DIY Seatbelt tensioner with simtools

[HoiHman]

A lot guys that have build the SFX100 are thinking about adding a seat belt tensioner to their rig.

Even though i would have prefered to use one the SFX100 servos doing the work, i have decided to order the parts servo seatbelt tensioner and will start building as soon as the parts arrive.

Parts needed:
Servos
https://hobbyking.com/de_de/turnigy...-servo-33kg-0-21sec-154g.html?___store=de_de1

Turnigy 5A (8-40v) SBEC für Lipo :
https://hobbyking.com/de_de/turnigy-5a-8-40v-sbec-for-lipo.html

Aduino uno
https://nl.aliexpress.com/premium/a...SB_20190116083039&origin=y&catId=0&isViewCP=y

Power supply:
Still need to order one, or if the specs of the servos allow it, i will use the 6v power supply of my GS-4

I will be copying the build from @Ringorian and @saxxon66 who have already a topic with complete instruction on the german forum.

https://forum.virtualracing.org/sho...ffer-100-Euro-(auch-Ohne-Motion-realisierbar)

Jochen en Micha, feel free to jump in and educate us

 

[Michal Burisin]

When I first started to think about G-Seat, I bought these servos.

Then I bought these.

I haven't tested neither in application yet I wont be using the smaller ones, as I won't think they would survive long, especially with G-Seat. As I don't have much time at the moment, it's going slowly... I made little bit progress on seatbelt tensioner, so hopefully I'll move it along further on the weekend.

This implementation in the thread you sent seems really finicky and I'm not sure how long it will last and how strong it is. If I imagine SFX is moving your whole body and also GS4 is pushing your body forward, that movement will just tear the servos arms from the housing They are really small and meant for small RC models.

I would recommend to take inspiration here.

I think very important is some railing to guide the movement of the seatbelt, as on xsim thread. Then it also allows possibility for user to tighten the belt without loading the servo arm - it eliminates also side loading which will be important. This way you will be able to make it really tight and G-Forces responsive. Moving the seatbelt with bowdens seems really nice.

With attachment like in that thread you posted, you wont be tightening the belt at all, you will have to have it lose and let the servos tighten it for you, but there will be big lag in response.

These are just couple of quick notes from top of my head. In my thread on gseat you can find materials I plan to use - linear railing, linear railing carts, bigger servos, but it's not low budget for sure but I think if you want to implement something to match SFX100, I'm afraid you won't get much out of basic implementation like that.

 

[saxxon66]

These servos are not that strong, but they work surprisingly well with respect to their power.
But they are noisy. Like all RC-Servos due to their gears.
Using only the servo shaft is not the best, I changed the design by using bearings to relieve the servo and keep the "roll" in place and it makes them very rigid.
I can lift my rig only by pulling one of the seatbelts, it is very solid in that respect.

I think this mechanic, using a "roll" is the way to go.
No need to translate the motion from rotation to linear by using a screw/nut + guides/rails.
Just roll the belt up, it will give a linear motion.

This combined with a stronger/faster servo using torque mode and a load cell is my end game.
So you can apply a small amount of torque to make sure the belt will always stay in contact with your body. You can solve this mechanically by using a spring to keep the tension.

 

[TimothyDDX]

Looking forward to this project!

 

[sjb266]

me too!

 

[metalnwood]

I know this is the next thing I will be doing although I already use simfeedback and simhub so I will investigate if it is possible with simhub, which I think it is.

 

[anton_Chez]

I use Simhub, too, for tactile. It's a super good program. I still need to donate to it. It's alleviated the need for Simcommander for me which I always hated running. Sim Shaker Wheels was better but Simhub combines the best of both programs into one highly functional and FREE package. I love the way it tracks your sessions for each title, fastest laps, sectors, avg speed, all very nice additions that I've never really used before until now.

Not sure how it is used for the belt tension but I would love to see it.

 

[anton_Chez]

Also, can anyone recommend where to look for a genuine harness? There are so many fake ones out there. I know we don't need it for the safety aspect but one thing that's always bugged me about mine is that it's comfortable, but the shoulder straps come loose with even a little movement and have to be pulled tight like every lap. I am pretty sure it's because my harness was like $100 and the better ones go for like at least double that. I'd hate to spend heaps more on another one though and have the same thing happen.

In race car videos those straps are bulletproof. They don't move AT ALL and if they do slightly loosen it's after laps and laps, not half a lap. My lap belt stays tensioned but the shoulder straps bounce with the seat mover and it's annoying when they loosen.

