Any ideas for a simple inertia controller ?

[Dtsteam]

I'm a bit of a numpty when it comes to electronics, so I would appreciate a bit of help. Does anyone have a plan for a dead simple inertia controller ? I just want something that gives a soft start and stop, ideally something I could make on a bit of stripboard or a tag strip.

 

[gregh]

This is a simple as it gets. There's no overload protection and I'm assuming less than 25V dc input. Choose diode current rating to suit your locos.

The value of 4.7k resistor and of the capacitor C2 determines how long it takes to start up / slow down. Try 470uF if the 1000 takes too long.
The 2N3055 transistor will get hot. It needs to be mounted on a bit of aluminium as a heatsink, say 1mm thick and 100 x 100mm.

 

[Tony Walsham]

Greg,
Just a suggestion to help make the output bullet proof.
I would put a big diode on 2n3055 Emitter. 2n 3055 transistors don't handle kick back very well over 20 volts.
Also I would substitute it for an MJ15003. They have the same pin out. I have never had one fail with a diode fitted. They do cost a bit more but I think it is worth it.

 

[Dtsteam]

Thats excellent Greg., thanks very much. I use old lap top power supplies, so can I omit C1 as well as the Bridge rectifier ?

Could you extend the discharge cycle of C2 by, say, breaking the connection between the 4.7K resistor and the pot once C2 has charged ?

 

[gregh]

Thats excellent Greg., thanks very much. I use old lap top power supplies, so can I omit C1 as well as the Bridge rectifier ?
Could you extend the discharge cycle of C2 by, say, breaking the connection between the 4.7K resistor and the pot once C2 has charged ?

Yes, leave out diodes and C1.
Not quite sure what you mean by extend the discharge cycle. The train 'start up' time is when the capacitor is charging up through the 4.7k. The 'slow down' time is when the C discharges through thr 4.7k. If the train is running and you break the connection between 4.7k and C2, it will take a much longer time to slow down. C2 is then discharging through the 2 transistors, so the time depends on their gain and the load current. As the transistor gains (beta) vary widely, it's hard to predict how long, but will probably be minutes.
At the min quoted spreadsheet gains of 100 and 20 for beta, and 1 amp train, the current from C2 will be 1/2 mA, so will take about 30 sec, but transistor gains are always well above quoted values.
Or are you trying to get different startup and slowdown times? If that is waht you want, put a diode in series with 4.7k with current flowinginto C2. Then parallel those two with a diode facing the other way and larger resistor.

Also note that your max track voltage be 1.5V less than your input supply due to the drop in the two transistor base-emitter junctions.

 

[Dtsteam]

Excellent ! I think I'm beginning to understand this a bit better. Many thanks for your help.
Now to get some bits and build one...

 

[gregh]

Greg,
Just a suggestion to help make the output bullet proof.
I would put a big diode on 2n3055 Emitter. 2n 3055 transistors don't handle kick back very well over 20 volts.
Also I would substitute it for an MJ15003. They have the same pin out. I have never had one fail with a diode fitted. They do cost a bit more but I think it is worth it.

Sorry Tony, I missed your post somehow.
I must admit I don't remember I've ever had a 3055 fail. Is this failure mode due to back emf do you think?
Hopefully if David is just using a computer power supply, he's probably only 12Vdc.

 

[Tony Walsham]

Hi Greg.
Yes back EMF was the killer but usually only if the spike exceeded 20 volts. Some Bachmann locos could generate 60 plus volts. Other brands seemed less destructive.
A suitable kick back diode helped but I ended up spending the extra and using MJ15003 transistor. Big difference in price though. TO3 packages are A$2.50 retail compared to A$7.50 from Altronics.