What If? $8D (1227727 ECM) for TBI

[yelruhde]

Sounds like Syn/Async fueling? Is there a way to see in ADX data?

1 ms is going to be way short...

1ms is the time it takes to turn on. The signal from the ecm will keep it on as long as necessary...hopefully. I'll do some research on FET's and see if one will fit.

If I can get the timing belt in the daughter's mom's volvo and through the junkyard, then I'll take some screenshots while driving. It may not happen for a week or so though.

[yelruhde]

Here is one way to fix the injector driver problem. Response has improved, but running with the tbi mamcal is going to require some massive tuning. Feel free to flame the file.

[dave w]

Here is one way to fix the injector driver problem. Response has improved, but running with the tbi mamcal is going to require some massive tuning. Feel free to flame the file.

I'm not going to "Flame the File", instead I'm going to say "Thank You" for your hard work! Tuning is managable using data logs and Microsoft Excel spreadsheets. I'm thinking the tuning is the easy part. Figuring / doing the "Electronics", you did very good!


dave w

[RobertISaar]

that method will work perfectly fine.... fairly certain the injector drivers are FETs though, not regulators.

another method would be to do that outside of the ECM case itself, just use the existing injector trigger signal to fire off another FET, which has it's other two legs connected to the injector circuit and ground. a little less likely to kill something inside the ECM by accident.



also, a 7727 CAN be modified to run two injector drivers.... just need to rob some parts from another ECM to do so, then change the code to fire them in an alternating fashion.

[EagleMark]

I'm not going to "Flame the File", instead I'm going to say "Thank You" for your hard work! Tuning is managable using data logs and Microsoft Excel spreadsheets. I'm thinking the tuning is the easy part. Figuring / doing the "Electronics", you did very good!


dave w

Yup did good!

[JeepsAndGuns]

Ok, a little confused. Mainly because I do not know anything about the internal workings of the ecm's or the code running them. But what exactly has been done here? If this isnt another injector driver, to make tbi easyer, then what does it do? You state in the file that it will "add additional power capability to the driver already there" Whats the purpose/need for doing that?

[RobertISaar]

it is another driver... just that it's triggered at the exact same time as the original driver.

think of it like parallel wiring: if you have a single 10 guage wire feeding something, you can push something like 30 amps through it. now give it another 10 guage wire and you can support another 30 amps, for 60 total.

the existing driver can handle something like 8 amps safely(at least that's the largest it did in a factory application, the 90-91 corvette L98), now you're basically adding another 8 amp capability. most TBI injectors i've seen draw 6 amps before they hit "hold" mode, so with 2 of them, that's potentially 12 amps of draw.

[yelruhde]

Thanks, it went pretty easily once I was told how to trace a circuit board. Before that it was impossible.

The tbi piece that is soldered in carried the part number IR9508. Specs on the piece are scarce, it probably went out of production with the vcm 17 years ago. Best I could find came from the China market, which is what called it a voltage regulator. That info may not be correct. I never found a data sheet with current or frequency limits.

[1project2many]

Are you using the peak and hold jumper or running purely in saturated mode?
The proof that the capacitor isn't needed... because the engine starts and runs?

[daleulan]

I've not read through this whole thread, but to avoid a lot of blown-up electronics or injectors, I thought I'd write some of my TBI experience just in case it might save some frustration.... most TBI injectors (certainly the Delphi/Rochester ones) must be driven from a true peak-and-hold current source. The Rochester/Delphi TBI injectors are fast but they need 'correct' drive. Not so much of a requirement for the low-impedance multipoint injectors, but the TBI injectors are really a mechanical marvel. Well, diesel injectors are faster, but the TBI injectors are pretty good. The proper waveform ramps the injector to 4 amps then drops it down to 1 amp to hold it. If you try to drive it without the peak-and-hold jumpers in place properly then you can fry the FET, injector, or both. The 'usual' failure mode for a FET is a short-circuit (ie. full fuel flow). There is also enough inductive 'kick' out of those TBI injectors to seriously fry anything not designed to drive them. Those injectors don't have a good enough opening force-to-energy handling capability to just use a dropping resistor like some of the multipoint injectors can get away with.

The 1227727 base PC board and chipset has 'space' to support 'proper' TBI - two injectors worth of it. There's a calibration field that changes the pattern written to 3FFC (B93A or B91A) for the type of injection you are doing. What I would probably do is look to see if the second driver could be added easily and just solder in the parts. Most of the resistors, capacitors, and certainly a suitable FET and the zener diodes, etc. could be added. The calibration bit flip and a two-injector TBI MEMCAL (or a MEMCAL suitably modified) oughtta do the trick after the parts are added. If anybody is really interested in this I could take a run down to the junkyard, mod one up, and give specific instructions on how to do it. Then I'd test it on my bench (not a vehicle - my GM wagon has the 6.5L electronic turbo diesel). I don't really 'feel like doing it' but what the heck, if it helps everyone out on the forum get some of their EFI stuff running properly, it's worth it.

Also, the ECM generally has an injector flow constant. This generation of controllers don't always divide the number by two when you enable the second injector. So you need to adjust some constant (either fuel flow of the injector or something like that) when you switch modes like that. The software I write (not for GM) automatically does this calculation for you. Most of our customers are switching to multipoint but we have a lot of customers still running TBI on natural gas (which is actually a very sane thing to do even after TBI hasn't been able to meet gasoline emissions for almost two decades), and many of them are perplexed by injector sequencing on something like a two-injector straight-six application.....