 

[HugoB]

Also, can anyone recommend where to look for a genuine harness? There are so many fake ones out there. I know we don't need it for the safety aspect but one thing that's always bugged me about mine is that it's comfortable, but the shoulder straps come loose with even a little movement and have to be pulled tight like every lap. I am pretty sure it's because my harness was like $100 and the better ones go for like at least double that. I'd hate to spend heaps more on another one though and have the same thing happen.

In race car videos those straps are bulletproof. They don't move AT ALL and if they do slightly loosen it's after laps and laps, not half a lap. My lap belt stays tensioned but the shoulder straps bounce with the seat mover and it's annoying when they loosen.

You probably already saw my post on Xsim, but this harness has been approved by a user of the SeatTime harness tensioning system who uses it himself (Australia shop for you):

https://vsport.com.au/product/sabelt-6-point-saloon-harness-3-steel-series-fia8853-2016/

It has 3" straps so should give a nice broad pressure point, compared to 2" (got that tip from SeatTime).
This is a racing quality FIA approved belt, but still doesn't break the bank like some that cost double, only because of lightweight spacecraft materials. This one has steel clamps, but weight is not a problem for us If you want lightweight, it costs more:

https://vsport.com.au/product/sabelt-6-point-saloon-harness-3-silver-series-fia8853-2016/



SeatTime said it's best to get a six-point harness because the tensioning system pulls the straps up, and without the crotch belts, your buckle keep being pulled up... I've searched low and high, and this is probably the cheapest high quality 3" six-point harness.

But maybe somebody has found a better alternative?

In Europe there is a cheaper one too, but blue only, this one uses metal clamps, even cheaper than steel:

https://www.sandtler24.de/sabelt-renngurt-4353b.html?number=4353B

 

[Chris_I]

Let's go lower in price: https://www.oreca-store.eu/fia-6-pl-3-3-oreca-vintage-harness.html (blue ones are the cheapest).
They build LMP cars soo their harnesses should be good too...

It's FIA approved soo can be used on track.

 

[anton_Chez]

Thanks for the replies dudes. I see that they use pretty much the same buckles as is on mine. They might be different, but just look the same. I don't know. I have a feeling what's happening is because of the way the harness is mounted to the frame at the rear of the seat. All the way down basically to the floor. I think if we mounted it like they do in race cars, to the bar right behind the head of the driver, it would keep it much tighter for longer. When the seat bounces around it allows the buckles to unseat themselves where the tension is and they come loose. You can test this by running it without motion on.

I have a feeling if I lash out 3 - 4 hundred bucks on a harness, which I already have a 5 point 3 inch one here I'm using, that it's going to end up exactly the same.

I wonder if there is something I can wrap inside the buckle to thicken it up and hopefully cause the belt to grip more? Maybe some tape, or some abrasive material. Sort of the same idea as the couplings. Make it rougher than it is.

 

[HoiHman]

Since i only send 70 euro's i'm seeing this project with the small servos as an in between until a stronger solution will available.

Why, do i think this could actually work really good on the GS-4?

The braking effect of the back panels makes them move back under braking. It's simulates your body moving away from the back of the seat because of the braking forces.

This frees up the space needed for the seat belt tensioner to do it's work.

Needless to say i will only use the seatbelt tensioner under braking and in while doing corners.

 

[Michal Burisin]

These servos are not that strong, but they work surprisingly well with respect to their power.
But they are noisy. Like all RC-Servos due to their gears.
Using only the servo shaft is not the best, I changed the design by using bearings to relieve the servo and keep the "roll" in place and it makes them very rigid.
I can lift my rig only by pulling one of the seatbelts, it is very solid in that respect.

I think this mechanic, using a "roll" is the way to go.
No need to translate the motion from rotation to linear by using a screw/nut + guides/rails.
Just roll the belt up, it will give a linear motion.

View attachment 286630

This combined with a stronger/faster servo using torque mode and a load cell is my end game.
So you can apply a small amount of torque to make sure the belt will always stay in contact with your body. You can solve this mechanically by using a spring to keep the tension.

Brilliant design! Really simplifies everything. As I don't have any experience with motor control I went with off the shelf servos, which are usually just 360 degrees rotation which is not enough for this roll up design, so I didn't consider such different solution, but seems really best for it's simplicity.

