9v motor working/safe/cutoff temperatures ?

[papluh]

Long story short (the long one is at the end)

A 3 carriage 9V train on a loop over 5x15 baseplates - running long term in a display window. Originally with just one 9v motor as rear one in the locomotive (from set 60198: Cargo Train), but was not reliable and getting stuck in most remote cornet after like 2 hours of runtime. Adding a PF motor to the front (powered from the 9V one) and a second power connection (about 1/3 of loop apart) made it semi reliable, gets stuck maybe once in 1-2 days. 
The 9V motor is warm to the touch when running, was a little hotter when stuck. 
Took a FLIR one to check on the train, there is quite a disparity in heat - max and distribution - between the PF motor in front and 9V in the back. 
- is such heat difference normal ?
- Or is the 9V motor on the way out ? 
- at what temperature should the thermistor trigger (rise of the resistance curve ? )
- the PWM source measures current, the train with just 9V motor was taking 140-240 mA between straights and curves, with 2 motors it is 190mA-290mA (hard to say how much goes to 9V only). Are these currents something that should explain the heat ?
- the motor was opened before deploying last year due to pickups not having contact. It was cleaned so there shouldn't be much of a mechanical resistance to generate heat. 

- could it be that the DC motor itself is the source ?


Now the long story...
I'm not a 9V train expert , but we have been running a 9V track on our exhibitions for years. One of our member provides a basic Metroliner and we lay down some 10-20 meters (depending on site) of collected tracks powered by original controller(s). Displays usually lasts weeks in local venues, power is on a AC switch with timer of around 3-5 minutes. So far so good, once the initial setup with replacement of some misbehaving track pieces the train works unless external accidents happen. 
Last 2 years we made a winter display for local shop and things don't seem to work anymore with 9V trains.
2020 one - round 3x8 baseplates, original controller hidden and connected in the mountain set on a 15 minutes on/off socket timer, 9V motor in the tender of the 10254 Winter Holiday Train ... ran well for some test at home (like 2-3hrs) and even there for like 2 days. Started to get stuck in curve next to winter fire station (cca furthest from power connection). Due to placement I had to use battery loco to push it, but only after it had a rest. Motor felt warm. Removed the tree rotation, last carriage - helped a little maybe, still got stuck each day or two during the 6 weeks on display. Later I have used the display on pop-up events - it still got stuck after some time of uninterrupted use, even with less carriages. I had to return the track to a member as I don't own any and it happened even with completely different track. 
For 2021 one display which is lot bigger (5x15 baseplate track) I dumped the winter train and went with 60198: Cargo Train stock loco (part of request was to use as much available stuff as possible) with carriages made by member to be on metal axles bogeys and not too heavy. My only 9V motor was used in the rear of the loco. Power was now through a custom PWM modulator (9V and 65-70%), originally connected at the train station, switching by remote control by staff.  It ran for hours in storage ... only 2 days on site and they reported it started to get stuck after an hour or two, just as it passed the tunnel and the hill. As a fix I have added second motor, PF one to the front connected to 9V one, a second power cable (just long enough to reach end of the straight before hill. Now it gets stuck maybe once in a 2-3 days. 

Edited November 11, 2021 by papluh
pic rearange

[XG BC]

the 9v motor has to get a little warmer than the pf one in theory because the pf one draws current from its pickups too thus generating more resistive heat

[zephyr1934]

18 hours ago, papluh said:

Last 2 years we made a winter display for local shop and things don't seem to work anymore with 9V trains.

That is a large loop for a single track power connection. There is a lot of resistance in the track connections, and it is 2x since the current has to go out and come back.

Try two (or more) power connections to the track, all off of the same controller. Try to evenly space them, even better, figure where the highest draw will be (probably in the corners if everything is flat) and put the power connections there. Even with light trains you might want to consider using two 9v motors, so if one hits a dead spot the other will have electrical contact and will push/pull it through.

[davidzq]

What XC BG said is correct, as heat is a byproduct of electrical.

 

Another factor is actually your PWM power supply.

The PF motor features a capacitor to help smooth the voltage.

The 9v motor does not, so the voltage spikes are dissipated as heat by the motor. It's not enough to effect the motor, especially because it has thermal protection, but it does bring the temperature up.

