DIYPNP Idle Issues
 

I know idle tuning has been covered MANY times, but I am having a couple issues that are getting quite frustrating, and per the various threads, should have been easily fixed. The issue is regarding my AC compressor in closed loop idle (This is a DIYPNP on 3.3.1 installed in a 2000) . First off, when trying to tune the PID fields, it is not reacting as it should be. In the idle test mode, I found my min and max steps, and everything went well. After putting in the values, I found that very little changes when playing with the I value. When the PID values are all at 0 the car idles at roughly 1000 rpm (850 is my target), and once I input "10" for the I field, the idle drops to 900. From there I can enter values up to the max 200, and the idle stays at 900. When changing the P value, nothing really seems to happen, and nor does when entering in the D value. I am able to get a good idle (even though the numbers appear not to change anything); however, once the AC comes on it goes downhill. The RPM's fluctuate from 300RPM up to 1200, and keep going until the compressor turns off. Nothing I can change seems to reduce the initial dip and oscillation. Under "AC idle up" control, I can put in any values for the duty/rpm adder with no effect; however, when changing to "high" polarity, my RPM's when the AC is off do in fact increase about 500 (still changing duty/rpm change nothing). I am confused at this point and nothing is really making sense.

FWIW, I don't think it is related, but my CLT and IAT sensors are reading off, and I cannot control my radiator fan. My CLT reads around 130 when the car is at ~200, and IAT was reading 70 yesterday when it was 20 out (On initial warm up, before any possible heat soak). I programmed them in the correct units and bias per the diyautotune website. I can also verify with the stock ECU, the CLT readings are accurate. For the radiator fan, I wired it and verified with a multimeter there were no cold solders, and no matter the entered values, the fan is ALWAYS on.

Thoughts/Ideas? I tried contacting DIYautotune 3 times, but haven't gotten any replies back :/. Part of the reason I purchased from them was because I heard their after sales support was fantastic. Maybe it's just a bad time of year for them? MSQ and screenshots uploaded!

Recalibrate the sensors using the 3 point curve generator with matching bias resistance in Tunerstudio. In my experience DIY's calibration numbers were inaccurate for the GM open element IAT sensor.

CLT and MAT should agree within a few degrees of the ambient temperature when the engine is "cold" (steady state thermal equilibrium with ambient).

I assume you mean this option? Since the resistance/value is not linear, how would one go about re-calibrating this table?

Choose your sensors from the Common Sensor drop down list. Make sure the bias resistance is correct and click write to controller.

TS makes the new table for you by calculating the coefficients for the Steinhart-Hart equation from those 3 points.

If you are interested in viewing the most recently burned table then look in the TS project folder for std_ms2gentherm.log and open it in notepad++.

I think I must be misunderstanding something. When selecting "mazda" (it's the stock sensor), it populates similar values. My car is at operating temp (as per my oil temp gauge), but the CLT is still reading ~130. If I take that and enter in the 2490 resistor bias, I get 300+ degrees. Am i missing something?

Use the RX-7 sensor for CLT.

With RX-7 CLT selected, I am now reading 10 degrees lower than before. So 120 when it should be at 200.

If you sure the coolant temperature is 200 F and the bias resistance is correct then you could try removing the sensor and bench testing it in water. Record 3 data points and enter the values in the curve generator.

So after getting my 3 values, I found the oddest phenomena (I don't live in Colorado, this is actually happening). When I go into TS and change the CLT values to get accurate readings, my AFR's, battery voltage gauge, and clt/IAT readings start bouncing all over the place similar to a poor ground. The signals all become very jittery. As soon as I re-calibrate the CLT to the normal spec, everything acts as normal. What's going on here? Why can't I just change my thermoresistors so that they read normally?

What were the data points?

I can't seem to find my notepad at the moment, but I'm pretty sure I had
F | Ohms
----------------------
32 | 12000
--------------
70 | 6300
--------------
200 | 400

and I just used the 2490 resistor bias as diyautotune listed as I don't have all the required info to calculate this (and I assume it should be similar?) Weber Bias Calculator

Why would this change my AFR's etc? Shouldn't these readings be on a separate circuit altogether?

it didnt.

