A Guide to Electric Fan Retrofit / Swap

[oddball]

First, big thanks to joegr for posting signal traces in this thread - wouldn't been able to do this without that info. Also, go check out the excellent threads by hite337 on the pump bypass and immusicman on his conversion for background info. There are many other threads that touch on this topic.

Warning: USE AT YOUR OWN RISK! FIT AND FUNCTION ARE NOT GUARANTEED! NO WARRANTY, EXPRESS OR IMPLIED, IS PROVIDED FOR THIS MODIFICATION! This is working OK for me, but we'll see what the Texas heat brings next summer.

Goal: replace gen I hydraulic fan system with the gen II electric. Want to keep the intelligent fan control functional. Yes, this can be done cheaper with aftermarket fans, but the controls are typically lacking. This integration allows the PCM to select the fan speed from very slow to very fast depending on conditions. You can also get to my full flickr feed of LS mechanical work w/ high res copies of the pics. No pretty pictures, sorry.
www.flickr.com/photos/ostermannia/sets/72157625387968918/with/5176894689

Stuff you need:
Gen II fan and wiring pigtails: note that 03/04 has two connectors while 05/06 has one. Wiring, colors and everything else is identical.
Alternator (possibly)
two sets of two pin weatherproof connectors
6 gauge wire, ring and butt connectors
18(ish) gauge wire and various connectors
Provisions for control mechanism of your choice (see below)

Electric fan basics:
The power and ground should be connected at all times. No relays or switches. Get a 60 amp fuse. The fan pulls 43 amps at full blast.
Big red and big black wires are obvious. The small green wire feeds power to the in-fan computer and is on in run and start for the stock install.. The white/black wire is the signal line that carries a PWM that instructs the fan on what speed to run. The really nice thing is that if the fan computer can't figure out the signal, it'll fail-safe to full speed. The signal line can be grounded, hot, floating, whatever, without damaging the fan. So if all else fails, just run the sucker full blast. Fan will only run when power is applied to the computer. Although, if it's bootstrapped, then power removed with the signal still connected, it can sometimes remain running. My initial install just left the signal wire floating. That is not a very good choice, however.

Other fun notes: fan pushes a significant ripple into the car's power rail when running full blast. Looked like about 1V @ 2khz (ish). Audio dudes beware.

Power:
If you have a bunch of accessories/options, then you'll need a stronger alternator. The stock is barely adequate for the car with all options. Big stereo? Get an alternator. Seat heaters? Get an alternator. My car does not have the heaters and only the stock radio, no other gadgets. See my other thread about my experience with DB Electrical alternator vs stock.

Removal:
Remove old fan and hydraulic pump. Remove all hoses. Plug cooler inlet/outlet. I had to drop the a/c condenser to get to the outlet on the cooler. That took an hour alone. You do not need to disconnect the a/c drier, although you may need to loosen various hose mounts to get enough movement.

Connector and alternator:
While alternator is off, cut off the old actuator connector and add your own two pin connector. I did this after everything was back on, and it's REALLY hard to solder between the engine and the frame. Keep track of which pin is which. Green is power, Brown is signal.
Install alternator and new belt. My belt was 100 1/4" Gates - normal stock part at my local store.

Fan modification and installation:
See hydraulic fan vs electric Front:

Back:

Note that the hyd blades are much larger. This means it can spin slower to move the same amount of air, but is harder to turn. So the electric will spin faster - and be louder - compared to hyd. Electric does not have water pump mounts, but that's OK. A/C drier mounts just fine.
**ALL PROTRUSIONS ON THE PASSENGER SIDE MUST BE GROUND OFF!** Not the mounting tabs - the stuff for things to mount to the shroud. Anything that sticks out towards the engine. We spent a couple of hours trying to wedge it in there in various conditions, finally just ground down everything. For the Gen II, the DCCV actually mounts on the shroud. All that crap just bumps into things.
Make sure to get the tabs into the slots on the radiator. This is also much easier with the upper rad hose disconnected.

I ran the ground to the body bolt under the air box:

Be sure to clean under the washer to make sure it's bare metal. Hit it with a wire brush.
I pulled the positive along the big wire harness in the front. Got a MAXI fuse block from Wal-Mart - after trying three auto sound stores and Best Buy - and a 60 amp fuse:

Routed through existing harness:

Tied it into the big lug on the fuse panel:

I used 6 ga for the power lines to match those on the harness. The feed line to the fuse panel is a 4 ga, so it should be fine with this extra load. Again, YMMV. Be careful, monitor it for overload.
Run the signal and fan control power lines along the front and towards your new connector.

