Rolling resistance: Plastic vs Metal axles

[Moz]

I can't find an existing thread, so while we can still sort of get the older metal axles I thought I should make sure the new ones really are improved. So I made a simple test track and did some simple tests :)

TLDR: my wagon with one pair of metal axles is significantly faster than the one with all plastic axles.

Red    Yellow
16.67s    19.8s
16.86s    19.9s
24.57s    31.97s

For the last set I reduced the slope of the test track to make the difference even more obvious. I'd love to hear ideas for modifying my test setup to get better results. I plan do swap axles so I have all-metal vs all-plastic, and I will also do an R40 wiggle track to see if that's dramatically different.

 

 

The test track is one of my 3x9 tabletops, with the left hand end 10cm or 8cm above the right hand end. It's ~2.3m long and very flat. The release is via a pin-in-hole mechanism so it should be fairly consistent between runs. The big issue turned out to be my camera not being great so I will need to set up better lighting when I do more tests.

The wagons are 24 studs long with 16x1+8x1 or 12x1 bricks on one layer, then 6x1 bricks across those on top. There are photos in the video. They weigh 187.38g and 187.35g, with less than a 1x1 plate difference in weight.

The photos of the scales have had the display area curves blown out so you can actually read the display in the video. The scales are cheap and the display is faint in real life.

 

Edited August 27, 2022 by Moz
delete the image again

[Ashi Valkoinen]

It is not suprising that metal exles have less rolling resistance.

I remember when they introduced the full plastic wheels TLC made a manipulative video to show us why the new ones are better. They built up a track with only curved tracks used and run the same train from the same battery, only with the wheels differing. The battery in the train with plastic wheels lasted longer. 

But what happened?

The current plastic wheels can turn independently from each other. The older wheels, connected with an axle, have an extra resistance, which is called curve resistance - the sharper the turn, the bigger the wheel, the more resistance appears. This is because the two wheels turn the same speed (same rotation speed), but they need to run different distance on a curve (the outer longer, the inner shorter), so the outer wheel will be dragged or the inner slipped. This is causing extra resistance and this is the reason why TLC made the test video using only curved tracks.

In your experiment, you used straight track and gravity to accelarate - since metal axles have less friction and mechanic resistance at all and there is no curve resistance, they will perform better. Adding a 180 degree sloping turn to your experiment with other straight part behind and measuring part times of travel times would complicate the things, but with better results.

[zephyr1934]

Very interesting to see. I thought I had seen similar tests but looking back, they were probably roller bearing vs normal wheels.

6 hours ago, Moz said:

and I will also do an R40 wiggle track to see if that's dramatically different. 

Wiggle track would be interesting, but also doing an "S" pattern with 90° turn into 180° counter-turn into 90° counter turn all on R40 track. This would show if the independent wheels provide significant benefits on tight curves, and possibly if there is enough inertia to get across the lateral portion. Might be hard to keep the track steady though.

[Polarlicht]

I ordered 40 wheel sets (80 wheels and 40 wheel holder) from lego today. The old holders and wheels are still listed, but didn't find any metal axles. Doesn't matter to me, the new ones are cheaper and run just as fine, imo.

I plan to use all these Lego wheel sets for some bluebrixx rolling stock which have terrible wheel sets despite having metal axles!

Edited August 27, 2022 by Polarlicht

[LEGO Train 12 Volts]

Personally I switched to buying the train holders, train axles and train wheels from BB plus the addition of ball bearings (from ebay).
This is the best results in terms of performance (and probably also the cheapest if you buy in quantity)

[Moz]

43 minutes ago, LEGO Train 12 Volts said:

plus the addition of ball bearings (from ebay)

Cheapest I've found in Australia is $1.36 per ball bearing race which is IMO still unreasonable. But better than the $3-$4 ones from BB suppliers, or the $5+ onces from AliExpress (I am obviously using the wrong search terms)

I have some HDTracks BB parts coming, I ordered a set just to see if the idea is worth pursuing.

7 hours ago, zephyr1934 said:

"S" pattern with 90° turn into 180° counter-turn into 90° counter turn all on R40 track.

I have that but I need a really steep ramp to get the wagons through the last horizontal section so I've dropped back to 3 sections one way, 6 sections back, then 3 more to go straight down the ramp. I've also nailed the track down because the curves move a lot as the wagons go over them. And I'll dig out some 60337 wagons too. Once the sun comes up and I have enough light to make the camera happy I'll shoot some more video and collect more results.

I'm thinking the video will be less slide show than last time and more of a "methods" section, with a summary slide and the rest here. Unless anyone really wants to see another ten or more 30s runs down the track.

[Moz]

More tests done, and while I work on the R72 TrixBrix tests and some more 60337wagon tests, here's a 30s update:

Those of you predicting that the plastic axles would work better on curves... well done!

    Day 2    
    Red/Metal    Yellow/Plastic
R40 curve 1    
    6.63            4.4
    7.13            4.57
    8.16            4.6
        
straight
    3.1                3.53
    3.1                3.5
60337 wagons on straight track        
    3.1                3.74
    3.2                3.66
    3.1                3.74

 

[Moz]

Results with TrixBrix R72 curves (only 45 degrees of curve each way because that's what I have available).