Michael Gale actually did a nice summary of this Last week.

 

Another factor is weight distribution. The motor with the most weight above it does more work. Though the difference is typically negligible on most Lego locomotives.

 

Edited November 13, 2021 by davidzq

[papluh]

On 11/12/2021 at 10:30 AM, XG BC said:

the 9v motor has to get a little warmer than the pf one in theory because the pf one draws current from its pickups too thus generating more resistive heat

I imagine the metal wheels can add to a different overall heat profile by just the thermal conductivity vs plastic ones. 
Here I'm just theorizing as I don't have any measurements, but I'm  not sure how much heat would the internal circuits add, the pickups and the connector metal have quite some cross area. At least comparing to the cables that need to supply the same current from regulator (which then would be lot more warmer).
But yes, as the 9V motor bogey has to pass both current for itself and the PF one it's conductors will generate more heat than just the PF. The question is how much. The power source reports just about 0,3 A @ 9V max with both of the motors. 

Anyways it needs more observation ideally with a known good 9V motor. With the issues reported I just can't get anybody lend me one :D

On 11/12/2021 at 6:07 PM, zephyr1934 said:

That is a large loop for a single track power connection. There is a lot of resistance in the track connections, and it is 2x since the current has to go out and come back.

Try two (or more) power connections to the track, all off of the same controller. Try to evenly space them, even better, figure where the highest draw will be (probably in the corners if everything is flat) and put the power connections there. Even with light trains you might want to consider using two 9v motors, so if one hits a dead spot the other will have electrical contact and will push/pull it through.

Is there some common rule on track length vs power connections ?
Here my MSPaint masterpiece edit of the last version of the layout with some notes for reference (ignore the stuff in pink areas, that was for our last minute changes :D ) 

As I mentioned in the first "fix session" I added both the PF motor and a second power point. Hard to say which of them helped more.
I have been measuring the voltage on track in the problematic curve, on PWM 100% and 9V there was a drop of cca 0.4V as reference with original power point only.
The train runs at 70% PWM which should be in theory around 6.3 V on average over the wave, my digital multimeter seems to work with that (I should get more measurements for better calibration )
The voltage measured at the curve depends on the position of the train - if it was running and on oposite curve it was around 5.8-5.9V. As it was just before the curve it was ~5.3V with single power connection and ~5.6V with added one. 
One thing - when doing test runs with the train on a short track it would start moving just below 50% - which in effective current  should be less than measured in the problematic curve. Although I don't remember if I tested that in a curve which can add drag.

5 hours ago, davidzq said:

Another factor is actually your PWM power supply.

The PF motor features a capacitor to help smooth the voltage.

The 9v motor does not, so the voltage spikes are dissipated as heat by the motor. It's not enough to effect the motor, especially because it has thermal protection, but it does bring the temperature up.

Michael Gale actually did a nice summary of this Last week.

 

Another factor is weight distribution. The motor with the most weight above it does more work. Though the difference is typically negligible on most Lego locomotives.

 

Thanks for the post reference - I read it through. But I'm kind of puzzled, don't (some of ) the 9V motors have a capacitor ? That part between the contacts on the DC motor itself as seen for example here (I can't read the markings but people refer to it as capacitor). But even if its there maybe it is of wrong spec for my setup.
My setup is a AC-DC adapter (12V @ 2.5A ) - socket - DC-DC stepdown module set to 9V output - PWM module - 9V connector  (I planned a 2way switch for reversing but picked a too small enclosure :) ) 
I don't have an osciloscope to confirm but the 16 kHz in the specs of the PWM module puts it into the Very High frequency as per Michael's post - I'm not sure what other implications it has besides the switching being quiet. 

If I get there next time I could turn aroung the loco (9v in front, PF in back) to see if it makes a difference. 

[davidzq]

On 11/13/2021 at 3:33 PM, papluh said:

Thanks for the post reference - I read it through. But I'm kind of puzzled, don't (some of ) the 9V motors have a capacitor ? That part between the contacts on the DC motor itself as seen for example here (I can't read the markings but people refer to it as capacitor). But even if its there maybe it is of wrong spec for my setup.

