Saturday, August 18, 2018

The Great Sawzall Massacre of 2018

For a variety of reasons, most importantly to shorten my commute and get some free time back, I recently moved.  I have a nice new space designated for an industrial switching shelf style layout.  I'm still in the toying with ideas stage, serious planning will probably start in a month or so and, with luck, a bit of construction before the end of the year.  More on that in a coming post.

Selling the old house is of course a rather critical part of the proceedings, and having a train layout screwed to the basement walls is not a good way to try to sell a house.  So it had to go.  My weapon of choice was a sawzall type saw (porter cable tiger saw).  With that and a crowbar, pry bar. cordless screwdriver, and some coffee it took me most of a weekend to separate the layout from the house and reduce it to small enough pieces for a junk removal service to remove.

When I built the layout I had the attitude that it was going to be a lifetime layout, or at the very least not reusable.  So I built it from the walls in - I started by screwing the back frame member to the wall, then added the cross pieces, front frame, and braces.  Some sections were built during my massive overkill lag screw phase, some in my drywall screw phase, and some in my pocket screw phase.  Due to both the build in from the walls construction and the number of changes over time, there were screws that were buried under other framing pieces, fascia, etc.  In retrospect, that wasn't the best idea.  The sawzall and crowbar were required.

First step, remove the Palmer Industrial peninsula to free up floor space.  The trickiest part was excavating the head of one screw that held the frame to the door frame.  The fascia was vigorously glued over it, and I needed to leave the door frame in good shape.  I couldn't just pull it out, or try to cut the screw between the layout and the door frame because that would mess up the door frame.  A bit of prying and chiseling eventually revealed the screw.  Here's the Palmer Industrial after being separated from walls and floor.



I used the sawzall to cut it up into manageable pieces and stacked them in the corner.



By the time I got to this dismantling operation I had already started thinking about a new layout, so oddly enough one of the biggest moments of regret was seeing those nicely curved MDF fascia corners in the junk pile.  I had fun making those.

So, with the Palmer Industrial out of the way it was on to the O scale No-Name Industrial.  As I mentioned it was screwed to the walls.  The walls it was screwed to have a 10" or so ledge just under the layout height at the top of the foundation.  It would have been excruciating to try to unscrew the layout as is from the wall.  Cutting off everything except the back few inches seemed like the best approach, so I fired up the sawzall again.  In this photo the corner with the interchange has already been removed except for the back edge screwed to the wall, and I've sliced through the layout for the whole length a few inches from the wall.  All that's holding up in this photo is habit.


One tap on the front edge while standing out of the way in the corner, and the habit is broken.


The remaining bit attached to the wall.  What kind of idiot uses this many lag screws for a train layout!


A bit of backdrop removal, not too hard in spite of painted over screw heads.


Then it was relatively easy to remove the pieces from the wall and the lower end of the brace supports.  Chopping the remains into easily carry-able pieces and stacking them in the corners finished up the front wall portion of the layout.




The side wall portion of the layout was not quite as straightforward to remove.  The portion closest to the corner was built to allow scenery and track extending below the baseline elevation of the rest of the layout.  A sheet of plywood was screwed directly to the ledge in the wall, and then track/scenery was built up from there - in other words lots of buried screw heads.  The remainder of the wall was more normal construction.  My approach was to cut out the normally built portion and then nibble away at the remainder.

Hm, what's holding it up?
Maybe a couple more whacks?
That did it!
Next problem was the oddly built section.  I had built it up by screwing joists of the right height to the plywood underpinnings, then screwing the plywood roadbed to the joists, and then gluing homabed on top.  Lots of screws covered up by other layers.  I ended up just brute force crowbaring off the plywood subroadbed for the main and runaround behind Catania-Spagna - you can see some of the supports split across the middle.


Then I was able to cut it off just inside the ledge.


The weapon of choice.  That's the third or fourth blade - I kept bending the tips hitting things I couldn't see under the layout.


