feasibility of designing MR shock system for aftermarket 1

[bluly]

Hello all,

I'm new to these forums, and stumbled upon them while searching for MR information. I found a previous thread, but it's about 5 years old (and closed). I have interest in designing an MR system for my car (2012 Subaru Impreza WRX). I know this is still expensive technology, so working with an MR company such as Lord or BMI directly is not really an option. My first thought is to purchase some production replacement MR shocks, physically adapt them to fit, then design a control system to control them. Any thoughts? Has anyone had experience with MR shocks outside of oem applications?

Thanks!

 

[GregLocock]

Not me

Do you have a shock dyno?

My main thought is that if they are adjusting the viscosity on a millisecond by millisecond basis then the basic cal of the damper may be completely unusable, and that the modification may need to be made that rapidly, in which case you'll need a good idea of a strategy

On the other hand if the viscosity is just being treated like a gain, with some fairly simple rules for adjusting it then I suppose you might be in with a chance.

Sorry that's not much help.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[bluly]

I do have access to a shock dyno.

What I had in mind was not necessarily a millisecond by millisecond input, but rather a less sophisticated idea. The system would consist of 3 settings, of which would adjust the gain (much like a manually adjustable shock). They would also incorporating some sort of accelerometer so the suspension would remain soft in it's relative gain, but would stiff up under hard braking or hard acceleration.

 

[GregLocock]

Sounds doable. So basically setting 1 would replicate a soft shock, then your other settings would harden it up, and if the suspension was using a lot of travel then harden it up again.

Here's some curves for adjsutable shocks

http://fatcatmotorsports.com/FCM_Adjustable_Bilstein_Miata.htm

You could get a lot cleverer than that if you can pick up some signals from the Canbus, although that might get rather scary eg harden the shocks if the steering wheel velocity exceeds X seems like a good idea, but it'll give you understeer when you might not want it.





Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[nobog]

Arctic Cat did this on their snowmobiles 10ish years ago:

Utter failure, suppose to update the dampening at 1000 Hz, wires going to some "brain". 2 year market run.

or maybe this - mostly on Polaris snowmobiles:
A Fox position sensative shock - hard - soft in the middle - hard, these actually work, no wires, no knobs, I think there are 5 different versions on the market.

Jim

 

[gt6racer2]

Bluly,
MR does not lend itself well to your idea. The characteristic of the MR fluid is not velocity sensitive - such that the damper output under a constant control input looks more like friction than it does damping. This paper shows the characteristics

http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA426496&Location=U2&doc=GetTRDoc.pdf

As a result, to get decent ride/handling and sprung/unsprung mass control you need a real time controller, to control the damping force during each stroke.
A simpler approach, along the lines of Greg's suggestion, would be to start with a conventional fluid semi-active damper. Under a constant control input these will give an output similar to a passive damper, and then you could start to add on a control strategy. However, please bear in mind the control is not easy - you can create bad handling transients, especially if you wait for acceleration as you'll always be behind the input. ( Not to mention other issues such as end stop control, usprung mass control etc..etc..). It's not something I'd recommend to undertake.

 

[GregLocock]

Excellent stuff, the meat is fig 23 and onwards.

So, if you are interested in good ride then you have to adjust the current ~ hundred times a second, to approximate a circle from squares, and then you still have the rather large friction jump at 0 velocity to deal with.

However if all you are interested in is handling perhaps the task is less onerous.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[bluly]

Thank you for the information guys! I'll be sure to read over the posted paper and links.