Why don't ball bearings conduct electricity?

[mmagliaro]

Max,
Forgive me if you already addressed this.

But how are you measuring again?

From inside race to outside race?

Have you tried putting the bearing on a loco shaft and then measuring the outside race (casing) to the loco shaft?

It probably won't be any different.  But I was just wondering.

This is exactly what I did.   I ran a steel axle shaft through the center, and then measured the resistance between that shaft and the outer race.  I tried it with both a steel and brass axle.
I think I've measure the resistance of all the pieces of the these bearings now, and they al measure zero,
so clearly it is not oxidation on the surfaces, nor the materials themselves.    Nor does it improve with a conductive liquid like Neolube in there.   So I think the best theory is the pressure one.  The pressure has to be on the balls, but it must be spread over a few of them such that there is just not enough pressure at any one point to get a good contact.

[narrowminded]

What you are experiencing is to be expected especially when amp draws and voltage are so low, especially when we like to have engines that creep so slow.  In my tests in my small mining loco when running on DC, the super slow creeping current needed is down in the range of .7V at about 4-6 Ma.  In electrical terms that's low.  It is the lube film as much as anything that causes the problem and just not enough voltage to break through that.  Dust or dirt will do it too  A metal to metal wiper, very small surface area, absent lube, with a small surface area (small SI's for the low P affords relatively high PSI but with low force), is almost a must for these small voltages while still offering minimal drag and is what you see in one form or another in most reliable runners.  A lot of weight is the other aid for reliable pickup. What I have found is somewhere around 7 grams per wheel +/- seems to be decent for pickup purposes with a bare minimum around 4 grams and up to maybe 10 as a rough, shoot from the hip, working range.  The lower it is, the more track and wheel cleanliness matters but with a decent wheel pickup, the bearing joint doesn't have to be part of the problem.

Aside from previous life experience applying bearings, I have experimented with small ball bearings in my current project and found them to be noisy operating, not as free running as a spin in your fingers would suggest, GROSSLY oversized in capacity which is part of the oil film/ insulating action, and very vulnerable to the smallest debris which immediately puts a hitch in their giddy-up.  A plain bearing and a small axle diameter (1.2 mm?) affords enough strength for handling and plenty of bearing without getting the diameter, therefore bearing speed and friction, too high.  Add a pickup.  I've been there.   My tests in exactly this area are ongoing but ball bearings have been eliminated.  Between bearings, insulated bearings (for other reasons) and power pickup, the following materials have or are being tested and evaluated.  Delrin, Delrin AF, ceramic surface, TFE treated ceramic surface, hardened steel, high speed tool steel, brass, nickle silver, sintered bronze, phosphor bronze, beryllium copper, tungsten, and even gold are part of the ongoing experimenting/ testing but at least one that's been eliminated is ball bearings and at least in part for the reasons you've begun to discover. I hope that helps.

[mmagliaro]

narrowminded:
That is a lot of really useful information there.  Thank you for pointing all this out and reporting what
you've been finding with your experiments.

Although the voltage and current restrictions are not quite as bleak for me as the numbers you are citing.
Even with the coreless motors I use, while it's true that they draw only 5-10mA at 0.7v to run, typically
in my engines, especially with a gearhead, some dropping diodes or resistors, and an LED headlight,
the engines draw at least 30 mA at about 4 volts before they start to creep, and that can overcome oxidation
better than 0.7v.  You are right about that... I've had a few that start running under a volt and it
requires really outstanding electrical pickup to work right.

As for the ball bearings being noisy and getting hung up on dirt in the bearings: yes, I can see how those might
cause a concern.  Reading up on ball bearings, the general consensus in the bearing biz is that ball bearings are noisier than sleeve bearings.  But I don't know, at such low rpms on the axles (maybe 300 rpm, maximum), and on a switcher like I'm building, more like 100), I can't believe there will be much noise from them.

Dirt... hmmm you may have convinced me there.

