Puzzling it is. 

A couple of things. 

Thing one. Does the timing light clip have a directional arrow and does it always point towards the spark plug, including cylinder four? 

Thing two. Question: Let's say cylinder four is under compression, does that mean the circuit would go through cylinder one first, then reverse for the next spark? (I think so not, but I'm starting to go non linear here)

Robmack, You asked "Can you test the HES at ICU pins 5&13?" That's a great question! Can I? Will the LED test work there, too? I don't know exactly how I'd hook that all up, but maybe it could work. Question, though: Would I be getting spark at all if the ICU wasn't getting the hall signal? Doesn't spark timing come from that?

   Thank you for that diagnostic page! That could be really helpful. I'll dive in on that next time I work on the bike.

   re: resistance on the spark plug leads. These are the same leads that have been working fine for thousands of miles. Can the resistance on those fail? I'll check it, though. One more thing to take a look at. Though, the more I'm thinking about the spark plug timing, the more I'm thinking it's a dead end. I'll follow up on it, of course, but I think I must have just had some weirdness of my timing light. Especially since you're confirming what I thought: 1 and 4 fire at the same time. It's hard to think of a way that 1 could be firing normally but 4 would be firing only half the time. It must be my timing light.

   Plugs are NGK D8EA. They're new, but the same as I've been using for many years.
   I'm going to check the resistance on the coils, too. Such an easy test there's no reason not to do it.

GTJos,
   That's an interesting thought. I'll take a look. The fuel pump is always running just fine, even when the bike is messing up. But that's only one of the two outputs from the relay, right? And the other output is power to the injectors? I've heard the injectors clicking through a stethoscope while the bike is messing up. But you know what, I just realized I did that stethoscope test with only one injector. I'll take a look at that relay connection and also listen to the rest of the injectors to see if they're all clicking.

Daveyson, The problem seems to have become permanent. Early in the testing I had brief periods of the bike running normally for many minutes on end, but not in the last week or so. It's only messing up now.

How did I get readings for 9 and 14? I had the positive probe of my voltmeter stuck into the back of the ICU plug at 9 and 14, and the negative probe against the engine block. The readings  I got doing that were 12 volts as soon as I turned the ignition on, dropping to 9-10 volts while the starter was running.

I'm curious how you got your bike running without a fuel pump. Was that just gravity feed to the injector rail?

From your second post, yes my timing light has a directional arrow. I'm pretty sure I got that right. I even double checked after I got a weird reading. But I have had issues with the timing light from time to time. Using it on my R60, I have to hold the lead exactly in the middle of the the timing light pickup for it to register at all. I'm really starting to think the spark timing anomaly was just a quirk of the timing light (but I am still going to follow up on that.

As for behavior of the coils, my guess would be that each coil behaves like the setup on my boxer. Each coil sends out two sparks at the same time, once per revolution. When 1 is getting its necessary spark, 4 is getting a wasted spark at the end of its exhaust stroke. But that's just a guess. I'm sure someone here can confirm. 

Porto, You're totally right. The last time my bike had one of its brief moments of running well it started while I was checking for spark. I had plug 3 pulled. All of a sudden the bike started working. It was running a little rough for the lack of that one cylinder, but it was running. Not all that badly, really   So, yeah, even if I do actually have a problem with the spark on cylinder 4, the bike would still be running.

Okey dokey, I haven't been confident with any of my guesses, none of them fit because everything is working, except some questionable cylinder four weirdness, and that. It's hard to consider compression as intermittent. You got fuel, you got spark, you got clicking injectors (one,  three more to check) And check all four spark plug leads for resistance, clean and tight. 

When it cuts out while riding, do you think it was on corners, bumps, acceleration, deceleration? 

Coil swap is still on the To - Do list, I think. 

No good, back to the drawing board. Fuel, or spark? I'd remove the air filter and spray some fuel in, it only goes a few seconds after fuel, it's a fuel problem. (that starting fuel in a spray can stuff)

You got black or blue smoke out the exhaust, I spose nothing.