Eventually I managed to find all the screws and get the remaining plywood off the ledge.


The second pile of layout remains when it was all over.


And a few days later it's all gone.



That last bit of "backdrop" is where I "cleverly" painted the back drop right on the door for the bulkhead.

So, what did I learn from all this?

  • Even if you don't want to build a layout that can be moved to a new home, at least build it so it's not too difficult to remove.  Even if you don't have to move, you may want fewer obstacles to re-arranging portions of the layout.
  • Layouts do not need to be lag screwed to the wall!  The section just drywall screwed to the wall held up just as well with me crawling on it as the section lag screwed to the wall.
  • Pocket screws are cool.  But they can be harder to find when you're taking something apart.
  • Sawzall style saws are fun!

Sunday, February 1, 2015

Temporary (?) car routing for the Palmer Industrial

Despite my intention to ignore the problem of car routing and paper work for the time being, it has been hard to avoid getting distracted by it as I continue to experiment with operating on the track that's laid so far.  Today I decided to put a good temporary solution in place to stop the distractions.

For car routing, to start with I'll just make it up as I stage a new session.  I expect once all the track is down and I do some serious test operations I'll begin to get a feel for exactly what works well in terms of number of inbound cars, number of respots, etc.  At that point the next step in car routing will be a some dice to select random entries from a table of reasonable possibilities for each industry.  Perhaps several different sets of tables - one set for short/easy sessions, one set for long/challenging sessions, etc.  I'll work out the details when I get to that point.

For the paper work, the possibilities seemed to be hand writing a switchlist or PICL list (seems like a lot of tedium every session), rig up a spreadsheet of some sort to manually enter the list into and then print (possibly even more tedious), or resort to the tab on car approach.  The tab on car approach seemed like the easiest option, so that's what I went with.

Since all cars on the Palmer Industrial Park are either headed to a particular industry, or back to Palmer (off layout), I can use very simple labels on the tabs.  Maple Leaf, for example, gets tabs labels like "M1-5" for Maple Leaf track 1 door 5.  Plus a few extra tabs just labelled "M" for off spots.  Quaboag has three locations "Q - Dock" (one of the three spots on the loading dock), "Q - Lot A" (the first half of the rest of the track after the dock), and "Q - Lot B" (the second half of the track after the dock).  Cains and Trans Plastics are not spot specific at all, so they just have the appropriate number of tabs labelled "Cains" and "TransPlastics".  Any car without a tab is headed back to Palmer.

I made the tabs by setting up a table with entries for all the tabs in my word processor.   I printed them out on card stock, used a paper cutter with a roller blade to score for folds, perforate between tabs in one direction, and cut in the other direction.  I think it might have been easier overall to perforate in both directions.  Each tab is 1/4" x 1", with the last 1/8" on each end folded down.  The result just sits on top of a flat topped car very nicely.  For the Cains tank car tabs I'm going to see how well they stay on when curved to more or less match the tank curve.

Tabs on cars

A fan or direct sneeze would probably be bad

What it looks like from a normal viewing position
You can see it looks a little odd, but certainly no worse than the complete lack of scenery and weathering.  It should be easy for guest operators to understand too.  It's good enough for now.  It might be good enough for a long time, which is why the title of this blog entry has a question mark in it.

Saturday, January 24, 2015

DCC control

So far I've been doing my test runs on the N scale Palmer Industrial Park with the NCE PowerCab I keep on my workbench for programming and testing.  It works, but it's not a permanent solution - I wanted a radio throttle, and I wanted the PowerCab back on the workbench.  While I wouldn't mind cannibalizing the DCC system from my dormant O scale layout, I wanted something newer and more appropriately sized for N scale.  The O scale layout has a SystemOne command station and an NCE 10 amp booster with a large transformer in an even larger metal case.