Control: The Simple Way
Add the opposite end of your connector type onto the fan wires you just ran. Green to Green, Brown to White. This will give you full fan any time the ignition is in Start or Run. This is an excellent fail-safe. So regardless what else you do, have this as an option in case things go nuts.
Problems: In the cold the engine will have a hard time warming up. This can be very bad for you engine if you do short drives in the cold. Extra power load when not required.


Control: The Slightly Effective, but Stupid Way
Note: THIS IS REALLY STUPID.
I was ill-informed at the time and made this monstrosity:

Tie the PCM power and signal to 85 and 86 (either way). Also tie PCM power to 30. Tie Fan power to 87. Put in a diode from PCM signal to PCM power to protect the PCM when the field collapses.
OK, WTF did this just do? I found that when the PCM requests mid to high power, it can throw a relay. So when it does, you're now powering the electric fan computer, and it'll run full speed.
Problems: This is REALLY STUPID because the relay will chatter as the PCM changes its control signal. This will kill the relay. The rapid field collapses might fry your computer (although I ran like this for two weeks fine). The relay will die quickly. Mine would stick on. But, for cold mornings, this left the fan off while the engine warmed up. A quick whack to the relay when I stopped the car reset it. Annoying, but quick, easy, cheap, and slightly effective.



Control: The Very Effective, but Difficult Way
Why, just adapt the Gen I signal to the Gen II signal. Easy!
Ha.
Granted, I'm horrible with analog circuits, but I spent around 60 hours getting this to work. At least half of that was just me not being very good at analog work.
Gen I is a PWM, period of 10ms, ranging from 95% duty cycle for minimum speed to 10% duty cycle for maximum speed. The Gen I *switches ground* to make the PWM. Key note. If you hook up a standard multimeter you'll get some funny results (like a smooth resistance curve, which makes it look like it just changes resistance to ground, which can cause you to spent two weeks designing around that, just to learn that isn't what it's doing at all and getting really mixed up).
The Gen II signal has period 7 ms, ranging from 5% duty for minimum speed to 95% for maximum speed. Note that if you fall under ~5% duty then it'll lose the signal and run full speed.
I built a circuit based on the TI TL494. This can be done with any similar chip, and in lots of other ways.
Schematic:

I used a 50k pot on the voltage divider feeding DTC. It's set around 14k right now. Adjusting that changes when the output curve starts. I didn't get a perfect match (ie, 90% in = 10% out, 50/50, 10/90, etc), but pretty close. Changing this value lets you either delay when the fan engages - and reduce max speed - or engage sooner and get to max speed sooner. It's just a trade. I expect to adjust it again during summer. Upping the voltage divider by an order of magnitude might be a good idea. The PWM output resistor set can be higher - it was stable at 470 ohms, shaky above 600 ohms. The set of four 1k's is about 270 ohms (variance due to 5% resistors). Well within the 1/4 watt rating of the resistors, but may waste a bit more energy than required. The 1k load on the PWM input was also a complete guess. The input filter resistor/cap was determined by experimentation. That set is stable, but there may be better values. The timing resistor is the actual tested value I used. Any variation will change the output PWM frequency. I didn't try to test the timing capacitor for its true value, though. A scope is invaluable in testing.
I got a better curve match when using the feedback input, but that required an op amp voltage follower to buffer the PWM input filter and a different voltage divider. Way more complicated and used quite a bit more power, so ditched that idea.

And a few pics of the final piece. I added some rs232 pins for taps for testing.

I pulled ground from a body bolt on the shock tower then velcro'd the box to the side of the fuse box. Not the cleanest install, but easy to get to. Once I have a few months on it I may hide it under the fuse box.

Shazaam!
This is working so well I may actually do something similar for my '72 Cutlass.

search engine help: fan swap gen 1 I to gen 2 II hydraulic fan replacement

 

[Silver_LS]

Nice Job.

 

[De-marko]

Good write up!

Why don't you make and sell the conversion kit for us with instructions. Then we can go out and source our own fan and then install everything

 

[oddball]

Cuz you wouldn't want to pay me what I'd want to charge for that little board.

If someone's seriously interested I can make it for them.

 

[FASTLLS]

Nice

 

[owlman]

Cool! I enjoyed reading that. Ingenuity, engineering, and a happy ending

 

[StanE]

Great work - this is the sort of intelligent research and practical application information and advice that is absolutely invaluable to our forum membership.
This one shouldbe made a sticky.

