Atlas Relay not reliably throwing

[Maletrain]

Results of more diagnostic testing last Wednesday:

With turnout detached, the relay throws reliably.

Both of the turnout coils measure 5.5 ohms.  On the new and old relay, both coils measure 3.3 ohms.  So no indication of bad coils anywhere, and turnout resistance is higher than relay resistance.

The new relay still fails to throw completely in the same direction as the old relay.  But, it seems to make the electrical connection more reliably, even though it doesn't throw substantially farther.

Tried unscrewing the relay from the layout to see if it was bending slightly and mechanically binding - didn't matter.

So, at this point, I am questioning the DS64.  Maybe its capacitor isn't up to spec any more?  There are some other switches that I think it also throws, but they involve a diode matrix.  I remember them having some difficulties, but we never throw them. 

One possible work around is to use the connection on the other end of the relay, because I only need to make and break one circuit.  That would require me to swap closed for thrown on the servo controller.  But, it would not help the other turnouts on that diode matrix.

[peteski]

Diode matrix?  That was not mentioned earlier.  Does the DS64 throw multiple switches through a diode matrix?

Any chance you could temporarily hook up the switch machine and relay combo to another output of the DS64?  Do the other outputs drive similar setups, or just switch machines?

[Maletrain]

There is no diode matrix on the DS64 channel with the turnout plus relay. 

I am just thinking about what is on the other outputs, and I think this DS64 is also connected to a diode matrix for 6 turnouts that create a set of passing sidings under some non-removeable scenery.  But wouldn't that matrix be on the inputs?  Haven't thought about that except that it is another thing I need to trouble shoot because it hasn't been fully reliable. 

If the DS64 has a weak cap, that could explain problems on the other outputs, too.  They throw multiple turnouts, too, and not always reliably.

[peteski]

I'm still confused, especially when you mentioned that they are on an "input".  I am not very familiar with DS64. Are youstating thet the diode mirtix is used in conjunction with manual pusbutton switches which will direct the DS64 to throw a series of switches it controls?  If that is the case then you are correct - the diode matrix is irrelevant.

What I'm thinking is not that the decoder's power supply is weak, but the transistors on the output channel of the troublesome switch machine/relay are failing (getting weak in layman's term), so they cannot provide the voltage and current needed to throw both solenoids.  That is why trying another output of the DS64 might be helpful. Or just replace the DS64 of there is a spare one available.

[Maletrain]

I really don't know much about the diode matrix at this point, other than it is associated with at least some of the other turnouts on the same DS64, which I expect will complicate any direct comparison with the output for this relay + turnout combination that is the subject of this thread.  I will have to visit those turnouts too, but later, because they aren't perfectly reliable, either.

Because the DS64 needs to be programmed if I replace it, I would need to completely familiarize myself with how the Rube-Goldberg-prone previous member set everything up. 

So, my next step is to see if I can get the relay to reliably actuate the input to the Tam Valley Octopus III servo controller by simply changing which contact is used in the servo.   To do that, all I need to do is swap the red and white input wires on the relay and move the output wire to the other contact (now empty).  At least that is all I think I need to do until I try it. 

Next Wednsday.

[Maletrain]

After missing a week at the club due to a badly sprained knee, I limped in yesterday and tried swapping the leads to the relay to see if it still did not achieve full travel in the same direction, or if the partial travel depended on the inputs. 

I was pleased to find that the relay refused to fully travel in the same direction, independently of which orientation the inputs from the DS64 were made.  That tells me that the Atlas snap relays have some sort of mechanical resistance that is higher for the slug moving toward the end away from the input connectors than it is for the direction toward the input connectors.  It was the same for both the original and new relays, although the final resting position for the new relay was a bit farther along the path than for the old relay.

So, my immediate solution to this particular problem was to just also switch the output leads so that the contact I want closed is the one where the solenoid slug achieves full travel.

It may still be that the DS64 isn't fully up to snuff, too.  But, that will be a task for another day, when I look at what is happening with the other outputs from that DS64.