The NCE SB5 seems like a pretty ideal package for a small layout - it's got some limitations compared to the top of the line PowerPro CS02, but none of them matter for a layout like mine.  (I can't imagine I'd need more than 8 cabs or run more than 12 trains at once no matter how much I might expand the layout).  For further limiting the 5 amp output, an NCE EB3 (electronic breaker) floating around from a previous project seems perfect. For the radio throttle, reusing the RB02 radio base station from the O scale layout was a no brainer, but the throttle itself wasn't quite so clear.  I have a radio ProCab - it's very nice, but I like the feel of the cab04p better.  I have a cab04pr, but it's getting a bit long in the tooth, and I don't like the idea of pulling the back off all the time to flip dip switches to change the cab address (yes, it's that old) from one of the low numbers supported by the SB5 to the high address available to me for use on friends layouts and vice versa.

Bottom line: I went to my friendly local hobby shop and ordered a new SB5 and a new cab04pr.  They arrived yesterday.  Today I set the new system up.

Top to bottom - power supply, SB5, EB3
The SB5 and the power supply it comes with are mounted to the plywood between the end legs with velcro.  The EB3 is set for the lowest trip current (2.25 amps) and the fastest trip time (10 ms).  The coiled cord plugged into the cab bus that leads upward goes to the RB02 radio base station which is sitting on a box in the middle of the teardrop end on the backdrop.  It all works very nicely.  There was only one thing that irked me a bit - the plug on the power supply that came with the SB5 was a little too big around to let the track bus connector that also came with the SB5 plug in properly.  I shaved about 1/16" off one side of the power plug, which solved the problem.  In the detail crop below the power plug is actually sticking straight in, it just looks like it's angled down because of that 1/16" flat spot.

Shaved 1/16" off side of power plug
It's nice to have a permanently wired up system in place, with a radio throttle.  I think the SB5 was the perfect choice.

The one lingering question is what exactly will I do if I want to start up the O scale layout again?  While it is in theory possible to have two separate NCE radio systems share the same space, it seems like a pain.  I think the simple solution will be to retire the SystemOne command station, and run a control bus lead from the SB5 over to the O scale booster.

Sunday, January 18, 2015

Ground throws, test ops

Progress this weekend was limited to installing ground throws for the track that's already laid.  I spent the rest of the available model railroading time actually operating what there is of the layout so far.  And I enjoyed it.  Goal partly achieved!

In spite of the fact that I'm using Lance Mindheim's book "How To Build A Switching Layout" as a loose guide for this layout, I couldn't quite bring myself to rely on friction to hold the turnout in one position or the other as he recommends.  I like the positive action of the ground throws.

I used the caboose 222S ground throws intended for Atlas code 55 N scale switches.  They go in pretty easily - a little trench in the foam to clear the pins, and fasten them down with white glue and a couple spikes.  I realized after the glue had dried that it would have been smart to paint the blue foam in the trench black before hand.  Perhaps when I paint the track I'll be able to get enough in there to hide the blueness.

Caboose 222S ground throw installed (note trench in foam for pins).

Caboose 222S ground throw
Two things strike me about the second photo above, which I don't find very noticeable just standing in front of the layout (see the operation photo below).  First, that ground throw is huge and ugly!  Second, the proportions are vastly different from my O scale layout.  Below is a photo from my post on O scale ground throws.  When you stand back at a normal viewing difference though, the other aspect of the vast difference between my N scale and O scale layouts becomes apparent - the N scale is spacious with prototypically sized turnouts, the O scale is crowded with obviously tight turnouts (#5).

O scale Atlas O code 148 #5 turnout with caboose 208S cut into head blocks
I had a couple enjoyable test runs this weekend.  With half the track missing operation is obviously limited, but if I treat the last foot or so of the track around the end as if it was Maple Leaf / Trans Plastics with a total of 6 spots, I can get the beginnings of an idea of what full operation will be like.