Edit: yes the time my brick ran without a pump, I think gravity had a lot to do with it. It's not something you really want to do but more details are in the How - To section, the post is called "how to ride home when your fuel pump fails"

daveyson wrote:. . . everything is working

Yes it's funny, basically everything passed the tests 

Hey Bikesmith, you say you do your best thinking on a walk, I do that too. And sometimes on a ride, one great one was also to do with the gravity fed brick on a thread (years old) at motobrick called something like "Need electrical help, brick runs without a pump, what's going on" The thought bubble came totally out of the blue.

daveyson wrote:Yes it's funny, basically everything passed the tests

Everything?  It's time for me to fire up the grill and listen to the birds.

Well something's not right, but what? There's spark, pumps working, fuel returning, injectors clicking. I think the plugs are dry, so a few sprays of fuel is quicker than a fuel pressure test.

That's a quick test which might help to find out, is it fuel or spark?

Daveyson, "I'd remove the air filter and spray some fuel in, it only goes a few seconds after fuel, it's a fuel problem."That's a fabulous idea! I'll try that for sure! 

The bike has cut off while running three times now. Once was while going a straight, steady, bumpless 35mph in third gear. Next was a straight, steady, bumpless 20mph in second gear. Third was while idling in my driveway.

Laitch, I agree it sounds funny when Daveyson says "everything is working", but what's been especially frustrating about this round of bike repair compared to previous ones is that, other than the obvious fact of the bike not running, I just keep coming across "Yup, that's working and is not the problem" moments. Once I can isolate the problem, I'll fix it no problem. But I just can't isolate it. I actually had a dream earlier this week that I had replaced every single part on the bike. It was literally a brand-new K100. And it still wouldn't start. Funny, but not funny.

Oh, and regarding the leads. Sure, I know they're impermanent. I guess I should have phrased my question as "Can the leads fail in such a way that the bike is running great one moment and completely dead 2 seconds later?" I'd have thought they were more like a gradual-decrease-in-performance sort of item.

I would start fresh. Forget every thing you have done. Start again, keep to basics.
As for the leads, they can be stuffed completely and arcing in the dark like crazy and the bike will still keep running. If your bike runs ok but won’t restart or stops and won’t restart, that’s the problem stated from the start. Then either your loosing all your fuel supply at the rail or all the spark. Even with a trickle of fuel and crackle of spark it will start and run, very bad but enough to tick over. If the problem is extreme then you would still get a pop. You have nothing. Start again with the simple stuff. Ps in an early response I mentioned the load shed relay, I meant the fuel injection relay, the one with double load contacts. But you have replaced that. I was thinking of another thread at the time, sorry

True, back to basics, spark or fuel, fuel test say five sprays two seconds apart. 

What I'm thinking now, the quickest way to the solution is if I stop posting, so I'll try to make this my last post. I'll be watching with interest and only post if it's not a punt. 

If no change with the fuel spray test, could it be that there's not enough power for the ICU from pins 6 and 10 with the start button released?

Good idea. I’ll let some else answer your ICU question.
I don’t no and my head Will getting if I think about it lol. But don’t worry about your injector squirting. That’s not going to be your problem. Good luck. Can’t wait to here what resolves it. Mean while I need to get myself new coils lol

OK, I'm known for my willpower, so here's another post. 

The weirdness at cylinder four could be due to a faulty spark plug, lead or coil. Temporarily swapping the green/yellow wire and the brown wire under the coil might shift the problem to cylinder one. Maybe easier to swap leads 3 and 4 to see if it's the lead. Lead four should have 4 printed on it, if they're OEM leads. (if it's from a K75 that wouldn't be good, they're different, with an air gap, and couldn't have the number 4)

Bikesmith wrote:* Fuel. I can't see gas being squirted in the cylinder directly, but I have new fuel pump (which I can hear running), new pressure regulator. I can hear the injectors clicking. They might be dirty, but that seems like it would make the bike run rough rather than alternate between perfect and dead.

