Mitsumi small flat-can motors - not all alike?

[Atlantic 3279]

Following the loss of Mashima motors, one of the alternatives has been a Mitsumi flat-can type with a can about 19mm long, about 15mm "diameter" reduced to 12mm across the flats. A while ago now I bought a couple as a trial initially, then a batch of 15 more of these via eBay (remarkably cheaply) and I've used them with success in at least three J6 loco models as well as in tandem (mechanically linked) in an A5 tank engine.
More recently I've been working on a project (yet to be revealed) which involved use of two of these motors driving different groups of wheels without any mechanical link between the two groups The starting characteristics and running speeds of the two groups therefore needed to be reasonably well matched. I'd catered for that by means of gear ratios to suit the relevant wheel sizes. I was therefore somewhat dismayed to find that with the two drives assembled, checked for free-running and both run-in, I had one unit showing more reluctance to start as well as having a top speed that was lower than the other unit and lower than expected based on results with previous similar motors. The difference in starting characteristics was enough to result in one driven set of wheels initially dragging the other stationary set along the track when starting from rest! A lot of time-consuming checking, adjusting and head-scratching eventually lead me to the conclusion that the gear sets, wheels, axles, bearings and power pick-ups were fine but NOTWITHSTANDING MASS-PRODUCTION AND ORIGIN FROM THE SAME SUPPLY BATCH, THERE WAS A VERY SIGNIFICANT DIFFERENCE BETWEEN THE TWO MOTORS.
I did some quick checks on a couple of other motors from the batch and grudgingly concluded that I would have to change the motor in the "sluggish" unit, a task that was possible but not quick and easy. That solved the problem.
So what was wrong? Well, I've now bench tested all the motors remaining from my original batch, connected to a basic H & M clipper controller, with a voltmeter connected across the motor. I'd run each motor briefly before taking readings to clean brush contacts and free off any initial stickiness. Although I suspect that my voltmeter may read up to 15% high, the comparative readings were revealing:
10 motors consistently started to turn, without faltering, when the indicated voltage was 1.5V or less.
2 motors, one un-used and the "definitely run-in" one I'd taken out of my sluggish drive unit only started when the indicated voltage was 3 to 3.5V and (subjectively) sounded less smooth and less willing when running.

I also found differing resistance readings across the motors, when static, but those with the apparently highest resistance figures were not in every case the reluctant starters.

Another lesson learned.

[Hatfield Shed]

...NOTWITHSTANDING MASS-PRODUCTION AND ORIGIN FROM THE SAME SUPPLY BATCH, THERE WAS A VERY SIGNIFICANT DIFFERENCE BETWEEN THE TWO MOTORS.

And in case it is of interest, this variation is typically primarily due to the magnetic field strength and field distribution within the motor. The other most significant potential variables in the motor assembly such as wire conductivity, windings, armature balance, shaft and bearings finish and fit, construction and action of the commutator and brushes, are readily controlled within very small limits by long established mechanical engineering technique.

As a side note, it is one of the reasons I find DCC so useful. The adjustment available on decoders allows such motor variation to be fully compensated for, such that on track performance is completely uniform, despite very different motor performance characteristics. (One particularly recalcitrant specimen some years ago which well fitted your 'less smooth and less willing' description just had to be opened up. One of the magnets was not securely fixed within the motor casing, and was constantly shuffling laterally as every armature pole passed it.)

[Atlantic 3279]

That is certainly of interest. On reassembly and re-testing of the initially sluggish drive unit I was reminded of another point that can affect full satisfaction when using these Mitsumis. Additional control of armature end-float is sometimes required in order to make them run equally quietly both ways when a worm is fitted.

[Pebbles]

I have only so far discovered a couple of duff motors. This is not a problem, as due to their not being serviceable and their cheapness, I had decided to obtain more than I required.
As the motor shaft extends from both ends of the motor the end play can be reduced externally. Ideally this should be taken up at the non commutator end.
I would add that at various times quite interesting motors crop up on ebay; just get them whilst they are available.

[exile]

I am not at all surprised by the findings.

Some years/decades ago I had to source 2 matched Roco 0-6-0 chassis to build an 0-6-6-0 Mallet tank (nothing even slightly related to LNER). It took 4 purchases to find a pair that were even closely matched.

None of our model requirements are being manufactured to very tight tolerances (unless we are talking P4 wheels and similar) and this shows up when we need to pair chassis.

[Atlantic 3279]

Now that I know the potential variations in these motors it would not stop me from buying and using more of them, as long as they remain so inexpensive. All I have to remember to to do is to test the willingness of any further new purchases to start on low voltage, to run smoothly, and to attain full speed, especially if they have to be part of a matched combination or need to meet other criteria for a particular loco. Those I already have in stock are now all labelled...

[john coffin]

We have I think become a bit lazy with the whole reliability and cost factor in terms of the original Mashima motors,
they were made for the car industry, and have basically been replaced by electronics, plus of course the Japanese
motor industry is much smaller than it used to be.

Once you buy any thing from a new source you should of course check it and not assume it is correct out of the box.

Anyone who does any woodworking above basic levels knows that they need to "fettle" their brand new tools
even quite expensive hand planes need their blades honing and tuning up, why should model railways be any
different??

Just be thankful we are not back in the days of "k"s motors!!!!!!!!!!!!!!

Looking forward to the new model Graeme.

Paul