Isuzu G-series Twin Cam

I'm having a few "issues" getting my G180W running correctly, cylinders 3 and 4 are running irregularly, not a constant miss, but not consistently firing whether at idle, accelerating or at partial/full throttle openings. It's an EFI model and I'm new to EFI. The spark is fat and healthy. Compression test shows 160psi on all except 3 which is 150. Leakdown shows 8% an all except 3, which is at 14% (and leaking through the inlet valve). Once I get the tappets (shims???) sorted - hopefully all will be well mechanically, in any case the problem is with 3 AND 4 not just 3. I have written ISU069 about this and he has given me an extremely detailed reply, following this, I have decided that an upgrade to a Delco digital ECU and wiring harness is probably a worthwhile exercise, as well as a dizzy upgrade. Can anyone tell me what model Camira/Gemini/Commodore (or whatever else) I should look for to source the needed parts? Also are there checks I can carry out to see if the ECU is ok before I buy it?

Obviously the injectors may be partially blocked and I'll check this out anyway, but the tuneability of the more modern ECU sounds like it is worth it in any case.

Thanks very much.

The compression etc should be more than adequate. It may simply be an injector problem.

Poida fitted a Delco to his Gemini (blue) G180W. I think yours was a green G180W, but that shouldn't matter.


Though Delcos are commonly taken from the JE Camira (the last model Camira from 1987 which finally overcame its problems; usually 2L but also 1.8L), some prefer the Nissan (N13?) for its better or longer wiring (more options wrt where to place the ECU).

The 2 main essentials are to grab one from a matching transmission - either manual or auto - and take the harness with all its sensors and remount them onto the G180W. Once exception is the dizzy where only the HEI (ignitor) need be obtained - more info below)
Preferably also matching in capacity (1.8 vs 2L) though that shouldn't be important, especially if you remap your Delco to better suit the G180W.
I'm unsure about the injectors, but Delcos usually handle both hi-Z and low-Z injectors (addition or deletion of an injection driver component - I've forgotten the detail and it may simply be the addition of the hi-Z's injector resistor/ballast (which is likely to be the same as on the G180W) BTW, Isuzu use low-impedance injectors.

As to the dizzy, the superior RG dizzy is recommended. Hence no grafting of the older-style and inferior Camira & Nissan dizzy tops to a G-W dizzy shaft.
The electronic RG dizzy (RWD Gemini or any G-Z points dizzy fitted with FWD RB Gemini upper shaft and breaker-plate with reluctor sensor and (preferably) its blue ignition-coil) is easily modified for any G-series engine including the G-W twincams.
For a Delco however, the RB ignitor need not be obtained nor mounted,
Instead get the Delco's ignition module (aka HEI-4?) off the (N13, Camira etc) distributor. The RB's relector sensor takes the place of the Camira/N13 multi-pole sensor (with the red or brown wire to the + input, and grey or white to the "-" input if the HEI (ignition module).
The sensor wires can simply be extended to the HEI which can be mounted on the body. Ordinary wiring should be ok, maybe twisted, but it is unlikely to require shielded cable (unless too close to the HT ignition, especially CDI - but that usually isn't worthwhile in comparison to the RG dizzy with RB ignitor and blue IgCoil).


But really Poida should give you the good oil. He can explain exactly what he did and how it's going.
He is also aware of the cheaper tools available for remapping the ECU.

Thanks for that IZU069. Is Poida still active? I'd like to discuss his modifications if he's around. Thyanks Again

He's still alive...
But due to other crises, it seems scarce.

Keeping in mind mine is a blue top Gemini G180W, I found the engine wiring loom from the LD Astra/N13 Pulsar 1.8 litre MPI setups (they are the same setup rebadged) was more useful. The wiring in the engine bay is nearly identical to the 2.0 litre MPI JE Camira but the length of wiring from the bay to the ECU was a lot shorter in the Camira so I opted for the longer Astra/Pulsar loom. You'll still need to do some adapting though. I have literally relocated nearly everything electrical in the engine bay to suit the different engine and not have a clutter of electrical bits.

