Ok, I've been (finally) getting around to pulling balky switches out of my layout to see what's up.
Before pulling the layout apart, I followed Bob Nelson's advice and made sure I had 'good' (14 AWG) power and common using 16V from a Lionel 1033 going to each switch.
The first two that were still not switching crisply seemed to only need all sliding surfaces cleaned and lubed with Labeles grease (an 022 and a 711). My next switch with sluggish switching (a 711) did not improve sufficiently with the same lubrication. All wires seem to be intact and original. I don't have a meter to check conductivity. I did try applying 20V, which produced crisp switching.
I'm not anxious to up all my switch power to 20V. What should I do?
runtime
I cleaned off the contacts about 3 times and finally decided there was something wrong with the contacts. I grabed the post that moves the poings and moved it to one side. The sliding contact springs came up off the fixed contacts. It was clear that the spring pressure was insufficient. I took a paring knife and slid it under the springs near the center and lifted the spring up and bend the end down. This increased the spring pressure and now the switch is working very well. Just one more thing to check if you are restoring some 022 or 711 switches. I used a ScotchBrite pad to gently clean the contacts.
Bruce Baker
For enquiring minds, here's how the diode trick that Bruce advised works: The root-mean-square (RMS) voltage is what counts for powering things like incandescent lamps; so it's the standard way of describing AC voltages. The RMS voltage is the square-root of the average of the square of the voltage waveform, which in the usual case is a sinewave. The square of a sinewave looks like a series of positive pulses coming at twice the frequency of the original sinewave, that is, 120 instead of 60 hertz. When you put a diode in series with the lamp, the lamp sees only every other one of these pulses; so the average of the pulsed waveform is cut in half. Since the square-root of one half is about .7, the diode reduces the RMS voltage by about 30 percent. So 20 volts becomes 14 volts.
Bob Nelson
Yes, the diodes are to make the bulbs run cooler. The bulbs in the controller don't get as hot as the bulb in the switch motor since they are in series with the coils in the switch motor and they are in the open air, but the diodes should make them last a lot longer and run much cooler.
And, yes, alternate the diodes. I do this at random, not paying attention to which way I put the diodes in.
Alternate your diode polarities from controller to controller.
Rob
I modified some 022 switch controllers to add a diode in series with the light bulbs. I used only one diode. I cut the metal strap on the bottom in the middle using a Dremel tool and a cut off wheel. The I soldered a diode across the gap. I had to clean off the metal with a sanding disc in the Dremel tool so the solder would stick. I also moved the center wire to the stand where the lever is mounted.
BTW, I am running the switches using the 20V fixed voltage on a KW.
B
I have some test data on an 022 switch that I just opened up because it was intermittantly failing to throw properly. The sliding contacts had some carbon on them which came from the WD-40 I left on them after I cleaned them. I cleaned this off and reassembled the switch, and it still failed once in 30 tries. The train throws this switch every minute while it is running.
What I did not find is any erosion of the contacts. Some time ago, there was a post about erosion of the contacts on the newer versions of the 022 swiches. I am happy to report this is not a poblem on the PW switches.
I cured most of the intermittant problems I had with the 022 switches by putting cardboard under the track section that approaches the switch. Remember, I am running on the carpet, and since the switches are flat on the bottom, they don't sink into the carpet like the track sections do. Apparently the engine lead truck was bouncing when it first got onto the switch, and this was intermittantly causing the switch not to throw completely. I first discovered this problem with a 2046 Hudson.
You are both welcome.
You could discretely mark the turnouts with the polarity, e.g., + if the cathode is toward the lamp, - otherwise.
If you are using 1445 lamps, I suggest swapping them for the cooler 53s. That's what these are, although Radio Shack doesn't identify them by number: http://www.radioshack.com/product/index.jsp?productId=2103782
Use absolutely only one per turnout. You can install it either way; but, if you're modifying several turnouts that will be powered by the same transformer, do half of them one way, half the other. That way, you will minimize the DC component of the current that is drawn from the transformer, which will avoid core saturation and allow it to run cooler.
Hey runtime,
I was wondering when you would come back on the radar since you started this thread. I had been wanting to post all the modifications I was doing to the 022 switches so everyone could have the benefit of this knowledge. Now I would like to take some pictures to add to the post, but that will take a little while. Bob Nelson has added a lot of valuable information to this post, also.
I have a small magnetic dish that I got at Harbor Freight for holding the screws and other small parts. It is a big help.
