As others have said - you’re definitely exceeding the allowable working load of the belt. You can find in this document from SDP-SI on pg T-15 that the recommended working tension on 1" width of 5mm pitch HTD belt is 102lbf or 454N. Again note this number is for 1" width and you’re only using 15mm (0.59"). You can do some math to determine what your belt working tension is using the pitch diameter of the belt, wheel diameter and traction limited point of your wheel (based on CoF of wheels and weight on the wheel). Of course this load is meant if you want your belts to last a lot longer than they need to in FRC, and there’s some safety factor built in.

Some rough math:

Robot Weight: 150 lb
Wheel CoF: 1
Wheel Diameter: 4"
Number of wheels bearing weight: 4
Max torque / wheel before breaking traction = 150 lb / 4 wheels * 2 in = 75 in lb
Pitch diameter of pulley = Tooth Count * (5 mm pitch / (25.4 mm/in)) / pi
An 18T pulley has a PD of 1.127" or a radius of 0.564"

Max working load on belt = Max Torque / pitch radius
For an 18T pulley this means you’re loading the belt with (75 in lb / 0.534 in) which is 133 lbf of tension.

Additionally you’ll notice the belt that failed transmits all the torque to two wheels. The more weight you have on wheels drive off a single belt the more load you can apply to the belt before your wheels slip. If your weight is shifted onto the two wheels driven off that middle belt you may be putting half the weight of the robot on the wheels driven from a single belt allowing for 266 lb of tension!

You can help reduce this load if you run a belt from the gearbox output to the far wheel instead of using the middle belt to transfer the torque to the outer wheel. But really you want to run at least 24T pulleys if not larger. We typically run 24T pulleys on the gearbox output and 42T pulleys at the wheel.


One word: Chain.
Preferably 25H.
We had this problem in 2017 and 2018 and after trying multiple solutions came to the conclusion that belts just weren’t gonna cut it for the powerful drivetrains that my team likes to design.

This isn’t really accurate considering there are a variety of factors that come into play.

Plenty posts above mention this already, but all belt profiles, pulley diameters, and belt types have their own performance ratings you need to design around. You can also use a combination of belt and chain, or belt and gear drive, etc…

However, I also prefer chain b/c I’m a lazy designer who changes stuff a lot and leaves details to the last minute.

I was just saying that my team tried multiple different belt setups, thicknesses, pulley sizes. and managed to strip or break belts in all situations and we’ve had no trouble with chains. They’re also much easier to repair.

1 Like


While true, proper design using belts (and making sure they avoid metal in the system) shouldn’t need repair.


Akash already pointed out that this conclusion is not accurate but I want to reiterate - this statement is explicitly false. Not nuanced false, not false for some interpretations.


My brother walked through the math above, I recommend evaluating what you are doing based off of physics rather than anecdotal data.

paper: Belt vs. Chain Drive Evaluation - Team 234 is an old, but interesting study by FRC234 on the topic of chain vs belt. There’ve likely been some changes since then however they outline a fairly thorough test methodology that may be useful to emulate in a new revision of this document.

(Who can tell I’m sick of “well 254 does it so I have to” logic?)


I can find the pictures of the stripped and broken belts if you want… Not saying belts can’t be used in drivetrains I’m just saying that’s my teams 2017 and 2018 robots stripped and broke belts with multiple combinations of belts and pulleys so we decided to switch to chain which can hand much higher loads.

I can confirm that cody is telling the truth, 1073 broke belts on many occasions with many different variations in center to center and pulley size. We have found through our testing that chain can handle a higher load than belts.

Do I need to quote you saying that belts can’t handle the powerful drivetrains you like to build… again?

I’m not taking issue with your use of chain, I’m taking issue with blanket statements that chain can handle higher loads than belts. It’s factually incorrect. Is #35 chain going to handle more load than a 9mm wide 5mm HTD belt? Yeah, obviously.

But there’s a huge range of belts, a lot of variables, and to make a blanket statement like I keep seeing is wrong.

I’ve run belts without issue for years in many high power drivetrains as well as other mechanisms. Much larger, higher power systems function fine with belts.


Maybe you could start by being a little more GP this is the FIRST community not the bar down the street…

The reason chains work better with shifting transmissions is since there is a small amount of slack in the chain due to it being a lot of different pieces rather than a single entity it is able to handle the sudden jerk much better.
A 25H chain also has a higher tensile strength than a 15mm htd timing belt which is the type of belt most commonly used in FRC drivetrains.
I’m not saying belts don’t work I’m simply saying they are not physically as string as belts.