I use a HEI dissy housing that had a similar multi notch optical setup to what I have seen in 4ZE1 engines that I suspect is a G200W unit but I turfed the optical gear and use the reluctor trigger and pickup from an RB dissy. You will need to set it up as a fixed shaft with no vacuum or mechanical advance, the Delco takes care of all advance. I run it through an EFI module so it controls the spark and timing in start and limp modes. The programming manages the spark the rest of the time. Basicly you set it up the same as it is on the donor wiring loom car but make your own reluctor dissy.

To get a Delco tuned you will need to invest in a few items but they are not overly expensive and there is a heap of info on http://www.DelcoHacking.net forum. You will need a EEPROM chip and adapter to replace the original MemCal program chip and an interface to link to your lappy to real time tune. Swapping to a wide band O2 sensor for programming helps heaps as well. Refit the standard O2 sensor when you are happy with the tune.

I use an E30 BMW fuel pump bracket, anti hammer valve and surge pot with the pump from the E30 as well. A bog stock Z200 fuel filter does the job and also fits in the BMW bracket clamp. I mounted the whole lot in the space behind the rear axle on the back wall of the diff hump. It's below the tank pickup so no need for a primer pump, just make sure it's full of fuel before starting it first time. Replace all your original fuel hoses with EFI rated hose and clamps.

I used none of the original EFI parts except the injectors, fuel rail and pressure regulator. Everything else is adapted from the Delco system, including the throttle body from a VN Commodore. You need to fit everything the Delco expects to see or it will give error warnings. You can switch stuff out once you have the tune gear on hand. I would highly recommend buying a genuine Holden JE Camira manual or a VN/VP V6 Commodore egine manual or an LD Astra or N13 Pulsar factory manual. They are the bible for understanding the workings of the Delco system. Everything you need to know is in those manuals.

Im a bit late to this thread, but I am running delcohacking.net gear on my g200w in my gemini, and recently at the bellett nationals got the same tuning running on a bellett gtr with only very minimal mods. It seemed the VE table (pressure vs air map that is used for the basis of fuelling) transferred fine as the heads between the g180w and the g200w are the same. Really on this car I just disabled the speed sensor error (as it had none), adjusted the injector rate to compensate for the capacity changed, and increased the entire spark table 1 degree at a time on the street until it felt like it doing some good power. Then, It had some trouble with high idle, so I watched the idle air solenoid on the laptop, closed up the throttlebody untill it was at about 30 steps, then cut the return spring on the throttle down a little to tighten it up as it wasnt closing the last fraction. The car was sweet, and the owner was stoked.. called it his "new car" after that. Haha.

Couple of pics, and the tune here:
http://gemisa.net/forums/viewtopic.php?f=24&t=3630

Im also on delcohacking (im admin there) under the same name if your still keen on this project.

Note the delco wont drive low impedance injectors (low-z) but camira/pulsar/astra/commodore injectors swap in fine.
Also, I probably would not use the VN throttlebody, as its pretty large and is likely to flow too much air at idle, requiring mods to reduce the airflow down enough to get idle happening ok. I'd go the TB from a 1.8 or 2.0 camira/pulsar/astra. Having said that, if its working for poida....

You can also see some details about my g200w here (though its a bit light on about the efi - note that although it started off as kalmaker, its now ose12p from delcohacking): http://gemisa.net/forums/viewtopic.php?f=22&t=28 or some video of it at a local hillclimb here: http://youtu.be/vJeGPy4sAmA

I was discussing the lo-Z issue with Poida recently.
I did have a schematic showing that both loZ & hiZ could be used. I though that was for the 808 etc but can't be sure. Maybe one day I'll find the info again.

I don't currently have the info in the 808's switching MOSFET. As I recall, it was capable of driving lo-Z injectors. (Or was it a transistor, and am I confusing my VN Delco??)


Thanks a lot for your info, I'm sure it will put many prospective Delco migrators at ease. (I think Delco is winning. I'm sick of pointing out the shortfalls of other aftermarket systems... multiple maps indeed!)


Nice looking GT too! And great to see the RG dizzy - the last GW Bellett I saw (at a BCCV meeting) had a graft dizzy - and it was a Camira/VN top to boot (as in "poor choice" or lots of expense & trouble).
Now, as to that single-circuit brake master...

