The Great A/C FAQ
#42
Boost Pope
Thread Starter
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,339
Total Cats: 6,793
The A/C switch, A/C Amp, and all the relays are on the IG2 feed- nothing has power when they key is off. Unless some wires got crossed, I don't see how this could kill a battery.
#43
The quality of the wiring in the test vehicle is.... somewhat less than questionable. The fans have since stopped working, too. This is taking a lot of digging. So far, no current along any of the jumpers I put in.
The only real issue would be is somehow the ac amp supplies power. VERY unlikely.
_______________
Edit: While not conclusive, several issues were found with the wiring (based largely on incorrectly sized vampire clips), so again everything was checked an no issues found - so I still stand by that it's a reasonable way to set it up, your mileage may vary.
The only real issue would be is somehow the ac amp supplies power. VERY unlikely.
_______________
Edit: While not conclusive, several issues were found with the wiring (based largely on incorrectly sized vampire clips), so again everything was checked an no issues found - so I still stand by that it's a reasonable way to set it up, your mileage may vary.
Last edited by AbeFM; 08-22-2008 at 12:59 AM.
#44
Boost Czar
iTrader: (62)
Join Date: May 2005
Location: Chantilly, VA
Posts: 79,688
Total Cats: 4,113
thinking out loud....
is it possible that simply triggering the second fuel/spark map could a plausible solution? a/c on fuel map is richened in those cells and spark advanced slightly, might be enough to compensate for the idle dip.
is it possible that simply triggering the second fuel/spark map could a plausible solution? a/c on fuel map is richened in those cells and spark advanced slightly, might be enough to compensate for the idle dip.
#46
Not only can the idle code be altered to accept extra input (though many set ups are "out of pins" and can't DO any more I/O, the idle code itself is KNOWN to be wrong. It isn't a PID controller, hence it's irresponsive to step functions. Some people have rewritten this, but you MS-I folks are out of luck, find someone who likes writting assembly.
#47
Not only can the idle code be altered to accept extra input (though many set ups are "out of pins" and can't DO any more I/O, the idle code itself is KNOWN to be wrong. It isn't a PID controller, hence it's irresponsive to step functions. Some people have rewritten this, but you MS-I folks are out of luck, find someone who likes writting assembly.
What's the story on the folks who've written the current idle code? When you say it's "wrong", what do you mean?
Edit: I mean can you elaborate on what it lacks, what is it if not a PID controller? and what would be gained by it being one?
#48
Basicaly, a PID controller has three parts. There's a target input value, a measured input value, and a set output value.
The difference in the target and the measured is the error, and you tweak the output to try to get the error to zero.
P - proportional. You set the output proportional to the error. If you're off by a lot, the output is large, if you're off by a little, the output is small. When you want to go 70mph but you're going 20, you step on the gas hard. When you want to go 70mph but you're going 60, you step on the gas less hard.
I - Integral. You add up the error every time through the loop, and you can get a big correction over time, even if the immediate correction is generally small. If you're going 54 you wait a while to see if you hit 55, but after a while you realize you're not getting there, and you give it more.
D - Derivative. You look at the rate of change, from this loop to the last. If you're losing 15 mph per second, you might give it a little gas BEFORE getting to your target so you don't over shoot.
The MS PID code I've seen on the MS-II (I have no idea what the MS-I uses, so maybe I'm falsely accusing here), but on MS-II the P term is really an integral, the I term is a double integral, and the D-term is some sort of integrated derivative.
The result is that if you make your P-term anywhere near as large as it should be, you get oscillations. If you don't make it big enough (and D, as well), sudden impulse changes won't react how they should, and your motor stalls.
I'm still a fan of just shutting off the A/C with a relay whenever RPM dips below some set point.
The difference in the target and the measured is the error, and you tweak the output to try to get the error to zero.
P - proportional. You set the output proportional to the error. If you're off by a lot, the output is large, if you're off by a little, the output is small. When you want to go 70mph but you're going 20, you step on the gas hard. When you want to go 70mph but you're going 60, you step on the gas less hard.
I - Integral. You add up the error every time through the loop, and you can get a big correction over time, even if the immediate correction is generally small. If you're going 54 you wait a while to see if you hit 55, but after a while you realize you're not getting there, and you give it more.
D - Derivative. You look at the rate of change, from this loop to the last. If you're losing 15 mph per second, you might give it a little gas BEFORE getting to your target so you don't over shoot.
The MS PID code I've seen on the MS-II (I have no idea what the MS-I uses, so maybe I'm falsely accusing here), but on MS-II the P term is really an integral, the I term is a double integral, and the D-term is some sort of integrated derivative.
The result is that if you make your P-term anywhere near as large as it should be, you get oscillations. If you don't make it big enough (and D, as well), sudden impulse changes won't react how they should, and your motor stalls.
I'm still a fan of just shutting off the A/C with a relay whenever RPM dips below some set point.
#49
Yeah, I know what a PID controller is. Do you know why it was done with all the terms integrated on the MSII?
This sounds like it would just be P then, no?
P->I
I-> double I
D-> P
I've looked at the MS1 idle code, but haven't really scrutinized it.