[narrowminded]

I was aware that ball bearings would be oversized but for simplicity, I was wanting to use them for the layshaft in my unit.  Being a four wheel truck, I added compensation (axles swing independent of each other for pickup on irregular track)  but not by hanging the compensated axle on the worm shaft or using the thrust from it to hold the independent axle in place, the common method for this feature when it is included.  I didn't want the additional friction that that adds at all times and then shows up further as a higher speed in one direction than the other.  The answer that seemed logical and simple was to use ball bearings as they could handle the rotation as well as thrust and the shaft could be pressed or loctited to the inner race.  No separate retainer, no separate thrust faces, simple, huh?  Well...   It was not satisfactory at all as described.  The speed was much higher in that position in the gear train which, as you noted, didn't help with the noise but it was the dirt and friction that made it fussy operating.  It also wasn't easy to clean assuredly.  When you look at the radius from C/L where this action occurs it helps to understand why this happens.  It has a relatively huge lever over that axle C/L.  With the limited HP available I can't afford to give much away to mechanical loss. In HO with all of the space that affords for bigger motors, bigger everything, maybe.  It could probably overcome it adequately.  But in this stuff...   I'm also working the worst scenario first, Nn3 or basically Z, then only a four wheel truck for pickup, a task that's tough at all levels, evidence all of the complaints with small power.   Once all is resolved, scaling up and eight wheel chassis should only get easier. 

While we're at it, SS has the issue mentioned with the oxide.  That's how SS works, the oxide skin sealing the surface and it's not conductive.  Nickel has some of this same problem (and an important alloying element in SS).  No surprise the issue is there.   While nickel alone is pretty corrosion resistant and a reasonably good conductor, in a normal atmosphere it does oxidize and then, that oxide is very tough, unlike brass and copper whose oxides do not conduct electricity either but in its favor, the oxide is fairly easily polished off.  Gold is virtually perfect for the oxide and conductivity BUT... is costly, soft, and when plated, relatively easily wears through.  The wear's greatly improved when hardened with cobalt but it's not a panacea either.  What you get is great conductivity with virtually no oxidation to deal with so the contact preload can be very low while still conducting well, buying back some of the wear problem, but then the servicing and fits are extremely critical and with rough handling and the abuse that things see in the course of regular maintenance it invites failures that are self induced by the mechanic.  I'm still fooling with that but I'm finding that it may really not be necessary if the pickups are well designed and are reasonably easy to service when finally needed.  The "over oiling hurting performance" discussions, a known problem well accepted as true, has these issues at the core of the problem.  Good luck!

[garethashenden]

Max, I haven't followed your threads too closely, but I have the general sense that you're trying to make the best model possible. Can I ask why you have decided against the 2mm Association style split chassis with plain bearings? From my point of view that is a proven system that has been used for decades on countless locomotives. Without meaning to offend, it strikes me that you're trying to reinvent the wheel...

[mmagliaro]

I am pretty much building a split frame with bearings.  The only deviation here is I was considering
using ball bearings instead of the traditional bronze bearing blocks, because it seemed to me that the
friction would go super-low.   

The fact that a split frame with fixed sleeve bearings has worked well for years is a good reason to copy
the design, but not a good enough reason, I'm afraid, for me to not wonder if there is something even better.
(Is the 2mm Association really responsible for the original concept of a split frame with plain bearings?  We
have seen that idea going back at least as far as the Kato 4-6-4 from 1978.   Did the 2mm folks come up with
that idea before that?)

narrowminded: If I do try this, I would not be pressing the axle shafts into the inner races of the bearings.
I want the axles to be able to slide back and forth through the inner bearing holes.

[garethashenden]

As far as I am aware, the split frame chassis goes back to before the Association was founded. The association was founded in the early '60s and there are models built in the scale to what would become the standards dating back to 1928. They used split frames even then. http://www.rmweb.co.uk/community/index.php?/topic/67296-1928-model-engineer-exhibition-silver-medal-winners/

[Lemosteam]

As far as I am aware, the split frame chassis goes back to before the Association was founded. The association was founded in the early '60s and there are models built in the scale to what would become the standards dating back to 1928. They used split frames even then. http://www.rmweb.co.uk/community/index.php?/topic/67296-1928-model-engineer-exhibition-silver-medal-winners/

@garethashenden Wow that is some fantastic info there!  For the WIN!

[Lemosteam]

@mmagliaro , certainly surface contact area has a bit to do with this? When you think about it there would only be two single points of contact per ball available for conductivity: one at the inner race to the ball and another at the outer race to the ball. Even if the inner and outer races were made EXACTLY to the identical radius of the ball, there would still only be two arc lines of contact, but we all know this is impossible with manufacturing variation.

Under load you might only have two balls in contact with the inner and outer races.

