Solution To Flickering Fan Speed 2 Or 3 A/C PART 2

[pasadena_commut]

I believe they changed the voltage requirements on the circuit board of the unit

Fine, but did they do that by changing the ratio between a couple of 1/4W resistors or is there a completely different circuit? I suspect the former. If all they did was change a few reference voltage points the easiest way to do that is by changing the values on the passive components (resistors, maybe a diode) in that part of the circuit. That is the sort of upgrade that can be applied to existing boards.

If we had the circuit diagram for this controller it would be easy to determine, but I have never been able to find one.

 

[pasadena_commut]

If we had the circuit diagram for this controller it would be easy to determine, but I have never been able to find one.

Found a diagram of the overall circuit on page 44 of the wiring diagram manual. Unfortunately it shows the majority of the climate control unit as "circuit", which isn't too helpful. On the plus side it does show that there is a single wire from the climate control unit to the blower resistor pack, where it attaches to the Blue/Yellow wire. (Schematically, in practice the connection is probably nearer to the control unit). So this wire must be the sensor that is giving us problems. This gave me a clue as to where to measure (at the connectors on the resistor pack under the dash) and what might be needed to fix this. It was a pretty cold day so positions 1,2, and 4 were working, and 3 was going in and out.

Blower    Wire             Voltage at      AC lit?
Position  color            Blue/Yellow
1           Blue/Yellow     .45              Y
2           Blue           1.04              Y
3           Blue/Red       1.15              N
4           Blue/white      .89              Y

When the voltage fluctuated below around 1.10V AC came on for position 3. The voltage was measured from the blue/yellow wire at the resistor pack to a chunk of bare metal in the center console. I'm not sure where the .45V is coming from on the first measurement, since in that switch position Blue/Yellow is switched directly to ground. In the other positions it is being pulled down by whichever wire is grounded, but current through the resistor pack from the sensor, and also from the blower back through the wiring, is pushing it back up.

If the blue/yellow wire was to be shorted to ground the AC would work in all fan positions. Unfortunately, the fan would also be blowing on low even when the switch was on 0, and it would slightly change the speed of 2 and 3. A better method would be to use two diodes to pull down on the signal line (Blue/Yellow) form Blue/Red (position 3) and Blue (position 2). The problem here is, assuming that the voltage at each of the grounded wires is also .45 (I didn't measure it yet), there is a .7 to 1V drop across a forward silicon diode if there is much current flowing through it, which adds up to more than 1.10V. The only diode I had laying around was a big power diode, and putting it in the circuit in this manner did not change the voltage on blue/yellow more than it was already fluctuating. Guess I'll drop by the electronics store and see if I can find a better diode for this.

One might also wonder why the original controller was set to <= 1.1V, when they had around 12V to work with. Surely they could have set it as <3V and then this problem would never have arisen.

 

[pasadena_commut]

Hmm. I wonder what would happen if one put a 10K resistor between Bue/White (position 4) and Blue/Yellow (position 1/sensor). I did measure the voltage on Blue/White, which is tied to the blower, on position 1, and it was around 9V. So if the switch is in positions 2 or 3 the new resistor would act like a voltage divider for the sensor line and should force that voltage way down. I do have 10K resistor, stay tuned...

 

[pasadena_commut]

Hmm. I wonder what would happen if one put a 10K resistor between Bue/White (position 4) and Blue/Yellow (position 1/sensor).

Answer, nothing. The change in the voltages, if any, was in the noise.

So I measured the voltage on each wire when the switch is positioned to ground it with these results:

Switch    Wire       Voltage
Position
1           Blue/Y    .45
2           Blue      .85
3           Blue/Red  .85
4           Blue/W    .81

The diagram does not show it, but those measurements sure looks like there is a power diode between the fan switch and the ground. The currents go up with the switch positions, and judging from the wire diameters, quite a bit, and these voltages match the IV characteristic for a diode pretty well. (Keep in mind that the voltages were jumping around a bit the whole time.) [This suggests one possible origin for this problem. If the original design did not have this hypothetical diode then the original controller design would have been fine, with the sensor voltages all <1V. However, they probably found that they were blowing out controller boards from noise and put in a protective diode, but didn't modify the climate control board. It still worked with the diode in, but was on the hairy edge, and after a bit of component aging... ]

This also tells us that my diode idea from two posts back will not work. The "grounded' voltages on these lines are not low enough to pull the sensor line down far enough reliably.