NECR GP38 3844 Shoving to Maple Leaf / Trans Plastics
There are two ways I'm thinking of operating the layout and I tried them both.  One possibility is treating this as a separate switching RR that interchanges with NECR, using the front track in the above photo as an interchange track.  The other way is more like reality, and probably what I'll go with.  The front track is the spur off the NECR at Barretts.  NECR Train 606 (the Palmer local / Barretts switcher) starts the session by running in on that spur, and ends it by running back out on that spur.  The above photo shows the 3844 shoving 4 cars for Maple Leaf and 2 for Trans Plastics out of the runaround over to the other side of the layout, after having pulled back 6 cars from Maple Leaf.

Saturday, January 17, 2015

Keeping the goal in mind

Now that I actually have half the track down and can do a little switching, many potential side tracks (pun definitely intended) are beckoning.

  • Updating the software and tinkering with the motor control CVs once more in the one loco I have a Zimo MX620 decoder in.
  • Replacing less exotic decoders with Zimo decoders in the two locos I will probably actually use on the layout.
  • Body mounting all couplers
  • Figuring out a way to cram a big enough capacitor into a loco to keep decoder and loco running for just long enough to get past a dirt spot or a contact problem on the loco itself.
  • Finding/writing software to automatically generate the PICL list I'd like to operate from.  (Robert Bowdidge's Switchlist comes tantalizingly close, but so far I haven't got it to generate the traffic I want quite the way I want it)

In the past I've succumbed to temptation and spent, for example, several months of my limited hobby time doing nothing but trying different decoders and configurations.  It was fun in it's own way, but what was the end result?  I had a spectacularly smoothly running O scale switcher I could amaze my friends with, and my layout was no further along.

This time I'm trying to avoid those distractions.  Lance Mindheim has written quite a bit about model railroading (good blog, good books), and one of his recurring points is basically pick your battles - don't waste time on stuff that doesn't get you closer to what you're ultimately trying to achieve.

Time to remind myself what my goal for the N scale Palmer Industrial Park is.

Goal: Reasonably realistic appearance and enjoyable operation.

It's all a matter of compromise.  I think spending time on the trackwork, on things like the point jumpers is definitely on the path to the goal - those are things that are pretty hard to fix after the fact.  Putting the best decoders in the locos now isn't really on the path to the goal - it's trivial to swap out decoders later, it doesn't need to happen now and what I've got is good enough.  Metal wheels are more of a tossup - I've decided to go ahead with putting metal wheels on everything now on the grounds that the guys in the operating groups I'm in all seem to agree that plastic wheels contribute to dirty track problems.  Dirty track problems get in the way of enjoyable operation.  Etc.

So, today I stopped myself from downloading Zimo decoder updates and asked myself what can I do this weekend to make the most progress toward my goal?  Thinking about it like that, it's obvious - order the wheels, and work on the rest of the track.

Wheels ordered, track work in progress.

Sunday, January 11, 2015

Wiring, first test run on the Palmer Industrial Park

Today I ran the bus wire, connected the feeders to it, fixed a screw up on a turnout, and did a bunch of track cleaning and test running.  I'm pretty pleased, and eager to get the track going on the other side of the layout.

The feeders (#22 solid) are connected to the bus wire (#14 solid) with scotchlok 905 connectors.  The scotchloks are insulation displacement connectors, very easy to use and reliable if you choose connectors that are rated for the wire sizes you're using.  The 905 connectors are rated for #14-#18 on the run, and #18-#22 on the tap.  Perfect for N scale, I think.

Scotchlok 905, with #14 bus and #22 feeder

905s installed on the layout
The screw up I discovered after powering up the layout.  The closure rail of one turnout was unpowered.  Presumably the reason is I had had to shift the stock rail a little bit to get the notch far enough along so the points would seat properly.  So I had to solder a feeder on it in place, and it looks pretty ugly.  I don't think it will be too obvious once the track is painted and ballasted though.  Overall I'm shooting for a "good enough" appearance, not proto 160.  I will definitely be more careful to check absolutely everything about a turnout before installation in the future.