Thanks Antus. Good info and pics.

The problem is that they lack the peak and hold circuitry. Kalmaker used to make a low impedance driver (and I have one), and have used it with success but its covered in tar and I cant see how they did it. Its not really worth the hastle of hacking ways around the problem though, as for NA engines there are heaps of options available from the wreckers, and if you want to start doing serious power the 60 or 90lb siemens deka injectors on ebay are high impedance, work well and dont cost the earth like brand name aftermarket jap injectors tend to. the delco can run single fire at low flow rates (programmable switch point) which means it can fuel at idle happily even with the 60 or 90s.

Actually after writing above I was wondering if that Delco used injector drivers. Man - it's all so long ago. But then IMO it should be one driver per injector.
But peak and hold shouldn't be too difficult even as an add-on - eg, after hitting ~16A drop back to ~4A (for batch fire!). For sequential I'd get a bit more sophisticated, and that's probably a mere $4 PIC or two.

Years ago I said I'd never use lo-Z injectors - not that I recall the specifics. However it was emphasisied again when an "expert" was describing that Haltec monitored voltage to compensate injectors for voltage drops. Not that I necessarily trusted his reasoning - despite his qualifications he was obviously unaware of what ECUs were capable of (and as I found later, had been doing to some degree at least 5 years earlier) as well as simple inexpensive circuitry to solve such voltage problems.


Anyhow, it's not an issue for me at the moment. And since cheap hi-Z injectors are available - especially as many might update to a more modern fuel rail to get rid of the old hose-clamped injectors - it's not a big deal for them.


When (or if!) I find that Delco lo-Z info I'll update this thread & post. (Lo-Z required resistors/ballasts to be added of course! But they are common, else four x 6R8 or whatever they are. Or maybe better still - current limiters.)

yeah, but the advantage of low-z is the quicker opening time, if you add resistors/ballast you loose that, so you might as well just run high-z to begin with. The default delco calibrations normally have acceptable compensation tables in them, and even when you do an injector swap your usually fine with the default compensation (though at about 10v other things in the car start struggling). The default opening delay tables are close enough for any injectors - but if you do have data about opening times from the manufacturer you can use it. And then there is the single/double fire if you go for large injectors and idle injection time starts getting so small that the injectors cant accuratly deliver... set the ecu to fall back to single fire which doubles the opening time, and stops the injectors needing to do the impossible.
A bit more info that backs up the '808 cant do peak and hold here: http://www.delcohacking.net/forums/view ... 45&p=22274

That link refers to the 749 but I assume that that does do sample & hold from its reference to sense resistors that need to be upgraded (ie, to R05 2W). Is the 749 an 808 equivalent? All I remember is "not that number".
I thought it was a 126 or 162 or ...
... ah...

Maybe the 1227165? (One of the few cct diagrams I still have...)
The 165 cct shows a 2W ~R1 sense resistor for U12 (16034984) that is between GND & INJSENSE#. U12 INJOUT switches Q1 INJ# to INJGND#. (Is INJGND# connected to INJSENSE# for Lo-Z?)

Anyhow, that too suggests peak and hold capability.

Not that I care, it's not a current issue (gotta love them puns eh?). And when the time comes, I'll recover my old info else confirm with others.
Besides, there is a relatively simple add-on if it isn't P&H capable.


Incidentally, my Isuzu Guru used twin injectors on his G180W Gemini Coupe, but that was long ago AND in his Haltech days. (He also later discovered the Delco!)

But I'll leave it there. There is enough info in this thread for GW conversions to Delco.

Cheers.

I couldnt link directly to the end of the thread, but in in the last post vl400 who is the main 12p delco software developer states that he could not make the '808 do peak and hold on the bench. He said it just sucked 8 amps no matter what and never reduced the current, even when he tried to electronically convince it to do so with the mentioned tricks. Since he has programmed 12p, and the fuel controller at chip level, I take that answer as absolute. For the record the '165 is an american ecu that is idential to our '808 except it has the high speed comms chip on it, where as in australia they removed it (and we add that functionality back on as its really useful for tuning.) 12p does run on the american 165s with high speed coms with the ecu in stock form, if you can find one. They are no longer so common there.