Harbor Freight also has a digital multi meter for $8 that is plenty good enough for working on trains. If you are trying to debug an electrical problem, a meter is a big help. And learn how to use it and learn what E=I*R means. Or V=I*R. (Ohm's Law)
You can buy drills and taps at Harbor Freight that are plenty good enough for working on trains, and they are relatively cheap.
RockIsland,
Don't look at the switch repair as a complex problem. It isn't. It is a number of simple steps taken one at at time. There is very little you can do to mess up the switch as long as you know how to solder. The only thing that comes to mind that you can mess up is putting the screws in the wrong holes. The screws that attach the motor to the switch are always 4-36, except in one switch I have, which is very old, and it has 8-32 flat head screws. I also found some newer versions of the 022 switches have metric screws. These are very small diameter and very fine thread. The screws that attach the plastic motor cover can be 4-36, 4-48, or self tapping screws. Don't mix up the self tapping screws and try to use them where there should be a 4-36. When you reinstall self tapping screws into the plastic covers, make sure the screw finds the original thread. If you don't do this, you will most likely lose enough plastic that the screw may not hold. This goes also for the screws that hold the bottom cover on the switch. Don't over tighten the screws as you may break them in half, and then you will be talking to yourself. Oil the two screws that attach the motor to the switch, unless you never expect to take them out again. I have had one of them seize up due to corrosion.
The 4-36 screws are no longer standard. During WWII, the standards for fasteners were revised, and 4-40 and 4-48 were designated as the standard. 4-36 screws are still available on the Internet, and Skycraft in Winter Park Florida has a large bin full of them. Skycraft charges about 20 cents a dozen, but I don't know if they will ship them. 4-48 screws are hard to find also.
Bruce......what I want to know is: who was the knucklehead who said toy trains aren't brain surgery. This stuff makes brain surgery sound simple and quick. I agree with runtime. Holy smoke!
Jack
IF IT WON'T COME LOOSE BY TAPPING ON IT, DON'T TRY TO FORCE IT. USE A BIGGER HAMMER.
While I am thinking about it, here are some more things to do to the 022 switches while you have them apart. Reflow the solder on the two tabs on the sliding contacts that mate with the pins on the switch itself. I could not get a reliable contact without doing this. You have the soldering iron out anyway, and this only adds about 30 seconds to the task. I found it works even better if you clean off the new solder with a wire wheel in a Dremel tool.
The 042 manual switches are a good source of parts. I have had to replace the switch points on 3 switches, and the 042s were used for parts. The screws, nuts, switch points, bottom cover are the same as the 022 switches. The switch points are die cast zinc, and sometimes they go bad. You should be able to buy these very cheaply at a swap meet.
Check the pins. If the pins are rusty, pull them out and clean them on a wire wheel. If they are loose, tighten the rail opening before you re-inset the pin. Check for 0 ohms between the pin and the rail.
I have been pulling out the steel strip that goes through the frog. It makes the train run smoother through the switch. Some of them have been bent and cause the train to derail. They were apparently installed to reduce wear on the frog, but after examining many switches, I never saw enough wear to worry about.
I just fixed a switch with a broken lantern holder. I have a change to the procedure to replace a lantern holder that I posted earlier. When you are drilling the pilot hole in the rivet for the 4-40 screw, don't drill it all the way through until you have removed the rivet from the switch motor. If the drill goes all the way through, you will damage the wire for the lamp. After the pilot hole is part way through the rivet, remove the rivet and drill it through and then tap it.
Wow! Just checked back in on this thread that I started.
No wonder I love this forum!!
But...boy do I have a lot of work ahead of me.....oh well, it's a hobby isn't it.
The resistor will get hot, because whatever fraction of your transformer voltage is not across the lamp will be across the resistor. For example, if you size your resistor to reduce the lamp voltage by 30 percent, 70 percent will be across the lamp and that 30 percent will be across the resistor. They will both carry the same current; so the power dissipated in the resistor will be 3/7 of the power dissipated in the lamp.
