Author Topic: 28mm vs 32mm  (Read 1802 times)

quixoticgeek

  • Mostly Harmless
28mm vs 32mm
« on: February 25, 2019, 08:13:47 pm »

Am pondering ordering my new summer tyres. I've been using 28mm as previously they only went upto 28mm. But I'm now considering going to 32. Science has said that 28mm can be faster than 23 and 25mm, but does that mean that 32mm should be faster than 28? How far up does that go?

I'm after long distance comfort with minimum of rolling resistance. 28 or 32?

J
--
Beer, bikes, and backpacking
http://b.42q.eu/

LittleWheelsandBig

  • Whimsy Rider
Re: 28mm vs 32mm
« Reply #1 on: February 25, 2019, 08:25:41 pm »
Narrow is better aero, wide is better rolling resistance. You mostly finish in the back half of brevets. I'd go wider.
Wheel meet again, don't know where, don't know when...

jiberjaber

  • ... Fancy Pants \o/ ...
  • ACME S&M^2
Re: 28mm vs 32mm
« Reply #2 on: February 25, 2019, 08:29:23 pm »
Why stop at 32, I'm now running 35c Schwalbe G-one speeds TLE, nice comfort & speed combo.
Regards,

Joergen

bludger

  • Randonneur and bargain hunter
Re: 28mm vs 32mm
« Reply #3 on: February 25, 2019, 08:30:49 pm »
I ride both 28 and 32. 28 on my fixed deliveroo / work bike, 32 on the Sunday best.

I really like both sizes but 32 is very much more comfortable imo. If you're not looking to set records it is absolutely worth the extra mm, particularly when you're hauling cargo. You feel absolutely rock steady on them even on marginal surfaces.
Bikepacking bargain basement: reviews of high value kit great for the tourer, bikepacker and randonneur on a budget

https://yacf.co.uk/forum/index.php?topic=109048.msg2312359#msg2312359

Kim

  • Timelord
Re: 28mm vs 32mm
« Reply #4 on: February 25, 2019, 08:35:03 pm »
Narrow is better aero, wide is better rolling resistance. You are mostly finish in the back half of brevets. I'd go wider.

Agreed, if clearances permit.  It's a marginal change in performance that could make a decent improvement to comfort.

Dunno how much you can usefully extrapolate, but they tested Marathons in a variety of sizes, and concluded the 37mm was the sweet spot, but mostly because the 32mm version didn't have a comparable construction to the larger sizes: https://www.bicyclerollingresistance.com/specials/schwalbe-marathon-32-37-40-47

(I run the 40mm version on my tourer.  The extra width is useful when the surface tends towards the COR end of the spectrum.)
To ride the Windcheetah, first, you must embrace the cantilever...

Re: 28mm vs 32mm
« Reply #5 on: February 25, 2019, 08:44:44 pm »
on a bike with a springy fork I quite often run a narrower tyre at the front than the rear, so you could try that, either 28f/32r or 32f/35r. You could try 28f/32r for the price of one 32mm tyre by trying it with a used 28mm tyre at the front, and then buy another 28,32,or 35mm tyre depending on how you get on.  Needless to say the aero penalty of a wider rear tyre is less than that of a front, and the load at the rear is more easily borne on a wider tyre, so it all makes sense.

cheers

Re: 28mm vs 32mm
« Reply #6 on: February 25, 2019, 08:48:59 pm »
I am running 32s and no slower than 28s, mind my average speed is not above 20 mph. I have basically gone back to the width I was running in the 80/90's on my old racing bike.