Strange that the sense resistor exists.
If the ECU can read that voltage value, then it should be possible to code.
If not, it may be a combined hw & sw mod.

And if still no to all that, then external circuity - which may be easier anyhow. I don't see that the ECU needs to know - especially if it isn't monitoring the current profile already. Just update the parameters, else tables.

delco/holden/gm cheapened out with the 808, no low-z, no sxr uart chip for high speed comms, lower quality (but still acceptable) pcb. Thats why the design looks like it might be possible but doesnt come up with the goods. Also note that its not voltage - its current, which is a lot harder (read more expensive) to control. Sure, anythings possible with enough time and effort, but as there are pleanty of cheap and capable hi-z options, I still consider the answer to a newcomer to efi to be "high-z only". Then if anyone wants to take on the challenge and make low-z work properly (current control), feel free. The geeks will applaud. If it works with ballast so they open slow and dont draw current with or huge power transistors that can run them with 8 amps solid for their open time instead of giving them a break once open when they should be hitting the 2amp hold state, not so special or correct. But it does save getting a set of garden variety injectors off nearly any random wreck with comparable individual cylinder capacity and the right connectors. Most the OEM injectors at the wreckers are bosch who have been good with providing flow rates for pretty much all their models, and the mount options are standard issue so its not a hard swap.

The hi-Z has higher inductance; the lo-Z should hence open quicker.
The resistors are only to drop the current to the hi-Z's normal required current and doesn't slow the current's rise time, but sure - overdrive it without resistors and it will open quicker.

But hence why I prefer current limiters (two linear components). The P&H can be added to that by various methods, or the lot done using "digital" methods.

And these days transistors would not be used. MOSFETs that sink >80A are cheap, and easier for straight switching applications. (I thought the 808 used FETs and not transistors - as per your link. It's an easy mod anyhow.)


I was only aware that Holden omitted certain components like the UART, but not that there were any significant firmware changes since that would add cost (much like the move from the locked optical dizzy & mapped ignition of the G200W I-TEC Piazza to the mechanical dizzy of the 4ZCt-1 of the Aussie Piazza, though as discussed on Bellett.net, that was seen more as an inability of GMH to handle CPUs etc).
Nevertheless, with no PCB changes and most other components in place and (AFAIK) of the same values, surely the locals can be updated to the 808 firmware?
But I see no reason why the software can't be modded given what I said earlier. And if the 808 did have P&H...

But as you said, the hi-Z works, is plentiful, etc. I merely refer to my "traditional" view that I would use lo-Z rather than the OEM-cheaper hi-Zs. Plus whatever Delco publication I saw showing the option of lo-Z with the mere addition of the ballasts (which I thought was 808, but can't be certain; but Delco tended to have all possible features designed from year dot - as I recall they were very thorough in their original research etc, hence their exceptional versatility and performance as many have since discovered).


Are geeks now people with simple circuit abilities? LOL - it used to be those into software!
But like Haltech's voltage compensation, IMO some things are better solved with circuitry. If not then, certainly now with the mass availability of cheap power solutions.

You mention software, but the delco is so old that it has a hardware fuel controller, and hardware spark controller on board. The main cpu only runs at 2mhz. It samples the inputs, calculates everything it needs to, sets the direct output (fans, tcc if auto, nitrous or water injection if your running 12p and need it haha) the it offloads the next fuel event and spark event to the external hardware. That is why low-z and current limiting is a hardware thing. The cpu only has control over the injection via the FMD chip. Running ballast will limit your current, and thus will reduce the injector opening time. People do it, and technically it does work. But if you do, remember that any factory injector opening time data you may have will no longer be correct.

I don't understand the last - factory injector opening times are for the particular injector whether lo-Z with resistor or hi-Z vs voltage.

And a lo-Z (with resistors of course!) still opens faster than a hi-Z. (Provided they are the same technologies - ie, a newer hi-Z may open faster than an old lo-Z).
POST EDIT - a quick google and it seems the latter has changed - hi-Z is now more favoured inter alia for its faster opening times and lower heat generation. Hi-Z certainly was the inferior when I was involved, but that was last millennium except for a short foray circa 2005. And as to why I selected the lo-Z, well, that's back with my old stuff as well (read: as yet unrecovered data).