 

[hite337]

good stuff right there. nice work oddball.

43 amps is a huge draw on ur alternator. i didnt realize the factory electrics were rated that high. im glad i went with the flex-a-lite...only 18 amps so no need to upgrade my factory alt. but im sure the factory electric looks better when installed.

btw...how much of a pain was the hydraulic fan pump removal? lol

 

[oddball]

Thanks for the comments!

I did some testing on the alternator compared to the DB electrical and posted that in the other thread linked in the writeup. It is a huge load, but the stock keeps up. Fingers are crossed.

Hyd pump had one frozen bold. Had to grind the head off. Same problem on the power steering pump, too. Several hours of my life I wish I could have back.

I was worried about getting enough air movement for the Texas summer with AC with an aftermarket. I bet MO gets fairly toasty, too. We'll see how it acts this summer.

 

[hite337]

the bolt behind the hydraulic line on the top of the pump was frozen on mine. i broke 4 10mm socket and 2 7in socket extensions before i took a punch and scored the top of the pump and hit it with a cold chisel and it split. then i got the bolt out with vise grips. several hours i'd like to have back as well

yes very warm in MO. my fan kept up well this summer. it didnt overheat at all. i will be upgrading my alt soon just to be safe. it had hyd fluid leaking on it from the pump so its just a matter of time before it goes i'm sure.

 

[oddball]

Just a follow up that during the 100+ degree days here in TX with the A/C blowing around 50 degrees, this swap keeps the engine around 230, which seems to be right about where the computer wants it. So I'm declaring success.

For anyone interested in a different spin on this, I was able to drive the Gen2 rad fan from an Arduino. More expensive than the TL494 method, but would've been easier (and cheaper considering the number of parts I fried the first time around)

Here's some sample code. NOTE: THIS IS EXAMPLE ONLY AND IS INCOMPLETE! USE AT YOUR OWN RISK! MAKE SURE YOU FULLY UNDERSTAND THE CODE AND YOUR SPECIFIC SCHEMATIC FOR APPLICABILITY!

/* 
   LincolnFanControl
   
   USE AT YOUR OWN RISK. FOR REFERENCE ONLY. THIS CAN EASILY DESTROY PARTS AND CAUSE BODILY HARM.
    
   Radiator Fan:
     Used a radiator fan from a '03 - '06 Lincoln LS. They utilize a PWM with period 7ms to 
     select the fan speed, entirely variable from 0% to 100%. This signal is to instruct the fan's
     circuitry ONLY - a high impedence input. Note that an input signal of less than ~5% duty 
     results in the fan "losing" the signal and defaulting to full speed. In Arduino speak, the
     minimum safe duty value is 50 (out of 1024).
     Use Timer1 for high accuracy on frequency.
     Note that the fan can operate with a 5V PWM signal.
     
     This uses the library near the bottom of this thread:
     http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1238817116
     I saved it as TimerPWM.h and TimerPWM.cpp under libraries/TimerPWM then inported into the sketch.
     Note that the code as-is on the thread isn't quite right. Needs #ifdef/#define protection of the 
     header file.
     
     The get_rad_fan_duty() function needs to be updated to reflect your chosen input. One example
     is the 2000-2002 Lincoln LS which has a PWM input to the hydraulic fan control solenoid which
     varies from 100% for minimum speed to 0% for maximum speed. Feed that into an R/C to get a smooth
     voltage input.

     WARNING: The A0 RAD_FAN_INPUT **MUST** BE LIMITED TO NO MORE THAN 5V! A direct R/C would give up
     to 14+V which would quickly kill the Arduino. A complete implementation will, at a minimum, require
     a divider circuit to bring the span down to 0V...5V. It may even require a voltage follower to 
     buffer the R/C output.
       
     
*/

#define DEBUG
#define BLINK

#include <TimerPWM.h>
TimerPWM  wiggler;

#ifdef BLINK
int led_state;
#endif

// Pin definitions
#define RAD_FAN_OUTPUT         9
#define RAD_FAN_INPUT          A0


#define MAX_DUTY                1024  // Maximum (ie, 100%) duty


#define AC_FACTOR               30   // percentage to increase fan speed when A/C is engaged
#define RAD_FAN_MIN_DUTY        50   // minimum duty (out of 1024) - less thanfa this and the fan loses signal and goes to max
#define RAD_FAN_PERIOD          7000 // 7000 ms = 133 hz = LS Gen2 fan signal period



int rad_fan_duty;

void setup() {
#ifdef BLINK
  pinMode(13,OUTPUT);
  led_state=HIGH;
#endif

  pinMode(AC_CLUTCH_INPUT,OUTPUT);
  pinMode(RAD_FAN_OUTPUT,OUTPUT);
  pinMode(RAD_FAN_INPUT,INPUT);

  analogReference(DEFAULT);

  rad_fan_duty=0;
  wiggler.initialize(1,RAD_FAN_PERIOD); /* Rad fan expects 7ms period */
  wiggler.pwm(1,RAD_FAN_OUTPUT,rad_fan_duty,0);  
  wiggler.start(1); // ENGAGE!
  