R72 is less dramatic but plastic axles still handle the curves better. When I get the ball bear races I think I'll test hose vs graphite lubricating the plastic axles.

I could do a wiggle track but I don't really see the value as it's not something I use in actual layouts - I get the feeling it's something people do for house-size tracks because they inevitably have a big pile of curved track left over.

    Red/Metal    Yellow/Plastic
R72    
    5.33  (that's a bad run, discard it)
    4.07            3.77
    4.33            3.84
    4.17            3.77
                
R72 60337 
       5.1        4.4

https://imgur.com/a/PZ1Oo1I   A few photos, one showing what's left of my layout after I removed the tracks for these tests.

Edited August 28, 2022 by Moz
clean up

[Moz]

Rolling Resitance 3: Graphite

More trials, sorted by time:
W   4.4
W   4.5
Y   4.57
Y   4.57
G   5.3
G   5.6
R   5.9
R   7.3 (barely made it through, was the very first run so may be dubious)

Red and Yellow are as before. Two new wagons with same weight to within 0.1 grams, but with graphite added. For the metal axles I tried to get some into the axle hole on the wheel as well as on the axle where it goes through the holder. I think the former is likely to be more important, so we get the wheels rotating independently more easily.

G = grey, plastic axles with graphite
W = white, metal axles with graphite.

Ran on the R40 curved track from before.

That helped, but the plastic axles are still better on curved track.

Any more ideas?

[brummer]

What sort of meta axles did you use  ?

The typ used in the last years, metal axles but plastic points at the wheels,

or the old typ with long meta axles and metal points at both ends ?

The older runs better.

 

 

Torsten

With this old metal axles we do the record for the longest only pulled lego train, long time ago....

 

 

[Moz]

1 hour ago, brummer said:

The type used in the last years, metal axles but plastic points at the wheels, or the old type with long meta axles and metal points at both ends ?

The recent ones, my Lego trains go straight from the old blue rails to what I can buy this year. I'm more a Technic fan but long covid means I have too much Lego time and not enough enthusiasm for yet another Technic car.

I might have to see what I can do about your record though 😉

Edited August 28, 2022 by Moz

[zephyr1934]

16 hours ago, Moz said:

Those of you predicting that the plastic axles would work better on curves... well done!

The interesting bit is that the plastic axles on R40 reduce the time by almost half compared to metal (and that is with much about 1/3 of the track being straight) whereas the metal only save maybe 10% on straight compared to plastic. I'm sure exact results would change as you make the cars heavier but the general trend probably doesn't change.

So for the target customers (kids on R40 loops that have more curves than straights) the split plastic axles really do improve things.

 

6 hours ago, Moz said:

I might have to see what I can do about your record though 😉

Definitely impressive for 9v, but only if you limit the number of motors and transformers. As soon as you allow for "distributed power units" it simply becomes an unlimited number of short trains coupled together (doable with 9v but difficult, easily done with PF and PUP).

With power only at the front you are likely to stringline the curves (e.g., at 1:34 in the video). It looks like the "record train" in the video had about 4 9v train motors at the front, but it is still impressive given the forces they overcame. Make the cars heavy enough to prevent stringlining and the magnetic couplers will pull apart, strengthen the coupling and you go back to strignlining. You could counter this with negative banking on the curves. It quickly turns into a very complicated optimization problem.

 

[LEGO Train 12 Volts]

22 hours ago, Moz said:

Cheapest I've found in Australia is $1.36 per ball bearing race which is IMO still unreasonable. But better than the $3-$4 ones from BB suppliers, or the $5+ onces from AliExpress (I am obviously using the wrong search terms)

I'm talking about 100 pieces shipped for less than 24 euros, that is 0.24 per bearing

Ebay ball bearing

[XG BC]

this was also done with the old metal axles:

plastic axles might have even melted in the worst case. the speeds where achived by running 9v motors at 20+ volts (something you wouldnt even think about doing today considering current pricing of these motors)

[Moz]

53 minutes ago, LEGO Train 12 Volts said:

100 pieces shipped for less than 24 euros

Thanks for the link. Although given the limited advantage of the metal axles in tests so far I think I'll hold off until the HDBricks sample pack arrives and I can test those.

Same seller lists in Australia, but only 10 sets of bearings... for 1/4 the price of 100. A different seller has 50 for ~$US18. If I want some I think I'll contact the seller you linked and ask for 100 :)

Edited August 28, 2022 by Moz
add ebay info

[Ashi Valkoinen]

14 hours ago, zephyr1934 said:

The interesting bit is that the plastic axles on R40 reduce the time by almost half compared to metal (and that is with much about 1/3 of the track being straight) whereas the metal only save maybe 10% on straight compared to plastic. I'm sure exact results would change as you make the cars heavier but the general trend probably doesn't change.