When you use torque mode + load cell sounds really great! So if the motors would be set for example to apply 30 kg.cm at some load under heavy breaking, it would roll up seat belt and pull againts body and as the torque would decrease, the body which was squeezed would overcome the hold torque and actually pushed the motor back to its original position? Do I understand it correctly?

With regular motor, you probably have to account for increasing radius of the belt that's actually rolled up, but it's probably no issue.

I will finish my design with conversion to linear motion as I'm quite close to finishing. But this seems as really interesting solution which would be very exact due to the load cell usage. I will have to account for G-Seat paddle movement in Seatbelt tensioner movement profile.

I received Nema 23 stepper motors, so I will start experimenting as these would be possible to use with seatbelt tensioner design as yours, but it might be overkill for me. Also I think I won't be able to program something as complex to do torque mode on these motors as the code for steppers is hard to find in general and it might not even be possible to do torque mode on stepper, will try to read little bit on topic.

Really nice idea

 

[saxxon66]

Stepper motor codes for arduino or teensy are all over the place.
Get a teensy it's much faster.

And here is a nice example for a load cell
https://www.instructables.com/id/Arduino-Tension-Scale-With-40-Kg-Luggage-Load-Cell/

With a stepper you need to poll the loadcell and let the motor turn untill it shows the tourque you want. Than release the load by turning to the other direction.

 

[Michal Burisin]

Stepper motor codes for arduino or teensy are all over the place.
Get a teensy it's much faster.

And here is a nice example for a load cell
https://www.instructables.com/id/Arduino-Tension-Scale-With-40-Kg-Luggage-Load-Cell/

With a stepper you need to poll the loadcell and let the motor turn untill it shows the tourque you want. Than release the load by turning to the other direction.

From what I read, most of the code for steppers is designed for CNC machines as it's most often used with arduino etc.

The problem is that you set position the motor starts to move, but it ignores incoming data meanwhile.

Here I found some code, that should work with SimTools and should not have this problem. The author writes it works, but very basically. I will try it on weekend, hopefully it won't be that much trouble as you suggest.

I will also look at the Teensy, but I'm not sure about SimTools integration.

Ahh, I understand now with load cell, so still postion control, but driven by load cell data, nice.

I have built load cell handbrake and it was really easy to do. Thanks for info!

 

[Ringorian]

I am prepared

 

[HoiHman]

Well lets say that @Ringorian last post and a quick chat i had with Jochen killed my current aspirations for using the small servos for a seat belt tension project.

I will wait for the second controller support of Simfeedback, until then i going to tryout a passive solution. Just ordered some bearings and a set of springs. The goal will be to build the basis for an active seatbelt tensioner, but use springs instead for now.

Ofcourse in my case things will be more complicated because i have mobile rig.

 

[BluePotato00]

Question about seatbelt tensioners. I understand that they get tighter as you brake, since when you brake, the pressure you feel from the belt/harness increases.

But couldn't they also work the other way?

When we're accelerating, we feel like we're being pulled back into our seat, which a seatbelt tensioner getting tighter could emulate. Am I completely off in my thinking here?

 

[dedraro]

I fell in love with this seatbelt tension system

When I have finished the SFX-100 build, I will try to do something similar, I do not know yet whether with scn or geared motors, what makes me doubt about a scn6 is the noise ... and the cost
Do you like a seatbelt system like this?

 

[anton_Chez]

I agree with the thoughts on acceleration. A passive system will just react to the movement of the chassis, so there's not a lot you can do to make it feel a certain way other than make sure your connection point is the best it possibly can be. For a full chassis motion platform that means connecting the shoulder straps to the ground, or something that is fixed under enough weight that it doesn't life when the movement is causing it to tension. I suggested in the other thread that the least destructive way to one's floor would be to place a length of steel or something similar under the rear two actuator feet. Then, connect the shoulder straps to a bracket in the middle of this piece. This way you don't have to permanently bolt something to your floor and the weight of the chassis will be more than enough to keep it planted under tension.

The active solution will need to be tuned to achieve such a feeling, if it's even possible. I've not played with Simtools but I'd say you'd need to create a profile and give the command to tension under acceleration G load. Maybe slightly more progressively that the braking tension, which should be a pretty swift tug.

I have an in between solution, for now. I mount my shoulder straps to the chassis, behind the seat. The V3 seat mover gives me a similar feeling to a chassis mover that has bolted straps to the floor, seeing as the seat platform moves independently of the chassis. One day I'll switch to the other method and give it a crack. I'd say it's going to be a lot better in terms of tension produced compared to what I have now. But it's also more work.