Chasing the Hapy Electrical Gremlins (final)

Well, it may have taken a while, but with today's post, we are done with the re-wire of the "fuse box harness". I did not mean for this to become such a large endeavor, but that's what happens when you lose focus for a few minutes and cut some wires too close to the fuse-box for you to re-integrate them. I was asleep at the switch and missed my usual release time, so it's a few hours later than usual this week. Hapy last day of the worst US presidency in modern history; may the transition to the new guy be a smooth reflection of 200+ years of our democracy in action.

I realize that for many readers, this series has not been terribly interesting. I get that. Electrical sucks. However, if you were thinking about doing a swap, I think this series helps demonstrate how much of the swap work is done with a multi-meter, and how little is done with a shop-crane / engine hoist (or a floor jack).

Today, I will go over my steps to determine what went wrong in my system test in part 5. Apologies for no pictures of my testing. I'm not sure what would have been picture-worthy, except, maybe, my "fused wire", but that's not terribly interesting, I don't think. I mean, it's just 2 wires, each with a male wire connector on one end (to plug into relay sockets) and a female connector on the other. The two wires are then joined with a bladed fuse plugged into the female connectors on the 2 wires. So: male connector - wire - female connector - bladed fuse - female connector - wire - male connector. Easy peasy.

The T-12

Before I get into it, I realized after posting part 6 that I put a picture from when I was integrating the T12 wires into the front-to-back cable, but I didn't give any context nor explanation. That picture was illustrating that once I cut the cable, it wanted to spring back through the engine compartment and over the driver-side axle. Resisting that spring-effect does not help one make solid wire connections. So, I held the cable in-place with a vice-grip. Some of the ShadeTree solutions are just funny enough to be photographed.

Isolate

I mentioned in part 6 that I removed the diode between the RUN and START relays. This isolated those 2 circuits. I continued down this path by pulling all of the relays and fuses from the new fuse box, disconnecting the two always hot (30) circuits from the battery-top fuse box, unplugging the "D plug" from the alternator and removing the connection to the reverse switch. So, now, the engine and new fuse box are completely isolated. My plan to find the issue is from here, we slowly add things back in, testing each planned connection first.

Original Bus Loom

I started with the main bus. This wasn't touched, but still, if we can eliminate that huge variable, so much the better. So, I turned the key to run, flipped on the running lights and then the headlights. I even flashed the high-beams. Everything was working great at this point, but nothing in the old loom was changed. I did discover that one of the grounds on the front (front-is-front) of the dash was falling off, so I tightened the connector with some needle-nose pliers and kept moving. I felt confident that the test I had just done was not going to trigger a battery discharge. I did not have that feeling for anything I spent weeks modifying, so the rest of the tests below were much more careful. The next step is re-integrating the donor stuff.

New Fuse / Relay Box

My process was pretty much the same for evaluating and connecting the donor stuff and new fuse box: Identify the circuit, and before you connect it, check the amp draw. Pretty much, regardless of the reading, I then tested the connection by bridging the gap with a fused wire. I would select a fuse that was appropriate for the circuit being tested. In the case of the red always-hot wire going to the new fuse box (relays 109, RUN and START), I used a 20A fuse, and it didn't pop. To be fair, this test really only proved that this red wire was not itself going to ground independently of the key going to "run"... which wasn't happening. And, the relays were all pulled when I did this. So, this test was not as meaningful as I thought it was when I did it. I connected the circuit and verified voltage at the relay 109 socket 30,  before moving on.

Once I had the main power to the 109 relay, I popped it in place and the tested the RUN relay socket. Similar to the process I described above, I checked the amp draw between pin sockets 30 and 87. Then, I made the connection between them with a (20A) fused wire. Once validated, I only put the relay in once to confirm that turning the ignition key caused the same behavior. It did. After that, I did not plug in the relay simply because the ignition key is 15 feet away from this fuse box, and I just did not want to have to keep going out from under the carport into the pouring rain. Besides, this kept the testing isolated, as I continued to trigger RUN tests with a fused wire jumping from 30 to 87 in the RUN socket for the rest of the fuse box tests. I validated that socket 30 of the START relay socket was present. Yep yep.