A relay might work though. Wire a low current relay in parallel with positions 2 and 3, and have that switch in a diode from blue/yellow to ground. With the right diode that should be well below 1.0V and should reliably switch on the compressor.

Failing that, we are back to the controller board. Does anybody have a controller board out of their car at this time that they could take a picture of and post the image?

 

[pasadena_commut]

It was nice and toasty today so neither fan speeds 2 or 3 would work. I temporarily added the following circuit to the system (with test clips for wires/attachment):

blue/yellow wire -> switch -> [anode] 1N4004 diode [cathode/band] -> ground

With the switch open all was like before, however with the switch closed I measured the following between the blue/yellow wire and ground:

Switch    Voltage       mA   AC Works?
Position
0           .75       2200   (not tested, oops)
1           .45          0    Y
2           .76         60    Y
3           .77         70    Y
4           .74         32    Y

So success in terms of resuscitating the AC in positions 2 and 3.

With the switch closed position 0 blew somewhat less strongly than on position 1. Stupidly I did not try turning on the AC in position 0 with the switch closed, but I imagine that it should work there too. It would of course be even more futile than running the A/C on position 1, which in LA at least, doesn't cool the car enough to notice.

The downside to this fix is the 2.2A through the new pathway if the fan switch is off. (I don't have a lot of confidence in that number given the crappy DC ammeter I used, but a better one with a limit of 500mA maxed out, so it was definitely more than that.) It isn't a huge amount of power IV -> 0.75 * 2.2 = 1.65W, but it was enough for the diode to heat up a bit, and that part is considered a 1A part, and is only rated for 3W continuous. There are power diodes that bolt through holes to their ground/heat sink, which would be perfect, but you aren't going to find one at the local Radio Shack. The wires didn't heat up appreciably, even though they were very thin, and only rated for low voltage electronics work.

The circuit diagram does not show a diode in the return path for the fan switch, but electrically there does seem to be one there, and I assume the engineers had a good reason for adding one. For that reason I think we probably need one in the alternate path too. That said, if we are to believe the rest of the circuit diagram there is a diode in the control unit on the sense line, and no other connections to it, so it might be safe to connect the blue/yellow wire through a switch directly to ground. That would eliminate the hard to find power diode.

 

[DeCaro]

Hello, unless I missed something, I dont believe I read this issue.

I have been having all the same flickering issues for some time now. My A/C only works on the Number 1 setting. After reading this thread, I pulled apart my console to pinch the connectors for a temporay fix. Upon disconnecting the switch form the connector, I revealed this. See attached Pic. Does any1 think it could be a problem with the Blower Motor Resistor. I read a little about this in other forums.



Uploaded with ImageShack.us

 

[pasadena_commut]

Upon disconnecting the switch form the connector, I revealed this. See attached Pic. Does any1 think it could be a problem with the Blower Motor Resistor.

Probably not.

I don't have the wiring diagram in front of me, are the burned up ones on thick wires or thin wires? If the thin ones then maybe a Resistor pack problem, but I still doubt it.

The resistor pack is wired like this:

blower - 4 -R1- 3 - R2 - 2 - R3 - 1

The fan selector switch connects points 1->4 to ground. So even if the resistor pack was completely blown (all resistors open) position 4 should still work and presumably the power connector is rated for that much current. Now if resistors R1/R2/R3 were shorted there would be too much current for the ratings on wires 1,2,3 and the wire itself might get quite hot, and MAYBE that might be enough to melt the housing like you see. But more likely the insulation would have melted off the wires.

Since you say it works on position 1 that tells us that R1/R2/R3 are not open.

Three possibilities come to mind:

1. the connector was slightly loose or installed dirty, and the resulting arcing melted the
housing.