You indicated back in the mists of time that if you kept the starter button pushed down, the engine would run until maybe 400-700 rpms then would stall as the starter shut down then would start running then would stall again, repeating that cycle ad nauseam while the starter button was pressed. Is that right? Regardless, the engine wasn't alternating between perfect and dead; it was alternating between idle and dead. 

One reason it keeps stalling could be it can't get enough fuel to run faster than idle because the injectors are suppling only enough fuel to provide an idle but no more. When the engine needs more but doesn't get, it stalls until the air/fuel ratio balances. Another reason it keeps stalling could be the hall sensor signal indicating increasing revs drops out completely after ≈700 rpm instead of being picked up by the FICU to increase fuel to keep things running. Then there are the coils, which are simple to check. What's your data on them?

Here's something fun and exiting to do. You've indicated you don't have x-ray vision to observe fuel in the cylinders. That isn't surprising. Remove the injector rail with injectors attached, aim them into a bucket large enough to accommodate their flow—don't let fuel splash onto the electrical components of the engine—then turn the ignition to On and press the starter button for the usual length of time it takes for the engine to run and stall a few times. Keep a gasoline fire-rated fire extinguisher handy in case you get distracted. Here's a Bricker doing a similar test but likely for a shorter duration.

It would be nice to record the results on video, post them to YouTube and post the link here so we can enjoy the show and critique it.

Yes take a video, I’d love to see it.

All right, I'm back on this!

First up, here's a video of what it sounds like when the bike tries to start: https://photos.app.goo.gl/FSbqXAejtxqr1QJy8

After I filmed that, I tried the test of pulling the injector rail and squirting the injectors onto cardboard. They squirted, but I didn't get a chance to see how vigorously they squirted because... here's the really weird part: While the injector rail was pulled the bike started!   Running poorly, obviously, but it started. It kind of sounded like it might have kept going, had I not killed the bike as fast as I possibly could. Explosions coming out of holes right next to where injectors were squirting everywhere seemed bad to me for some reason.

So I have no idea what to make of that. Could it possibly be that air actually is the missing ingredient? Sure doesn't seem likely. But, man, I really did not expect the bike to do that!

Oh, you know what? While I was pulling the injector rail one of the injectors popped out of its clip and came out of the rail, spilling some gas until I could get it back in place. I bet some leaked into the cylinders. I bet that's it. And that lends credence to the idea of gas being the problem. I need to do the starter fluid test!

But first, I found a problem. Not necessarily THE problem, but a problem. While doing the injector test, I noticed a small gas leak. Just a few drops, but still not good. Right at the top of the hose from the tank to the injector rail. While feeling around for the source I found that the top inch of that hose is sticky, the clamp has permanently deformed the rubber, and black rubber goo came off of it on my fingers. Nicht gut! The hose is less than a year old, too! Nice Continental fuel line. Not ethanol resistant?!? Seems crazy, but it must be. The thought crossed my mind that if there's black goo mixing with the gas, that could be causing problems. I don't see any way it's making the gas unburnable, but maybe black rubber goo is clogging the injectors?

So, I'm going to run over to the auto parts store for a new hose. I'll be back.

It is quite possible that the liner of that fuel line is collapsing and limiting the fuel flow.  In any event, it certainly won't hurt to replace it with the correct grade. 

I wouldn't take much of a stretch of imagination to see that the fuel flow is so severely limited that there would be just enough in the rail to start the bike, but not enough flow to sustain the running.

Okay, big clue!
  With a generous squirt of starter fluid into the airbox (filter removed), the bike will run for about 30 seconds! Not super well (though not badly), won't come off idle, but it runs!

  I also checked for clicks on all four injectors. All are clicking rapidly and in time with the engine. When I hold the starter on they click very fast. If I let the engine die, they slow down with it.

