If you look at the throttle switch carefully, you'll see this:
1) An idle switch (activates the special hand tuned idle mixture circuit, really, this hand tuned idle mixture was the entire reason D-Jet was used, made the idle clean enough to pass the early 70's emissions tests)
2) A switch that closes as the throttle opens (via friction), it opens as the throttle closes
3) Once that is closed, it slides across an alternating 'sawtoothed' set of contacts. Two circuits, one for each of the two injector batches. As the contact slides across, it's basically adding an extra firing cycle, the same thing the distributor contacts do.

So basically, as you put your foot down, you just add in fake 'engine rotation' signals, fires the injectors more, acts as a fairly low tech acceleration enrichment mechanism.

If the injectors aren't firing when the engine is running, and they're not firing when the throttle is opened, that (within the realm of reasonable probability, i.e. simultaneous unrelated failures) eliminates the distributor contacts, and the wiring to them, as suspects.

It likely means that the ECU is receiving signals to fire that batch of cylinders, and is either unable to send the signal (internal failure*) or the signal isn't getting there.

Later model injection systems generally have the car supply 12v+ to one side of the injectors, and the ECU grounds the other side of the circuit to open them, but D-Jet does it the opposite way. The ECU sends out switched current from the ECU (not 12V, I think it's 5V, or less) to one side of the injector, the other side is permanently grounded. On the 4 cylinder B20E - each injector has it's own individual wire from the ECU, even though they're fired in batches of 2. And all ground wires are grounded at the same point (on the manifold).

Here's a 164 injection wiring diagram:
http://www.164club.se/service/el75.jpg

Looking at that you can see how almost comically clumsily they modified the 4 cylinder ECU to handle 6 cylinders. Along the top edge of the picture, just left of center, around the larger '81', is the injector wiring. You can see that the ECU still has 4 outputs (2 batches). However, two of the wires (#4 and #6) go to single injectors, as in the B20 installation. The other two outputs (#3 and #5) each fire TWO injectors, via a split in the wiring.

Now, that wiring diagram also shows all 6 injector grounds going to a common source. Without really looking at the wiring loom itself it's hard to say if all the wires are really spliced into one before the ground eyelet which is bolted to the manifold or not.

But, just from looking at that wiring diagram, the whole wiring aspect is starting to look a little dim. I don't think a wiring fault could produce the symptoms described.
1) Since the distributor triggers and throttle switch triggers both fail to fire that batch of injectors, it's unlikely to be a wiring or trigger switch issue
2) The hard wiring scheme doesn't seem to support any way in which 3 cylinders in a single batch could be incapacitated. 2 of them, but not 3. There's a faint possibility that they're also grounded in the same grouping as the batches, but there's certainly no reason for them to have wired them that way.

It really does seem to leave it up to an ECU problem, where it just isn't firing one of the batches because of an internal fault.

They really are large scale, non integrated circuit, analog electronics though. There are probably people about who know how to fix them. And probably people who haven't really ever seen a D-Jet computer, but who know electronics well, and could have a good chance of cracking the case open and finding what resistor/capacitor/transistor failed, and replacing it.

Although a cheaper bet is still probably just finding another 164 ECU. They're rare as can be, but they also seem to rarely fail, so it's a case of low supply teamed with low demand - they're not likely to be ridiculously expensive.


And worst case scenario? Replace that cake pan ECU with a programmable Megasquirt ECU! That has certain other benefits.
--
'63 PV544 rat rod, '93 Classic #1141 245 (now w/16V turbo)