The Great A/C FAQ
#21
Elite Member
iTrader: (11)
Join Date: Jun 2007
Location: Overland Park, Kansas
Posts: 5,360
Total Cats: 43
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Well, glad it's sort-of working for you. ![Biggthumpup](https://www.miataturbo.net/images/smilies/biggthumpup.gif)
Since I never heard back from anybody on the 90-93 purge valve hack, why don't you try it on your car? You have a '95, right? If it's an early '95 then the wire to the purge valve itself is exactly the same- yellow / red at pin 2X. If it's a late '95, then yellow / red should be at 4T. Jumper that to the green / black wire at 1Q, and disconnect the hose between the purge valve and the charcoal canister. I'm really dying to know if this helps the situation.
![Biggthumpup](https://www.miataturbo.net/images/smilies/biggthumpup.gif)
Since I never heard back from anybody on the 90-93 purge valve hack, why don't you try it on your car? You have a '95, right? If it's an early '95 then the wire to the purge valve itself is exactly the same- yellow / red at pin 2X. If it's a late '95, then yellow / red should be at 4T. Jumper that to the green / black wire at 1Q, and disconnect the hose between the purge valve and the charcoal canister. I'm really dying to know if this helps the situation.
![Giggle](https://www.miataturbo.net/images/smilies/gay.gif)
#22
Boost Pope
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/smod.png)
Thread Starter
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,339
Total Cats: 6,793
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
If that doesn't do it, we can experiment with a bigger solenoid on a larger port. Actually, the EGR valve might not be a bad idea here. the 1.6 never had EGR so I can't experiment with that.
#23
Elite Member
iTrader: (11)
Join Date: Jun 2007
Location: Overland Park, Kansas
Posts: 5,360
Total Cats: 43
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
That's the idea. The capacity of the purge valve (and the port it connects to) is quite small, so it's not much of a vacuum leak, but it may be enough.
If that doesn't do it, we can experiment with a bigger solenoid on a larger port. Actually, the EGR valve might not be a bad idea here. the 1.6 never had EGR so I can't experiment with that.
If that doesn't do it, we can experiment with a bigger solenoid on a larger port. Actually, the EGR valve might not be a bad idea here. the 1.6 never had EGR so I can't experiment with that.
Seriously, are you one of those "Think Tank" people for your company?? Just sits around all day thinking up ideas?
![big grin](https://www.miataturbo.net/images/smilies/biggrin.gif)
#25
Elite Member
iTrader: (11)
Join Date: Jun 2007
Location: Overland Park, Kansas
Posts: 5,360
Total Cats: 43
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Okay I pulled one of the front vacuum plugs off my intake and went for a drive with the A/C cranked.
Seems that leak is just what the doctor ordered, didn't stall on me at all or even stumble. We'll see what happens tomorrow, but I believe that purge valve wired up would make this work perfect!
Seems that leak is just what the doctor ordered, didn't stall on me at all or even stumble. We'll see what happens tomorrow, but I believe that purge valve wired up would make this work perfect!
#26
Boost Pope
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/smod.png)
Thread Starter
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,339
Total Cats: 6,793
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I experienced a minor epiphany today which I will share.
Even though I do not have A/C, I have found then when coming to a stop, when I release the clutch and shift into neutral, my idle tends to dip down to about 700-750 before recovering. When the cooling fans are on, this dip is even more pronounced. Here is an example:
![](http://img28.picoodle.com/img/img28/4/5/27/thraddax/f_Dipm_d52a9dd.gif)
I spent some time tinkering with the idle settings. Namely, this screen:
![](http://img33.picoodle.com/img/img33/4/5/27/thraddax/f_PWMIDLEm_68db756.gif)
The way these work is that whenever you are on the throttle and / or RPMs are higher than the calculated idle speed plus the idle activation adder, the idle duty cycle is at the "closed" value.
When TPS is closed and RPM falls below that point (which, for a 1,000 RPM idle plus a 500 RPM adder would be 1,500 RPM), the IAC DC goes to the "Minimum" setting and then begins seeking idle. It stays at the Minimum setting until idle actually falls below the idle point minus the deadband value (950 RPM, assuming 1,000 idle and 50 deadband) and only then does it begin working its way up to seek a stable idle. Hence the dip, and the overshoot that often happens on recovery as IAC races upwards to prevent the stall.