I repeated this testing process with each of fuses 1-3 (11/15, 12 and 29 respectively) in the new fuse box, and then the connection with relay 109 before inserting relay 109. In each case, the steps were to check the amp draw first, then insert the fuse. Next, I repeated that process with fuses 4-6 (32, 34 and 43 respectively). I had a moment where my relay 109 was humming during the investigation of fuses 4 or 6 (34 / 43) which I tracked down to my not having had two grounds (one for the ECU and one for the main harness) re-attached during testing. Now, after the fact, I think this could have been a big deal, but as I worked through these steps, it was not apparent.

At this point, I had all 6 of the fuses installed and the 109 relay. When I bridged across pin sockets 30 and 87 of the RUN relay, the dash-pod would illuminate, the odometer displayed and idiot lights would come on. After a few seconds, most of the idiot lights turned off. The glow-plug light flashed though, and I realized that I had not yet tested the glow plug relay. I repeated the process for the glow-plug relay (180) and last, plugged the relay in. Once power was applied via the RUN socket, everything was the same except the glow-plug light winked out with the others. I had not yet wired in the always-hot (30) for the glow plug, though, so that was next. Same amp draw and voltage check process, same result.

Front-to-Back Run Relay

I figured testing the RUN and START relays was next. First, I plugged in the alternator "D" plug and tested everything again. As before, everything was as it should be. As I worked through this stuff, I suspected it was one of 2 things: either it was related to the "other 2 relays" I haven't really documented -or- it was because of this plug, since I split this signal to the front of the bus. So, my next test (of the RUN) relay from the ignition switch would be done first without the "D plug" clicked in.

 

This test was rather simple: stuff a standard Bosch-style relay ( I re-used one of the "53's" from the donor) and turn the key to run up front. I had already done this once before so this should not have been dramatic, and it wasn't. The donor electronics did the same thing as before. Feeling confident, I figured that the starter circuit would be just as seamless. That was not the case, and it was here that we found our cause.

On a side note, the relays with a "53" stamped on them appear in a few places in the TDI relay panel, but they also appear in the little relay box for triggering the coolant glow plugs. So, I figure these should be up to the task of carrying some decent current. Not that either of these do that, but the RUN relay needs to be on so long as the key is on, so I would think that takes some stamina. Oddly, in the 10 years since I first did this, the old Radio Shack basic 12V automotive relay never failed me.

Starter Relay and Starter

I started this test like the others: test amp draw, bridge the 30 and 87 relay sockets with a fused wire and see what happens. The starter grunted like it was gonna try to start the engine, so I quickly pulled the wire and figured things were fine. I put another spare donor "53" relay in the socket and turned the key to RUN and then to START. This time, what happened was what happened during my full system test: a pop followed by everything going dark. Concerned that the starter was blown, and that was causing the problem, I pulled and tested the starter.

I have an old battery sitting on my garage floor for things like this. I ran a set of jumper cables from the battery (negative to starter mount tab, positive to positive post on the solenoid. Then, with my fused wire, I jumped from that positive post to the start trigger and the start fired right up. 

Since the starter tested fine, I slapped it back in. While I was on my back working the long mounting bolts back in, I remembered that I had split the START signal at the ignition just like I had split the RUN signal. At the time, I had not yet hosed the wiring in the donor fuse-box with my aggressive wire-cutting, so I was intending to keep the original starter trigger from the ignition, but to also send a start signal to the donor ignition switch wiring so the ECU knew when a start was being attempted. This all made sense at the time, though now in retrospect I realize that the ECU doesn't need to know you're starting. Later, after this effort mushroomed, I connected the starter signal to the T6 like the donor engine was triggered, and forgot about the old start signal wire. So, now that none of that old stuff is being retained, that old start signal was going somewhere... and possibly shorting us out. 

Start Trigger

I very much wanted to test this theory. First, I confirmed that I had not screwed up anything new: get the battery hooked back up, verify it's at 12V or higher and reset for the prior test with the RUN relay, etc. Turned the key, and we're looking at a code-free dash-pod again. Nice. I turned off the key, and cut the wire from the start side of the ignition key to the original bus starter wire, and reset the key to RUN. Back at the fuse-box, I tested the START relay and the engine tried to start. Hoping that we may have found our root cause, I went back up front and turned off the key.