  So here's what I'm thinking: This seems VERY likely to be a lack of fuel in the cylinders, since a squirt of starter fluid makes the bike run. I definitely have gas in the injector rail. I'm getting a positive flow of gas back into the tank from the regulator. It's possible the pressure in the rail is low, but it's a new pump and a new regulator. Besides, wouldn't it at least idle if the pressure was low? I've heard these bikes can even run without a pump 

  So, say I have gas at proper pressure outside the injectors and not enough inside the cylinders. The injectors are clearly getting some signal, since they're clicking. Maybe they're not being held open long enough per cycle? Duration of injector opening comes from the FICU, determined by what? Air flow meter? I checked that that was good. Temperature sensor? Replaced it. Timing from hall sensor via ICU? The bike can run for 30 seconds with a squirt of starter fluid, so I'm pretty sure it's getting a proper hall signal. Maybe the ICU is not passing timing on to the FICU correctly? Maybe, but I'm getting clicks on all injectors that are timed to the engine. I've also cleaned and checked all connections to and from the FICU. Is there anything else the FICU is using to determine length of injector pulse? Something else to check out?

  This is really starting to seem to me like my injectors are clogged by black goo from my disintegrating fuel hose. But I certainly have a history of feeling certain about wild guesses  , so what do you guys think?

If the rubber parts in your tank were deteriorated and turned to black goo, and the pump circulated that stuff into the fuel rail, then it is most likely that the filter in each one of the injectors was contaminated.  Not enough to stop fuel flow but enough to reduce it.  Having an injector overhaul may be in order.

See my last reply above.

You might want to replace the filter screens in the injectors when you do the fuel line.  You can buy them at the auto parts store for about 10 bucks.  Take one of your injectors with you so they can see what they are going in.  You can pull out the old ones by screwing a sheet metal screw into them and pulling on it.  The new ones just press in.

Okay, I was feeling good about the injectors being the issue. I was (and still am) going to replace the filter screens.

BUT... Just for the sake of being thorough, I decided to check resistance on the ignition coils. Resistance on the high-voltage terminals on 1/4 was about 11K. A little lower than what the troubleshooting guide suggests is ideal, but not bad. Resistance between the high-voltage on 2/3 is... Infinite. I checked it over and over, wondering if I messed up the test. But no, that coil has infinite resistance.

Welp, I think I found my problem.

Follow-up question: I'm now looking for a replacement coil. There are not a lot out there. Beemer Boneyard doesn't have any. Ebay has just one in the US right now (https://www.ebay.com/itm/183758813746?fits=Make%3ABMW&hash=item2ac8e0fa32:g:L4UAAOSwax5Yrd2Q), but it says it's for a 92 or later. Will that work? From the photos it looks very similar to mine. It looks like it has a different connector than mine, but if I can make the connector work, will the coil work otherwise?

It looks like MaxBMW has a new one, but its 2.5x the price. The ebay one would be great if it'll work.

Good news. It sounds like you had more than one problem. One bum coil won’t stop the engine, it will just drop it back to two cylinders. How ever if the other two cylinders are low on fuel pressure, then yes it will stop running. Reading the recent posts suggests all your fuel lines need to be replaced. Also clean the screens, replace filter. Clean the injector screens as well. Sounds your nearly done

Deleted.

Bikesmith wrote: Ebay has just one in the US right now (https://www.ebay.com/itm/183758813746?fits=Make%3ABMW&hash=item2ac8e0fa32:g:L4UAAOSwax5Yrd2Q), but it says it's for a 92 or later. Will that work?

The eBay candidate should not be used. It's from a Motronic system, not an LE Jetronic system. It's primary resistance is ≈0.31Ω compared with your old Brick's primary coil resistance of ≈2.2Ω. Resistance that low is likely to eventually cause ICU damage.

Bikesmith wrote:Welp, I think I found my problem.

It's too early to start the party. From what I've read, there's more investigation remaining. Check the fuel pressure and the injector output. It's probably time to recharge your battery and I don't mean trickle charge.

Thanks for the heads up about that coil!

What were the primary resistance numbers?

Pretty much spot on. About 2.8 each.

Bikesmith wrote:Pretty much spot on. About 2.8 each.

Spot on according to the troubleshooting guide is 2.6Ω. There's an ≈8% difference, which is acceptable.