Re: 28mm vs 32mm
« Reply #7 on: February 25, 2019, 10:19:59 pm »
In relation to long audax rides I've always considered comfort as a significant performance factor, taking it to the extreme with 42mm tyres for the 2017 LEL.  This approach has always worked well for me but I'm faced with a challenge in April that is "Bill's Easton Connection" 300k with 4950m total climb.  I know from my average speed on big tyres and associated wheels and bike that I would not complete within the time allowed and so have been putting together a bike that will at least give me half a chance of finishing.  28mm tyres on comparatively light weight wheels on a classic steel lightweight frame with a Cambium C17 carved saddle. It's a revelation to me, fast and comfortable, though not quite the "armchair" comfort that I'm used to.  So I'd say yes, go for bigger tyres if improved comfort is a relevant performance factor for the rides you're planning to do, but do watch out for the small incremental reductions in speed that heavier tyres at lower pressures on wider rims will bring.
Most of the stuff I say is true because I saw it in a dream and I don't have the presence of mind to make up lies when I'm asleep.   Bryan Andreas

Re: 28mm vs 32mm
« Reply #8 on: February 25, 2019, 10:22:22 pm »
You can't say flat out that wider tyres are faster than narrow ones, as what's fastest is all a matter of trade-offs.

Comparing tyres like for like - i.e. the same model in different sizes, such as 23, 25, 28 & 32 GP4S or Gatorskins, you can say that...
a) At the same pressure, a wider tyre has less rolling resistance and a higher aero resistance.
b) The same tyre has a lower rolling resistance at higher pressures than lower ones.
So, does a 23mm GP4S at 120 psi (8.25 bar) have a lower rolling resistance than a 32mm GP4S at 80 psi (5.5 bar)?
I don't know, but you can look at the various tyre testing results on bicyclerollingresistance.com, and come to a reasonably well-informed judgement.

However tyre rolling resistance is only half the story.
Roads aren't uniformly smooth, and as you ride over them, there's vibration of both the bike and the rider. The rougher the road, the more the vibration, and the lower the tyre pressure (with due regard for the chance of snakebites) the less the vibration. The bike being vibrated doesn't amount to much, but when the rider's body is vibrated, most of the energy used to do the vibrating is lost as heat in the body tissues. That energy has to come from somewhere, and the source is the kinetic energy of bike and rider. If you freewheel onto a stretch of cobbles, it's pretty obvious how quickly you lose speed compared with a similar length of smooth road.
Unfortunately, human bodies don't lend themselves to standardised scientific testing, so what information there is available is more ad-hoc, but it's enough to show that these losses can be very much larger than tyre rolling resistance - between 150 and 200W on particularly bad surfaces.
https://janheine.wordpress.com/2016/08/08/the-missing-piece-suspension-losses/
https://silca.cc/blogs/journal/part-4b-rolling-resistance-and-impedance

I interpret a lack of comfort to indicate that vibrational losses are getting high, and pick my tyre as being wide enough to allow pressure low enough for good comfort on roads at the rougher end of normal, without being at risk of snakebite punctures.
In my case, that's 32-35 mm tyres at 70-55 psi (4.8-3.8 bar), but weight, body composition and riding style will mean that someone else may be better on different tyre sizes or pressures.


Re: 28mm vs 32mm
« Reply #9 on: February 26, 2019, 09:18:49 am »

Dunno how much you can usefully extrapolate, but they tested Marathons in a variety of sizes, and concluded the 37mm was the sweet spot, but mostly because the 32mm version didn't have a comparable construction to the larger sizes: https://www.bicyclerollingresistance.com/specials/schwalbe-marathon-32-37-40-47


While quoting BicycleRollingResistance the very recent comparison of 23/25/28/32 GP 5000 tyres might be more illuminating, I am convinced

https://www.bicyclerollingresistance.com/specials/grand-prix-5000-comparison

Re: 28mm vs 32mm
« Reply #10 on: February 26, 2019, 09:43:02 am »
Unfortunately, human bodies don't lend themselves to standardised scientific testing, so what information there is available is more ad-hoc, but it's enough to show that these losses can be very much larger than tyre rolling resistance - between 150 and 200W on particularly bad surfaces.
https://janheine.wordpress.com/2016/08/08/the-missing-piece-suspension-losses/
https://silca.cc/blogs/journal/part-4b-rolling-resistance-and-impedance

Both of those articles discuss relative amounts of watts without mentioning either speed or absolute numbers of watts, so both of them belong in the bin.

It is *so frustrating* we could so easily have hard data on this stuff, but it’s impossible to find any that hasn’t been filtered through pseudoscience quacks who remove all context and meaning.

