My new problem : the power supply. I thought that the LM317 provides short-circuit protection ; I read that in fact it does ... by defuncting in some cases ; my problem is still the short circuits caused on a Rokuhan turnout which would not switch properly, I wanna damage neither the turnout, nor the rolling stock. So I am looking for an electronic circuit breaker, which could be used at the output of a CONSTANT voltage source (my PWM will be done by a microcontroller, so I don't need a standard train controller ; and 7.5v could be used both for traction and turnouts) ; there are some standard switching power supplies, in plug or box shape ; unfortunately somebody said that they manage short circuits correctly ... but not too often, say once or twice per day, and that the only power supplies which are not too easily destroyed are those for lab use. A circuit breaker in a former post is only intended for DCC use. So, what can I do ? Does anybody have schematics for such a protection (or a link to a hobbyist page) ? Or is there an IC intended for this use ? Or should I simply use a standard train controller, set to maximum ? Thanks for every help ...
Yes, 3-terminal regulators provide short-circuit protection for themselves only. You need a multi-pin device for load current limiting. It's been a long time, but one I can think of is the uA723. It may not meet your needs but it's a place to start. All the switching regulator chips I worked with had current limit/shutdown but I can only remember the manufacturer, Plessey.
Thanks Mark. I read that the LM317 could be wired as a current limiter, too. But what I want is a real BREAKER. I found a system which triggers a thyristor, here : www.sonelec-musique.com/electronique_realisations_disjoncteur_001.html R1 acts as a current-sensing shunt, an adjustable amount of this voltage is directed to the base of Q1, the transistor saturates if this voltage is about 0.7v ; C1 is for a triggering delay. I think the same effect could be achieved with a LM311 comparator. The thyristor-relay combination could be replaced with a RS Latch built from CD4000 series, with two NOR-gates (so the trigger input would be active high, for the transistor version with a pull down resistor) or NAND-gates (active low for the LM311, with a pull-up). The only problem is that reset would still be manual. I heard of some systems which, after a shutdown, send automatically very short pulses to detect if the problem is still present and, if not, reset automatically ... But in worst case, a breaker with a manual reset would be more "boring" but better than a limiter, all the way.
Last Edit: Apr 5, 2017 16:09:37 GMT -5 by alberich
It trips in exactly the mode that you are trying to detect, overheating of the coil.
To set the premise: your issue is not the voltage, nor the current, but the power applied to the "motor" over time.
The PolySwitch (brand name from Littlefuse) otherwise known as thermally resettable fuse/switch, basically integrates voltage and current (watts) over time, and will go into a higher resistance mode if the limits are exceeded. I think this will work perfectly.
Thank You Greg for that answer, but what's about the reaction time ? Although I wrote my post into the topic for locomotives (due to the reference to the breaker for DCC), the system should protect from the short-circuits in the turnouts (frog) too, not only from short-circuits inside the locomotive overlapping two sections of opposite polarity. As I wrote in the topic about the turnouts, I don't want to melt the springs (the contacts between the upper shell and the PCB) like our German modeller friend (but he used DCC, where voltage and current are higher). A turnout is rather difficult to replace between other track sections ! I thought of a solution suggested, using power routing to a section of track long enough to stop the loco before it comes to the turnout ; unfortunately I decided meanwhile to 99.8% to run two RDC's coupled together, so the first car would stop on the section but the second would "push" it, beeing still powered. So I don't know exactly what would happen ...
Last Edit: Apr 5, 2017 16:53:06 GMT -5 by alberich
Post by Greg Elmassian on Apr 5, 2017 17:08:53 GMT -5
I'm only proposing this for the turnout motors. It's the best solution for the Rokuhan turnouts I know, besides the capacitive discharge power in the first place.
Your issue protecting locos and derailments is another one with very different requirements, and I suggest you use other methods. I have ideas and ways I solve these issues on my own, and I have them in G scale also at many more amps, but melted traces and wires in locos from derailments is an issue there.
Also, I deal only with DCC for track power, makes things a lot simpler, so my solutions would be of no help to you, sorry.
Yeah the uA723 is an ancient regulator but more controllable than the LM317 or LM78XX. Giving your discussions a bit more thought, I have a few ideas for you to consider.
The example you linked to suggests you're looking for an analog solution for shutting down the supply. Continuing along that line, if you are still using the L298, I'd use the device's Enable pin to control throttle output. If seems that the circuit you linked to would work with adjustments to component values. You could look into using a silicon controlled switch (SCS) instead of an SCR. The SCS is the same as an SCR but with an additional connection that gives you both an "on" gate and an "off" gate, an analog R-S flip flop. You could connect the extra gate to a timer to allow for a timed automatic retry.
