The rivets will loosen over time.

I had some Atlas "Snap Switches" that had the points loosen while I was trying to install them on a layout!  Their "Custom Line" of switches were much better, but I think any rivet designed to let the points move will loosen up in time.  Just look at how thin the metal in the rivets is.

These are Custom Line,

These are Custom Line, thankfully. I guess I'm not surprised that they get looser, but you'd think that even so, if the problem is that they're loose, they'd make contact intermittently. These have basically stopped conducting electricity at all.

Same Issue

I ultimately switched to Peco just for this frustration. Not interested in discussing turnouts though.

I would though really like to know how you did the jumpers and some photos might help too if you can.

I suspect this might work on Walthers turnouts also. Have a bunch of those sitting in a drawer.

Jack

poor conductivity

It's probably a combination of wear, corrosion or oxidization, and fine gunk build up over time. Soldering jumper wires from adjacent stock rail or closure rail seems like the most permanent fix.For a short term relief you might try a contact cleaner like they used on tv tuners or slot cars in the old days( if you can still buy the stuff)......DaveB

what I did

Diagnosis method: use a voltmeter, and check rail-to-rail voltage at the base of the turnout, meaning the stock rails at the extreme point side of the turnout. This nearly always checks out, on our DCC layout about 14.5v AC. Then check the same between the diverging closure rail and the diverging stock rail. Similarly, check the straight stock rail and the straight closure rail. On degraded turnouts, one or both of these often turns up basically 0-2v AC. I also check on the far side of the frog, ie from the diverging stock rail to the diverging frog point rail and the straight stock rail and straight frog point rail. You'll expect this, because the whole reason you're doing this probably is that locomotives, even large ones with lots of contact points, stall on the turnout. Inevitably, if the closure rails turned up dead, you can switch your meter to a continuity and you will find no continuity from, say, the diverging closure rail to the diverging point. These are the two items that are joined by the infamous rivet. Similarly for the straight closure rail and straight point. Usually I've found that if one joint is dead, they're both dead, but occasionally only one has failed. (I suspect that in these cases, the other one is about to fail anyway, so I fix both rails while I've got the tools out.)

Fix method: If this pans out as I've described, you can solder jumpers to redress the problem. You could directly attack the literal failure by soldering a jumper from the point to its associated closure rail. The jumper goes from one side of the rivet to the other side. If you're running traditional DC and need to maintain power routing, this is probably your best bet. Or, at least in a DCC environment, you can go from any rail that has the same polarity to the isolated rail. For example, you can go from the diverging stock rail to the straight closure rail. They have to be the same polarity since they don't short when the switch is set for straight through and they are literally in contact with each other.

It's been pointed out to me that brand new Atlas Custom Line Mark IVs (which are as scarce as hen's teeth at the moment) have these "jumpers" built in from the factory, on the bottom of the rails. But this doesn't help my ancient Mark Is and Mark II's. My guess is that most newer "DCC friendly" turnouts have these jumpers too, but I don't know as I have switched to building FastTracks for myself. I build them with a solid piece of rail serving the function of both closure rail and point, so there is no joint to degrade.

two bad ideas.

The only other way I can think of would be to replace the tie next to the rivet with a PCB tie and solder that tie to the stock rail and the point rail (nailing down the end that formerly rotated on the rivet). Might not work if the point rail isn't long enough or your turnout throwing mechanism isn't robust enough to bend the point rail. Probably harder to do than soldering a jumper, unless you pull the turnout, in which case you'll be wanting to replace it with a full fast tracks turnout. 

So you asked for another way, and that's one. I doubt it's a better way though. 

If you want to make the jumpers a little less visible you can run them from stock rail to point rail but underneath the table instead of along the rail or between the ties. That's another idea that's probably not as good as what you did. I figure if the looks of the jumper are bugging you that much you'll be replacing the turnout with a fast tracks one when and as you get them built. So the jumper is just a band aid.

You are definitely in my

You are definitely in my head...

Most of these problematic ones are either ballasted firmly in place or in extremely inaccessible locations, or in one case, both. The inaccessible ones of course we don't care what it looks like. And yeah, if I'm gonna pull them up, I probably won't be putting the old, failing one back down, although my jigs are all code 83 and most of this is Atlas code 100. Replacing them with FT not only means a boatload of code-83-to-100 rail joiners, it also means more mixing rail types. Unfortunately ME/Atlas is possibly the least reliable combination due to the relatively wide disparity between the rail profiles. We've already got way too many operational problems, so this seems like a mixed blessing.

I did run the jumpers, feeder-like, from the side of the rail, under the roadbed, and back up to the other side. I've been trying to put them on the non-aisle side of the rail, but as noted, some of the locations are extremely hard to work on so I punted on them

Frankly, this layout has so much accumulated technical debt that I'm pushing to tear down and build new. The biggest reason is that a full 40% of the track is hidden (!!!) and some of it is astonishingly inaccessible and unmaintainable, which is how we get to a discussion like this one in the first place.

I was wondering if something like Deoxit Gold might break down some of the problem. But I suppose that even if that does fix the oxidation problem, it probably doesn't address the wear factor.

