When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
My experience is that the PID don't need to change much with target. They are picky, especially I, which can easily wind up under (2) conditions:
1) before EBC really takes over (below waste gate) If you call for more boost than the system can deliver, during a pull from low RPM
2) sudden punch, before turbos spin up
Thus, my view is to limit I and also leave a little bias on the table. If closed loop called for 45 bias, I set open loop to 40. That holds down the wind-up and still allows I to correct you up to target over a few PID cycles.
I use a lot of P, and a fair amount of D, little I (already discussed) The D is not really as an anticipation to get you to target faster, but the opposite, to account for fast increasing boost (spin-up) to actually subtract some output to break the fast rise to control overshoot.
Weather change has more effect, as cold outside air causes the turbos to spin up faster.
Tuning the boost table is definitely a two person job. It is very hard to keep an eye on all the details when the car gets up to warp speed.
And, logging and pulling over takes forever.
I have enlisted a friend who happens to be a pretty successful tuner.
I also made these:
The copper tubing is bolted to the center intake manifold stud.
I kept the hose diameter small to eliminate the low drone.
A quick test on the freeway with induced knock was pretty successful.
The portion of the otherwise clear thick wall hose threading through the engine bay has a sheath to mask noise.
Does it overshoot with open loop?
Does it overshoot at WG pressure?
Keep in mind that most people use EBC to improve their spool. For installs that don't really need to improve spool, EBC is not quite as valuable. In my install for instance, I use the EBC to manage the compound to twin switch-over, not so much early spool.
I don't remember the setup very well, and am too lazy to scroll back, is this a sequential (compound) configuration? Do you need to control the compressors independently?
Out of curiousity, is there any benefit to the the air tube for listening to knock, as opposed to using a knock box etc with the OEM knock sensor?
Less noise? more accurate? or just cheaper than an amp?
(I know my old Adaptronic had an inbuilt amp for the knock sensor, so you could just plug headphones into it)
5th gear, I floored it at 1800 rpm. Could not go any slower at the time, because freeway.
All the way to 7000 rpm.
Definitely getting there. Looks like we will need a couple more tuning sessions.
My friend is an accomplished tuner who used to stay away from MS. He had a point, because most MS units here are really questionable DIY units. Someone I know lost an engine recently due to crappy output transistors in one such unit.
Once he went over my setup he was pretty impressed.
So, I drove, he had the laptop, much discussion and enjoyment ensued.
I am attaching a short log file from last night.
Here's a synopsis of our first tuning session:
* We can definitely harness boost using open loop - it was much easier with two people. No more pulling over every few minutes to make changes...
* Unfortunately boost tapers down after 5000 rpm or so. Looks like I need to reinstall my 44 mm exhaust turbine in the large turbo.(I have a 42 mm turbine right now)
* My det can idea is a complete success. You get none of the ear splitting engine noise, but any knock is heard very clearly.
* We pulled back fuel in a lot of cells, corrected some timing in upper rows (timing was really too aggressive)
* Max injector duty cycle is 62-63%, so lots of headroom
* Boost is instant regardless of load, rpm, or volume of music in the car.
* I I increased the redline up to 7500 from 7100 (I had lowered it back in the day so I would not damage the supercharger). I'm sure I can go a bit higher since I have dual valve springs, but better stay on the safe side of things..
We will work on:
* Maintaining boost rock solid at 16 psi all the way up to redline (I need to install the larger exhaust turbine first for that)
* Smoothing out the "hand shake" between the turbos (there is a tiny, minor dip in boost during the initial hand over)
* The timing table, for power obviously, but without risking the car.
Edit: I have no clue why font size changed while I was typing this up...
Out of curiousity, is there any benefit to the the air tube for listening to knock, as opposed to using a knock box etc with the OEM knock sensor?
Less noise? more accurate? or just cheaper than an amp?
(I know my old Adaptronic had an inbuilt amp for the knock sensor, so you could just plug headphones into it)
I personally want this style DET cans to confirm my MS3 Knock sensor settings are operating correctly with the stock '99 sensor.
Hakan, I'm glad to know you got good results from mounting Cu tube to the IM. Much easier than a mounting below the IM. I will go through the firewall where the Cruise Control cable used to be.
It's a vacuum actuator, used to open/close the HP-LP bypass.
In other words, this is the gizmo that switches the turbos.
I spoke with a knowledgeable parts guy and he said they go pretty often. They are not cheap.
I promptly replaced it.
So, rather than waiting for the damn thing to break (and leave me with - gasp - just one turbo), I am designing and building a remote mount electric version of this thing.