Re: 28mm vs 32mm
« Reply #11 on: February 26, 2019, 09:49:56 am »
there are no absolutes; the relative contributions of aero and rolling resistance losses vary from one bike, one rider, one road surface etc to another. Even if all these things are the same then the speed at which you ride can tip it one way or another.

'Hard data' is actually very difficult to obtain; wind tunnel tests are (for many reasons) unlikely to generate real numbers and so are Crr tests; both necessarily contain approximations to reality that have a significant effect on the results; there is no 'so easily'.

  Arguably the only thing that 'belongs in the bin' is the idea that a single set of numbers might accurately represent any given cyclist....

cheers

bludger

  • Randonneur and bargain hunter
Re: 28mm vs 32mm
« Reply #12 on: February 26, 2019, 09:57:01 am »
I think two other important variables are missing from this chat; braking performance and loaded stability particularly on corners.

I think wider tyres have big big benefits for both of these factors especially at lower pressures.
Bikepacking bargain basement: reviews of high value kit great for the tourer, bikepacker and randonneur on a budget

https://yacf.co.uk/forum/index.php?topic=109048.msg2312359#msg2312359

mattc

  • n.b. have grown beard since photo taken
    • Didcot Audaxes
Re: 28mm vs 32mm
« Reply #13 on: February 26, 2019, 10:07:17 am »
However tyre rolling resistance is only half the story.
<snip>
I interpret a lack of comfort to indicate that vibrational losses are getting high,
...
I'm liking what you've written there! There is definitely a hard-to-measure part to all this, and those 2 phrases sum up the situation as well as all our other waffle - well, to my mind, anyway :)
Has never ridden RAAM
---------
No.11  Because of the great host of those who dislike the least appearance of "swank " when they travel the roads and lanes. - From Kuklos' 39 Articles

quixoticgeek

  • Mostly Harmless
Re: 28mm vs 32mm
« Reply #14 on: February 26, 2019, 10:42:46 am »
 622x32 Gp5000s it is then!

Cheers all!

J
--
Beer, bikes, and backpacking
http://b.42q.eu/

Kim

  • Timelord
Re: 28mm vs 32mm
« Reply #15 on: February 26, 2019, 12:11:54 pm »

Dunno how much you can usefully extrapolate, but they tested Marathons in a variety of sizes, and concluded the 37mm was the sweet spot, but mostly because the 32mm version didn't have a comparable construction to the larger sizes: https://www.bicyclerollingresistance.com/specials/schwalbe-marathon-32-37-40-47


While quoting BicycleRollingResistance the very recent comparison of 23/25/28/32 GP 5000 tyres might be more illuminating, I am convinced

https://www.bicyclerollingresistance.com/specials/grand-prix-5000-comparison

Even better.  I hadn't spotted that, as I don't pay much attention to tyres that aren't available in useful sizes.
To ride the Windcheetah, first, you must embrace the cantilever...

Re: 28mm vs 32mm
« Reply #16 on: February 26, 2019, 02:59:34 pm »
there are no absolutes; the relative contributions of aero and rolling resistance losses vary from one bike, one rider, one road surface etc to another. Even if all these things are the same then the speed at which you ride can tip it one way or another.

No no no. When they talk about something "costing you 150W" do they mean 50W vs 200W or 650W vs 800W? That's what I mean by needing to know absolute numbers of watts (or tested speed, which is a rough proxy) to figure out how the data might apply to your own riding.

Quote
'Hard data' is actually very difficult to obtain; wind tunnel tests are (for many reasons) unlikely to generate real numbers and so are Crr tests; both necessarily contain approximations to reality that have a significant effect on the results; there is no 'so easily'.

This bit I agree with you on. I suspect the actual reason we have no hard data (apart from very artificial lab tests) is that in anything resembling real world conditions the differences between similar tyres are too small to measure. But no one publishes articles about that.

Re: 28mm vs 32mm
« Reply #17 on: February 26, 2019, 03:24:27 pm »
there are no absolutes; the relative contributions of aero and rolling resistance losses vary from one bike, one rider, one road surface etc to another. Even if all these things are the same then the speed at which you ride can tip it one way or another.

