Digi7ech

iTrader: (1)

Like the title says.

My car has an Rtek 1.5 so it starts the secondaries at 3500rpm.
I have a Walbro fuel pump and run 12.5 AFRS under boost until the secondary turn on.

It jumps up to 13+AFR's for a second during the transition.

If I start the car and start driving under vac it will go over 3500 RPM with no problem.

Once I decide to hit boost and pass 3500 it will either do a back fire(sounds almost like a rally double fire/whatever exhaust).
OR
It will just give a big hesitation there(I mean BIG like car surges).

After a few minutes of driving this will slowly diminish until it just does slight bucking/surges at the injector transition.

THe car also under full boost will sometimes surge/hesitate unless I let off of full throttle a tiny bit.


So what do you think it could be?


I've done the TB mod

I have a good solid ground on the rear housing(even sanded off the housing paint)

I have not done the extra ground thing yet.

Then some have said my TPS could be out of whack.


The surging and recent backfires have kind of got me scared to boost around the secondary turn on point since this is a brand new engine and a rebuild is the last thing I want to do.

WAYNE88N/A

How far back in the exhaust system are the backfires?

Still got the air pump on her?

Digi7ech

iTrader: (1)

Completely emissionsless. No cat no airpump.

I know I shoot flames but that is during decel and high rev's.

These back fires/studders happen at the exact injector switch.

It will usually do a bad backfire/hesitation for the first 2-3 boost related transitions. After that it just becomes bad surges.

I will check my spark plugs but they are only about 3-5k miles old.

I'm almost thinking the Rtek 1.5 is doing this but I don't have any previous Turbo experience to check against.

I'll try the extra grounds this weekend but don't expect it to fix it.

Also a side note which I think Wayne remembers.
I have a S4 TII arnesws which is high impedance. The secondaries are low impedance so I put in the resistors. I had a previous rear primary injector problem but have resolved all bad wiring problems.

Second note.
At idle(600-900rpm) I pull 8-10 vacuum!!! Very not good for a newly rebuilt engine.
Stock port and everything. although the throttle plate could be stuck open a small bit at idle which could answer that part.

Digi7ech

iTrader: (1)

I do know for a fact that the car goes lean at the injector change since me and JRat did a few wide band runs. No matter what psi (5 or 10psi) I would go VERY lean and then go back to rich/good boost conditions.

BTW i run in the 11AFR's at idle.

J-Rat

iTrader: (2)

Umm.. D,

Your gonna have to do the grounding mod. No way around it. The hesitation was verified on the LM-1.

And your idle is still low vac? Maybe we need to tune your idle..

hondahater

iTrader: (2)

is your water thermo sensor pluged in??? I lost my hesitation witch was exactly as you described after

a. lost all emissions (you've already done)
b. plugged my water thermo sensor back in


Good luck with the grounding. After I regrounded everything including making new grounds for the ecu I still had a bad hesitation (no help). if you said your main ground was fine then that should be all that matters as well as all the other grounding points are clean.

Digi7ech

iTrader: (1)

Hmm looks like I've got a job for the weekend. **** I have to work this weekend.

Well I guess next weekend then.

WAYNE88N/A

Dude, you got a leak somewhere. When my EGR block-off plate was leaking, she'd go really lean at the secondary crossover also (not to mention the little hesitations out of nowhere). Check that block-off plate, and any others you may have.

If you have access to the ECU, read the 4 pins to ground near the ECU. If you get less than 1 ohm (hot AND cold engine bay), it's probably not the grounds...Everyone forgets about the most important ground on the car (for engine/ECU performance, anyway), which is the big terminal near the starter (with the rear rotor housing ground a close second).

I wish I could convince some guys about the misdirected attitude about these grounds on our cars, but I've about given up trying

TulsaTurbo

I recently have worked on 2 turbo rx7's that other mechanics gave up working on because they could not find or fix the problem.

The problem:

major backfiring and bucking at 3500 rpm and above under a load ( boost ). Almost sound like if the timing went way out of wack.

Solution:

Bad Knock sensor (replace it )

No more problems and a happy rx7 owner...

The RX7 maintenance manual leaves much to be desired with their instructions on checking the knock control system. See manual, page 4B-52

NOTE: Knock sensors should be considered regular maintenance items. If you have over 100K miles and have never replaced the knock sensor, you can probably consider it bad.

================================================== ==

Q) What does a knock sensor do?

A) A knock sensor is a specially designed listening device for car engines. It listens for engine knock, or detonation (pinging). Detonation can be very bad for an engine and is the result of the fuel/air mixture exploding too quickly instead of burning evenly and rapidly. This is due to either bad fuel, ignition timing which is too high, or built up carbon deposits in the engine (which increases the compression ratio). The sensor is small and consists of a piezoelectric sensor that listens for knock by detecting pressure. It is very sensitive and can be considered worthless if dropped. The vibrational pressure is then converted into a voltage and sent to the ECU for evaluation.


Q) How do I know if my knock sensor is bad?

A) Symptoms of a bad knock sensor include a sluggish engine, poor acceleration and poor fuel economy. Knock sensors rarely fail outright and more often get “soft” over time and cause false signals to be sent to the ECU, which thinks the engine is knocking when it’s really not. Thus, the ECU will reduce ( retard ) the ignition timing to the engine.

hondahater

iTrader: (2)

do you have a picture of what the knock sensor looks like??? I've never seen this knock sensor??? Isn't it in the door or something?

