Mosfet Short Circuit Protection

[electrosteam]

I reported in thread "Loctited Wheel Slipping on Axle" in General Chat the effects of a Mosfet failing short circuit on my 0-B-0 boxcab electric locomotive BOB.

Has anyone developed an effective protection device or scheme to prevent this happening in the future ?

Note that I need protection per Mosfet as the two motors have separate drives and are mechanically coupled.

I will begin investigating fuses (high speed ?) and circuit breakers.

John

[isc]

I'm a bit out of date with electronics, but I think you'll find that if Mosfets are anything like Transistors, they are faster than any fuse, so I think you are looking for an electronic protection circuit. You my be able to find something in one of the electronic hobbies sites on google/ or what ever. isc

[electrosteam]

Thanks for the quick reply.

The protection is not required to protect the Mosfet, it is needed to limit the let-through energy into the motor causing the mechanical damage reported.

Perhaps a shunt Mosfet to divert current away from the motor operating on a detected high current.

The fundamental problem is two motors mechanically coupled through the coupling rods - a design flaw that is obvious now.

My next loco may have two motors, but it will have the two motors geared together, then to a layshaft, then to the coupling rods.
(I have a thing about rigid framed electric locos with rods !)

John

[peterseager]

John

I read this yesterday evening and been turning it over in my mind during today. My thoughts are:

As has already been pointed out the MOSFETs themselves can only be protected by current limiting in the controllers themselves.

You do not say what caused the MOSFET to go short circuit. Did it itself fail or was it forced by an external happening?

The use of a Miniature Circuit Breaker (MCB), as used in mains consumer units, would protect the motors if suitably rated and at the same time provide an ON/OFF switch. In your case it would need one MCB for each controller/motor. Some MCBs can be linked for use on three phase circuits and that would be useful with your setup.

If you now have a failed controller could you not replace it with a single controller driving both motors? I single MCB would then surf ice.

My Hymek has Qty 4 150watt 12v motors running off a single 24V 100 A Parkside controller. The MCB is Type B rated at 50A. Higher amperage ones are available but not off the self locally. But I have never had any nuisance tripping with the 50 A MCB.

Another club member has a Class 8 shunter. He has a 4QD controller running 4 motors geared to the axles which are coupled like yours. It runs successfully.

I assume the drive train from each motor is reversible, eg not a worm drive? The latter would cause problems if power was lost to one motor.

Peter

[electrosteam]

Peter,
The loco was moved at low speed from the depot to the carriage yard.
When reversed to back up to a train, it accelerated at maximum rate. I hastily turned the master switch off.

I assumed I had bumped the throttle control as I operated the reversing switch as the controls are fixed in the rear of the loco and I was standing, so I checked all the settings and tried again.
A typical very busy time in the yard as we got ready for public running - I did not put enough thinking into what was happening.

The instant I operated the master switch the loco locked solid with at least one wheel turned on an axle.
I had to remove a siderod to roll the loco back to the depot.

It is possible that the reversing switch was thrown a bit early, but I am well aware of this risk and I am always very careful not to do it.
I think it was a normal failure due to either chance, or less than optimum rating and snubber design - probably the latter.
The second motor still operates fine.

I will do a re-check on all the ratings and the snubbers.

My thinking has turned to protective functions - keys on the wheel/bush and bush/axles joints, high strength Loctite with reduced gaps and improved cleaning as a start.

Then there is the possibility of an active protection scheme.

The HST train in the UK and the related XPT in NSW are fitted with a very large short-circuiting contactor across the output of the traction alternator.
If a flash-over is detected in a traction motor, the contactor comes in to quench the current flow.
The engine and alternator are designed and rated to do this a specified number of times before overhaul is required.

An equivalent scheme would be a second Mosfet detecting excessive current in the switching Mosfet and short-circuiting the motor terminals, or to simply take the battery directly to ground.
The Mosfet would be rated to hold the battery maximum current long enough for a MCB to trip open.
This stops excessive motor current and results in positive isolation.

John.

[peterseager]

John

I have put circuitry in the remote console to stop inappropriate commands.

The loco can only be "started" if the reverser and throttle are in their Engine Only (EO) and minimum positions. Starting applies 24V to the controller control circuits and the sound system.

The throttle has a simple threshold detector that senses movement of the throttle from minimum and sets the forward/ reverse relays in the controller. Thus reversing can only be instigated at this point.

I cannot see that short circuiting the motor is going to do anything. In full size the contactor is being used to extinguish an arc inside the motor and presumably crowbars the supply to the motors off. This is not the problem here. I think a well designed controller with current limiting matched to the starting current of the motor(s), MCB to protect main wiring and motor(s), fuses to protect low power circuits and a well designed console are the way forward together with improved wheel security as you mention.

The Class 8 shunter mentioned in my previous post suffered a problem after an overhaul. The connection to one of the four motors became reversed. It operated for a while before it was decided that it was not up to scratch. On investigation the reversed connection was found. The motor was eventually replaced but everything else was OK.

Peter

[electrosteam]

Peter,
All your suggestions are well founded and I will think through any proposed re-design carefully before committing.

As an aside on reverse motors, Comeng in Sydney built a class of 1500 Vdc Co-Co electric locos of about 1800 kW ( one was arranged Bo-Bo-Bo as a demonstrator ).

The first duty for one of the locos was a passenger train north.
The loco was failed by the driver because of "severe vibration" at high speed and full throttle (series parallel and resistor style control).
All tests (light engine at low speed) by Comeng said it was OK.
The railways insisted an engineer travel on the next run.
As reported, under full throttle as the transition went to all parallel at speed, "all hell broke loose".
The engineer was hanging out of the cab watching a bogie reacting very badly to good (for us) track.
Inspection revealed one traction motor was wired backwards.

John

[electrosteam]

Further investigation of the snubber revealed a crack in the glass envelope of a 33 Vdc 1 W Zener diode wired across the Mosfet.

This explains the observed behaviour and subsequent testing with a multimeter that seemed to indicate an intermittent fault.

Replaced with 2 x 15 Vdc 5 W Zener diodes in series, each with a balancing 15 k Ohm resistor in parallel.

Although the other motor driver tested OK, its snubber was changed as above.

Conversation later with another builder produced agreement that a Zener at 1 W might be borderline.

Isn't experience wonderful !
John.

[peterseager]

Thanks for completing the story. A nice simple solution!

Peter