the data doesn’t appear significantly different to before from a launch aspect.
Here’s a video of the launch.
Here’s a video of the run, the wheel hop is obviously quite bad. BUT, the OSG has completely fixed the snappiness mid corner transitioning to power of the Torsen, which was very helpful. That was only my 4th (and final) lap of the track, which all looks very samey and isn’t the easiest to learn!
After the sprint, I took the car to my local track (Brands Hatch) and ran some laps to get a feel for the setup. Now with zero rake (rear raised 10mm from before!!!!), I had to make rear ARB full stiff and the balance on power was still a tad pushy, rather than wildly oversteery like it was with the torsen. If anyone is thinking about getting an OSG, just look at the trace below of TPS around a lap, and guess which is the OSG…. Even on the street, it’s so much nicer. Yaw gain is now a linear function of TPS, rather than nothing, nothing, nothing, LOTS!
So in summary, the OSG
· Hasn’t improved the launch wheel hop noticably
· Has completely transformed the car, and it now handles like I had expected a supermiata big grip kit car to
Now to try and solve the wheel hop...
Before I got there, I wanted to see whether the diff had fixed the inside wheel oscillation, so I went back to the same roundabout as before, and tried to replicate the manoeuvre I plotted earlier. The short answer; sort of, the long answer is that the inside wheel still starts to oscillate like the torsen, but the initial amplitude doesn't worsen like it does with the torsen. I don't think the OSG would be able to vibrate like that, so that led me to believe that it's either driveshaft twist or some form of compliance allowing the entire upright (and so the abs sensor) to have a relative angular acceleration.
I've no videos from Hethel (not allowed to film at a development facility) but there was still a lot of wheel hop at Coventry Motofest (this is a ridiculous event, they close the main ring road of a fairly large city in the UK, and have a Sprint around it!)
So, did the diff fix the wheel hop? No…. The Coventry data was in the damp with 205 AR1s, so it was never going to be a good launch, but the data doesn’t appear significantly different to before from a launch aspect.
Here’s a video of the launch.
Here’s a video of the run, the wheel hop is obviously quite bad. BUT, the OSG has completely fixed the snappiness mid corner transitioning to power of the Torsen, which was very helpful. That was only my 4th (and final) lap of the track, which all looks very samey and isn’t the easiest to learn!
After the sprint, I took the car to my local track (Brands Hatch) and ran some laps to get a feel for the setup. Now with zero rake (rear raised 10mm from before!!!!), I had to make rear ARB full stiff and the balance on power was still a tad pushy, rather than wildly oversteery like it was with the torsen. If anyone is thinking about getting an OSG, just look at the trace below of TPS around a lap, and guess which is the OSG…. Even on the street, it’s so much nicer. Yaw gain is now a linear function of TPS, rather than nothing, nothing, nothing, LOTS!
So in summary, the OSG
· Hasn’t improved the launch wheel hop noticably
· Has completely transformed the car, and it now handles like I had expected a supermiata big grip kit car to
Now to try and solve the wheel hop...
#17
In my experience drag racing once upon a time and with trying to launch a Miata the tire used is about 75% of driver in how bad you get wheel hop. Any kind of compliance in the control arm bushings are 20% and the last 5% is everything else. This may sound dumb but blowing the tires off a bit on the launch may be smoother with whatever tire you're running, the fans will like it at least.
#18
Junior Member
Join Date: Dec 2015
Location: Brisbane, Queensland, Australia
Posts: 214
Total Cats: 67
Harder launch...
^^ echoing a Bronson’s comments. Launching harder for more initial wheel spin is a way to get past wheel hop when setting off.
I have the spec s tune OSG largely due to the additional cost of buying Japanese via America with Australian dollars and postage. I also bought it when my Torsen started to fail (260hp). It’s rather good out of the box.
I have the spec s tune OSG largely due to the additional cost of buying Japanese via America with Australian dollars and postage. I also bought it when my Torsen started to fail (260hp). It’s rather good out of the box.
#19
Junior Member
Thread Starter
Join Date: Jul 2017
Location: Sussex, England
Posts: 223
Total Cats: 47
In my experience drag racing once upon a time and with trying to launch a Miata the tire used is about 75% of driver in how bad you get wheel hop. Any kind of compliance in the control arm bushings are 20% and the last 5% is everything else. This may sound dumb but blowing the tires off a bit on the launch may be smoother with whatever tire you're running, the fans will like it at least.
^^ echoing a Bronson’s comments. Launching harder for more initial wheel spin is a way to get past wheel hop when setting off.
I have the spec s tune OSG largely due to the additional cost of buying Japanese via America with Australian dollars and postage. I also bought it when my Torsen started to fail (260hp). It’s rather good out of the box.
I have the spec s tune OSG largely due to the additional cost of buying Japanese via America with Australian dollars and postage. I also bought it when my Torsen started to fail (260hp). It’s rather good out of the box.
Thread
Thread Starter
Forum
Replies
Last Post
apexanimal
Suspension, Brakes, Drivetrain
24
08-17-2019 08:10 PM