And if the 808 had P&H implemented in hardware, you are saying that that Aussie hardware has been changed, yet it still has the sense resistor? IMO that's strange. And unexpected.

But if the CPU is manufactured at mask-level, I can understand why $FFFE can't be remapped and different firmware/software implemented. Not that that negates any hardware hand-off.

But all that should be answerable on delcohacking, though I know some there are very hardware or circuit limited. (The last I recall was someone discussing a single injector resistor instead of the usual single-ended parallel combo! Hopefully that was later rebuffed - diplomatically of course.)

And it should be a new thread. It's superfluous to this OP, though IMO his questions have been answered and hence it may be ok to continue here. But that too IMO is useless without appropriate 808 software and circuit detail.
And the whole thing is pointless if the hi-Zs are now the better product for whatever reasons. And if voltage is still an issue, I might still add an SMPS for constant voltage with a fallback to raw 12V upon failure, but that's another historical issue. That is no longer complex nor expensive (say $12? DIY; or $20 for an ex-USA standalone?).

more power means more magnetic force. power is watts and watts is volts * amps. So assuming the car voltage is sitting at 13.8 (for example) and a hi-z injector is pulling 4 amps, and a low z is pulling 8 amps when it opens, then the low z should have twice the magnetic force moving the pin to open the injector, thus faster. But due to extra heat, then you need to drop the current to 2 amps which is all thats needed to hold it open. Put a resisor on there, and you reduce the power to half or so at opening, then waste power when your holding. But overall you dont cook the injector or the drivers. But you get half the power in the coil and half the magnetic pull.

For the 808 - yeah they left resisors and capacitors on the pcb and removed others. They dont always remove the interconnects, just the main ICs as they are a bigger expensive on the producion line.

The cpu is not masked, the code is in the memcal and it is replacable. It jumps to what ever address is stored in the reset vector at 0xFFFE. The cpu does not controll the drivers on GPIO though - it tells the FMD IC how many milliseconds to open the injectors for, and when, and then its up the FMD IC to do it. If the FMD IC doesnt have the abaility to do current control, then you cant do peak and hold low-z properly. Im not sure if its a different FMD, or other missing hardware external to it. But I do know that when you try the documented method to enable low-z, it doesnt even try on an '808. Thus even if you grade the drivers, you are missing the needed control circuitry.

Though you are not correct regarding "more magnetic force" (and it contradicts your "adding resistors slows..." comment if that was in comparison to hi-Z), these days it's academic.

It seems the current (pun!) preference is for hi-Z. That change occurred after 2005. Kinsler in 2008 stated "High resistance injectors typically respond slower than low resistance injectors".
But now, hi-Zs seem to be quicker. I presume mainly due to lighter construction or materials.
No longer is it a case where only lo-Zs can supply higher fuel rates, or are considered more reliable in operation.
And there is still the hi-Z's better linearity.

So technology-wise, I'm happy - hi-Z is the way to go. Only if I decided to use the original Isuzu injectors would I use lo-Z, but I see that as unlikely since I want an o-ringed system.

It has been amusing reading the contradictions - like which type runs hotter. (That should be such a simple calculation but I suspect these EFI geeks aren't good at electrical basics. One even claimed that the lo-Z's resistor was to "trick the ECU" into thinking it was a hi-Z injector. I wonder how many of her/his readers blew their ECUs?)

But it has been nice finding some practical advances since I was last involved even if for me it is still not a priority.

I still wonder why the Delco's sense resistor and why your (indirect) link to that comcast upgrade increased the resistor rating when it merely needed to be bridged. But I assume the former gets back to my aforementioned early research and implementation by Delco (ie, if it was intended to P&H but was never implemented), and the latter is the difference between engine and ECU and electrical geeks.
(Apparently few ECUs incorporated P&H for lo-Zs though that may have been an aftermarket ECU comment.)


And thank's for confirming the external firmware, not that anything else would make sense (wrt to hacking etc).
But as I said, there are various methods of current control, though I reckon an external add-on is the simplest. Ooops - that's merely academic!