I've installed the GReddy kit in my GF's RX-8. The kit was purchased used from a member of this board. The 3 fixes were done to the kit and Tim's map was loaded. A GReddy BOV was fitted and seems to work well. We've used a canscan to check to be sure we're not running too lean. The car seems to run pretty well except for some surging around 4k - 4.5k RPM and some similar issues when the throttle is closed. Still drivable though and feels pretty quick. The car has the R flash and the emanage has 3 dongles.

However... the coolant temp gauge (stock RX-8 gauge) reads dead cold whenever the engine is running, and the idle will never settle below around 1500 RPM. The oil temp gauge works fine. When the engine is warm, if I turn the key to ON, the temp gauge will read as normal. Then when I start the car, the gauge will slowly move down to dead cold. (Slowly = 5 seconds or so.)

I have double checked that all factory connectors are in place and nothing was left disconnected. I checked to see if the temp gauge matches the reading that canscan takes, and it does. Before the car is started, it will read normal, then fall when the car is running. I checked to make sure the pressure sensor is working by using the map trace function in the support tool. No problems there- it reads very close to what the boost gauge reads. The emanage harness looks fine. The only thing I noticed that looked weird was that there is a white wire not connected to anything (see pics, links below). However I searched around here and found that someone else had the same, so I don't know if that could be related to the problem or not. It seems the temperature dongle is always switched in, and/or maybe it's defective since it seems to simulate a temp of 62.4 degrees (according to canscan) instead of the ~160 I've read about here.

Does anyone know if there is a wiring diagram for the GReddy RX8 emanage harness, that shows all connections and all wire colors, that I could check against to be sure all the dongles and everything are connected correctly? Since the kit was purchased used, I will never be sure until I check it for myself. I called GReddy and they told me they'd send a diagram, but they never sent it. I took a few high-res pictures which there are links to below.

Thanks,
Ken

Top side of the harness.
Detail view of the top side of the connectors.
Back side of the harness.
Detail view of the back side of the connectors.

 

The fast idle indicates that the temp dongle is always on. I had this problem initially with my setup. You need to be careful here. The car can overheat if its idling fast in traffic. For me, the idle would stick at 3000RPM after I went into boost the first time. The cause was that the VTech map was set incorrectly and wouldn't 'let go' of the dongle. I did end-up overheating so please get this resolved before that happens to you.

MD

 

Thanks for the warning. I'll make sure she keeps a close eye on the oil temp.

Can you go into more detail about how you fixed the problem?

 

I corrected the values in the vtech map.

 

The map I have is yours, I think. (SeveredInjector380map13 or something close to that, don't have it in front of me right now... is that the latest?) Can you tell me what the VTEC values should be, so I may double check it when I get home?

Thanks.

 

Ok, the map I am using is "Tims380SeveredInjectorMap13.GSC". Is that indeed the latest? And what should the VTEC values be? Mine shows all zeros UNTIL I click on one of the zero... then it changes to 1500. Different behavior than the other maps. Correct, or not?

 

Mine:

 

Ok, clearly I have a different software revision or something. However I did try to apply MM's updates and got a bunch of weird errors and had to re-install from the CD that was included in the kit. My software is version 1.1. Could this have anything to do with it? Perhaps I approached it wrong. Should I uninstall my support tool and try to install MM's "v1.2/1.4 Full Software & Firmware"?

 

My software is v1.2, firmware v1.4

 

I figured out why the car hesitates like crazy... well at least partly. When accelerating, the map trace shows that the RPM the emanage is reading is literally all over the place. There is something wrong with that RPM signal. Ugh. Something else I'll spend the night on this forum searching about.

Regarding my temp gauge problem, sure enough, the software update was the key. I had a heck of a time getting the software update to work, but eventually got it. Then I went out to the car and hooked up the emanage and ran the support tool, and had some kind of communication error. Took me several computer restarts and stupid serial-usb dongle config changes and finally I got a successful emanage update. Loaded the map and went out for a spin. Temp gauge works now as does the idle. Go figure. Everything was pointing me to come wiring problem. I still do have a dicrepancy between the wiring diagram Benny just sent me and what my harness actually is. You can look at my pics posted above and compare to this:

 

I'm sure you already know this, but the canscan reads from the car's narrowband. It's not going to be accurate enough for this application.