No no no. When they talk about something "costing you 150W" do they mean 50W vs 200W or 650W vs 800W? That's what I mean by needing to know absolute numbers of watts (or tested speed, which is a rough proxy) to figure out how the data might apply to your own riding.

a) in neither link do they say anything about something 'costing you 150W' and

b) it wouldn't apply to your riding unless you can measure absolutely how rough the surfaces are and how well coupled your mass is to the rest of the bike.

It is quite unrealistic to expect hard data that you can use directly  and I don't think that they particularly misrepresent what they are trying to say. Their examples are just that; examples.  I note that Anhalt wrote that he planned more detailed tests and that it would generate a lot of data; a massive understatement if ever there was one.  I'm not sure he has published anything since then, and it is over two and half years ago now.

cheers

Karla

  • car(e) free
    • Lost Byway - around the world by bike
Re: 28mm vs 32mm
« Reply #18 on: February 26, 2019, 05:34:03 pm »
there are no absolutes; the relative contributions of aero and rolling resistance losses vary from one bike, one rider, one road surface etc to another. Even if all these things are the same then the speed at which you ride can tip it one way or another.

No no no. When they talk about something "costing you 150W" do they mean 50W vs 200W or 650W vs 800W? That's what I mean by needing to know absolute numbers of watts (or tested speed, which is a rough proxy) to figure out how the data might apply to your own riding.

You're completely right and Brucey is dead wrong.  The Heine article twice quotes "up to XXX Watts" numbers without telling us where they got them or what they're a fraction of, then tells us that it clearly shows we should buy YYY tyres.  It's not even pseudoscience, is a classic advertiser's trick.  "Buy my washing powder, it will make your shirts 100% whiter! (if you put them through the wash 1000 times.)

As for the Anhalt article, that's rubbish too.  They were doing Chung tests on an outside road and varying the pressure in the tyres.  For anyone who knows the typical error in a Chung test and the losses they'd have been getting through their tyres, the kids through the tyre would basically have needed to halve for then to be able to detect it.


Re: 28mm vs 32mm
« Reply #19 on: February 27, 2019, 12:40:59 am »
When they talk about something "costing you 150W" do they mean 50W vs 200W or 650W vs 800W?
The (first) article does say that the speed was about 17 mph, that being the maximum that the test rider could maintain for the requisite number of runs and consistency.
Maybe 100W vs 250W?

Re: 28mm vs 32mm
« Reply #20 on: February 27, 2019, 06:44:08 am »
you can't have it both ways; on the one hand there is a demand to have 'absolute numbers' so that they are relevant and on the other hand folk slag off the Chung method and similar, even though that (and similar)  are amongst the few methods likely to give real world results.  [BTW my understanding is that the Chung method produces results that are statistically significant after repeated runs, and then only if you can keep your errors under control.]

 No one thinks that riding over rumble strips is  representative of normal riding (and therefore why worry especially about absolute numbers?) but it does illustrate what is going on.  All Heine has done is illustrate -for the benefit of those who have never noticed previously-  what every thinking cyclist has known for years, i.e. that whenever you are being jiggled around on the bike it is probably costing you energy.  Just holding  the handlebars differently can change this, which rather underlines the futility of wanting 'absolute numbers' to play with....

My advice is that you carry out your own tests -on your bike, on your roads-  to see what is going on and whether it might be significant or not.  My take on it is that provided you are not demonstrably slower using the wider tyres (eg during coastdown and rollout tests), you are probably better off overall in any kind of distance event.  This relies on another imponderable (which may vary greatly from one rider to another) which is that after a long  day in the saddle "more comfort = more speed".

BTW there are further points which should be acknowledged and may be signficant for some riders/events; of these an important one is to acknowledge that you only get to go down the road at all because there is a torque transmitted through the rear wheel.  Most Crr testing blithely ignores this altogether, in that the only torque transmitted though the wheel is enough to overcome rolling resistance, not propel the whole bike, i.e. the real force is probably about x10 or more higher. The real situation is much worse than that, because the effort is rather 'pulsey' in nature.  If the tyre slips at the contact patch even slightly, or the transmission/tyre carcass deforms under torque loads  in such a way as it doesn't conserve energy, every pedal stroke could be costing you.   