Let me datalog for you guys, this doesn't make any sense to me either. Give me a sec.

I tried the measured values on ms3 with jimstim and there are no problems albeit clt went up 13 F.

You may have some hardware problems.

So here are 3 data logs.
1. On the RX7-CLT settings
2. On the benchtest settings (Really squirrly huh?)
3. After RX7-CLT settings

Since you guys say that the circuits are not tied together (which does make the most sense), the other thing I can think of is that since this took my temp readings from 130F to a more proper 170F+, is there something that would switch when the car reached operating temp? It just doesn't make sense.

Try setting fixed field min and max in MLV so the signals do not look so "squirrely".

http://imagizer.imageshack.us/v2/640...0/713/6zhz.png

After spending more time toying around with different numbers, I found that there is no issue changing the numbers, the issue happens only once the coolant >~160/170. I can change the values in the table to whatever I want, and as long as my coolant reads under this number, everything works fine. As soon as the car gets above this, the ECU occasionally disconnects from the laptop, my voltage/afr readings go crazy, etc. Is there some setting that's only engaged once the car is warmed up? This obviously isn't a calibration issue, this issue lives somewhere else.

Your ego correction activates when CLT >= 160 F. Try setting ego authority to zero.

You have the stock narrowband o2 sensor which could be the problem.

I suggest leaving ego off until you have a wideband, and in the meantime do not tune with the narrowband.

I didn't see that the EGO was enables and set to a narrowband. I have the MTX-L (Would never have even tried to tune without a WB02.) I will disable EGO, then see if this solves the issue!

I tried disabling EGO control, along with using the single WB02 setting, both with the same result. I can however give the specific number everything goes downhill. 175 degrees is the magic number. As long as the CLT reports back under 175 everything is fine, at and above is the issue.

When CLT >= 175 F the cpu sends a self destruct signal to your car by PA0.

http://imagizer.imageshack.us/v2/800...0/811/40ng.png

I tried disabling Pa0 and I found that problem went away, thus it was safe to assume there was a jumpering issue. I stumbled upon someone else’s jumper diagram as provided by Matt at DIYautotune. This is what I changed mine to:

AC:
1p -> input 1 in
input 1 out -> pe1
PA0 -> relay 2 in
Relay 2 out -> 1S
1S -> 1I
2.2k resistor in (R14?)

Radiator fan:
WLD -> Relay 1 in
Relay 1 out -> 1R

Clutch Switch:
Clutch Switch -> DB15 connector 11
DB 15 connector 11 -> input 2 in
input 1 out -> FLX

After doing this, I reflashed my ECU because I wanted a fresh start (probably no point in doing this...) and these are the problems I am now having.
1) My AC compressor is always on, whether or not the AC button is pressed, or the fan is in the off position. As soon as I turn the fan on (with the AC button in the off position) cold air comes right out. The same result is with the button pressed on. This doesn't make sense to me....
2) My Tachometer goes to 0 RPM when I step hard on the throttle with Tacho Out setting set to off. If I turn Tacho Out setting on, the RPM gauge moves very slowly, then occasionally oscillates randomly from 2k-6k.
3) When the car is sitting warm at idle, if I go to start it, it will start strong, try to stay alive, then die. I must wait for the car to cool before it will start well on it's own. If I give it some gas to stay alive long enough to read AFR's, it will be at 16-17; however, if I start the car from a cold start and let it warm up, it will idle right where I want it to ~14.7.

Attached is my current tune file (I know the whole VE table is not tuned, but the bottom end/idle looks good enough).

I tried opening your tune under a new project with your current firmware that was on your .msq and got so many errors I didnt feel like reading them. Send me all the details of your build and Ill try to get you a decent start up map while Im at work. Then we can go from there. Otherwise I feel like were pissing in the wind. Also reflash to the newest firmware again. Its freakin TITS and BEER.