Remember how I removed that diode between the RUN and START signals? I confirmed that the RUN side of the ignition switch continues to deliver 12V when the key moves to START at this point. I simply measured the voltage at RUN when I turned the key to START. Since the START position was now only sending 12V down to the relay socket, nothing happened. We did still have 12V at RUN, so the diode is not part of the new wiring.

Back at the bus rear, I replaced the donor "53" relay into the START relay slot, plugged the alternator "D" back in and returned to the front of the bus. I turned the key to RUN... all the usual lights lit up. I turned the key to START... and Hapy started. He sat idling while I danced around a little bit. I did notice, however, that the battery light on the donor dash-pod was still on. So was the light on the bus dash. So, while Hapy starts and idles, some issues remain. Okay, we'll figure that out... Also, after a few minutes, Hapy's idle increased from bouncing around 900 to... 1200 rpm firm.

Next!

Ugh, the dreaded 1200 limp mode is still around. I'll be working on that next, as well as hooking up the reverse lights, and looking into those other 2 accessory relays that apparently were not part of the problem. I have some documentation to mop up with regards to the diagram, so I'll be spending a few evenings getting that done too.

I will have different blog posts and titles for all of that stuff (may not post on the diagramming), since this was really about simplifying the wiring. You could argue that this "Electrical Gremlins" series started because I couldn't get Hapy to start with any consistency, and now I can. So, that's it for today. Thanks, as always, for following along-

Chasing the Hapy Electrical Gremlins (part 6)

Continuing the saga of fixing the front-to-back (or original bus to donor engine) wiring, today I will cover some of the bus install joy. Consider that the positive battery cable has been disconnected this whole time, and the battery has been on a float-charger. I strongly suggest that you do the same any time you are playing with your electrical system.

Make It Fit

how it looked before

When I left off last time, I had accidentally cut wires from the donor fusebox that I ideally would not have. This led to a complete re-do of the fuse-box wiring harness. After spending my holiday break addressing that harness, and test fitting, I was about ready to start installing it. I plastic-wrapped the clumps of wires, and took the whole thing out to Hapy. I spent a few minutes looking at how things would fit and it quickly became apparent that I could have shortened the wire-clump going to the oval T10 by a couple of feet. Oh well. I stuffed the T10's, the custom T12 and the collection of leftover wires down through the hole in the bottom of my spare tire well. I should probably point out that when I first did this conversion 10 years ago, I cut a larger hole down there, approximately 1 inch across by 2 inches front-to-back.

With the harness pretty much where it needed to be, I removed the dash pod from it's mount and plugged in the blue and green connectors. The dashpod set back into place, and was bolted back in quickly, like end-to-end in 15 minutes. Next, I plugged in the computer and pressed it into place on the plastic holder/stand. The ECU really didn't move much once I had it set in, which, when we consider the thickness of the cables heading into it, really doesn't surprise. I will be applying bungy-ties eventually to hold it in place. This left the fusebox. I test fit it to make sure the wiring lengths were good and then let it sit so I could finish the wiring of the glow-plug relay.

Last Few Wires

how it looks now

During the harness rewire work, I left the glow-plug relay and related heavy wiring in the bus. I had cut the trigger wires (label-label-cut), but the hotwire supply and to-the-glow-plug wires were left alone believing that splicing wires that thick should be minimized. So, I had to re-integrate that circuit. The wires were clearly labeled, and long enough so the splicing was easy, and the relay holder clicked right into it's designated slot. All that remained above deck was tying down the ground ring terminal from the larger ECU plug; I leveraged the fusebox mounting screw after confirming it was a solid ground. The final result is pictured on the right here. Considering what it used to look like (top picture), it is hard to imagine these are actual before-after pictures.