That is the thermistor (overload protection).

On PF/PU motors, there is a capacitor, with the thermistor being on the motors wire lead.

Edited November 16, 2021 by davidzq

[papluh]

6 hours ago, davidzq said:

That is the thermistor (overload protection).

On PF motors, that is how/where the capacitor is mounted, with the thermistor being on the motors wire lead.

borrowing a static pic from here the thermistor should be in series (due to how it works) and in 9V motor should be the disc thingie between contacts on DC motor and one of the pickups (orange highlights). 
I was originally referring to the part in green highlight. But looking at it more and more the packaging seem to be that of a diode and not capacitor. 

Anybody retrofitting capacitor to 9V ?
What parameters to use ?  (i guess the max capacity depends on the PWM frequency as just to smooth the pulse edge not counter the whole up-down pulse, right ? )

to be clear, my formal electrical engineering education ended up with mandatory introduction classes at uni, some 20y ago  :D 

[zephyr1934]

16 hours ago, papluh said:

Is there some common rule on track length vs power connections ?
Here my MSPaint masterpiece edit of the last version of the layout with some notes for reference (ignore the stuff in pink areas, that was for our last minute changes :D ) 

I'm not sure what you are showing with "PWM" but with the power point in the lower right the problem spot makes complete sense, the train is both the furthest from your power supply and in the most curves.

No rule, more along the lines of "if it slows down too much in the far corner, add a power connection" and keep adding connections until you are happy with whatever speed drop is left. Back when we would regularly do 9v shows I would have had two power connections on a loop of this size. Probably the best spots would be A3 and P3 to get the power closest to the curves where the physical drag will be greatest, but I bet you'd be fine if you put them at H1 and H5 to keep it simple.

[papluh]

2 hours ago, zephyr1934 said:

I'm not sure what you are showing with "PWM" but with the power point in the lower right the problem spot makes complete sense, the train is both the furthest from your power supply and in the most curves.

PWM is the power source (using PWM) with 9V connector.

The cables to power points are the 9V rail connectors with wires replaced due to insulation rot. The shorter one was conservative in length :) for the longer one I used all the wire I had purchased at the moment and it was not enough :D Once I source more of the wire I'll fix some of the 9V extensions cables to be lot longer to move it further down the track and have as few connections as possible. 
Another challenge is that the orange section was made a LUG member including terrain (as the only module) and I'd have to look if I can pass the cable under the tracks without breaking the MILS like base too much .
The move to P3 should be possible, there is a small module of track that is not pin connected and could be modified :)
Let's see what changes in the behavior.

[davidzq]

11 hours ago, papluh said:

borrowing a static pic from here the thermistor should be in series (due to how it works) and in 9V motor should be the disc thingie between contacts on DC motor and one of the pickups (orange highlights). 
I was originally referring to the part in green highlight. But looking at it more and more the packaging seem to be that of a diode and not capacitor. 

Anybody retrofitting capacitor to 9V ?
What parameters to use ?  (i guess the max capacity depends on the PWM frequency as just to smooth the pulse edge not counter the whole up-down pulse, right ? )

to be clear, my formal electrical engineering education ended up with mandatory introduction classes at uni, some 20y ago  :D 

Oops, you're right, the disk is the over-current thermal protection.

The black component is a TVS diode for over-voltage protection.

Been a few years since I've opened up my motors and my memory got mixed up 😛

 

Anyways it should be easy to add a capacitor. Simply solder it to the bussing on the motor.

[papluh]

10 hours ago, davidzq said:

The black component is a TVS diode for over-voltage protection.

No worries, I'm refreshing (and adding to) my knowledge on the go too.  
This seems to be used as a flyback/snubber diode
Which leads me to a question - how can that (efficiently) work on a setup that can be easily reversed, either by control and/or by build ? I mean the principle of a diode as a discrete part is that it has different behavior on each side of polarity

10 hours ago, davidzq said:

Anyways it should be easy to add a capacitor. Simply solder it to the bussing on the motor.

Yep, soldering is easy. Choosing wrong capacity could make either the capacitor or the PWM ineffective. 
First step would be to look what is used for PF motors, but then their PWM base frequency is lot lower than what I have (around 1kHz compared to 16kHz).
Time to look into some equations. 

