The electronic timer circuit in this article has a number of serious problems. As an Electronic Engineer, I would not recommend building the circuit as presented in the article.
While the idea is good, the circuit as presented is far from ideal and potentially hazardous due to the acknowledged overheating problem.
With some simple modifications, the deficiencies can easily be rectified and the operation of the circuit improved at the same time, at the same time using fewer and/or cheaper components.
Overheating
The most serious problem, as mentioned in the text, is that resistors R1 and R2 get quite hot during normal operation. Worst case scenario, if the unit was mounted under the layout in a location with restricted airflow, is that the overheating could cause a fire and destroy the layout and house.
The overheating problem is entirely avoidable, by modifying the circuit slightly.
As designed, R1 and R2 form a voltage divider which charges capacitor C1 to approx. 25% of the supply voltage, i.e. about 4 V. There is no reason not to charge C1 to the full supply voltage, in which case R2 is simply not required. Circuit changes required are:
1. Change R3 from 470 ohms to 1800 ohms, to adjust for the higher voltage on the capacitor.
2. Change the capacitor from 10,000 uF 16 V to 2,200 uF, 25 V. This maintains approx the same time constant when discharging.
3. Remove and discard R2.
The operating time may well be considerably longer, so a smaller value of C1 could be substituted if this is a problem, or alternatively, use a smaller value of R3. If the time is 50% more than required, change either C1 or R3 to a value approx. 2/3 the present value (i.e. 100%/150%). It is probably easier to change the resistor to fine tune the time delay as capacitors are not available in as many different values as resistors.
R1 can remain as designed, as its purpose is simply to limit the inrush current when the capacitor starts charging.
A further benefit of the above changes is that the charging time is decreased to 100 milliseconds or less, so a simple normally open (NO) push-button could be used in place of the more expensive version with a changeover (or double-throw) contact for activating the circuit. The push-button should be connected between R1 and C1. R3 can be permanently connected to C1. The NO push-button only require 2 wires rather than 3.
High Quiescent Current Draw
As designed, each circuit draws approx. 160 mA from the 16 V DC supply, for no benefit. If 10 such circuits were installed, a total of 1.6 amps would be wasted, and there would probably be insufficient current to operate an uncoupler (each uncoupler requires 3 amps when energised).
The changes described above to cure the overheating also fix this problem, as the quiescent current draw is zero.
Diodes Close to Max Current Rating
Diodes D1 and D2 are rated at 3 amps, which is the normal current draw for the Kadee uncoupler. While this is probably OK, diodes rated at 5 A could improve long term reliability.
Diode D2 Connected Wrong
As designed, D2 is connected in the wrong place to perform the intended function of protecting transistor T1 from back-EMF when the solenoid is deenergised. Both D2 and T2 are vulnerable to damage from back-EMF.
The coil should be wired direct between the positive power supply rail (cathode of D1) and the collector of T1. D2 wired should be wired across the coil, with the cathode of D2 connected to the positive power supply rail.
Should Not Solder to case of Power Transistor
It is not recommended to attempt to solder a wire to the case of the power transistor, as suggested in the article. It is difficult to solder to and requires a high-wattage iron, as the case acts as a heatsink. If you do get it hot enough to solder to, the internals of the transistor may be damaged from the heat, as the temperature required for soldering is considerably more than the transistor would be subject to in normal operation.
The correct way to connect a wire to the transistor case is to solder a small solder tag (available from electronics stores) to the wire and bolt the tag onto the transistor case using either of the mounting holes. Use a shakeproof washer under the nut so the bolt stays tight.
T1 must be a Darlington Transistor
Although it is not stated in the text, T1 is a "Darlington" transistor which means it actually contains 2 transistors wired in cascade to increase the current multiplication (gain) factor. Other transistors such as 2N3055, which look the same and have similar current ratings, will not work in this circuit as their gain is way too low. The quoted Dig-Key item is actually a 2N6284 transistor, if you want to try to obtain it from other suppliers.
Electrolytic Capacitor Polarity
The suggested method for identify the polarity based on lead length is not as reliable as checking the outside of the capacitor for the prominent black stripe with a string of "-" (minus signs) printed on it. Not surprisingly, the wire nearest the stripe is the negative wire.
It is a bad practice to base the polarity on the lead length. Although OK for new capacitors, it doesn't work if you take a capacitor from the "junk box" which may have had its leads trimmed previously.
Description of Transistor Operation is Wrong
In the "How it works" section, the transistor is described as being "open" which allows current to flow in the solenoid. While based on a plumbing analogy, this terminology is wrong for an electrical circuit.
Current flows when a switching device such as a transistor is ON (closed circuit) and does not flow when it is OFF (open circuit).
Also, the transistor is incorrectly described as having "3 inputs". This is also wrong. It has 3 connections (or leads). Strictly speaking, only the base is an input and the emitter and collector are outputs.
thanks Richard,
I just got November MR today and was just reading about the timer and had some questions about the circuit. You have answered my questions and provided a better circuit. A new circuit diagram would be helpful, but I can work out what to do from your description.
cheers
Alan
Alan Jones in Sunny Queensland (Oz)
Someone brought this up in the general forum last week and many of the same concerns were expressed. The whole thing is quite a 'backwards' way of accomplishing a time delay. The 'frontwards' way would be to have the push button charge the cap and the coil to switch on while the charge leaked away from it. I don't have a diagram handy but a 555 timer set up for a one-shot would work wonderfully.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
Hello Alan
Glad you found my comments useful.
I use the Kadee undertrack permanent magnet uncouplers and have hinged them so they drop down when not required. I use a small push-pull control knob on the fascia, which is intended for model aircraft use, but it gives a neat result. The electromagnetic ones are pretty ugly and require a hefty power supply.
PS I am in Perth so just around the corner, so to speak.
Thanks Randy
I didn't see the earlier post. With the alterations I suggested, it would work the "frontwards" way you describe.
A 555 timer would be a more elegant solution but would require quite a few extra components. The intent of the MR circuit is not bad in that it does the job, just, with a minimum of components.
richardstallard Thanks Randy I didn't see the earlier post. With the alterations I suggested, it would work the "frontwards" way you describe. A 555 timer would be a more elegant solution but would require quite a few extra components. The intent of the MR circuit is not bad in that it does the job, just, with a minimum of components.
I bolded the area of your comment I think is important as to why it was proposed. So many today are used to buying something already made vs making or modeling. Many times I get more info from MRR mags of the 70s to 90s than today, especially with electronics. I appreciate your post very much and just hope more become comfortable with building electronic assemblies which can help their layouts.
BTW, I'm just now trying to figure out the best design for a speedometer with display that doesn't have to interface with JMRI/Decoder Pro. Any suggestions on where I could find a good one?
Richard
Hi Richard,
I too have some hinged magnets, but bought an electromagnetic one when the Auzzie dollar was high. A 3 amp power supply is pretty hefty and could easily be a source of enough heat to start a fire so I'm going to try your circuit.
BTW I think you and I are further apart than any of our mainland US model railroaders