Next, I moved down into the engine compartment. It was an unruly mess (see bottom picture). I had shoved the harness through the hole, and left everything else as it was. My first steps were to plot out where the various wires would go, and that started with the always-hot lines for the 2 donor relays (109, 180). The 180 supply line was a hideous hack I did when I had no money, but it worked. I have more resources now, but more importantly, I have a long enough wire of the right size. Once I determined how long a run I needed so the wires could run along the top of the engine compartment straight back from the spare tire well, turn 90* over the top of the engine hatch to the battery with some flex in the wires, I spliced in wires with ring terminals for the battery-top fuse box. Next, I moved to the remaining little wires for things like the alternator "D" signal and the coolant level.

wiring T12 to cable

A few of these little wires needed to be extended. I re-routed the coolant level 2-wire cable. It had previously run along the front of the engine. Now, it runs along the top of the engine compartment hatch like the rest of the signal wires from the engine. I wrapped the little wires into another plastic wrap, and tucked the cable into the wire loom holder along the top of the engine compartment over the hatch. I followed that with the power supply wires. Finally, I plugged in all of the T10's, the T12 and the 2-wire glow-plug harness. Believing everything was set, I removed the float charger, and started testing amp draws from the battery post to the battery connector. The draw could be measured in milli-amps, so I figured it was safe to put the battery cable on the battery and see where smoke came out. I'm kidding; with the care I took for the wiring, and the fact that 95% of the changes I made were on that one harness, everything should be fine. The only changes I made that were not on that harness were the always-hot supply lines and tying signals from the rear to the front for things like a low coolant light.

Of Course It Didn't Work

below deck unruly mess

Feeling relatively confident, I put the battery cable on the post. No clicking, no other noise. So far so good. Interesting... the dash-pod was not showing the odometer like it used to. Slightly puzzled, I went to turn the key to run. The starter tried to spin, so I knew I had something wrong. I pulled the starter relay, and tried again. This time, I heard a slight pop noise, and the battery discharged. Clearly something, or many somethings, are wrong. Neat.

Well, that's it for today. I am sure I will figure out what I did wrong, and I hope it will be a minor, simple thing. The wiring was well labelled, and the other modifications I made were small. I can tell you that I pulled the diode between the start signal and the run signal. I've decided that if I need to keep that, I will move it to the ignition switch and out of the fuse box. I will keep posting on my progress until I figure out what's wrong. I don't think it will take many hours of labor, but finding the time for those hours is the challenge. Thanks, as always, for following along-

Chasing the Hapy Electrical Gremlins (part 5)

Continuing the saga of re-doing the TDI-related wiring in the bus. Recall, this all started because I was not getting consistent signals at the computer from the accelerator pedal, 20 feet (7m) away. I got carried away with some wire cutting and suddenly, rather than just replacing the front-to-back wiring, I was replacing the fuse box. Today's post is about the wiring surgery.

Expose the Harness

I started simply: set the main harness on the kitchen table and then spend hours trying to figure it out. Many of the wires were still hidden within stretches of cloth tape, so that was the first task: cut that tape off. This is simple, but messy because that tape remains tacky. This black tacky ick gets all over your fingers, and sticks to things, like your kitchen table. My wife loves me. I found that Goof Off was very helpful getting the sticky off of the wires (and the kitchen table). Once all of the wiring is separated from the wraps, and wiped down, we can start thinking through the various wires.

Wiring

My first order was to consider how much length I needed between the different pieces. For example, the distance from the bottom of the fuse box to the dashpod is less than a foot. In between them, I have the pass-thru hole below deck. I figured that the cleanest look from up-top would be for only the connectors to be above. So, that moved the T10 connections below, but I wanted them close. So, I shortened the wires from the main harness to the T10 connectors so they were all close together. Into this mix, I
added what I labeled a "T12". I purchased a 12-pin male and female connector to go between the front-to-back cable and the main harness. This is so that I can disconnect it like the T10's. Once the T10's were at length, I started integrating the T12. Not all of these connections go directly into the main harness though. Reference the post listing the T12 here. Circuits 1-6, 8, 9 and 11 tie into the main harness. That left 7 and 10 for the next step: the fuse box.