What's each plug wire's resistance from its coil terminal connector to its spark plug terminal connector? Check then replace one wire at a time.

Deleted.

From what you are reporting about running on starting fluid and the coil resistance, the rational course of action is to replace the fuel line to the rail, clean the rail and replace the injector filters.  If you haven't installed a NAPA Gold 3032 yet, do that too. 

Once that is done, see if it runs.  Then you can decide on spending the big money on the coils.  If you get it running you will now have the luxury of watching eBay for the opportunity to get a good deal on a set of later model coils. 

I don't know what kind of meter you are reading the resistance with, but in my experience with testing electrical stuff for 50 years, any reading below 20 or 30 ohms is pretty much a wild ass guess.  At 2 ohms the accuracy/error of the meter can be as much as two and a half times what you are trying to read.  Add to that the resistance in the connections in the probes, dirt, corrosion, etc. and about all you can tell is that you aren't dealing with an open circuit. 

In my experience, coils normally fail as open circuits, the fact that you are reading some resistance tells me that is not your main problem right now.  There is the possibility that heat is causing the failure, but the best way to troubleshoot that is with a running engine.

True, that makes sense. I realise the cylinder four reading is a side issue, but if there's a spare mini (it's a quick check) and the problem can be shifted to cylinder one, that's a real thing.

Deleted.

Deleted.

All credit to you Bro,we try to help each other when we can.  The written word has no emotion,, had PSF placed a smiley face  or maybe  would show his intended sense of use.   Misunderstandings can cause more problems than a smack   ever can Pease and goodwill to all of us 

daveyson wrote:It's still baffling since you say the injectors are spraying. I spose your trying to get hold of a pressure tester. In the mean time as a simple test swap the primary wires for cylinder one and four coil to see if the problem shifts to cylinder one.

This is a much faster check than chasing around trying to measure resistances. Even better, if the problem shifts it can only be a coil or HT lead issue. If it remains you are looking at a coil or plug issue. If it is only one of either 1 or 4 it totally rules out Hall sensor fault because both cylinders spark off the one sensor, so 1 of them sparking clears the Hall sensor off the list of suspects as the coil is getting its input signal.

This would be one of the first trial and error tests I would run in trying to hunt down a one cylinder problem or a non start problem. Find the errant cylinder by the age old method of feeling the header temperature.

Deleted.

Since Bikesmith told us the engine will run for 30 seconds on starter fluid and that the fuel line is breaking down, it says, to me anyway, that his problem is fuel related.  Chasing plug wires and coil resistance is not going to address the possibility that the engine is not getting the proper amount of fuel to run.  He needs to follow that line of investigation to completion before going to a possible spark problem.

Aply, KISS and the old saying If a hammer wont fixit its Electrical

All right, I have an update for you guys:

The bike is running!!!! 

I replaced the intake filters on the injectors and it made a difference. Not a magical change, but the bike would run more like 6-8 seconds rather than 1-2. Since it was seeming like a gunk problem, I decided to replace the fuel filter again.

While doing that something caught my eye. It was hard to say for sure, but the gas in the tank looked orange. I found a glass jar and compared a sample of fresh gas against a sample from the tank. Look at this: Control gas versus What came out of the tank.   

It's not just the color. If you look closely you can see a ton of black specks, too.

So I drained the tank, replaced the fuel filter and the hoses inside the tank, and put in a load of fresh gas. The bike started running! But not great. It alternated between running pretty well and just barely idling. It would run quite well for thirty seconds and then suddenly start running roughly. Still idling, but low (7-800rpm). And if I touched the throttle while it was running rough it would die (but start right back up again).

But the more I ran it the better it got. I've run the bike for probably an hour, hour and a half now, and it's been getting better and better. It's still idling a little low, around 900 rpm (I generally keep idle set between 950-1000). It's still a little rough, just not quite as monotonously smooth as usual. The idle wanders a little. I think it might be smoking just a touch (hard to say for sure. If so, it's really subtle). I took it out for a ride and noticed it hesitates on hard acceleration. No problem with slow, smooth acceleration while riding and no problem with hard, sudden acceleration with no load.