2. Dissimilar metals in the plug and socket, resulting in galvanic corrosion, then increased resistance at the junction and local heating. (This would be incredibly stupid, but look up the history of aluminum wiring in houses...)

3. The blower is partially shorted, resulting in too much current on the higher wires in the normal positions. The lower wires are protected from that extra heating, to some extent, by the blower resistors ( I=V/R).

The most likely resistor pack problem is one of the resistors blowing open. When that happens positions below that do not work (no electrical connection to the blower), and for the positions where the fan blows the compressor will never light (because there is a sense line that must be pulled low attached to line 1, and it will be floating).

 

[pasadena_commut]

OK, have the diagram in front of me. The wires facing into that are

[FONT="Courier New"] 
Gnd 1 2
 *  4 3
[/FONT]

So too much current through the connector on position 4 (high). The resistor pack is completely out of the circuit in position 4, so it didn't cause this. Unclear why only 1 works, 2 and 3 look like they are in OK shape, but maybe those wires are shorted or burned through behind that plug?

Things to check:

1. Resistance across the blower. (I don't know what it should be, but it must be in the manual - see the sticky at the top of this forum.)

2. The blower fuse. The diagrams shows a 15A, make sure that somebody didn't stick a larger one in there to "compensate" for a partially shorted blower. We all know the type, "Gee, the blower fuse keeps opening, it still blows though, lets just put in a bigger fuse..."

3. Check the switch side of that connection. If it isn't too messed up you may be able to clean it off. Otherwise, replace the switch when you replace that part of the wiring harness.

It may not be easy to pull out the existing wire, but you could leave it in place and cut it off at both ends. Then put in another cable (if you can find those ends somewhere) with similar gauge automotive wire, and there is plenty of space under the dash to route it the 3 feet or so it needs to go, tacking it down along the way with cable ties. 5 of those wires run straight to the blower resistor, but the L/Y one (fan position 1) also makes a connection to the L/Y wire on connector G1-01 (the big one on the back of the climate control unit.) That last wire is just a sensor line, so I think a wire tap should be good enough for the little bit of current it needs.

The wiring diagram for all of this is "Heater and Air Conditioner" in the shop manual.

 

[DeCaro]

Thanks for your help. I will update my findings when I am able to.

 

[DeCaro]

You know what? I forgot to say that after reconnecting the switch, after it's been out of the car for a day, all positions, except for #3 work consistently. #3 is intermittent and the switch gets warm.

 

[pasadena_commut]

You know what? I forgot to say that after reconnecting the switch, after it's been out of the car for a day, all positions, except for #3 work consistently. #3 is intermittent and the switch gets warm.

The FAN works intermittently or the compressor light?

If the latter, get a piece of wire and with the switch in position 3 temporarily short from the L/Y wire at the resistor pack connector to ground (any bare metal on the chassis.) The A/C light should come on solid and the A/C should blow cold. That isn't a fix, but it tells you the problem is in the controller board and not the resistor pack or even the wiring. Look back a few posts to see why this works.

 

[jdwk]

So I got an updated unit from ebay as suggested by another forum member. Turns out the jackass at the junk yard just cut the cables instead of pulling them off the vents.

I thought maybe I could pull the cables off the control unit and swap them with mine. This turns out to be pretty much impossible. The plastic prongs are a one way only affair.

So my only solution was to swap the circuit board from the updated module, which is a BL8F D, vs my BL8W D. The circuit boards had exactly the same part number on them though. In fact, they looked 100% identical except, the two black plastic pieces below the switches (I think these are just bump stops for the A/C and defrost switches), were gray on the updated module.

The only other difference is in the top right corner of both circuit boards, mine has a 1, and the new one has a 2. So maybe some resistance value was changed that is not discernable with the naked eye. I was too lazy to go around and measure them all to compare.

Well anyway, I put it all back together and it's exactly the same behavior. A/C light is steady in 1 or 4, but flickers and turns off in 2 or 3.

So I am pissed, but realize the circuit board really isn't the problem, it's the connector. So I take them both apart again and this time swap the connector which is one piece with the fan control.

Now, the A/C is steady on any fan speed, but it still turns off in between speeds. I guess this is normal behavior. I don't like it, but my car has always done it, so I guess I just have to live with it.