So for the target customers (kids on R40 loops that have more curves than straights) the split plastic axles really do improve things.

 

Yes, it works until the target customers don't buy the additional packs of straight tracks to improve their home layouts.

Unfortunately there is no public data about how do train sets sell and how do additional train track packs sell - do people buy additional track with their big train set, do they buy it later or how many percentage of them never buys any additional track for the layout?

What is the lifetime for a plastic-pin connection, what is the lifetime for the metal axle?

I'm running older metal axles (for the running bogies of 9V era, slotted metal axle) for an eternity by now, under 900-1200 g tran coaches (225-300 g for each axle!), and they show no sign of wear.

Will the plastic last so long, when the main title is "environmentally friendly"?

Edited August 29, 2022 by Ashi Valkoinen

[Toastie]

12 hours ago, Ashi Valkoinen said:

I'm running older metal axles (for the running bogies of 9V era, slotted metal axle) for an eternity by now,

Same here - however, eternity is reduced to about 20 years. Which in modern times is actually eternal, I guess.

I may add that the shorter metal axles (40 mm) along with the close hole wheels (# 57878) do perform favorably as well (at least regarding my demands). Should you guys be interested, there is a related thread in the community forum.

I simply do not believe that the split axles, without any support in the center, will take the load from above in the long run. It just does not make sense. Shear forces in an all plastic design will eventually win. I believe. Not the point for TLG: They reverted to the short-run-maximum-profit approach. Which is not bad at all: In 20+ years, these "axles" will be dead cheap on BL, should BL still exist. Or not even show up. Metal axles though will.

Best,
Thorsten   

[dr_spock]

I think having the left and right wheels being able to roll independently of each other would be useful in reducing wear and tear around curves. We have a barely R40 able steam engine that we run at our events for a few years now. The fixed technic axle drivers have grooves worn into the flanges where they rub against the rails in the curves.

[Ashi Valkoinen]

13 hours ago, Toastie said:

Same here - however, eternity is reduced to about 20 years. Which in modern times is actually eternal, I guess.

I may add that the shorter metal axles (40 mm) along with the close hole wheels (# 57878) do perform favorably as well (at least regarding my demands). Should you guys be interested, there is a related thread in the community forum.

I simply do not believe that the split axles, without any support in the center, will take the load from above in the long run. It just does not make sense. Shear forces in an all plastic design will eventually win. I believe. Not the point for TLG: They reverted to the short-run-maximum-profit approach. Which is not bad at all: In 20+ years, these "axles" will be dead cheap on BL, should BL still exist. Or not even show up. Metal axles though will.

Best,
Thorsten   

The only reason I wrote "eternity" is that I had many trams and passanger cars (each of them has at least 200g load for each axle) running on a permanent display between 2017 and 2021 (with some lockdown months). These cars run 2-3 km (real km!) a day, imagine what running distance they have for years.

[zephyr1934]

I didn't say the independent plastic wheels are superior in general, just the limited target scenario of a single kid with a single city set for the short period of time that the parents are still paying attention to the new purchase/gift.

What I'd love to see is roller bearings for technic axles.

 

 

[Toastie]

58 minutes ago, zephyr1934 said:

just the limited target scenario of a single kid with a single city set for the short period of time that the parents are still paying attention to the new purchase/gift

Amen. Exactly that. A few hours max ...

Best,
Thorsten

[Moz]

I got the HDBricks ball bearing bricks, and I'm a bit unimpressed. One bearing arrived in pieces, and just generally the Lego axles+wheels I have don't stick together very well so picking up a model generally means half the wheels fall off. I suspect I need to 3D print some holders that have an exterior section like the Lego wheel holders do. It makes testing rolling resistance very tedious as I have to find all the wheels and washers then reassemble the test vehicle after every test.

Does anyone have advice?

Upside is that the BrickTracks R104 points are very nice. Injection moulding really does give a better finish than even the decent 3D printers that TrixBrix use. In Australia it's cheaper to buy BrickTracks stuff from Europe because postage from the US is extremely expensive (more than the cost of the tracks in any combination I can find). Now if only they offered more than simple points :)

Edited September 3, 2022 by Moz

[Polarlicht]

I recently got a brick built set of a BR 18 201 from a different brand. It comes with one passenger car, with wheels and wheel holders which are exactly as the current newer Lego ones,

and let me tell you, these have almost zero rolling resistance.

Also i have no problem using the newest Lego wheels and holders. They might have slightly higher rolling resistance than the old ones, but not as much to worry about IMO. And i run some pretty long trains.

Edited September 3, 2022 by Polarlicht

[Toastie]

3 hours ago, Polarlicht said:

They might have slightly higher rolling resistance than the old ones

What are you referring to: Old = metal axles and wheels with through hole + New = metal axles with wheels and no through hole?

Or is New referring to the split plastic axles?

Best,
Thorsten

[Polarlicht]

I mean

Old: both metal axle versions

New: current full plastic

I have all 3 (axle through, no axle through and current) and they all feel roughly the same.

Edited September 3, 2022 by Polarlicht