[dave w]

I've not read through this whole thread, but to avoid a lot of blown-up electronics or injectors, I thought I'd write some of my TBI experience just in case it might save some frustration.... most TBI injectors (certainly the Delphi/Rochester ones) must be driven from a true peak-and-hold current source. The Rochester/Delphi TBI injectors are fast but they need 'correct' drive. Not so much of a requirement for the low-impedance multipoint injectors, but the TBI injectors are really a mechanical marvel. Well, diesel injectors are faster, but the TBI injectors are pretty good. The proper waveform ramps the injector to 4 amps then drops it down to 1 amp to hold it. If you try to drive it without the peak-and-hold jumpers in place properly then you can fry the FET, injector, or both. The 'usual' failure mode for a FET is a short-circuit (ie. full fuel flow). There is also enough inductive 'kick' out of those TBI injectors to seriously fry anything not designed to drive them. Those injectors don't have a good enough opening force-to-energy handling capability to just use a dropping resistor like some of the multipoint injectors can get away with.

The 1227727 base PC board and chipset has 'space' to support 'proper' TBI - two injectors worth of it. There's a calibration field that changes the pattern written to 3FFC (B93A or B91A) for the type of injection you are doing. What I would probably do is look to see if the second driver could be added easily and just solder in the parts. Most of the resistors, capacitors, and certainly a suitable FET and the zener diodes, etc. could be added. The calibration bit flip and a two-injector TBI MEMCAL (or a MEMCAL suitably modified) oughtta do the trick after the parts are added. If anybody is really interested in this I could take a run down to the junkyard, mod one up, and give specific instructions on how to do it. Then I'd test it on my bench (not a vehicle - my GM wagon has the 6.5L electronic turbo diesel). I don't really 'feel like doing it' but what the heck, if it helps everyone out on the forum get some of their EFI stuff running properly, it's worth it.

Also, the ECM generally has an injector flow constant. This generation of controllers don't always divide the number by two when you enable the second injector. So you need to adjust some constant (either fuel flow of the injector or something like that) when you switch modes like that. The software I write (not for GM) automatically does this calculation for you. Most of our customers are switching to multipoint but we have a lot of customers still running TBI on natural gas (which is actually a very sane thing to do even after TBI hasn't been able to meet gasoline emissions for almost two decades), and many of them are perplexed by injector sequencing on something like a two-injector straight-six application.....

I am interested in a "How To modify a '7727 for TBI". I've have good soldering skills. I could use some help with the TBI code.

dave w

[JeepsAndGuns]

I have a 7727 that's bad. I tried a few times to get it to work, it never would. If anyone thinks they can salvage it for parts, they are welcome to it. I opened it up and looked for burnt/melted parts, and nothing appeared damaged. I even have a 7747 I can donate to the cause, its unknown condition, as I no longer have a way to test them.

[yelruhde]

Are you using the peak and hold jumper or running purely in saturated mode?
The proof that the capacitor isn't needed... because the engine starts and runs?

I am not using a peak and hold jumper, mostly because I can't find any consistent info on which pins to jump. Please share what you have. Also, peak and hold appears to be more of a current limiting method than something done to control a flow profile. Maybe it does both.

The capacitors are not wired in because I can't get a good reading from them. I am calling them capacitors because they are not resistors or the diodes and there aren't any other choices. I'll post a picture after work.

This is my first ecm project and its on my daily driver....so i won't do anything to it until the information behind the action makes sense. Hopefully that will make it easier to figure out what happened when things go wrong.

[yelruhde]

I don't have proof of what this is. 0 ohms across in both directions and the capacitance readings can be 10 nF or the same as shorting the leads.

I don't mind doing the work to add them to the circuit but need to know what it is first.

[1project2many]

Dale's post describes the importance of using peak and hold mode. Driving TBI injectors in saturated mode is very likely to kill them. The 4A ramp / 1A hold is a method of getting the injector to full flow quickly then holding it open without burning it out.

The picture above appears to be the injector current sense resistor, .103 Ohm, located at the RH corner of the board next to two green capacitors. The board is silkscreened R76 at that point indicating the component installed is a resistor. Next to R76 is a location for R93, the sense resistor used with the second injector driver. My camera batteries are dead now but you can see a set of modified sense resistors installed in those locations at Bill Shaw's injector driver upgrade page: http://home.comcast.net/~p-928/749mods.html


The injector jumper enables the current sensing peak and hold driver circuitry. To locate the pins where the driver should be connected you could use the wiring diagram for the turbo Sunbird as a starting point (Note: diagrams on this site are for Syty). Jumpers in the 'Bird are connected between C13 and C15, then between D5 and D6. Then you could use the 7730/7749 to 7727 pin map at http://www.exatorq.com/ludis_obd1/triplugecm.html to locate the corresponding points which are as follows:
7730 = 7727
D5=D8
D6=D4 or D10
C13= D1
C15 = D5

You need to install the sense resistor and complete the circuit before applying the jumper.

I don't understand what you're saying by "control a flow profile." The time between injector closed and injector open, while the pintle is in transition, is a time of variable flow and should be reduced to a minimum. P&H does this and should be utilized.

I'm sure there's more I want to say but my kids are up now and the three year old is starting to play with these disassembled ecm's so I've got to go.