Oh, and since the bike is running it's much easier to run the ignition coils through their paces. I tried removing each spark plug in turn. The bike could run without any one spark plug, though it can't run on two cylinders. Thinking maybe I detected a difference between removing 2 or 3 versus 1 or 4, I tried switching the coils. I really don't think there's a difference. Also, I have a tester thingy that you can set to any gap and hook a spark plug lead to. I tried setting that to about a 0.060-inch gap and all four leads could still make a spark on it.

So I'm now thinking I'm going to try just running a couple tanks of gas through this thing and see if it's still improving. I might replace the filter one more time after a tank or two of gas (they're $6, so I don't mind getting a few). Worst case I think is that that craptastic gas damaged the new pump and I might have to replace it again. I doubt it, though. I tried squirting the output of the fuel pump into my glass jar and it's an impressively forceful flow. It sure seems right now like just running a bunch of clean gas through this should clear things up even more.

Oh, and I noticed that the new fuel lines I installed inside the tank were sticky and leaving black smudges on my fingers after just a couple hours submerged, so I guess I'm not done with fuel lines   Do I need to find something specifically submersible? This doesn't seem like it should be hard.

So, at the very least I am well on the way to a fully functional bike. It sure was nice to get to ride it today! I'm open to suggestions if anyone thinks there are still things I should be looking at, but the way the bike continued to run better and better today while running clean gas through it had me very encouraged. It's funny that I've come full circle back to the first thing I thought was wrong. Because the bike first died seconds after getting gas, I initally thought that bad gas was the problem. I guess I just didn't go far enough down that avenue of investigation.

Thanks again for all your help! You guys are the best!
   -Jon

Bonus, good work, progress.

On this side of the world, fuel colour varies from suppliers, could be clear, yellow, red, but the black stuff, that's a problem. You want submersible hose for inside the tank. You flushed the tank, but flush the rail as well. Some crap might have got into the injectors. On YouTube there are clips about how to clean the injectors.

It might not be easy to flush sticky rubber out, but worth a try. I went to a junk yard once and got injectors from an '80s model BMW car, the part number was almost the same as the brick injectors, it worked well on a brick. I think the part number difference related to the length of an unimportant part of the injector. I also got injectors from a 96 Hyundai (Edit: Elantra, second generation) that looked the same, but I haven't tried it yet.

Bikesmith wrote: Do I need to find something specifically submersible? This doesn't seem like it should be hard.

You need submersible, fuel injection hoses—specifically, SAE 30R10 rated, 8mm or 5/16" hose fastened with fuel injection-type hose clamps as illustrated below. It isn't particularly inexpensive hose; MaxBMW has it for a reasonable price if you can't find it elsewhere. You need to thoroughly clean the tank if there is residue in it. Many of us have used white vinegar for that task.

I suspect you still have some crap in the injectors that probably got past the screens.  Not a big deal because the screens should stop anything that would block the nozzle.  A bit of run time should shake it loose and push it out.  You might want to throw some Seafoam or Techron in the gas tank after you clean it just in case there is any varnish in the injectors from sitting over the winter.

When you replace the filter in the tank, be sure to use the submersible SAE 30R10 fuel line.  The stuff that comes with the filter is not the correct grade, it's for use outside the fuel tank.  Even though the liner of the line that comes with the filter will probably handle the fuel, the outer shell softening will result in the clamping force on the filter's hose barbs being reduced, allowing some loss of fuel delivery because of leakage.

You are a long way from us but I redid the tank inners in my 83 RS back in March and replaced everything inside the tank. 

The fuel lines for inside the tank are marked NBR/NBR so they are fuel resistant on both inner and outer.

If you get stuck Hondas use the exact same 8/13 or 8/14mm so you can pick it up at a Honda dealer. You should change the tank to injector rail hoses [also 8/13 or 8/14mm] but they are not the same as the outer has to be UV protected. Again your Honda dealer can supply these too.

