OK, let's see what I can add to be helpful.
This loco was returned to my customer from another installer completely stripped down with even the running gear disassembled and removed from the wheels. I had to completely rebuild it and that took a bit of effort. I didn't realize for example that the link pin was squared so it would only fit 1 of 4 ways. I thought it was keyed and would only fit 1 way. But when I tested my partial reassembly of the wheels and running gear without the motor I had it bind. A bit of fiddling and comparing to a new loco showed me about the link pin and I got the running gear reassembled properly. (And turned right side up, I'd installed it upside down.) It ran OK on DC when I tested my rebuild. OK here means it ran forward and backward on DC, just a pass say back and forth on the 6' test track. I probably should have made 100% certain it ran perfectly on a real layout before I went further and did the install (Hindsight is so 100%
but now I'm wondering if somehow the frame itself got bent or some sort of other issue is involved. Something that makes the mechanics of the loco 'off.'
It makes sense that the thing is designed as a unit, a whole. So when I run with the boiler and its weight off I can see how the loco would be 'tipped' backward, the balance point of the loco moving back toward the rear of the engine. And when I use the pair of alligator clips not only does that provide 100% certain power to the frame but it also adds weight to the front of the engine. So I can see how that'd guarantee the loco would run like a Swiss watch, which it does. So I'll do all the rest of my testing with at least the weight sitting there where it belongs instead of just the naked underframe.
I think that would also explain how adding a .003 shim to the 4th driver would help, it'd counteract (at least somewhat) the tendency of the loco to tip backward since I was doing all my testing without the loco weight.
But...
I've since put the whole thing back together. And pretty much still the same miserable poor performance. Stops on completely straight track. The loco will almost immediately start again and move, go a few more inches and then stop again. Putting the shim back in was the only thing that seemed to make the loco run decent. Now I'm down to about 1 stop/start in 6 passes back and forth on my 6' straight test track. So things are improving. But it sure doesn't run super smoothly like the new one out of the case does on DC.
something is interfering with the pickup.
Rails to wheels -- checked
wheels to axles -- rarely a problem
axles to sideframes - - checked
sidefrmes to tender pickup strips - - checked
tender pickup strips back to sideframes - - checked
side frames to drawbar - - checked/replaced
drawbar to loco frames - - checked/replaced
loco frames to decoder wires ----> only thing left
you say you attach the jumpers to the frames
but to you attach to the decoder wires instead?
There really isn't any wheel to axle conduction in the design of the Mikado. Power goes right from the wheel to the bushing/bearing then to the frame. I also think seeing it runs perfect with the alligator clips tells me I have the power pickup for the decoder correct also. If that was bad/intermittent I don't think it'd make a difference if the loco was picking up power from the rails or my clips. So I don't see how attaching the clips directly to a decoder wire would show me anything more but I'd gladly try it if you help me understand what it might show so I can look for it when I try that test.
One of the first things I did was to add a second power pickup to the loco. I'd originally wired it just to pick up power from the contact strips in the tender, soldering the red/black wires to the center of each pickup. But when my test runs proved to be so bad I immediately added another pair of pickups wired to the lightboard in the loco also. So now the decoder is wired to both a contact strip and one side of the lightboard for each pickup.
So we're back to the drawbar. I've barely handled the new one. I put it on my way, with the brass contacts under the crosspiece of the truck. I then looked at Chris33's pic and tried it with the brass wire contacts over the crosspiece. No change so I put them back to the under side instead because that matched the way the contact was on a new loco I took from stock and compared against. That isn't much handling. I've put the loco up on the test track from the bench a dozen or so times. Again not too much handling. I've never lifted just the loco (or the tender) and let the tender (or loco) dangle so I don't see how I could have bent the drawbar.
I took a file to the nub on the underside of the loco when the drawbar attaches making it all nice and shiny instead of the blackened metal. I checked the tender cups for lint and made sure the tops of the contacts where they touch the brass tender strips are nice and shiny. But I'll redo all this again tomorrow just to make 100% sure that isn't the issue.
I ordered 3 drawbars. 1 for this loco, 1 for me (I've one here in pieces waiting for the GHQ PRR kit to be installed) and a spare.
I'll try the spare tomorrow but I can't believe I could have ruined the one I just put on yesterday (Friday) with the few times I've lifted the loco to my test track and back to the bench or when I installed it or moved the location of the pickup wires from over to under and back again on the tender truck.
But I really do appreciate all your folks suggestions of things to try and your attempts to help me. This really is quite the puzzler and I'm totally stumped and all your input has been so very helpful to me.
Again, thanks
Brian