Ugly feeder to fix screw up
With the wiring under control, and the track cleaned, it was time for an actual test run!  For the time being I'm using an NCE power cab (I don't want to subject code 55 track and N scale loco wiring to the arc welding power of the 10 amp booster on my NCE system for the O scale layout).  I have an NCE SB5 on order to power the Palmer Industrial.

I want to have only metal wheels on the layout, in part to keep the track cleaner, and in part to get a little more weight on the trucks and help tracking.  Turns out the only cars I have so far with metal wheels are some ExactRail box cars.  I guess I need get going on figuring out what metal wheels to use on all the Atlas cars I have.

The first test run was 5 boxcars and an MP15.

MP15 test run with 5 cars
Looks just the way I hoped going through that #10 turnout.  I like the look of MP15's, but an N scale one just does not have the traction necessary to work this layout.  The next test was more boxcars and a GP40-2.  It became apparent pretty quickly that the trucks on those ExactRail cars are not the most free rolling ones, and at least one of them has some sort of problem that causes it to derail on everything.  Also at least one of the couplers must have some flashing wedged into the coupler box - it's almost immobile.  Some more things for the todo list I guess.

GP40-2 pulls a cut off the interchange track...
Sets them onto the runaround track...

And finally (after running around) starts the shove to Maple Leaf Distribution
The appearance of the cars on the # 10 turnouts and the gentle curve in the runaround is exactly what I was hoping for.  It doesn't even look too bad shoving around the tight curve at the end.  Although those are only 50 foot cars.

All in all, I'm quite pleased with the weekend's progress.  Having got a taste of running now, I'm really looking forward to getting the whole layout operational.

Point Jumpers and Track Feeders

Some time ago I wrote up how I did point jumpers on my Atlas O code 148 turnouts.  For N scale I'm using Atlas N code 55 turnouts.  My rule of thumb for track is simple - if it doesn't have a wire soldered to it, it's going to be unpowered sooner or later.  Turnout points are electrically connected by a contact under the hinge point, and by contact with the stock rail.  They are exposed to dirt, ballast glue, paint, scenery materials - in short they are almost begging to have electrical issues if you don't take measures to make sure they're powered.

My solution for N scale is the same as for O scale - solder a jumper wire between the point and it's closure rail.  Even though it's a bit trickier to solder a jumper wire to a code 55 point without melting ties, resistance soldering and solder paste make it doable without too much trouble.

Resistance soldering runs a high electrical current from one probe to the other, and anything in between gets hot.  Assuming you pinch the wire and rail with the probes, this tends to put the heat exactly where you want it - right where the wire meets the rail.  The rest of the rail gets heated from there, so if you wait long enough the entire section of rail will get hot.  Since that would melt the plastic ties, you need to get the job done before that happens.  In other words, timing is important.  And since how long it takes depends on how fast the joint and solder get hot, power is important too.  You can do it with regular solder, but it's easier with solder paste.  I'm using Iso-Tip silver bearing solder paste (http://www.iso-tip.com/products-page/smart-paste-series/).  Solder paste is basically a mixture of solder particles and flux.  For the resistance soldering station I'm using the American Beauty 250 watt unit that micro-mark sells.

Here's what I do.  First, scratch up the surface you want to solder a little (I use one of these things http://www.micromark.com/scratch-brush,8058.html).  The scrub it with a toothbrush and some 90% alcohol.  Clean fresh metal makes for an easier joint.  Then I strip a section of stranded #18 wire and cut off about 1/2" pieces of single strands.  Same treatment for the wire - scrap it up a little with some fine-ish sand paper, wipe with a paper towel damped with alcohol.  Put a dab of solder paste on a scrap, and use a toothpick to put a little bump of it in the web of the rail at the spot you want the solder joint to be.  I do it between two ties, one tie away from the point hinge (so there are three ties between the two soldered spots).  Using some tweezers you can now set one of the single strands of wire across the point, with each end mushed into the solder paste a bit.

For the soldering I use the tweezer handpiece, with the tips bent a little to make it easier to get contact where I want it.