In illustation of this, everyone knows that with  suspension of any kind 'bobbing' is the enemy when climbing.  IME you can 'bob' on fat tyres just as easily as on springs.  If you do enough MTBing you also come to realise that a 'stabby' pedal stroke up a steep climb is usually a lot less efficient than a non-stabby one; with the former you can often see that the tyre is trying to break traction with every pedal stroke and this just costs you effort. [BTW in a comparable way, belt drives may measure low losses under constant torque but are pretty hopeless whenever the torque varies.]  The other thing that becomes apparent when MTBing is that if you don't need super aggressive tread, the life of rear tyres will most often be limited by the carcass. When they will wear out in the carcass, it usually  happens in such a way as they fail because of deformations induced by the torque loads passed through the tyre during climbing, i.e. the bias plies fail mostly in one direction only. The next time you ride an  offroad climb, take a look at what is happening near the contact patch of another rider's rear tyre; it may surprise you.   These things are also happening on-road too; although they are much less easy to see, they may be no less significant. For example  I don't think it is a coincidence that, on the road, I've never climbed well on fat tyres, and likewise quickly sensed that riding with a belt drive is like pedalling in porridge; the two things can feel comparable to one another, which if it is borne out in reality means that there ought to be an easily measurable difference, if only the right tests are done... ::-).

FWIW there is something to be said for training on bikes with suspension and fat tyres etc, even if you don't plan to use them during events;  speed per se isn't really important during training and  less fatigue may mean that you can train better for longer, and recovery is accelerated too. Educating your body so that you naturally pedal smoothly is certainly no bad thing if you are planning on riding distance events.

Anyway plaintive cries for 'relevant, absolute numbers' from a one-page article are IMHO just a waste of breath; reality is a lot more complicated than that and varies enormously from one rider/bike/road/riding position to another.

cheers





Karla

  • car(e) free
    • Lost Byway - around the world by bike
Re: 28mm vs 32mm
« Reply #21 on: February 27, 2019, 11:24:59 am »
Brucey, this is really simple.

The Heine article says "saves X Watts" but doesn't tell us the total figure they're being saved out of, i.e. what fraction of being saved.  It's as meaningless as a map without a scale.  Further, they could very easily have given that vital info - which is what grams was talking about - so the face they chose not to give it (but do tell us to buy their tyres) suggests they're up to something.

As for the Chung method, I've done it, and know the pitfalls.  I also know that professional testers, who really know their stuff and are using a technologically superior method, claim 2.5% accuracy on an outdoor velodrome.  I very much doubt Anhalt & co can get that accuracy - let's be kind and say they get 5% - which at 200W is a 10W resolution i.e. a massive proportion of total tyre loss.  The idea that they're accurately measuring changes in tyre drag is silly.

Re: 28mm vs 32mm
« Reply #22 on: February 27, 2019, 05:35:33 pm »

The Heine article says "saves X Watts" but doesn't tell us the total figure they're being saved out of, i.e. what fraction of being saved.  It's as meaningless as a map without a scale.  Further, they could very easily have given that vital info - which is what grams was talking about - so the face they chose not to give it (but do tell us to buy their tyres) suggests they're up to something.

This is what Heine says in the article  linked above;

Quote
...We tested various equipment on rumble strips to get a maximum value for the energy that is lost to vibrations. We found that riding on this “very rough” road can take up to 290 Watt more power than riding on smooth pavement at the same speed. So it’s true, vibrations can absorb a huge amount of energy. It was almost impossible to keep the bike moving at our testing speed on the “very rough” road. (Of course, in real life, you don’t ride on rumble strips, but the point was to see how much energy could be lost just by changing the surface roughness, and keeping everything else the same.)….