[XG BC]

2 hours ago, papluh said:

No worries, I'm refreshing (and adding to) my knowledge on the go too.  
This seems to be used as a flyback/snubber diode
Which leads me to a question - how can that (efficiently) work on a setup that can be easily reversed, either by control and/or by build ? I mean the principle of a diode as a discrete part is that it has different behavior on each side of polarity.

doesnt a tvs diode only breakdown over a certain voltage? if it is a bidirectrional one that shouldnt matter as it only becomes conductive if the voltage is too high

[papluh]

2 hours ago, XG BC said:

if it is a bidirectrional one

hah,thanks, good point.
That was my lack of praxis talking. And a fixation on the visible marking only on one end suggesting a directional part :)  

If I had the motor at hand I'd have read the marking first, but as it is still (hopefully) running at the shop :) I was just guessing from what i have found online. 
The only pic where 'zoom, enhance' revealed anything reads BZW maybe BZW04. Seems these can be bidirectional and it is logical if such one is used.

[davidzq]

Opened up one of my 9v motors for you.

Full markings are:

BZW

04P

15B

C605

So...TVS Diode BZW04-15B

 

I also opened up a PU train motor. The capacitor is a surface mount, with no label 🙁. It's labeled as C2 on the board. I don't have any equipment to test it's value. Traces to each motor lead.

But it also implies there's a C1 somewhere.... internal to the motor, maybe? But I can't open the shell without removing the pressed on gears.

One of the motor leads passes through a 509R resistor (labelled R1).

 

 

 

 

 

 

 

 

 

Edited November 16, 2021 by davidzq

[papluh]

19 hours ago, davidzq said:

Opened up one of my 9v motors for you.

Thanks a lot. 

As for PUp motor, the values would be related to the PWM base frequency in use. I'm wondering if that is the same as PF which should be 1150 Hz (as per sources found online). 
Once one of my first PUp motor burns out, I'd gladly open it up to look for C1. But so far they don't have much long term use as 9V ones, so hopefully lot of life left there .

[zephyr1934]

On 11/14/2021 at 2:32 PM, papluh said:

PWM is the power source (using PWM) with 9V connector.

The cables to power points are the 9V rail connectors with wires replaced due to insulation rot. The shorter one was conservative in length :) for the longer one I used all the wire I had purchased at the moment and it was not enough :D Once I source more of the wire I'll fix some of the 9V extensions cables to be lot longer to move it further down the track and have as few connections as possible. 
Another challenge is that the orange section was made a LUG member including terrain (as the only module) and I'd have to look if I can pass the cable under the tracks without breaking the MILS like base too much .
The move to P3 should be possible, there is a small module of track that is not pin connected and could be modified :)
Let's see what changes in the behavior.

What about putting the power connections around H1 and H5 to keep it simple while maximizing the track distance between the separate connection points.

 

Another thought, definitely inspect the track to make sure there are no loose connections, as can happen if the baseplates are misaligned and the track is connected.

[papluh]

6 hours ago, zephyr1934 said:

What about putting the power connections around H1 and H5

The 1st line is really close to the shopping window and not very accessible, definitely not the middle part. But I'll keep in mind such connection placement when experimenting with it on next display :)
This last video shows the placement https://imgur.com/VSfwDxt 
It was taken last week and it shows the slowdown coming out of the "problem zone" even with 2 connection points and 2 motors.

I though too about connection problems and pressed down any track I could reach there (generally still in reach from the side, about to D column), but no change. 

Funny thing is that I went there yesterday and the train was moving lot better, lot less slowdown in that curve. They did not touch anything as they are scared to break it :)
 I did not take a video as I was low on battery charge and in hurry. And of course did not had my FLIR. So not much real data to compare.
I'm wondering if it somehow fixed itself. Or maybe there is also something else. Like ambient temperature, it is getting colder here and even less sun. There is no direct sunlight there, but the corridor outside the shop has lot of skylights above and big entrance quite close so lightning and temperature conditions are dependent on the weather. Maybe a degree or two less and it is better ? 
Or the track has changed somehow ?