[mmagliaro]

John,
Absolutely, the limited surface contact has got to be the issue.   We already know (from testing the
races and balls individually with an ohmmeter) that there is no inherent resistance in the metals.
And there is no lubricant or oxidation present.

The very thing that makes ball bearings do wonderful things (limiting friction) is that there is only a contact
at a few curved points between the metals.

The most unbelievable thing about those 1928 2mm engines, at least to me, is the motor.
I didn't think it was possible to build a motor small enough to power a 2mm gauge engine
in 1928.  I really wish I could see the insides of those engines!

[Lemosteam]

Max, I'm sure they're for sale!   

Isn't everything?

[VonRyan]

John,
Absolutely, the limited surface contact has got to be the issue.   We already know (from testing the
races and balls individually with an ohmmeter) that there is no inherent resistance in the metals.
And there is no lubricant or oxidation present.

The very thing that makes ball bearings do wonderful things (limiting friction) is that there is only a contact
at a few curved points between the metals.

The most unbelievable thing about those 1928 2mm engines, at least to me, is the motor.
I didn't think it was possible to build a motor small enough to power a 2mm gauge engine
in 1928.  I really wish I could see the insides of those engines!

If you want to get technical, 2mm scale locomotives are larger than N scale ones.
2mm scale being 1:152nd and N being 1:160th, there is a bit of a difference there.
Although, the loading gauge in the UK is smaller than in the US, so it evens out.

[mmagliaro]

If you want to get technical, 2mm scale locomotives are larger than N scale ones.
2mm scale being 1:152nd and N being 1:160th, there is a bit of a difference there.
Although, the loading gauge in the UK is smaller than in the US, so it evens out.

But seriously, the difference between 1:152 and 1:160 is irrelevant when considering the craftsmanship and
the difficulties of building such a small-scale model, especially for something like making your own electric
motor... in 1928... yikes. 

[garethashenden]

The other interesting thread is the one on 2mm supplies available in 1956. http://www.rmweb.co.uk/community/index.php?/topic/87667-2mm-supplies-in-1956/

[nscaleSPF2]

Max,
I wouldn't give up on the ball bearings just yet.  There very well may be an issue with contamination, but I don't think that we have crossed that bridge yet.  If I were you, I would want to know just how much less friction the ball bearings have, relative to solid bearings.  I thought of a simple, accurate way to do this.

First, you need a test mule chassis.  This would have 2 side frames, drilled with 2 sets of holes for the ball bearings, and another 2 sets of holes for the solid bearings.  Make a couple of spacers so that the distance between the frames is what you want to have in the finished model.  Install both the ball and the solid bearings into the side frames, then use screws, glue, or solder to fasten the spacers between the sides.  Are you with me so far? 

Next, install a straight, solid axle into one driver (I wouldn't use your handmade drivers; they are too precious for this test).  Run the axle thru one set of the ball bearings, and install the opposite driver.  Repeat for the other axle.  Add enough weight to simulate the assembled weight of the locomotive.  You now have a rolling chassis.

Now you need a test track.  Take one piece of flex track, and glue it to the flattest board you have.  Better yet would be some counter top material, like Corian for example.  Put a fine thread thumb screw at one end of the board.  We are going to use the thumbscrew to adjust the slope of the track.

Put the rolling chassis at one end of the track, which you have adjusted to be perfectly level.  Use the thumbscrew to very slowly tilt the track, until you can very gently tap the rolling chassis and get it to roll the entire length of the track, at a constant very slow speed.  You have just measured the rolling friction.  The amount of friction is equal to the sine of the slope of the track (which is equal to the amount of extension of the thumbscrew divided by the distance from the thumbscrew to the far end of the board), times the weight of the rolling chassis.  You can increase this number by 50% to account for a third axle.

Now place the chassis at the end of the track and slowly increase slope of the track until the chassis just begins to move.  You have just measured the break away friction, and can calculate it as before.

Move the axles to the solid bearings, and repeat the above test.  Hopefully, at this point you not only have proven that the ball bearings have less friction than the solid bearings, but you know how much less friction they have.

You can use the mule chassis to measure the effect of ball bearing seals, and even wheel wiper materials and configurations.  Perhaps the wipers and the seals will offset the advantages of the ball bearings.  Perhaps not.  It would be good to run a test to find out.