With the final extra wiring removed, the T10's arranged and most of the T12 integrated, it was time to consider the fuse box. I had already defined where in the bus it would be, but it also needed to be located in the right place within the harness, so the wires were not stressed, nor flopping all around. I used the dash pod plugs as a reference point, and considered the plan above as I decided. Since the middle of the harness will be below-deck, it was okay for the wires to be a little long. Once the location was determined, I started thinking about the relays. The relay box has 6 relay slots, but 5 are designed for standard Bosch relays (30-85-86-87 pins, with a middle pin for 87a). That's great for my start and run relays, but not so great for the glow plug (180) or main power (109) relays. Those have particular pin patterns. So, getting them to integrate with the pre-formed relay slots would take some customization.

 

Integrating the Donor Relays

I had designated on my wiring diagram that relay slots 1 and 2 would be the glow plug and power relays respectively. In the original relay plastic, they used to slot into #10 and #12 respectively. Each of these original plastic retainers have a little cut out to allow the fit of a specific relay holder. Fortunately, #11 is exactly like #10 (while #12 is different). I realized I could cut #11 and #12 out of the original retainer, clean them up and super-glue those retainer into slots 1 and 2. With some careful cutting with a hacksaw, I removed the 2 holders, and managed to fit them into the fuse / relay box. In order to interact with the wiring for the fuses, I needed to clearance the top, front edge. Also, because of the super-large size of the 180 and 109 relays, as well as the size of the relay retainers, they are super-glued in along the bottom edge, leaving about 1/4" of overlap for the super-glue to hold. Still, super-glue is amazing stuff, and with just that little edge, these original retainers are absolutely held fast, even when I put a little wiggle-pressure on the wires.

Fuse by Fuse

With the original relay retainers in place for the 2 big relays, I was ready to get after the fuses. Now, recall that I had this bus running with about 7 of the original fuses still plugged in. I discovered, however, that I really only need 4 of them: 29, 32, 34 and 43. I still had fuses 11, 12 and 15 plugged in and wired up, though. I have ported all of them over. Former fuses 11 and 15 share a fuse in the new box (fuse #1) because they appear to not really be necessary, but the wiring was there, so they are wired up now. Fuse 12 (now fuse 2) is to supply power to the OBD-II port in the wiring compartment. I expect this to be consumed almost never.

 

The other 4 had been clearly marked during my label-label-cut process. I had left myself plenty of wire so I did not have to extend any of them, which was very fortunate. For all of the fuses, the "consume" side is the front or furthest-from-the-relay side. As I wired in, I set the wires down so they would exit the fuse / relay box towards the front. Next, I did the supply side of the fuses, wiring fuses 1-3 together and then over towards the "RUN" relay and fuses 4-6 together and then towards relay 109. In the wiring diagram, fuse 29 gets it's switched power from the ignition, not through relay 109 so this is a correct reflection, though a little weird. I think it is through this circuit that the ECU is notified that the key is in "RUN" so it should trigger relay 109.

Relays

So, the fuses are done, the relay retainers for 109 and 180 are super-glued in. Next, we get the other relays wired in. I actually did 4, but I am only going to write about the "RUN" and the "START" relays. The other 2 I'll get to later. RUN and START are basic Bosch-style relays. A signal arrives from the ignition switch into pin 85 or 86, that signal difference is compared against a ground at the other 85 or 86 pin. Once triggered, pin 30 is connected to pin 87. So, for our purpose, the pin 30 for both of these is connected to always-hot, the 86 pin is grounded, the 85 pin wires into the T12 (7 for start, 10 for run). Pin 87 is where things get interesting. Pin 87 from the START relay connects to T6 to send a start signal to the starter. Pin 87 from the RUN relay connects to fuses 1-3.

 

Similar to the relay set up I had before, I wired in a one-way diode between the START trigger and the RUN trigger. Recall, this is because some ignition switches will not send a 12V signal down the RUN circuit while sending one down the START. When this happens without that diode, the second you try to start the engine, the computer thinks you turned the key to off, so it won't start. This diode allows 12V across from the START side to the RUN side, but not the other way. With a multi-meter, you can check this: negative on one and then the other, you can read connectivity, but only one way. Very cool. Last, the reverse light signal will be integrated into the switched circuit, but that is the only bus-related cross-wire (with an integrated fuse. of course). I did that because switched power was needed, and the consuming wire for it going to the switch on the transaxle was less than a meter away.