A good way from you but a good source of parts is kpartsholland.com or tills.de and they will ship to you.

Hey gang!

  I bet you thought you'd seen the last of this thread 
  Not quite.

   So the bike is running. I put over 50 miles on it today. But it's still... weird. 

   Since my last post I replaced the in-tank fuel lines with submersible hose, replaced the fuel filter, and replaced all the injector filters again. I also did a super thorough cleaning of the tank.

   And now, The bike really doesn't like starting if it's been off for anything more than a few minutes. It'll start, but it will behave in a way very much like when it wasn't running, but not quite as bad. It idles low (maybe 700ish), sounds rough. Any touch of throttle will kill it. Seems to barely have enough power to keep itself turning over. It will do that for anywhere from a few seconds to 5-10 minutes, and then in an instant it starts running perfectly. 

   Once it clicks over to running well it stays that way. I can go for a perfect, trouble-free ride. I've ridden up steep hills, I've had it up to 70mph. Once it is in good-running-mode, you'd never know anything was ever wrong. It's smooth and strong and wonderful. And it stays that way until I turn it off and leave it for an hour. If I turn it back on after just a few minutes (gas stop), it maybe has a slight hesitation before coming up to full, smooth idling, but that might just be me over-analyzing normal running. 

  My gut feeling is to think I still have a bit of crap in the fuel system. Some chunk of whatever that settles into a critical spot when the bike is off, but is pushed into a 'safe' spot by the pressure of the running bike. Maybe one more thorough cleaning of the whole fuel system will clear it up? (the injector filters came in a bag of 50, so I have no problem with continuing to replace those   ). Maybe I just need to keep burning clean gas through the thing and it will take care of itself? I've put about 75 miles on the bike total since I got it running, so not even a full tank of gas.

  Oh, I did check the fuel line pressure. And it was while the bike was in its 'just barely running' mode. My gauge read 35psi. I know that's 1 lower than nominal, but that shouldn't cause trouble, right? Besides, once it clicks into perfect running, the fuel pump is keeping up with demand even while going 70mph uphill.

  Is there anything else this could be? What is the difference between a bike that was just running versus one that was running an hour ago (and is still pretty hot)?

   Yours in perpetual confusion, Jon

The fuel pressure is good.

In weird mode, does it improve with the help of a spray of starter fluid? Don't think you replaced injectors, but hard to imagine them instantly improving at the same time.

Not confident about this, as per usual for me in this thread, but only mention this cause it's a quick easy test; cylinder four had half the proper number of sparks. Does it get the proper number of sparks when the problem suddenly disappears?

How long have you owned this Brick and how many miles have you ridden it since you've owned it?

Bikesmith wrote:

   What is the difference between a bike that was just running versus one that was running an hour ago (and is still pretty hot)?

A sticky, lazy like vane in the air flow meter? 

Then the fuel injection computer would be tricked into denying a richer mixture for the extra load? But I think you tested that.

Or idle position switch tricking the computer into believing the throttle is at idle, so a touch of the throttle would kill it. I don't think that's been looked at yet.

The next time you do a cold start, disconnect the return line going to the fuel tank from the pressure regulator.  Direct the flow into a clean bottle. 

First, there should be a fairly good flow.  If not, it might indicate that the pump is having a problem priming itself. 

Second, put the fuel in a clear container and see if there is dirt or if water separates out when it sits for a couple minutes.  I ask because you haven't indicated that you have flushed the rail, and there may still be a bunch of crap in it. 

Have you added any Seafoam or Techron to the fuel yet?  It's possible that there is some varnish in the injectors causing some stiction on startup.  Cleaning it out may help the situation.

Have you tried spraying some starting fluid in the air intake when the engine is struggling?  If the problem is injectors sticking, the starter spray will make the engine run better. 

If the starter fluid has no effect, the problem may be an intermittent electrical connection.  This leads to a whole different troubleshooting path.

An air leak seems to be present. Is this possible. Air leaking in may cause the pump to cavitate. good luck.