 

Canscan reads the WB02 sensor...you can read the narrow band as well...which is a voltage signal. The A/F comes off the WB though

 

Is it at least accurate enough to tell me if I'm too lean? Right now whenever I read any boost, the AFR reads 12.05. I took that to mean that it is no higher than 12.05. It could be 10.5, but I won't know that because of the 12.05 limitation. I understand that the canscan is not going to allow me to tune properly, but I was hoping at least that it would tell me if I was going too lean. Is this not the case?

 

Here you go.

I don't think it's going to be useful for you. If the reading outside of 15 is undefined, then you're not going to know what's going on in a critical area of the afr range. You're also not going to know how rich you really are...

I like my AEM UGEO wideband. Innovate is also popular.

 

Oh, crap. Who wrote that quote?

I did MM's/MadDog's 3 fixes and I'm running his map. Other than that and a BOV, the car is GReddy-Turbo "stock."

Do I need to tell the GF to IMMEDIATELY stop driving the car until she buys a WB 02 for it?

 

^^ This is partly true...the best bet is to have a true wideband standalone system Like an AEM, PLX, or such...or one integrated into the ECU...like the Microtech one.

The Canscan does read the current output of the actual Bosch sensor in the car and doesn't rely on the ECU for conversion. So the lookup tables in the ECU are irrelavant. As a matter of fact the Canscan version for the RX-8 has a correction factor involved to correct the non-linearity of the stock sensor. ( wether this is NB or not is open to speculation )

The Canscan does work accurately down to 12.05 . I have run both a standalone ans a Canscan/Scanalyser on my car and they log very similarly. They do not read lower than that though......and the car will run down to 11:1 ish in NA form

 

I wouldn't sweat it - if you want to really know what's going on, get a wideband gauge... If you want to be careful, just stay off the boost till you get it installed.

 

Thanks for the advice everyone. Do you guys have any ideas on the RPM signal jumping all over the place?

 

The E-manage I had used to do that...it was a software version issue. I updated`it and it stopped. The aE-manage on the Greddy kit doesn't affect the timing...so It is either the RPM adapter not working or a software glitch. Did you set the TP sensor?

 

I am doing some searching and learning a few things. I just read one post from carbonRX8 that said, "This is the result of your acceleration map altering the rpm signal." MazdaManiac came in and said, "The other issue is the RPM dongle. Since it is wired to the leading coils only, it does weird things." This from this thread:

https://www.rx8club.com/series-i-major-horsepower-upgrades-93/strange-data-logging-90873/

I have not set the TPS. First time I've heard that. You mean calibrate it per the RX8 shop manual?

I think I have the latest software revision now, at least that I can find around here.

 

That R flash could be your problem. I believe that can interfere with the tach pickup.

 

It seems that it the case indeed. The question is, how do I fix it?

 

^ You will have a VERY hard time addressing this. The problem is caused by severe cross-talk / ground loops between the RPM extractor and the mazda igintion circuitry. This is why people get such poor performance from the GReddy EMS solution in combination with the R flash. Previous flashes apparently had a different ground plane or something similar.

GReddy attempted to address this through an RX8-specific RPM dongle. But, as seems to be par for the course, they did a $hitty job. It works for some folks and not for others. You could attempt to build a buffer between the two, as Jeff did. That would probably help.

You should also know that this interference is, in my mind, what is directly responsible for my blown engine and the blowing up of several others who have lost engines while using the R flash. In my case, I rigged-up the eManage ignition circuitry in an attempt to retard timing. The result was misfiring to such an extreme that the car was barely drivable, whereas prior to rigging the ignition circuits the car ran as good as stock (+100hp! ). I believe that the cross talk was so bad that I got trailing plugs firing before leading, resulting in massive detonation and blowing my seals.

As I see it, you have two options. 1) build a buffer circuit and hope that you do a good job. 2) get a better EMS. I decided to go with the IntX after realizing that the eManage was only going to get me so far (as far as I could go without igintion control). I've been VERY pleased with the results.

 

Thanks for the insight Tim. Jeff, where are you man? Can you give me the details of this buffer circuit?

 

Does anybody have any idea what the RPM dongle actually does in terms of signal conditioning? Obviously there is a clean RPM signal somewhere because the tach works properly. Has anybody put an oscilloscope on the output (or input, or both) of the RPM dongle? Is it a PWM signal, a variable frequency pulse train, what?

It would be impossible to build a buffer without knowing this information.