 

Last, I did the wiring related to relay 109, leaving the always hot circuit 30 to be determined once I got back out to the bus. Without that all the extra relay stuff, the hot wire had been cut out. I can now see that the actual demand on the 109 will be minimal (like 30A) in my implementation, so I will be replacing the old 2ga supply wire with a smaller one.

 

Once I was satisfied with the kitchen table wiring job, I took the whole contraption out to the bus for a test fit. Everything looked workable, so I brought it all back inside to wire-wrap. I used that plastic tube style available at Harbor Freight for all of the wiring clumps that run from a plug to the main harness, and from the fuse / relay box to the main. This left a few wires that still have label tags on them for things like the alternator signal and the coolant level. Both of those (alt and coolant level) are related to idiot lights I wired into the 6-wire cable when I did the oil pressure stuff (see Oil Pressure and Temperature - Part 4) so I will deal with those once I get out to the bus. All of these little wires have a partner in the engine compartment, so I will solve for them as part of the next phase: install.

That's all for today. This literally took me days. After many years of working at a place that shut down between Christmas and NewYears, I have grown very accustomed to having that week. In past years I have redone the seats in the MG or rebuilt a front end. This year, I spent most of that week doing the wiring, fuse / relay box stuff I just described. Yeah, this is a serious time killer. But, if Hapy starts on the first try, doesn't drop into that 1200RPM mode ever again and the wiring is no longer an eyesore, then it was all worth it. Thanks, as always, for following along. Please keep wearing your mask. These CoViD numbers are still very scary--

Did I Stutter?

As I expected from all this daily driving, a new issue would arise.  Today is all about that.

Stutter
On the drive into work on May23 the engine suddenly shut itself off 3 times.  First as I was pulling onto Walker Road off the 217, again on driveway off Jenkins and last in the parking lot while I puttered looking for a spot.  Some quick research showed me that the main power relay (Relay 109) really sucks on these ALH TDI engines.  Apparently, folks buy them in bunches, and keep spares in their glove box just in case these things happen.  So, I ordered a couple from IDParts.com ($10 each plus shipping).  In my classic impatient way, I bought another one at Discount Import Parts (DIP) on the corner of Hall and Scholls Ferry Rd (Progress exit off 217).

Relay 109 Symptoms, but not 109
Thursday night, I tried swapping the relays, but the same effect happened.  Friday night, I tried some additional diagnosis with Boo listening while I tried starting / runing him.  As we talked about possible causes over breakfast this morning, she hit on it: it starts fine, but when you try to just run him, he dies.

Dang, it was me!

solenoid over zebra
striped rear of bus

I drove back to the bus this afternoon.  I grabbed one of the ignition relays that I put in to spook the ignition key action (see Inch by Inch) and headed for Radio Shack.  I got another relay there, but their supply of wire ends has turned to crap, so I hit O'Reilly's for wire ends (I got some heat-shrink too, but forgot electrical tape).  Back to the bus.  I started un-taping the wires from their respective ends and found 2 loose ones.  First, was the juice from the battery headed for the "run" relay.  Second was the connection between the 2 relay send signal wires and the one-way solenoid.  That one-way solenoid sends juice down to 2 of the 3 major spots, so this was a big deal.  Once these were repaired, the bus is road-worthy again.  I intend to return to the bus with the heat-shrink (and a lighter) and really seal up all of those connections.  With all the use the bus has been getting, it is crucial to have the wiring done for keeps.  This serves as a good reminder to myself: fab with the long term in mind.  Don't just do enough to get it to work.   Do it right enough so that it will last a while.  I put it that way because it is easy to over-engineer things for "lasting a lifetime", when you're just going to change your mind about things a few years from now.

Well, there's still daylight left, so I'm gonig to take a celebratory drive down the way to Taco del Mar to celebrate with a burrito.  As always, thanks for following along and more next time...