I had previously been running the two values given by DIY in these cells- closed = 18, minimum = 19, and fidgeting with all the other values. Today, I spent some time looking back through some of my recent logs and found that while in a stable, warm, closed-loop idle, my IAC DC tended to be in the 23-24 range, and almost never below 22.
So I entered 22 into the minimum cell. It's a bloody miracle.
Now, the engine smoothly decelerates until it hits 1,000, and then immediately snaps into a stable idle. Virtually zero dip.
I suggest you guys try this. It won't help with the A/C coming on while idling (only the solenoid will help that) but it should help the situation when rolling to a stop while the A/C is already on.
Even though I do not have A/C, I have found then when coming to a stop, when I release the clutch and shift into neutral, my idle tends to dip down to about 700-750 before recovering. When the cooling fans are on, this dip is even more pronounced. Here is an example:
![](http://img28.picoodle.com/img/img28/4/5/27/thraddax/f_Dipm_d52a9dd.gif)
I spent some time tinkering with the idle settings. Namely, this screen:
![](http://img33.picoodle.com/img/img33/4/5/27/thraddax/f_PWMIDLEm_68db756.gif)
The way these work is that whenever you are on the throttle and / or RPMs are higher than the calculated idle speed plus the idle activation adder, the idle duty cycle is at the "closed" value.
When TPS is closed and RPM falls below that point (which, for a 1,000 RPM idle plus a 500 RPM adder would be 1,500 RPM), the IAC DC goes to the "Minimum" setting and then begins seeking idle. It stays at the Minimum setting until idle actually falls below the idle point minus the deadband value (950 RPM, assuming 1,000 idle and 50 deadband) and only then does it begin working its way up to seek a stable idle. Hence the dip, and the overshoot that often happens on recovery as IAC races upwards to prevent the stall.
I had previously been running the two values given by DIY in these cells- closed = 18, minimum = 19, and fidgeting with all the other values. Today, I spent some time looking back through some of my recent logs and found that while in a stable, warm, closed-loop idle, my IAC DC tended to be in the 23-24 range, and almost never below 22.
So I entered 22 into the minimum cell. It's a bloody miracle.
Now, the engine smoothly decelerates until it hits 1,000, and then immediately snaps into a stable idle. Virtually zero dip.
I suggest you guys try this. It won't help with the A/C coming on while idling (only the solenoid will help that) but it should help the situation when rolling to a stop while the A/C is already on.
#27
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I really appreciate you taking a look at this (and I will try it, as soon as I again have a clutch, exhaust, and fuel pump - all being replaced).
I experienced a minor epiphany today which I will share.
Even though I do not have A/C, I have found then when coming to a stop, when I release the clutch and shift into neutral, my idle tends to dip down to about 700-750 before recovering. When the cooling fans are on, this dip is even more pronounced. Here is an example:
![](http://img28.picoodle.com/img/img28/4/5/27/thraddax/f_Dipm_d52a9dd.gif)
I spent some time tinkering with the idle settings. Namely, this screen:
![](http://img33.picoodle.com/img/img33/4/5/27/thraddax/f_PWMIDLEm_68db756.gif)
The way these work is that whenever you are on the throttle and / or RPMs are higher than the calculated idle speed plus the idle activation adder, the idle duty cycle is at the "closed" value.
When TPS is closed and RPM falls below that point (which, for a 1,000 RPM idle plus a 500 RPM adder would be 1,500 RPM), the IAC DC goes to the "Minimum" setting and then begins seeking idle. It stays at the Minimum setting until idle actually falls below the idle point minus the deadband value (950 RPM, assuming 1,000 idle and 50 deadband) and only then does it begin working its way up to seek a stable idle. Hence the dip, and the overshoot that often happens on recovery as IAC races upwards to prevent the stall.
I had previously been running the two values given by DIY in these cells- closed = 18, minimum = 19, and fidgeting with all the other values. Today, I spent some time looking back through some of my recent logs and found that while in a stable, warm, closed-loop idle, my IAC DC tended to be in the 23-24 range, and almost never below 22.
So I entered 22 into the minimum cell. It's a bloody miracle.
Now, the engine smoothly decelerates until it hits 1,000, and then immediately snaps into a stable idle. Virtually zero dip.
I suggest you guys try this. It won't help with the A/C coming on while idling (only the solenoid will help that) but it should help the situation when rolling to a stop while the A/C is already on.
