I had been reluctant to buy the original SD70ACe because of the problems Carl Tuveson had described when the locomotive crosses an American Models switch. I finally ended up buying one and applying his fix of adding springs to the center axles, but now and then the locomotive still stalled on one of those switches. After studying the layout of contacts a bit, I swapped the positions of the axles 5 and 6 (counting from the front of the locomotive) on the rear truck only. The problem went away. I ordered a second SD70ACe and merely made the swap without adding springs, and it has run without a problem across the same switches.

That problem was one of electrical contact and geometric spacing of contact points. Swapping axles merely changed the spacing so that no contact would be lost across the AM switches. Your problem sounds different, although losing Railsounds intermittently is a symptom of lost electrical contact (as is flickering lights), even momentarily.  On Legacy AC power, both locomotives of mine run flawlessly. On DC power, they will frequently fail to respond at all to increases in power to run them forward. Sometimes they do respond, but run for several seconds, stop, start again for several seconds, stop, and continue to balk this way until power is turned off. Reverse direction is never a problem on DC. When the failure to respond or balking occurs, I start them very slowly in reverse, then switch direction to forward by changing the polarity. That always seems to work, and once the locomotives have "learned" how to go forward, they continue to work properly until I switch back to Legacy AC operation. (The ES44AC locomotives seldom show this behavior.)

Your symptoms are different from mine because you operate in Legacy AC all the time, and I don't use the smoke feature on the diesels I have. I have no problems at all in Legacy mode, but maybe the forward/reverse preference is a clue to something else. Railsounds cut out now and then with some steam locomotives, but that is always due to dirty track or wheels at a particular spot, and cleaning fixes the audio problem. I suspect you're experiencing a loss of electrical contact somehow, and it's difficult to determine if it's in the wheels or internal wiring. 

TOKelly-thanks for your reply. I read your first post back in March incorrectly-I thought you's swapped axle #4 with axle #5. I will try swapping axle 5 & 6 and report back. If that doesn't work, I'm going to try swapping the trucks from my ES44 onto this SD70 unit and see if anything improves as far as stalling.

I think the smoke issue is separate and more due to electrical components that are undersized for their intended usage....

I tried to find my notes on what I did, but I can't seem to find them. I just checked both SD70 locomotives and found that the front truck has the traction tired axle in the middle position, and the rear truck has its traction tires toward the front of the locomotive. I don't recall any problem in making the switch, such as interference with the Hall effect sensor. I looked at Carl Tuveson's pictures of the axles on the trucks, and it would appear the only change I made was to the front truck, not the rear truck. I apologize if I confused you--I confused myself on this one.

My BNSF SD70 had those issues on less than ideal track work as did my early SP U33c's, but with the wheelset swap that I described (search on "Better ops from first run AF U33c and SD70ACe") they've run much better ever since.  Unfortunately Lionel ran out of the required wheelset.  I haven't checked lately to see if they're in stock again or not.  

NEW DATA Point: This morning I decided to set up a loop of original Gilbert curve track to test out my SD70ACe and its operating issues. I cleaned the railheads with some Scotchbrite pad just to even the playing field. Power for the test was a Lionel GW180 transformer, only I was only running it in conventional mode. While the swivel of the trucks was just adequate for the Gilbert curves, the engine ran much, much better AND the smoke unit worked without any engine shutdowns. There were no Railsounds dropouts.

Data Point #2: My next test involved swapping the truck assemblies from my sweet running ES44 to my struggling SD70ACe. Once the engine trucks were swapped I put it on my American Models flex track loop to test it out. My SD70ACe now ran as sweetly as my ES44 did; therefor, my deduction is the problem is somewhere in the truck assemblies that came on my SD70. This engine now made it around my complete loop at Legacy speed step one without faltering, and, the smoke unit was operating and did not result in complete locomotive shutdown seen previously.

My next effort involved teardown of both of my SD70 truck assemblies completely to inspect for any obvious areas for electrical shorting, unfortunately, I could not find anything amiss as all three molded insulators between the truck block halves appeared to be doing their job. I reassembled everything and tried the engine again-while operation is somewhat improved, it does not mimic the operation it had with the newer ES44 trucks installed, even though they are the same basic truck design. For whatever reason, my spring-install effort on the center wheel sets  has not paid-off the same as Carl saw.

BTW: To correct an earlier mis-statement, the Hall effect sensor is on the rear truck block, not the front one.

SUCCESS!  Despite what I thought about the sprung center drivers not providing the expected improvement in operation, the more I ran the unit the better it seemed to operate this afternoon. Based on TOKelly's advice on swapping axles two and three, I elected to do that late this afternoon. It can be done as both axles are the same size, unlike the rear trucks.

That little swap was the final puzzle piece to get this cantankerous little piece of machinery operating like it should have five years ago when delivered.

Thanks to TOKelly and Carl Tuveson for providing the needed links to make this engine operational again. It will now crawl at Legacy speed step one around my layout WITH smoke operating if desired with no blinking of the engine marker lights. It even got through my American Models turnouts without issue now-this used to be one of the troubleshoots around the layout.

That's good news. I was worried I was leading you down the wrong path. The new spacing of wheelsets works just right to get across the American Models turnouts. The ES44s, on the other hand, work right out of the box. If Lionel makes an SD40 or SD40-2 as nice as these modern locomotives, I'm in.

TOKelly: I've had some luck improving electrical operation across my AM turnouts by taking a flat file and burnishing the outside edge of the swivel point rail + the inside of the stock rail where it rests. There seems to be some resistance there possibly from the weathering process on the rail during manufacture(?). I did this to both swivel rails and it even helped the SD70 prior to the last two days work on it.

The iconic SD40 would be sweet-much like the one announced around 1999 but never produced......

D&H 65: I also clean that part of the AM turnouts whenever I clean my track. It's an often-overlooked spot that can accumulate bits of ballast or rogue pieces of foliage as well as the material that fouls the top of the rails. A bit of bending the tip of the point rail now and then also aids in close physical and electrical contact. All this tidiness keeps things rolling--until someone comes over to "see the trains run." Then nothing works right!