I am still pissed at the junk yard for cutting the cables, that is bush league.

 

[pasadena_commut]

So my only solution was to swap the circuit board from the updated module, which is a BL8F D, vs my BL8W D.

Usually the revision is the last letter. Are those maybe the sedan vs. P5 boards both at revision D?

 

[jdwk]

Yes the one from ebay is from a P5, I have an MSP. What is odd is that the lights on both are exactly the same color, A/C and fresh air light are green, defrost and recirc are amber. According to MJO's post, mine should have an amber A/C light, but that is definitely not the case.

 

[Tzar177]

So I got an updated unit from ebay as suggested by another forum member. Turns out the jackass at the junk yard just cut the cables instead of pulling them off the vents.

I thought maybe I could pull the cables off the control unit and swap them with mine. This turns out to be pretty much impossible. The plastic prongs are a one way only affair.

So my only solution was to swap the circuit board from the updated module, which is a BL8F D, vs my BL8W D. The circuit boards had exactly the same part number on them though. In fact, they looked 100% identical except, the two black plastic pieces below the switches (I think these are just bump stops for the A/C and defrost switches), were gray on the updated module.

The only other difference is in the top right corner of both circuit boards, mine has a 1, and the new one has a 2. So maybe some resistance value was changed that is not discernable with the naked eye. I was too lazy to go around and measure them all to compare.

Well anyway, I put it all back together and it's exactly the same behavior. A/C light is steady in 1 or 4, but flickers and turns off in 2 or 3.

So I am pissed, but realize the circuit board really isn't the problem, it's the connector. So I take them both apart again and this time swap the connector which is one piece with the fan control.

Now, the A/C is steady on any fan speed, but it still turns off in between speeds. I guess this is normal behavior. I don't like it, but my car has always done it, so I guess I just have to live with it.

I am still pissed at the junk yard for cutting the cables, that is bush league.

All you needed to do was swap over you existing cables. Its easy. Theres a how to somewhere

 

[hurleyblink]

Just did mine only 1 would work before lol I could get 4 to work intermittantly I also disassembled the switch and lightly grinder the copper contacts and I'm solid green even going between speeds make sure you dont use rosin core solder it was a b**** on the larger wires. But this does work

 

[jdwk]

Swapping the cables is not easy, I don't care what kind of "how to" exists. Maybe if you blasted it with a heat gun to melt the plastic a bit, but the prongs are clearly dug through the cable's covers meaning they aren't coming off without ripping off the covers.

I e-mailed the wrecker and they apologized and actually refunded my money, fairly awesome customer service there. I'd feel bad if it was actually easy, but as much force as I put on that cable, any more would have just snapped the whole unit.

 

[jdwk]

Son of a b****! That lasted two days, it's now worse than before. 1 works, 2 and 3 don't work at all, and 4 flickers. And the high tomorrow is 99. Fantastic.

So word to the wise, BL8F D, is not a "no issues" revision. Looks like I'll be pulling the connector apart and cleaning cleaning the contacts and/or soldering.

 

[hurleyblink]

Just pull the plastic piece off cut the metal female ends off and solder. While your in there I would clean the switch off as well just to be sure it works

 

[jdwk]

Doesn't soldering the connector to the harness make pulling the radio bezel a huge PITA, or am I missing something?

I did take the fan switch connector apart and sanded down the contacts tonight. However the ball bearing (that I was not aware existed) got out of track so when I put it back together the knob wouldn't turn. So I took it back apart, and when I did, the plastic housing of the connector snapped in one spot. I didn't even notice until I put it back together again and realized it wasn't tight on one side.

Luckily I had two of them, so I grabbed the other connector, but this time hit it with a heat gun first to prevent it from snapping like the other one. I actually gave it a little too much heat gun and it's a little warped now, but at least it didn't snap.

I just re-used the same contacts I had sanded down earlier, and gave that a shot. It still likes to turn off in between 1 and 2 for a split second, but at least it was steady once it was fully in a number. Of course I thought this a few days ago as well.

The only plus here is I am getting very good at pulling the bezel now.