Atlas Relay not reliably throwing

[Maletrain]

Another Rube Goldberg type installation at my club needs work:

This time, it is a Digitrax DS64 that is throwing an Atlas snap switch (turnout) plus an Atlas snap relay (electrical relay) on the same output.  The turnout always throws.  The relay, which is the same solenoid setup, but with electrical contacts instead of a point lever, sometimes throws and sometimes not.  The DS64 manual says it should throw up to 3 Atlas turnouts on one output.

There is a slider on the relay, just like on the turnout mechanism, which can be used to manually change the position of the solenoid.  That tends to only make it part way when it doesn't work correctly.  And it only fails in one direction, not both.

The relay is located much farther away from the DS64 than the turnout, and a previous member had doubled the wires to the relay, apparently in an attempt to reduce the resistance relative to the turnout.

I tried increasing the duration of the electrical pulse from the DS64, but saw no change in the relay behavior going from 0.2 to 3.0 seconds.  That pulse comes from a capacitor discharging, supposedly starting from 20 volts.  No apparent effect on the relay performance.

So, I tried moving the relay to be near the turnout.  No better performance from the relay.

I tired a new relay (same type) and got the same unreliable performance.

Rather then buy and program a new DS(now)74 to see if that helps, I am thinking that I should try a better type relay, first.  All the output of that relay has to do is ground or not ground the wires going to a servo controller (Tam Valley Octopus III) that throws a servo turnout at the other end of this track segment.

So, any suggestions for a Mouser or Digikey item that would serve the purpose of this relay?

Edit:  I should have clarified that I need a latching type relay.  It just needs to mimic a panel switch that is either on or off, to control the input to a servo driver that is set to NOT use momentary control switches.

[peteski]

Maybe the switch machine's solenoid has for some reason a lower resistance (maybe some windings are shorted?) than the snap relay, so it "sucks" more juice out of that capacitive discharge unit?  Have you measure the resistances of both device's solenoids?

[Maletrain]

Good thought, Peteski.

I didn't measure the resistances.  I should for both coils in each.  But I won't get the chance until next Wednesday.

By then, I am hoping to have a better relay to try.

The thing that confuses me is that the failures are somewhat random.  But they  tend to come in groups.  Won't work for a while, then works for a while, then doesn't again.  And it doesn't seem to be related to how long the cap in the DS64 has had time to recharge.

[peteski]

I also assume that with the switch machine disconnected the relay will throw reliably (even when it its original distant location).

Randomness might be due to the problem being marginal, where sometimes the relay will throw.  There is mechanical friction involved in the solenoid's core moving through the solenoids.

[nkalanaga]

Maybe try a micro switch tripped by the "slider" on the turnout motor?  No extra power needed there.

[Point353]

How is the DS64 being powered - from the track or from an external power supply?

[Maletrain]

The DS64 is powered by its own PS14.
 
I did not disconnect the turnout to see if that affects the relay performance. 

I will spend some more time under the layout next Wednesday, assuming my back is fully recovered from last Wednesday by then.

Thanks for the suggestions of things to look into.

But how about some suggested alternative relays?

[Rasputen]

I wouldn't think you need any type of capacitive discharge power supply if you are using a DS64.  I would try just a 16V AC power source, and see if that works.  With a constant power supply, lengthening the pulse should then help the performance.

[Maletrain]

The DS64 has an internal capacitor and uses that for coil type switch machines.  It can also be set for stall motor machines, for which it does not use the capacitor.  It doesn't see to have a choice for just putting a straight 14 volts on the wires for a specific amount of time as set by Op Switches.

This morning, I measured the resistance of some salvaged Atlas snap switch machines in my parts stash.  There were 2 different types.  The ones that are the same type as the one installed on the club layout, as well as the relay there, measured about 5.5 ohms +/- 0.2 ohms, for both coils. They throw fine with 12VDC.

The older versions had much smaller manual tabs that moved over much shorter distances when they were thrown.  When I measured the coil resistances on those, I got very high readings - a few Kohms to hundreds of Kohms.  I suspected that they were dead, but, when I tested them, they both threw in each direction.

Next Wednesday, I intend to measure the resistance of the coils on the installed turnout and the old and new relays at the club.  If one coil of the turnout machine is much lower in resistance than the other, I will try replacing it with one from my stash.  I will also check to see if the relay throws reliably with the turnout disconnected.

Not much more I can do without getting to the club.  Hoping the scheduled service appointment for my home next Wednesday "morning" does not "go long" and keep me from getting to the club.

[Rasputen]

Sorry I mis-interpreted your original post!  You should be able to use the Digi-key search filters to find a nice Omron or Panasonic relay.

[peteski]

Steve, sounds like you did some research and have several options to try at the club.

I'm also o surprised that some of the older working solenoids have few kilo-ohms of resistance.  That would take a lot of winding with extremely fine wire. I also don't think that few hundred kilo-ohms was an accurate reading. Even few kilo-ohms seems too high. Or there is something else in the circuit affecting the reading.

In order to produce magnetic flux strong enough to move the solenoid's core it needs a lot of energy. That is current multiplied by voltage.  There are no shortcuts.  Low resistance will allow ample current to flow at low voltage to produce the flux.  With high resistance at the same voltage the current will be lower (Ohm's Law). There will be less "wattage" consumed to produce strong enough magnetic flux.  With 100k resistance the current flowing at the same voltage will be minimal.

[mmagliaro]

I am really leaning toward the idea that the two coils have disparate resistances (i.e. the switch being much lower than the relay.).  In a case like that, no matter how big the discharge cap is, and no matter how high the supply voltage is, the lower resistance coil is going to steal all the current.

It intermittently works, which sounds like the two coils are not too different, or it would never work.
Like, maybe the switch machine is 5.5 ohm, and the relay coil is something like 10.  If that's the case, a cheap quick fix might be to just wire a 5 or 10 watt power resistor of a few ohms in series with the switch machine to even them out. (I say cheap and "quick", because I have power resistors like that just lying around, heh).

[peteski]

Or just replace the switch machine at the turnout  . . .

[Point353]

Is there a spare/extra output on the DS64 that could be used to drive the snap relay on its own (and simultaneously with the switch machine)?

[Maletrain]

Point353, No spare outputs on the DS64.

Peteski, I agree that there is something wrong with the resistance readings on the older model switch solenoids in my stash.  Those machines were buried in "scenery" and there is still some stuck to them.  And the readings were pretty unsteady, even though I had cut and stripped the feeder wires to get new shiny copper to touch with my multimeter probes.

Let's not get hung-up on those old machines, because the ones in my stash that are like the ones in question at the club are giving pretty steady and consistent readings on my meter.