1 Like

I’m sorry, but you’d greatly help your case by including chain and belt strength data. Andrew is simply making the point that you cannot make a blanket statement like that. It’s going to be dependent on the belt and the chain. You’re arguing based on your experience that all belt is weaker than all chain.

By the way, I have a belt running the equivalent of 5 CIMs of power. Can your chain do that? Conversely, I’ve seen #35 chain fail in FRC drivetrain applications. Would a belt do the same? It depends on the unit… and I deliberately held back the size of belt I’m using.


I’ve definitely used chains in a situation where belts would have been a better choice, as well as the other way around. You need to consider the amount of torque transfer, pulley/sprocket pitch diameter, belt/chain speed, your tolerance for “stretching” vs capability for tensioning, and how much width you’ve got to work with. Different answers to these key numbers can result in significantly different answers for your best solution.

Along the lines of what @EricH was saying, pop the hood of your car. Are you certain that a single chain strand could drive your air conditioner and alternator as well as all of the other stuff on that serpentine belt - for about 120,000 miles according to my experience, with little if any re-tensioning?


Ask Volvo before you answer, Gus. Friend of mine had an old Volvo stop moving when the timing chain went. This was years ago, and the car was quite old then… That’s also the only car I’ve ever heard of with a timing chain instead of a timing belt.

Since when did disagreeing with a person on a technical matter and providing information and data to backup the disagreement become not GP…?

All Andrew said here is that your blanket statement of “we tried all forms of belts sizes and shapes and widths, and found that belts just aren’t going to cut it for the powerful drive trains my team designs” is false (specifically the “belts in general arent going to cut it for powerful drivetrains part”).

He also literally even said that he’s not disagreeing with you that chains can handle more load then belts. And you replied to that with a “well actually, chain can handle more load then belts”. Try reading more carefully next time, before you start throwing around the “not GP” card

1 Like

A perfect example

1 Like

Just to clarifiy, the serpentine belt is not the timing chain. The serpentine belt usually drives your accessories, AC, alternator, and sometimes waterpump. The timing chains are usually inside the front cover, and are what most over/dual overhead cam engines use. The timing chain connects the crank to the cams, which is what times the valves opening and closing.

The serpentine belt on almost all cars has a spring loaded tensioner. The timing chains usually have hydraulically loaded tensioners that run from oil directly off the oil pump.

Most of Ford’s motors use metal timing chains. I believe chains are on the majority of other cars, though some do use belts. In some cases, the timing chains even have to be specially coated or have a super-fine finish in cases like police cruisers where they idle for long periods of time.

Timing Chain:

Timing Belt:

Serpentine Belt and Tensioner


Whether a car has a timing chain or timing belt, it will stop when the chain or belt breaks. My 84 Mazda 626 stopped when the belt broke. Many current cars use timing chains in their engine and they are supposed to last the lifetime of the vehicle. No system is perfect. The Nissan VQ engine series from the mid- 2000’s have a high probability of the timing chain guides failing and causing catastrophic damage.

25H is an 25% stronger than 15mm wide 5mm HTD belts according to specs from SDP-SI and DID. However, it’s very easy to simply use a 24t pulley (~1.5" pitch diameter) over a 16t sprocket (1.27" pitch diameter) and end up with comparable strength. Going to a larger pulley like 28t is possible as well.
Different tooth profiles such as GT2 or GT3 will beat 25H chain in 15mm width. In fact, 5mm HTD has the same strength as 3mm GT3 for the same belt width.

Backing up your experiences with theory is good practice.


Let’s all take a breather and not take responses personally.

This can also be an excellent teaching moment to learn how to differentiate between anecdotal experience and objective facts when evaluating topics.

Both of these types of responses can have their place in [technical] discussions - if you understand their limitations and learn how to identify each one.

Some answers between “should I choose Option A or Option B” can have an easy and “better” answer on paper. But in reality, sometimes the “better” option may be harder to implement given the resources at hand.

To everyone with anecdotal experience: please just be cognizant that your experience may not be universally true. In fact, it may only be true for a small segment of use cases. It may even be isolated to just your own situation. But that’s okay, as long as you can accept this possibility and don’t try to pass off your personal experience as absolute fact. Perhaps you may even learn that the reason all your past attempts failed was because of an unknown or overlooked factor.

To everyone arguing with technical facts: just remember that sometimes the “better” answer can be a lot more difficult for others (in a potentially vastly different situation) to actually implement or follow. It’s also possible that the tone of a factually correct post can be off-putting to some, which can cause an emotional response to the tone, rather than content, of a message.

There’s a lot of “outrage culture” in the world now. Let’s try not to add to it.