So I had to do vcg's on the '95, and of course the no-brainer is to replace the spark plugs during that job. But after 24+ years, all the underhood wiring is getting very stiff and brittle. I'm careful as can be, but knowing that the subharnesses for the ignition coils are no longer available even from Japan, I decided to build my own.Then I thought....maybe I should go with a more modern CoP 'smart coil' approach.
The VH45DE came from the factory with 'dumb' CoP packs, meaning the ignitor (really a switching transistor) is external to the coil module. So the VH45DE ignitor (also used on early SR20 DE as an aside) is not needed in a smart coil setup, but there are some harness considerations, more on that in a bit.
Performance VH advertises some kits for this, and I quickly decided on the Volkswagen/Audi Group (VAG) coils due to some previous experience with my A6 Avant. The ubiquitous 'long' Audi coils are too tall to use with the factory VH45DE ornamental covers, and I'm anal about keeping as stock a look (even if it's not....) as possible. So that meant going with the VAG 'shorty' coils P/N 036 905 715F. The deal with these is that they never appeared on cars sold in the US, so you'll likely have to source them from overseas. I took a chance and bought some from fleabay, they look like a genuine article, with most of the correct markings, however I suspect they are also counterfeits, because they all read "WAG" at the top instead of "VWAG". They seem to be working so far, which is only about 150 miles. I haven't noticed any negative side effects, and the off-idle throttle response is a little better.

OK, pictures!
Before:

Centered is the old ignitor, and the ornamental cover. You can also see some of the coil modules subharness "under" the active pipe and secondary TPS. Yeah, Active/TCS Q's are an even bigger pain to work on. Not to mention my '94 which also has the 'air cut' valve under the TB and on top of the ignitor....

My mockup board for laying out the new wire harness (an old valve cover):

The 4-pin grey connector I bought from Wiring Specialties. The 4-pin coil connectors for VAG are all over amazon/fleabay. The black barrel connector (single pin) is 12-ga Deutsch DTHD (got from 'pro-wire USA').

Partially installed after the new vcg and plugs had been replaced:


My wiring schematic:

Pin 1 ECU ground (green)
Pin 2 Engine Ground (Black)
Pin 3 +12v from Ignition switch (Red)
Pin 4 ECU trigger (White with color stripe)
Note that this pinout is specific to the shorty coils! The pinout for the long VAG coils is different. Why on earth they had to change escapes me, but it caused me a lot of wasted time and angst....

Here's what I did to bridge the ECU side 5-pin factory connector to the new set-up:

Remember I like the stock/stealth look....so I gutted an ignitor and soldered straight wires connecting the ECU side to the coil side. Note also the black wire which goes to ECU ground. The 4-pin VAG smart coils have an ECU ground (pin 1 from above), so I needed a way to tie that signal back thru the harness to the ECU ground. Some people would not bother with this step, thinking that to tie the coil's ECU ground to the engine ground is 'good enough'. In my experience signal grounds and power grounds serve different functions, but the choice is yours.
After that photo I also 'potted' in the backside of the ignitor shell with E6000 as an added measure of keeping the firing order from potentially getting screwed up

And the near final product, with the dummy ignitor and ornamental cover installed:

just between the active pipe and the secondary TPS you can see a loop wire. This is the #1 cyl ECU to coil signal, which I intended to use to set base timing with an inductive timing light, but the signal level is too low to trigger my 35 yo craftsman timing light. may work with other models.... I ended up timing it by placing a length of high tension lead between the #6 coil and the #6 plug, and clamping on to that.
Note that using 'method B' on page EF/EC 34 of the FSM, while novel, uses the 12v line, not the ECU trigger. My harness loop attempts to provide a way to use 'method C' on EF/EC 34, but recall that the signal is now ECU level, not transistor level from the ignitor as in the original design.
Anyway, it all works. Will this setup last another 24+ years and 150k+ miles? I doubt it...but at least I should be able to buy replacement coils pretty cheap.....

Wow!

3Q,

did you do anything to bolt the coils down using the factory holes or have you had no issues since install with these coming loose without any sort of hardware keeping it in place?
I'm nearly positive that at least one of my coils are defective (getting P0305 code) and am interested in doing something similar to avoid parts sourcing 25 year old coils!

The CoP's are friction fit only to the spark plugs/valve cover thru holes.
I look at it as similar/better than the old days of HT leads to the plugs.
Interesting that you have the P030x code. Shows that OBD-ii has many more parameters. My OBD-i G50s will never give such a code.....
Don't get overly committed to it being specifically #5 coil (P0305). The OBD-ii algorithm can really only detect certain things to make judgement (frex, it has no way to know from EGT, or O2 sensed AF ratio *which* cyl is actually misfiring). But its certainly a place to start.

3Q Jay wrote: ↑

Mon Jul 19, 2021 7:20 am

Don't get overly committed to it being specifically #5 coil (P0305). The OBD-ii algorithm can really only detect certain things to make judgement (frex, it has no way to know from EGT, or O2 sensed AF ratio *which* cyl is actually misfiring). But its certainly a place to start.

Actually it does on newer stuff, but I'm not sanguine making blanket statements about old ECU's with slower ECM chips. On anything about '02-up the ECM is actually measuring acceleration of the piston after firing, by way of the crank sensor. The difference from a piston that has no or low power is many microseconds, an eternity for a processor running at 40 MHz. So when you get a specific P030x and not a generic P300, you can be very sure of the ECM's diagnosis.

good info VStar. Technology over time can indeed improve our troubleshooting capabilities
Since his is a '96 (first year/only year) for OBD-ii on G50, I'd still be skeptical, but absolutely acknowledge that it may indeed be #5. Some other personal anecdotes on P030x codes from early OBD-ii--
case 1: 1995 U13 Altima. Threw P0302. Indeed I found some moisture on the HT lead inside the valve cover sleeve for #2. But all 4 were wet to varying degrees.
case 2: 1997 FGY33. Threw P0304 (pending). Symptom was an intermittent stumble under partial load. Replacing #4 coil did not resolve. But replacing all 8 did. My suspected root cause was the leaky CoP boots so common to older VH41DEs (my separate crusade to use a little dielectric grease on boots/porcelain (not the electrode terminal)).

3Q Jay wrote: ↑

Mon Jul 19, 2021 10:21 am

Since his is a '96 (first year/only year) for OBD-ii on G50, I'd still be skeptical, but absolutely acknowledge that it may indeed be #5. Some other personal anecdotes on P030x codes from early OBD-ii--
case 1: 1995 U13 Altima. Threw P0302. Indeed I found some moisture on the HT lead inside the valve cover sleeve for #2. But all 4 were wet to varying degrees.
case 2: 1997 FGY33. Threw P0304 (pending). Symptom was an intermittent stumble under partial load. Replacing #4 coil did not resolve. But replacing all 8 did. My suspected root cause was the leaky CoP boots so common to older VH41DEs (my separate crusade to use a little dielectric grease on boots/porcelain (not the electrode terminal)).

Like I said, I'm not at all confident that early ECU's, which were very slow compared to modern stuff, could use the modern techniques. One thing I will say is that the modern ones all behave comparatively. If all the cylinders are down on power, say, from a 1-tooth timing chain jump, they won't produce misfire codes. That happens only when one or two are significantly off from the others. So while I don't know the exact algorithm Hitachi uses, I can tell you empirically that it's definitely comparative and not absolute. With an older ECU using indirect methods of measurement, I suppose in a situation where all the boots were leaky and all down on power to some extent, it might well flag just the worst offenders.