Digi7ech

iTrader: (1)

hmm. By luck I happen to have an extra knock sensor some where in my garage.

The sensor itself is a small octagon style sensor which screws directly into the housing.
It's the size of a golf ball and has one wire coming out of it.

TulsaTurbo

I have attached an image from the manual.

walken

iTrader: (6)

Where could you get a new knock sensor w/o getting ripped off?

hondahater

iTrader: (2)

thanks man! I do appreciate that greatly

TulsaTurbo

Here are some prices:

AutoZone : SU6600 : $105.99
O'reilly : S8680 : $118.99
Napa : ECH DKS596 : $139.00

Other options: Used:

Junk Yard : ????
Someone on a forum with a spare ???

TurboIIRyan

I'd strongly suggest looking at your tps. this sounds like something that happened to me, it turned out the TPS was so out of whack it was causing major hesitation and back firing when decel.

Rex4Life

This is all good advice but I would like to know how to read the knock signal to find out if the sensor is bad? It's easy enough to check that it works per the FSM, but much more difficult to see if it is giving out erroneous signals that the ECU sees as knock (and retards the timing).

I also have a moderate hesitation at 3500 rpm with the Rtek 1.7. It was barely there before I swapped and now after the Rtek I have a much more pronounced hesitation. I've been doing a lot of research on this lately and can tell you for SURE that no one has completely solved the riddle of the 3800 rpm hesitation. Interesting to note that the FDs also have the same exact problem but I think there's is at 2000 rpm (?).

The root cause is known as you have already measured--the engine is going lean due to insufficient fuel and this causes a "hesitation" or "fuel-cut". Before I go any further know that if you are running leaner than normal for whatever reason, the hesitation will most likely be severe. And if you are running richer for whatever reason, the hesitation will not be as bad. It is important to recognize this as "richening" things up can appear to solve to the problem. Unplug the TPS and you'll see what I mean--also an SAFC can be programmed to minimize the hesitation.

The million dollar question is: why is there not enough fuel? Basically the problem is a timing issue--not ignition timing--but injector opening time. Apparently the ECU programming was not programmed with a tolerance to longer injector opening time. So what causes longer opening time--lower voltage across the fuel injector coil. Remember the voltage across the coil is both a function of input voltage and the ground the coil sees. This is why many people report improving their ground helps fix the problems. The ECU is decreasing the primary injector signal and just starting the secondary signal--problem is the secondary injectors take longer to inject than they should due to a lag in opening the injectors. So here you are left with half the fuel you need cause the secondaries haven't opening up yet. Remember this is a real short time we are talking about here, but it doesn't take much.

What should happen is the injector signals get faded together over a 500 rpm range, not this step like function. If Mazda is fading the signals then they are doing it over too short of an rpm range.

I'm working on some sort of a fix but it is gonna take me some time. If I find a cure you'll surely see a thread about it. Sure would be nice if Rtek could reprogram the ECU to fade the injector signals--that would be the ultimate solution.

I know this doesn't help you alot right now but I needed to vent. Oh yea one question, is the hesitation before the backfire or after? I find it strange the backfire would happen first.

Scott

WAYNE88N/A

Hmm, that's an interesting point, Scott- I wonder, since all 4 injectors are fed from the same single power source, that when the crossover occurs with the extra two coils coming on line (remember a coil acts as a direct short until the EMF builds up), the voltage in the circuit is dropping to the point that it's messing with the switching transistors?

A good way to test this theory would be to run seperate dedicated power wires to the two secondaries, using a different circuit than the main relay provides...

TulsaTurbo

Some additional information I ran across concerning the knock sensor and the rotary engine:

The 12A turbo also has a knock-prevention system. Mazda engineers found in developing the engine that knocking would only occur under certain operating conditions: in a 2500-3750 rpm zone when intake temperature exceeds 85 degrees celcius.

The system, comprising of ignition-pulse, intake-temperature and boost sensors, detects this knock-prone zone and feeds a signal to the emission-control and fuel-injection computers, which in turn call for ignition retard and more fuel (in the latest 13B turbo a more usual piezometric knock sensor is used).

Digi7ech

iTrader: (1)

wow some good info.

The hesitation has become more pronounced now that I am running stock fuel pressures(40psi). This goes along with your noting up there.

Once I get the Aeromotive FPR on there I will raise pressure back up to 50psi.

Now you guys are talking about having two seperate power sources for the primaries and secondaries. How could I go about doing this?

I don't fully understand the ECU signals but from what I know the injectors are just on a variable circuit.
The ECU keeps the injectors open loop until they are supposed to be used then they close it.
Now what does it use as the variable? Does it vary the ground or the power?

WAYNE88N/A

Man, that's just a theory I pulled out of my a$$, lol...

BUT, if you want to be the test bed, let me know, and I'll look through the schematics for ya and find a good alternate circuit with the same branch power source (like IGN1). That's gonna take some work, though. How easily can you get to your secondaries?

Tulsa turbo had an interesting concept going on there, also. Have you ruled that out yet?

Umm, just picture the ECU as being a simple switch, one switch per injector circuit, that happens to be on the ground side of the circuit. When the switch is electronically put to "on", the circuit is completed (because it now has a ground), and the injector is energized. Now picture the switch opening and closing tens of times a second....Make any sense?