Even though I do not have A/C, I have found then when coming to a stop, when I release the clutch and shift into neutral, my idle tends to dip down to about 700-750 before recovering. When the cooling fans are on, this dip is even more pronounced. Here is an example:
![](http://img28.picoodle.com/img/img28/4/5/27/thraddax/f_Dipm_d52a9dd.gif)
I spent some time tinkering with the idle settings. Namely, this screen:
![](http://img33.picoodle.com/img/img33/4/5/27/thraddax/f_PWMIDLEm_68db756.gif)
The way these work is that whenever you are on the throttle and / or RPMs are higher than the calculated idle speed plus the idle activation adder, the idle duty cycle is at the "closed" value.
When TPS is closed and RPM falls below that point (which, for a 1,000 RPM idle plus a 500 RPM adder would be 1,500 RPM), the IAC DC goes to the "Minimum" setting and then begins seeking idle. It stays at the Minimum setting until idle actually falls below the idle point minus the deadband value (950 RPM, assuming 1,000 idle and 50 deadband) and only then does it begin working its way up to seek a stable idle. Hence the dip, and the overshoot that often happens on recovery as IAC races upwards to prevent the stall.
I had previously been running the two values given by DIY in these cells- closed = 18, minimum = 19, and fidgeting with all the other values. Today, I spent some time looking back through some of my recent logs and found that while in a stable, warm, closed-loop idle, my IAC DC tended to be in the 23-24 range, and almost never below 22.
So I entered 22 into the minimum cell. It's a bloody miracle.
Now, the engine smoothly decelerates until it hits 1,000, and then immediately snaps into a stable idle. Virtually zero dip.
I suggest you guys try this. It won't help with the A/C coming on while idling (only the solenoid will help that) but it should help the situation when rolling to a stop while the A/C is already on.
#28
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
But remember, there's no trouble with the idle code. Nope. none at all. Certainly it's working perfect if you need to add extra valves, and disallow the thing from closing the valve more than the stable value.
Damnit how many times do I have to argue this?
Whatever, I will write my own at some point. :-) Now that there's a "release" version of the MS-II code, it might be worth trying.
Damnit how many times do I have to argue this?
Whatever, I will write my own at some point. :-) Now that there's a "release" version of the MS-II code, it might be worth trying.
#29
Boost Pope
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/smod.png)
Thread Starter
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,339
Total Cats: 6,793
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Boy, you MS-2 guys sure are pissy about your idle code.
It works fine on the MS-1, you just need to set it up correctly. My point is that, for my car, 22 is the minimum stable value for a warm engine with no load. All values below that are unstable, so setting the minimum to 22 is the proper configuration which, until now, I'd been lacking.
![wink](https://www.miataturbo.net/images/smilies/wink.gif)
#31
Elite Member
iTrader: (11)
Join Date: Jun 2007
Location: Overland Park, Kansas
Posts: 5,360
Total Cats: 43
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
![wink](https://www.miataturbo.net/images/smilies/wink.gif)
The only way I could get it to idle for crap was to fatten up the idle mixture and take all the timing away.. Like 10 degrees and its rock solid 900rpm without missing. If I bump the timing up to 15-18 degrees at idle it sits there missing like crazy, even with more fuel.
#33
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Does it care if you turn on your headlights, roll up power windows, and have the heater fan blowing at the same time? Sudden changes screw up the "PID" loop, since it's not really PID, all the terms are busy doing other things - meaning they are at utterly wrong values, and the whole thing fails to be as adaptive as it should be.
No worries, I might try to come up with something better here soon....
No worries, I might try to come up with something better here soon....
#34
Elite Member
iTrader: (11)
Join Date: Jun 2007
Location: Overland Park, Kansas
Posts: 5,360
Total Cats: 43
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Does it care if you turn on your headlights, roll up power windows, and have the heater fan blowing at the same time? Sudden changes screw up the "PID" loop, since it's not really PID, all the terms are busy doing other things - meaning they are at utterly wrong values, and the whole thing fails to be as adaptive as it should be.
No worries, I might try to come up with something better here soon....
No worries, I might try to come up with something better here soon....
![vash](https://www.miataturbo.net/images/smilies/vash.gif)
#35
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Joe this is some great and really helpful information. I've used some of your settings and I've gotten my car to actually run pretty well. The only problem I'm experiencing is when I initially turn on my A/C, the car will stall. It won't stall coming down from an idle though with the new settings I got from your post. I have the MSPNP9697 on a '96. Do I need to do the method of the resistor? I'm trying to figure out if the above mentioned method is for the standalone ecu, PNP, or both. Thanks bud!