D-term is some sort of integrated derivative.
P->I
I-> double I
D-> P
I've looked at the MS1 idle code, but haven't really scrutinized it.
#50
There's a control theory term called a "z-transform", which it looks like they applied, but they got "alpha, beta, and gama" confused with Kp, Ki, and Kd. Basically, they dropped some of the corrections, so they never subtract the contribution from the previous D-term... You'll just have to read the code. Since the MS-II stuff is a rewrite, we might be talking about something which doens't apply.
#51
Well, glad it's sort-of working for you.
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.
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.
#53
Boost Pope
Thread Starter
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,339
Total Cats: 6,793
Assuming that it's wired as drawn, the vacuum leak should only operate when the compressor and fans operate. They are all on the same control line. If the pressure or temperature safety switches open, the everything will shut off, including the vacuum leak. Where exactly did you connect the solenoid?
#54
Well I *think* I connected the red/yel to the wire at 1Q, but now I suspect I must've gotten it wrong because actually my A/C just doesn't work anymore (the switch doesn't turn the compressor on). I assume I toasted something. This would explain the behavior I described in the previous post. :( The whole setup worked for a little while and then stopped and now 1Q is stuck at 0v i think (from what I could tell, 0v means compressor is on, 12v means it's off?)... but the compressor never turns on.
Maybe the switches that drive 1Q aren't rated for the kind of current that the solenoid draws?
The only other caveat is that I'm using an MSPNP, so I didn't actually have to do any of the other wiring described in this post since my A/C worked anyway. I only had a problem with the idle droop when the compressor kicked on. Of course now that my A/C doesn't work I guess I don't have an idle droop problem anymore! :P
Edit: the A/C doesn't work regardless of whether or not I connect 1Q to red/yel.
Edit2: I had this totally wrong, actually my compressor is ALWAYS on except for when the safeties turn it off, I think my A/C switch broke. (see post below)
Maybe the switches that drive 1Q aren't rated for the kind of current that the solenoid draws?
The only other caveat is that I'm using an MSPNP, so I didn't actually have to do any of the other wiring described in this post since my A/C worked anyway. I only had a problem with the idle droop when the compressor kicked on. Of course now that my A/C doesn't work I guess I don't have an idle droop problem anymore! :P
Edit: the A/C doesn't work regardless of whether or not I connect 1Q to red/yel.
Edit2: I had this totally wrong, actually my compressor is ALWAYS on except for when the safeties turn it off, I think my A/C switch broke. (see post below)
Last edited by holy driver; 09-01-2008 at 09:43 PM. Reason: I had this wrong, compressor is ALWAYS on
#55
I just want to confirm, will the instructions below also work on the MSPNP?
Chapter II: The 1.8 NA
Fundamentally, this system is exactly the same as the 1.6, with separate relays for clutch and condenser, and of course the main fan is driven by the ECU on these cars. I'm going to assume that you've already wired the fan driver of the MS to control the main fan on your car. Doing this mod has the side-benefit of giving you parallel cooling fan operation.
As before, two mods in one. The "simple" mod, in green, will operate the A/C whenever the button is pressed and the pressure and temp safeties are satisfied. The mod in red/purple disconnects the A/C at WOT, and requires a relay driver on the MS, configured exactly as in the 1.6 writeup.
Fortunately, the schematic and wire colors appear to be identical for all cars '94-'97, at least insofar as my Haynes manual is concerned. The only apparent difference is the positions of the various wires at the ECU connector.
In either case, note that both fan relays are hard-wired together, and the diode is there so that when the MS turns on the fan driver only the fans come on (not the A/C clutch) but when the A/C button is on, the clutch and both fans are energized.
Chapter II: The 1.8 NA
Fundamentally, this system is exactly the same as the 1.6, with separate relays for clutch and condenser, and of course the main fan is driven by the ECU on these cars. I'm going to assume that you've already wired the fan driver of the MS to control the main fan on your car. Doing this mod has the side-benefit of giving you parallel cooling fan operation.
As before, two mods in one. The "simple" mod, in green, will operate the A/C whenever the button is pressed and the pressure and temp safeties are satisfied. The mod in red/purple disconnects the A/C at WOT, and requires a relay driver on the MS, configured exactly as in the 1.6 writeup.
Fortunately, the schematic and wire colors appear to be identical for all cars '94-'97, at least insofar as my Haynes manual is concerned. The only apparent difference is the positions of the various wires at the ECU connector.
In either case, note that both fan relays are hard-wired together, and the diode is there so that when the MS turns on the fan driver only the fans come on (not the A/C clutch) but when the A/C button is on, the clutch and both fans are energized.
#56
If you read the first post NONE of these instructions are garunteed to work on ANYTHING.
These are just ideas from looking at the schematics.
What the PnP, not being "open", does internally, no one knows.
You can try it, but I personally wouldn't unless I was prepared to have something break, or to do enough testing to ensure things don't. Do you feel you understand electronics to that level?
I thought the PnP already ranthe AC?
These are just ideas from looking at the schematics.
What the PnP, not being "open", does internally, no one knows.