Squeeze the jumper wire and rail web between the tweezer tips right at the solder paste blob.  After some experimenting I've found that setting the 250 watt unit to 90 and using the foot switch to hit the power for 1/2-3/4 of a second is about right - you get a nice solder joint and don't melt anything.  You'll need to experiment a bit to find the right power and timing for whatever soldering unit you have, so practice on some flex track you don't mind wrecking before you start mangling turnouts.  You may mangle a turnout too, but get the worst of it over on the flex track.


And a closeup of that - note that this attempt was before I went with the longer jumper to put the joint between ties, but it does give you the idea of what it looks like with the wire mushed into the paste.
A closeup of pinching the jumper into the solder paste jut before soldering



Finally, assuming everything went well, clean any residual flux from the paste off with the toothbrush and alcohol.  Here's what the end result looks like if you got everything right.

A nice point jumper

Closeup of a nice point jumper
Note that you will almost certainly not get it right the first time - the power and timing are critical for one thing.  It takes a little practice.  In the above photo you can see a little fuzzy in the background my practice turnout (kind of destroyed) and a piece of flextrack I used for practice as well.  Both for soldering point jumpers (you can practice on flextrack just soldering the strand into the web), and for track feeders.  You will destroy some flextrack and quite likely a turnout or two getting the process right.  Use a magnifier to inspect the joint - sometimes a joint can look OK to the naked eye but when you look closer it's not so nice.  Here's one example of what happens when you get the power and timing off (this one looks a little questionable to the naked eye, but with a magnifier it's pretty ugly).

Bad power and timing, jumper too short
Also notice in that photo that I was trying to solder right over the tie - not surprisingly that makes it more likely that you'll end up melting the tie, which is why I ended up going with a slightly longer jumper to put the solder joints between ties.  It is possible to re-flow and fix joints like that one if you get it right the second time, but if you keep re-flowing eventually you'll wreck the turnout.

You might expect that adding a jumper like that on each point would stiffen up the pivoting action of the points.  Surprisingly, it doesn't - the points still slide side to side with no noticeable additional resistance, and then stay where you leave them - there is no tendency to spring back.

I did consider other ways of powering the points.  Here's one I got as far as trying - soldering feeders to the bottom of the point hinge pins.

A rejected approach
There are two problems with that approach.  The first is that the hinge pin itself does get hot enough to melt the tie a little, and it's at possibly the most critical spot in the turnout to not melt.  The other problem is the wire needs some space to move, and I could quite see how I would keep it unrestrained after gluing ballast down.  I think the point jumpers are better all around.

Enough of point jumpers, on to track feeders.

In O scale I did my feeders after the track was laid using the resistance soldering unit with the single tip probe to solder the flattened end of a #18 wire onto the back side of the rail base.  Very inconspicuous, and easy to not melt ties in O scale.  For my N scale layout though, I couldn't imagine doing that without melting a lot of ties, and also the pressure required on the probe would probably mush the track into the foam roadbed a bit (3/8" homabed in O is a lot less susceptible to mushing than the blue foam insulation I'm using for the N scale layout).  So I decided to solder the feeders to the bottom of the rail before laying the track.  I decided to use #22 wire for feeders - it's big enough to carry the necessary current for the relatively short distances from the bus (~12"), and it's small enough to fit between the ties without touching them.  Also I was able to find scotchlok connectors that are rated for #22 solid on the tap, and #14 solid for the bus.  I used the scotchlok connectors on my O scale layout and liked them, so in part selecting wire size depended on finding the right connector.

The process of soldering a feeder on is simple.  Cut out a bit of the pastic to expose the bottom side of the rail if there's not already a convenient exposed spot.  Scratch it up with a needle file.

Scratch up the rail for soldering
Strip a little bit (3/32 or so) off the end of the #22 feeder, and bend it over 90 degrees with some pliers, squeezing just hard enough to flatten the end a little bit (but not so much that it gets brittle).