Parts are on the way.
Details when I get the parts...
Had a great couple track days until I had bucking/popping/backfiring/stalling problems...
First one happened at around 5500 rpm out of turn 10 at Istanbul Park F1 track. Tach momentarily dropped to zero, huge explosion followed by a cloud of brownish smoke and it felt like I hit a wall.
Made it to the pits, cooled everything off while I checked things over, and the car started fine after about 30 minutes.
The drive to the garage (about 20 miles) was a nightmare. 8 to 10 foot flames, huge "explosions", bucking and jerking with the tach dancing up and down... (mostly down)
I was not calm enough to think clearly, and just wanted to get to the garage. After a ton of trial and error, I found an rpm band where the car would keep running uneventfully, and kept it as steady as possible between 2100 and 2600 rpm.
Cars behind me would hit the brakes and scatter every time the exhaust POPPED. I started shooting muffler internals at one point. My car was like a wounded pirate ship unloading its single cannon aimlessly.
By the time I made it to the garage, I had pretty much no exhaust left. Just a shell of the muffler, and some piping bits.
(I honestly did not mind the demise of the exhaust, because I will be replacing the entire thing with a completely new custom design very soon)
Since the tach was dropping every time the car was bucking (ignition cut), I thought it was the crank sensor. I had a spare, so I slapped it on.
Of course the car fired up right away... I had to be somewhere in a hurry, so I took off.
The car started acting up again, but not as violently. I thought "hmm, cheap *** sensor has crapped the bed already". Both the one I removed and the new one I installed are aftermarket Chinesium "Mazda 323" sensors. They were all I could find in a pinch when I managed to break my perfectly good crank sensor a short while back.
Long story short, I was stranded today. This time around, the car would run at a very specific (but different) rpm band, between 1500 and 1650, and any fluctuation would result in the engine dying. I live on a very twisty steep uphill road, and getting the car up that hill was impossible.
Then, I discovered something pretty unsettling: The damn crank sensor was moving towards or away from the crank wheel no matter how tight the single M6 bolt is wrenched. The metal protective ring embedded in plastic spins pretty freely, and since the bolt exerts all its force on that metal ring, the sensor is not properly secured... So I concocted a washer setup to keep the sensor from moving around.
And, since the car had cooled down and would fire up with the securely mounted and properly gapped sensor, I grabbed my laptop and started driving and logging.. The car warmed up, and bam. It's all in the logs.
I logged the engine and took a composite log, as well. They are pretty short, but document the issue - albeit not as violent as what I lived through on the track and during the drive home after that.
Glancing at the composite log as I was driving made me realize the culprit was the CAM sensor. It is a Mitsubishi EVO 8 /Carisma sensor, about 6 months old...
Came home, popped the hood, and lightly tapped on the cam sensor with the back of a screwdriver, and the engine started to miss / hesitate / almost die every time I "gently" tapped on it.
I am perfectly aware of the plug/pigtail issues the spec Miata guys had discovered a while back. In their interpretation, such symptoms followed by a gentle tap meant momentary loss of contact in the harness plug.
So, I believe I had two separate issues:
The engine bucked because of the cam sensor, and the violent action knocked my crank sensor out of adjustment. Or, I am completely wrong.
Please look at my logs and help me unconfuse myself.
I had already secured the crank sensor in place.
After replacing the cam sensor with another Mitsubishi unit, the car purrs like a kitten once more.
On a related note, the almost 10 mile drive to the parts store was hysterical.
My wife and one of our dogs were both in the car, it was during the rush hour, and I had to pull over, pop the hood, remove the sensor and cool it every mile before I could continue on the freeway.
The day ended very badly tho. The vet (a specialist professor) told us cancer has spread pretty much all over Eddie's body since the surgery back in February.
No way the poor thing can go through a major surgery again. He will be 12 come New Year's.
Ahh man, I'm really sorry to hear that. We're going through a similar thing with my girlfriend's cat right now and it absolutely sucks. Hang in there, it's never easy :(
The day ended very badly tho. The vet (a specialist professor) told us cancer has spread pretty much all over Eddie's body since the surgery back in February.
No way the poor thing can go through a major surgery again. He will be 12 come New Year's.
Eddie:
Eddie quietly passed away this evening. he was next to me on the couch all day.
His condition had gone from bad to worse in the last 48 hours. He tried to eat a couple bites last night, and vomited blood four times after that.
His vets knew his condition and they told me there was nothing to do. They said this was the end.
I cried my eyes out in the middle of the living room.
10-26-2019, 06:00 PM
1