I don't think anything is being misrepresented here, and I repeat that wanting relevant 'real numbers' out of a one-pager is not a realistic expectation;  well not unless you want to be badly mislead.... ::-)

Quote
As for the Chung method, I've done it, and know the pitfalls.  I also know that professional testers, who really know their stuff and are using a technologically superior method, claim 2.5% accuracy on an outdoor velodrome.  I very much doubt Anhalt & co can get that accuracy - let's be kind and say they get 5% - which at 200W is a 10W resolution i.e. a massive proportion of total tyre loss.  The idea that they're accurately measuring changes in tyre drag is silly.

The devil is always in the detail.  As I mentioned previously there are circumstances where you can apply statistical methodology to test data and find results with a higher level of confidence than you might expect.  For example carbon dating is notoriously inaccurate on any one test but if enough tests are done on material from the same source, you can be about ten times more accurate with the dating.  I don't know if that approach is applicable to Anhalt's data or not, but I wouldn't just go blithely saying "it definitely isn't", not without a few more checks and balances.

In point of fact you can pick holes (big ones) in most test data without trying too hard. One has to accept that, at any one time, it is very much a 'best guess' as to what makes you go faster at all and in particular choosing where to spend your effort/cash on so doing is an art in itself.

 I attended an interesting lecture given by Tony Purnell a while ago and he explained his methodology for choosing the areas to work on that might  improve the speed of our track cyclists prior to the Rio Olympics.  He managed to produce a single mathematical expression for 'speed' and inputted realistically achievable changes to each of the relevant parameters, and saw what effect that had on 'speed' (in the chosen events). Interestingly (even in fairly short events which require acceleration) he identified that saving weight was unlikely to change much; i.e. you could spend a lot of time and effort for very little gain. In fact the single thing that made the biggest difference (for the least effort) was changing the transmission efficiency; simply using larger chainrings and sprockets (than trackies might otherwise choose) improved the transmission efficiency by up to 0.5% and that was the bulk of the winning margin in many events where medals were won.  There was also a last-minute change of tyres, literally days before the event, I think, as new data came in.  He was asked if the same approach might yield good results in road events and he basically said 'no' because there were too many extra variables.

cheers

Re: 28mm vs 32mm
« Reply #23 on: February 27, 2019, 09:19:08 pm »
The Heine article says "saves X Watts" but doesn't tell us the total figure they're being saved out of, i.e. what fraction of being saved.  It's as meaningless as a map without a scale.  Further, they could very easily have given that vital info - which is what grams was talking about - so the face they chose not to give it (but do tell us to buy their tyres) suggests they're up to something.
A small modicum of googling finds a preceding blog of Jan Heine's, which does give the absolute wattages
Adjacent to rumble strip = 183W
On the rumble strip = 473W
https://janheine.wordpress.com/2016/06/14/suspension-losses-confirmed/

If you want the full data, you've probably got to pay for the relevant Bicycle Quarterly back issue. Expecting everything in every blog entry than may refer to the results is a bit daft, IMO.

ElyDave

  • Royal and Ancient Polar Bear Society member 263583
Re: 28mm vs 32mm
« Reply #24 on: February 28, 2019, 08:50:08 am »

The devil is always in the detail.  As I mentioned previously there are circumstances where you can apply statistical methodology to test data and find results with a higher level of confidence than you might expect.  For example carbon dating is notoriously inaccurate on any one test but if enough tests are done on material from the same source, you can be about ten times more accurate with the dating.  I don't know if that approach is applicable to Anhalt's data or not, but I wouldn't just go blithely saying "it definitely isn't", not without a few more checks and balances.

In point of fact you can pick holes (big ones) in most test data without trying too hard. One has to accept that, at any one time, it is very much a 'best guess' as to what makes you go faster at all and in particular choosing where to spend your effort/cash on so doing is an art in itself.


Notoriously innaccurate is not the term I would use.  High uncertainty, or possible error range, perhaps which are different from accuracy, and as you point out are reduced by repeat testing to give a higher confidence overall. There are statistical methods, including those suitable for relatively small populations that can estimate the uncertainty in a set of data.  I occaisionally need to use them on gas samples.

It's quite annoying really that "papers" such as these are not required to provide an error or uncertainty estimation to get published
“Procrastination is the thief of time, collar him.” –Charles Dickens