A bit flattened on the end that will touch the rail
If you put the track on a little block (like you see in the photo with the file above) with the spot for the feeder just a bit off the end of the block, it makes it easy to use the soldering tweezers.  Put a dab of solder paste on the rail where you scratched it up, mush the wire into it, squeeze the tweezers down, and hit the power.  I find the same setting of 90 for 1/2-3/4 second works for me here too.  It's a little harder to not melt ties a little when soldering right up next to them like this - even if you're fast with the power, the residual heat from the wire, the joint, and the tweezers themselves can melt a little.  On the plus side, a little melting here is nowhere near as bad as it would be right at the point hinge.  On turnouts after I add feeders to the stock rails I also solder one onto the frog power eyelet using a regular old soldering iron and solder.  I'm still not sure if I'm going to power my frogs but it's a lot easier to solder the feeder at the bench than it would be with the turnout glued down to foam.  Finally, of course, clean up the joints with alcohol and a toothbrush.

Feeders soldered on
If you get a good solder joint it's pretty strong.  If you get a cold joint the wire may pull off as you shift the turnout around on the roadbed.  So inspect those joints carefully!  I'm willing to count on the internal connections in the turnout to power the rest of the rails (closure and the two little stubs off the frog), a decision I might regret later but for now it seems like a good trade off in retaining the strength and unmeltedness of the turnout vs. potential electrical trouble.

One final bit of soldering, this time rail joints.  On curves I've found it much easier to maintain a smooth continuous curve in the rail across joints if the rail joint is soldered.  Doing this requires a little test fitting and thinking, to make sure you get the soldered joint in the right place without having to cut out more ties after you curve the flex track.  I use the resistance solder unit again, but with the carbon jawed plier handpiece this time.  I tried mushing solder paste into the rail joiner, sliding the rails in, and soldering, but it turned out to be virtually impossible to get the rails to butt up end to end (not all the paste would squish out from between them).  In the end it turned out to be easier and better to use traditional solder.  As always, scratch up the rail surfaces that you want the solder to stick to and clean them off.  Slide the joiner on centered on the joint.  Keeping the joint straight is a little tricky but not hard - I lightly clamp the track half off the edge of my bench with the back edge pressed against a ruler to keep it straight.  Brush on a little good quality liquid flux (much better than paste flux).  If you use too much flux you'll eventually end up with a glaze of flux on the carbon plier jaws and they won't conduct electricity well enough to solder with - if that happens a little sandpaper will clean the glaze off.  Squeeze the pliers onto the top of the rail and the bottom of the rail joiner.  Be careful not to shift the rail position as you clamp down on it - I brace my hand on my leg (not the same leg I'm using for the foot switch).  Hit the juice.  The foot switch is critical here unless you have more hands than I do.  The power setting of 90 worked well for me here too.  Be ready with some nice thin solder at one end of the joiner, preferably on the outside of the rail.  As soon as the solder flows into the joint a little quickly move it to the other end of the joiner, and as soon as it flows there take your foot off the switch but don't move the pliers!  Wait for the joint to cool a little before you move the pliers.  If your eyeballs are into middle age or beyond, an optivisor definitely helps in all these soldering processes.

Soldering a rail joint
Unclamp everything and use the straight edge to make sure you didn't shift the rail a little with the pliers and introduce a kink while you were trying to do everything at once.  If you did end up with a kink, re-clamp, use the resistance pliers again to reflow the solder and let it straighten out.  You will probably get a little tie meltage, you can either ignore it or if it's to the point of looking bad cut off the offending ties.  If you melt too many of them from too many reflows, consider cutting the joint out, trimming both pieces back, and trying again.  Once you get the track pieces soldered together, you can solder the feeders on the bottom.

I was a little surprised that it actually turned out to be easier to get a good straight clean solder joint on N scale code 55 than it was with O scale code 148.  The bigger rail takes longer to get to the point where the solder melts, and it holds heat longer - that means you have to hold everything still for longer.  And since the rail is stiffer, you need to clamp against a longer straight edge.

That's about it for soldering and track.