But, if you put a single 1N4001 (not a zener diode!) in series with your lamp instead of a resistor, the diode will act like a switch. It will turn the lamp on for one half-cycle of the AC voltage, 1/120 of a second, then off for the next half-cycle, then on, then off, and so on. During the half-cycle when the lamp is off, there will be no current flowing through the diode, so it won't dissipate any heat. During the half-cycle when the lamp is on, current will flow, but there will be (almost) no voltage across the diode, so again it won't dissipate any (significant) heat. So the diode stays cool. Having its voltage turned on for only half the time has the effect of reducing the voltage at the lamp by 30 percent, so the lamp lights to the same brightness and puts out the same power as it would have in the first example above, using the resistor; but there is no resistor to get hot, just a cool diode.
You can find 1N4001s at Radio Shack: http://www.radioshack.com/product/index.jsp?productId=2062589 . The higher-numbered diodes are the same, just with higher voltage ratings than the 50-volt 1N4001.
I recommend that you not use a number-1445, if that is what you were planning. As I tried to show above, the number 53 is cooler at any voltage.
Unsolder the wire on the solder tab that is under the constant voltage pin. I recommend that you solder the tab to the constant voltage pin while you are adding the diode. I have found several of the constant voltage pins not making a good connection to the solder tab. I polish the constant voltage pin and the tab with a small wire wheel on a Dremel tool. This makes the solder flow better. When you do this, you will have a nice ball of solder to use to attach the diode. Cut the diode leads until they are about 1/4 inch long. Then hold the diode by one lead with a pair of needle nose pliers, and solder the other lead into this solder ball. Then add a bit of solder to the free lead and solder the wire to the lamp to it.
While you have the cover off, you might want to solder the tabs which are on the outer two posts to the posts. I have found these to be high resistance on some switches. If you have a meter, you can check them with the meter to see if they are OK.
Also check the wiring and solder. Some of the switches I have had very poor solder connections. I found one that did not have the wire to the lamp soldered to the lug at the constant voltage pin. I have found about 4 of them with the coil wires going the wrong place. The two wires at the center of the coils should go to the constant voltage pin and not to the spring that comes from the center rail. If you find one like this, check my post as to how to fix it.
This is all a lot of work but well worth it. I am up to 67 022 switches now, and it takes 2-3 hours to fix each one. Last night I had an engine running around part of the layout that has 24 switches, all of which are thrown by the train as it goes through the switch. I ran the train for over an hour, and never had a switch fail to work properly.
A 1N4001 (up to 1N4007) is suitable and easy to find. A diode will generate very little heat, whereas a resistor will. Put one diode in series with the lamp. Eighteen volts will be reduced to about 12 or 13 volts (but don't try to measure it--most voltmeters will give you the wrong number with this circuit).
wow....you know some stuff about these switches!!! I am kinda new at all this so when you talk about the different parts of the switch i am totally lost. Is there a book out there that explains this stuff with pictures? I would love to try and resurrect an old switch but reading all this seems very intimidating...:(
The 18-volt lamps that often get used in attempts to reduce lamp power are 1445s. The number 53 lamp runs cooler than the 1445 at any voltage.
Lamp ratings are a tradeoff among voltage, current, efficacy, and lifetime. The ratings of the two lamps are
#53 14.4 V 120 mA 1728 mW 1 mscp 1000 hr #53 18 V 136 mA 2442 mW 2.18 mscp 69 hr#1445 14.4 V 135 mA 1944 mW .7 mscp 2000 hr#1445 18 V 150 mA 2700 mW 1.53 mscp 137 hr
The numbers in italics are values that I calculated using the rules that for incandescent lamps current varies as the .55 power of voltage, light output as the 3.5 power, and lifetime as the -12 power. The 1445 is actually rated at both 14.4 volts and 18 volts; but that doesn't give it any particular advantage over the 53 in terms of power dissipated at either voltage, as you can see.
For screw-based lamps, the number 52 is cooler than the 1447:
#52 14.4 V 100 mA 1440 mW .75 mscp 1000 hr #52 18 V 113 mA 2035 mW 1.64 mscp 69 hr#1447 14.4 V 135 mA 1944 mW .7 mscp 2000 hr#1447 18 V 150 mA 2700 mW 1.53 mscp 137 hr
18 volt bulbs are going to run pretty hot on 20 volts. That is why I put a diode in series with the lamps on my 022 switches. It cuts the power significantly, and the bulbs are cool enough to touch without leaving skin on the bulb. I like the high voltage to make the switches operate reliably, but I didn't like the melted lanterns. Hence the diodes. My earlier posts tell you where to put the diodes.
Our community is FREE to join. To participate you must either login or register for an account.
Get the Classic Toy Trains newsletter delivered to your inbox twice a month