Ok try this map and use these jumpers for AC

Add 2.2k resister to r14

1Q>Input 1 IN
Input 1 OUT>PE1
PA0> Relay 1 IN
Relay 1 Out > 1J and 4S.

Post results. This map is intended to fix your output settings and correct the AC Idle up setting you had before as well as turn off EGO control until you get a more developed tune. IMO I never use EGO control. Never really had enough reason to turn it on

Are you sure about 1Q, 1J, and 4S? Not trying to doubt you, but judging by DIYautotunes page these aren't the AC/Rad fan I/O's (DIYPNP MegaSquirt installation for the Mazda Miata). If your sure, I'll go change that!

Sorry I meant 1P, 1S & 1l. I was helping another person with a 95 earlier.

I already have this jumper config (going through relay 2 instead), so I assume I can just plug and chug with the tune? Or should I disconnect the WLD -> relay 1 in, relay 1 out -> 1R? (Radiator fan circuit)

I really dont think it should matter. you should be good to go then. Nothing in it will hurt your car. Im using the same ign, fuel and sequential settings you already had. Try this map. It should work out as long as your wiring is as you said.

So tried the tune and a few odd things are happening:
1. Once the car warmed up, I think I hear some ping at idle (I didn't try driving it). AFR's read ~14 and timing is only around 15-17.
2. My AC compressor idle up is not working properly, and on top of that, it appears that the compressor is still "always engaged." No matter which way I turn the AC button, or fan position, there is no change in idle and the air is always ice cold (it's ~70 degrees here ambient temp). I tried searching, but everyone with a similar problem seems to have the opposite where the AC comp will not engage.....
3. It didn't look like VICS was configured, and why isn't coolant based on CLT temp? Also, why shouldn't I be running CL idle?
*edit* here is the log file.

Weird. the reason why I took of Closed loop is to eliminate your idle fluctuations or eliminate closed loop being the case. Closed loop idle is not something you tune tell the very end. I havent even fully perfected my closed loop idle yet just because I am still playing with hot restart enrichment. Once thats done than Ill tighten up closed loop a little more. As for the AC Compressor I think you have something going on with your jumpers. Could you post HQ pictures f the front and back of your board. Maybe at this point you may even want to roll back your firmware since you said it was fine before. I wonder if when you update your msq after updateing your firmware if you didnt update it correctly. Could you also post your msq from when the car was running correctly?

The car has always been running on 3.3.1, and it hasn't ever been running correctly. The thermoresistors have always read off, and the AC idle up has never worked properly. I just went for a drive, and it actually seemed to be working pretty well. The Tach read properly, the AFR's stayed pretty constant, and the radiator kicked in (I changed it to turn on >175. The only real problem I am having right now is the AC, and to my understanding, the ECU does not control when the AC turns on and off. The compressor kick in right at start all the time, then never turns off. Also, AC idle up is not working, so I assume it is not getting a signal. I could post up pics of the jumpers, but I think there is too much going on to see anything; it's hard enough holding it in my hands. Do you/anyone know how the AC clutch decides to kick in and how it send a signal to the ECU? I'm thinking there may be an actual hardware problem....

Okay, so as I was buzzing out the circuit, I found what I had and was recommended can not work. First off, that radiator circuit will not work. WLD is a relay, so putting it in series with relay 1 will never control 1R as 1R never get's input voltage. 1R must go directly to WLD. Problem 1 solved. Problem 2: 1P needs to control 1S and 1I, but 1P also needs to go to PE1 to tell the ECU when to idle up. So here is my issue; how do I connect 1P to the IS-1I circuit and relay 2? drawing for reference, please pardon MSpaint. What goes where? And what does R14 do?

Okay, how about this? This way the compressor relay kicks on correct and I also get to keep pe1 for AC idle up. Do I need to remove R14?

PM Sent

So after looking into this further and playing with the multimeter, I found the last schematic was wrong too. Here is the correct diagram. This appears to be working, I will update with any issues.