Powering points

I’ve been bitten by flaky point power in the past, so I started soldering point jumpers on all my turnouts a couple layouts ago.  Doing it before laying the track is easier, but it can be done after too.  I used a resistance soldering tool, but I think you could do just as well with a regular iron and low temp solder.  And probably with a regular iron and regular solder if you practice a bit beforehand to get the timing and technique right.

O scale point jumpers:  https://rices-rails.blogspot.com/2011/04/point-jumpers.html

N scale point jumpers:  https://rices-rails.blogspot.com/2015/01/point-jumpers-and-track-feeders.html

Lance Mindheim’s approach:  https://lancemindheim.com/about-us/me-turnouts/

On the little test layout I’m currently building I’m going to do it Lance’s way with low temp solder and see how that works out.

Rivets....

I am not familiar with the Atlas riveted point rails, I looked one up on ebay  to see were the rivets are. Just throwing stuff at the wall here....no way to possibly solder part of, or the whole rivet? Then there is the pivot issue?  Any way to solder a jumper wire  at the very inside corner of the tab, coming up through the road bed? The longer it is, the more flexible.  In my situation, I got to solder jumpers on all my turnouts before they were installed. I ended up going under the turnout.

Went pretty fast once I got the technique down. What was the PIA,  digging a hole in the road bed to accommodate the wires. I'm going to do a new layout, but this time I'm going with the solder on the side of the rail jumpers.

Woops, that's a bit fuzzy, here's the other side.

 For me, by the time weathering is applied, ballast is put down, a switch is a crowded place anyway, I'm good with the three foot rule on what your going to see.  I'm using 24 gauge appliance wire for the jumpers. I think the longer the better for flexibility, but still, you don't want an eye sore chunk of wire. As far as throw goes, probably throws a little stiffer than no jumpers, but I sure had a hard time discerning any difference between jumpered  or not.           (Shinohara switches)

  "I think the longer the

I think if I had to do it that I'd use a drill bit the size of the space between ties then pick a spot between the point rail and it's adjacent stock rail and drill down thru the roadbed. That would let me push a jumper wire down thru the benchwork then bend it in a U and back up so I could solder one end to the  base of each rail. The U would be hidden but allow flexibility for the points to throw.  There would be a hole left between the ties but maybe a piece of  foam "weeds" could hide it? .....DaveB 

Powering Points

I've used Lance Mindheim's approach on my Shinohara switches.  I also add jumpers to the stock rails.  Admittedly it's a pain, but I think it significantly adds to the electrical dependability.

yeah, in advance is clearly better

I am not going to bother on most of my Fast Tracks turnouts. They're solid points so there is no joint to wear out. If the rail breaks, I know what to do anyway! But I'll be making some three-ways that need hinged joints, so I'm going to solder in the jumpers before there are even any ties to get in the way.

Unfortunately most of these at the museum are ancient history and obviously nobody looked into the future on this issue... so I guess I'll be doing more jumpers.

Non-conductive Points

The switch machine may be the problem and the answer to the problem.  I use the Hankscraft switch motor which applies constant pressure on the point against the stock rail.  I have had no problems and some of the switches have been down for 15 or more years.  These are handlaid O scale switches with rail joiners as hinges but the theory is the same regardless of the scale.

Hope this helps.

Ed

@DaveB

I'm seeing were your at,  a bond wire that is "hanging down".   After soldering all of my switches, it's surprising how short the wire could be and still let the switch operate freely. In the first pic of the under side jumpers, those are as short as you can get.  They are soldered from end of joiner to end of joiner. (with out soldering the joiner of course) I never had a problem with the operation of these switches in track, but, I did solder some jumpers that were 4 ties wide  vs the 3 ties wide, and they were easier to operate. (I extended the jumper towards the frog, to keep the point rail part of the jumper as close to the heel as possible, the heel moves the least in the switch operation.)                                                                                                                      Taking your idea in a different direction, how about bringing a feeder wire up through the road bed at the heel of the switch, and soldering there. I think even just that little bit of wire exposed from roadbed to the rail is enough flex.  Hardly any more work, If you drilled a slightly over size hole when drilling for the wire, giving more room for the wire to move,  that could give a little more insurance to avoiding a problem. The scale of all of this is pretty small, it's only the point of attachment at the heel of the rail that has to move, and not very much at that.

turnout conductivity issues

I had this problem. I found that occasionally a combination of dribbling white vinegar to contact and rivet areas , and re-staking the rivet usually helped restore performance. Simple maintenance is better than completely ignoring it until it doesn't work!

  "how about bringing a

Hi Jerry,  Yeah I think the closer to the heel of the points the better it would be, less sideways movement of the wire. I'd just  stay just far enough away from the heel to not solder the point hinge closed. .....DaveB

Hidden trackage

Actually, hidden trackage needs to be the most reliable trackage on your layout and if it’s not, you’ll be sorry later, as you point out. Every scrap of rail in hidden trackage needs to be powered. On turnouts this often means points with jumper wires and feeders or jumpers to the closure rails, always. Relying on just stock rail contact to power the points / closure rails is a strategy for failure.

But then I’m preaching to the choir, you clearly know this now, as your comments indicate. But for anyone who is contemplating a layout with hidden track, these preventative measures are a must if the track will be hidden and you don’t want to be sorry later.

just far enough...

 Right!  You have hit the nail square on the head!