One concern I have with this approach is that with PWM you're going to have switching currents on the order of:
which many cause false triggering. For a data point: on my test bench, using the Snail PWM throttle driving an AZL locomotive I see as much as 10mA switching current.
A digital solution may be the easiest way to monitor current and provide overload protection.
Thanks for Your answers ... 1) The idea with a timer (oscillator with short pulses) to try retriggering seems obvious, it could be used with all systems (including the RS-latch) ... About the current pulses, I think one can leave a safety margin (and use the delay capacitor), without producing immediately "electronically assisted smoke signals" ; I think the permanent short-circuit current would be much higher and could so trigger the breaker safely. 2) I saw that the uA723 (from 1972 ! ) is still available, even at Conrad France ; unfortunately it has no "current overflow" output, signalling when the limiter becomes active, and which could trigger a breaker. 3) The thing that amazes me most is that there is no IC with all the necessary functions, to build a power supply with only one IC ; I saw some interesting circuits by Maxim for a 2-ICs solution, unfortunately I could only find them in SOIC package. 4) If there is a readymade "digital" solution other than the limiter for DCC, or a dedicated real DC power supply (without PWM throttle and pulses for lighting), other than lab power supplies ( ), all tips are welcome ...
I saw an example with uA723 here : electronic.de.loisir.pagesperso-orange.fr/ua723.html Considering the low current needed by locos, I think I could discard both transistors and directly connect pins 10 and 2 (uA723 is rated for 150mA). So I will order all parts and test ; in the worst case I will "shoot" an uA723 ... About PWM, progressive acceleration etc : on my former layout in N, the trains ran at only 2 speeds : full and slow, with pure DC (and a supplementary correction for "downwards", 3% slope ) ; the inertia did the rest, and it was sufficient for me ; so I wonder if I am not again drifting away, to the "gaseous plant" ...
The most recent idea : a circuit breaker using a simple relay (not yet tested) ; the "voltage reduction" diodes should only be used if Vcc is too high for the coil and could be replaced with one Zener diode in case of a "very big" difference ; I think that this circuit breaker should not be triggered by power "spikes", only in case of a real short circuit ; and I read that reaction time of relays is less than 50ms. What do You think about it ?
Last Edit: Apr 16, 2017 0:53:17 GMT -5 by alberich
I want to shut down power supply to the tracks if a short circuit is caused by a locomotive running into a turnout that could be set to the wrong position by a malfunction (but I think this is not very likely to happen ... ) So the whole "chain" where the short current could run through would be protected : the PCB in the locomotive AND the contact systems in the turnouts ! And additionnally the power supply would be protected, too ...
About BAZman's last post : perhaps this supplementary protection is unnecessary with some regulators ? Or are 2 precautions better than one alone ?
About protection for the coils : the principle itself of capacitive discharge protects them ; and I hope it's very unlikely that an electrolytic capacitor becomes short-circuited under normal use conditions (I saw it once, more than 30 years ago, but with a tantalum capacitor) !
Last Edit: Apr 16, 2017 15:48:44 GMT -5 by alberich
With the gauge tolerance (or should I say more deviation tolerance), you are going to find that just flipping turnouts, steam locos running through a turnout and current surges from DCC decoders placed on the track, vintage trains with incandescent bulbs are all going to have transient 'shorts' or surge currents. Then you get into the higher currents of the MTL F7 and märklin permanent magnet motors building a threshold much closer to the trip current, whether the 'average' current or the *much* larger transient currents (from the *much* higher transient voltages (50-200 volts is NOT uncommon), compared to the *much* lower currents of the coreless motors.
BTW: the DCC Specialties PSX series has a CV that can be set to enable Manual reset.
To BAZman : 1) I will use only DC locomotives, NO DCC ! My layout wil be wired in "traditional" (old fashioned ? ) way, WITHOUT ANY DCC DECODERS , but with capacitive discharge (like the Kato system) for turnouts : 7.5v and 1500uF for each turnout, and perhaps PWM speed control (or variable voltage ? ). 2) And no locomotives with bulbs, no Märklin or steam locomotives : only one passenger train with 2 RDC's in MU configuration, two freight trains with a GP7 and a GP9, all locomotives built by AZL (cars by MTL and "Full Throttle" ), so I assume they have coreless motors. 3) The turnouts will be powered by their own supply (together with the 6v relays, these via 2 or 3 silicon diodes in series to reduce the voltage) ; and the microcontrollers will need a third supply. The only devices powered by the "trains" supply will be the trains themselves, and the detecting circuits for the track sections (with LM 339/393) ; and power to the tracks will be off during "direct/reverse" switching by the relays and flipping turnouts, to avoid transients. 4) So the most problems You mentionned won't occur !