You can try it, but I personally wouldn't unless I was prepared to have something break, or to do enough testing to ensure things don't. Do you feel you understand electronics to that level?
I thought the PnP already ranthe AC?
#57
Let me reword that... has anyone else has had success on the MSPNP and the parallel fan wiring mod with AC?
The MSPNP has AC but not parallel fans.
I wired 1L and 2S (94-95 miatas) together per a thread on miata.net but the AC clutch is on all the time, even with the fan and AC switch off. I was wondering if the "simple mod" in green in the diagram will correct this issue?
Thanks in advance.
The MSPNP has AC but not parallel fans.
I wired 1L and 2S (94-95 miatas) together per a thread on miata.net but the AC clutch is on all the time, even with the fan and AC switch off. I was wondering if the "simple mod" in green in the diagram will correct this issue?
Thanks in advance.
If you read the first post NONE of these instructions are garunteed to work on ANYTHING.
These are just ideas from looking at the schematics.
What the PnP, not being "open", does internally, no one knows.
You can try it, but I personally wouldn't unless I was prepared to have something break, or to do enough testing to ensure things don't. Do you feel you understand electronics to that level?
I thought the PnP already ranthe AC?
These are just ideas from looking at the schematics.
What the PnP, not being "open", does internally, no one knows.
You can try it, but I personally wouldn't unless I was prepared to have something break, or to do enough testing to ensure things don't. Do you feel you understand electronics to that level?
I thought the PnP already ranthe AC?
#58
Heh - ok then. Er, if nothing is going to 1J, 1Q - no diodes, wires, etc, it really should work awesome - and the AC shouldn't be effected at all - again, barring strange things in the PnP. I'd say, as a start, verify you've ONLY connected 1L and 2S (I'm trusting Joe's numbering here)... Another easy test. With them disconnected, and the car on (running or not), ground out one, then the other, of those pins. Have a friend poke his head under the hood. One, then the other, fan should turn on, independantly, and the AC clutch shouldn't move (when it's "on" you'll see it pull in, it's pretty obvious, you'll even hear a click).
If all that checks out, then you've got me stumped. :-)
If all that checks out, then you've got me stumped. :-)
#59
Well I *think* I connected the red/yel to the wire at 1Q, but now I suspect I must've gotten it wrong because actually my A/C just doesn't work anymore (the switch doesn't turn the compressor on). I assume I toasted something. This would explain the behavior I described in the previous post. :( The whole setup worked for a little while and then stopped and now 1Q is stuck at 0v i think (from what I could tell, 0v means compressor is on, 12v means it's off?)... but the compressor never turns on.
Maybe the switches that drive 1Q aren't rated for the kind of current that the solenoid draws?
The only other caveat is that I'm using an MSPNP, so I didn't actually have to do any of the other wiring described in this post since my A/C worked anyway. I only had a problem with the idle droop when the compressor kicked on. Of course now that my A/C doesn't work I guess I don't have an idle droop problem anymore! :P
Edit: the A/C doesn't work regardless of whether or not I connect 1Q to red/yel.
Edit2: I had this totally wrong, actually my compressor is ALWAYS on except for when the safeties turn it off, I think my A/C switch broke. (see post below)
Maybe the switches that drive 1Q aren't rated for the kind of current that the solenoid draws?
The only other caveat is that I'm using an MSPNP, so I didn't actually have to do any of the other wiring described in this post since my A/C worked anyway. I only had a problem with the idle droop when the compressor kicked on. Of course now that my A/C doesn't work I guess I don't have an idle droop problem anymore! :P
Edit: the A/C doesn't work regardless of whether or not I connect 1Q to red/yel.
Edit2: I had this totally wrong, actually my compressor is ALWAYS on except for when the safeties turn it off, I think my A/C switch broke. (see post below)
edit: i took out the heater control to get to the A/C switch, and as far as I can tell the switch seems to be operating correctly. Also with the switch disconnected, 1Q is still at 0v. I took out the megasquirt and found that pin 1Q on the MS seems to be shorted to ground. So I think I fried the MSPNP's magical A/C circuit. :( So... if you're trying to do the idle compensation with an MSPNP, you should probably think twice before diving in.
Last edited by holy driver; 09-02-2008 at 01:12 AM. Reason: more info
#60
Uhg! Really sorry to hear that!
Well, if the PnP people won't get pissed, get a picture inside the box, maybe we can help sort this out.
Also, is there continuity to ground, or just 0V?
None of the currents invovled are huge - then again, there are places (like using 200mA LED drivers to run fan relays?) where some of the conversion from "run a distributor chevy" to "run a modern motor" where the MS is a bit of a hack. It's not unreasonable it wasn't set to run the extra current from a relay though I'm still pretty surprised by it.
Well, if the PnP people won't get pissed, get a picture inside the box, maybe we can help sort this out.
Also, is there continuity to ground, or just 0V?
None of the currents invovled are huge - then again, there are places (like using 200mA LED drivers to run fan relays?) where some of the conversion from "run a distributor chevy" to "run a modern motor" where the MS is a bit of a hack. It's not unreasonable it wasn't set to run the extra current from a relay though I'm still pretty surprised by it.