#38
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
The moved the AC pins around for '01? Wow. Dorks.
Anyway, one more thing... Though there's a real chance of getting into some stupid oscillations, I thought about (as the MS-II will run a GPIO pin on TWO conditions) of triggering it on something like:
TPS < 90% && RPM > 650
The at least your motor should never stall. I thought about doing it off engine vacuum, but really, there's no great reason to use that. For those of us with uncontrolled idles, the bleeder valve is even more interesting.
---------
Aside: Huh, I didn't even think the amplifier could be anything but a ground. I guess in the diagram it was a box with only one wire, and that wire went to ground. Well, if it hadn't worked, maybe I'd have noticed.
Anyway, one more thing... Though there's a real chance of getting into some stupid oscillations, I thought about (as the MS-II will run a GPIO pin on TWO conditions) of triggering it on something like:
TPS < 90% && RPM > 650
The at least your motor should never stall. I thought about doing it off engine vacuum, but really, there's no great reason to use that. For those of us with uncontrolled idles, the bleeder valve is even more interesting.
---------
Aside: Huh, I didn't even think the amplifier could be anything but a ground. I guess in the diagram it was a box with only one wire, and that wire went to ground. Well, if it hadn't worked, maybe I'd have noticed.
#39
Boost Pope
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/smod.png)
Thread Starter
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,339
Total Cats: 6,793
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
The whole ECU pin numbering nomenclature seems to have changed in '01. The connectors are labeled 2, 3 and 4, with no apparent connector #1. Also, the color of one of the four wires in question changed. Very odd...
Not sure what you mean by the A/C Amp being a one wire device. Here's what I'm looking at:
![](http://img33.picoodle.com/data/img33/3/8/21/f_Image2m_abb6990.gif)
It's got three wires. (A) goes to the ECU through the pressure switch, (C) goes to switched +12, and B goes to ground through the A/C button on the dash. Couldn't quite figure out why the added complexity. But if you found that R/W gave a closure to ground when on, then good.
I like your idea about setting the double condition on the ECU for turnon- both TPS and RPM. You're right about there being a potential for an oscillation, though I'll bet that at worst it'll take one cycle to catch up. I'm thinking that the initial turnon will pull RPM down, causing IAC to open up while at the same time releasing the clutch. While IAC is in the process of causing the engine to overshoot high, the clutch will close again and the engine will be able to carry the load.
I still wish someone would report back as to whether or not the canister purge solenoid as a controlled leak works. I'm also starting to wish I had A/C. I'd forgotten how damn hot and humid it was here...
Not sure what you mean by the A/C Amp being a one wire device. Here's what I'm looking at:
![](http://img33.picoodle.com/data/img33/3/8/21/f_Image2m_abb6990.gif)
It's got three wires. (A) goes to the ECU through the pressure switch, (C) goes to switched +12, and B goes to ground through the A/C button on the dash. Couldn't quite figure out why the added complexity. But if you found that R/W gave a closure to ground when on, then good.
I like your idea about setting the double condition on the ECU for turnon- both TPS and RPM. You're right about there being a potential for an oscillation, though I'll bet that at worst it'll take one cycle to catch up. I'm thinking that the initial turnon will pull RPM down, causing IAC to open up while at the same time releasing the clutch. While IAC is in the process of causing the engine to overshoot high, the clutch will close again and the engine will be able to carry the load.
I still wish someone would report back as to whether or not the canister purge solenoid as a controlled leak works. I'm also starting to wish I had A/C. I'd forgotten how damn hot and humid it was here...
#40
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
haha - oh god yes, in FL A/C is not really an 'option'. It'd be like buying a house without a hurricane shelter.
Well, that wouldn't be impossible to implemnet. I've been trying to get Albert to step up to the plate, crack open his MS, and start throwing needed circuits at it:
VICS control
A/C override
PWM Idle (verses the idle screw he uses now)
But so far, he's just been painting and repainting his car. :-) Someday, though.
Well, that wouldn't be impossible to implemnet. I've been trying to get Albert to step up to the plate, crack open his MS, and start throwing needed circuits at it:
VICS control
A/C override
PWM Idle (verses the idle screw he uses now)
But so far, he's just been painting and repainting his car. :-) Someday, though.