#ifdef DEBUG  
  Serial.begin(9600);
#endif
}


int get_rad_fan_duty() {
  int duty=0;
  int request=analogRead(RAD_FAN_INPUT);
  duty=1024-request;  // invert it since input 100% = output 0%
  if(duty>MAX_DUTY) {
    duty=MAX_DUTY;
  }
  else if(duty<RAD_FAN_MIN_DUTY) {
    duty = RAD_FAN_MIN_DUTY;
  }
#ifdef DEBUG
  Serial.print("Rad fan duty: ");
  Serial.println(duty,DEC);
#endif
  return duty;
}



void loop() {
  int new_rad_fan_duty;

#ifdef BLINK
  digitalWrite(13,led_state);
  led_state=!led_state;
#endif

  new_rad_fan_duty=get_rad_fan_duty();
  // Only update the PWM if there is a notable difference in the duty
  // Changing the PWM causes a big blip in the signal, so don't want to do that every cycle
  if(new_rad_fan_duty>rad_fan_duty+10 || new_rad_fan_duty<rad_fan_duty-10){
    rad_fan_duty=new_rad_fan_duty;
#ifdef DEBUG
    Serial.print("Set rad duty to: ");
    Serial.println(rad_fan_duty,DEC);
#endif
    wiggler.pwm(1,RAD_FAN_OUTPUT,rad_fan_duty,0);
  }

  delay(500);
}

 

[oddball]

Whoops. There's some glaring errors in that Arduino code. Sorry. So if anyone wants to use it, please try to contact me or make really certain you know what you're doing. No copy'n'paste applicability here.

 

[tireman]

Congrats and thanks!

What a good article, thanks for all your hard work, and I am sure it was hard work especially on those bolts. This was for a V8, now if someone can figure out the V6? A pulley to replace the pump would be nice. I think I would just go the easy route with an aftermarket fan and temp sensor. Keep up the good work!

 

[oddball]

V6 should be the same, other than the belt routing. Just spend some time poking and see how the belt would be able to go.

 

[johnjoe]

I know this post is OLD!!! But I need HELP!!!! Im tryin to convert from the hydraulic fan pump on my 2000 Ls to a 2004 electric Ls. But everytime I i try to just connect the ground on the fan, and the positive to a source in the fuse box. (I have the black and white wire just capped off, and the green wire connected to the power) Everytime the fan cuts on the car stalls out. I NEED HELP!!! I would love to find out a simple way so the fan can turn on at a certain temp.

 

[joegr]

I know this post is OLD!!! But I need HELP!!!! Im tryin to convert from the hydraulic fan pump on my 2000 Ls to a 2004 electric Ls. But everytime I i try to just connect the ground on the fan, and the positive to a source in the fuse box. (I have the black and white wire just capped off, and the green wire connected to the power) Everytime the fan cuts on the car stalls out. I NEED HELP!!! I would love to find out a simple way so the fan can turn on at a certain temp.

Where are you tapping the power for the fan? I'd bet it's not a circuit designed for that amount of current. You may need to tap power at the starter.

 

[RigsLS]

oddball's motor,

WOW !!!!!!!!!! AMAZING

me so jealous

 

[johnjoe]

I'm tapping the power to a relay in the fuse box, I think it's relay #6, and of course the ground to a bolt on the frame

 

[joegr]

I'm tapping the power to a relay in the fuse box, I think it's relay #6, and of course the ground to a bolt on the frame

That's a pretty bad choice. It's a 15A circuit for the horn. I assume that your fan is 20A to 35A, more when it starts up.

 

[johnjoe]

I ran a 30A fuse holder in between the wires. So that it won't short circuit

 

[joegr]

I ran a 30A fuse holder in between the wires. So that it won't short circuit

It doesn't matter. You're trying to take more power from a circuit than it can deliver. As a result, you're killing power to the engine electronics. This is why the engine stalls when the fan starts. You have to find and use a circuit that can spare an extra 30A or 40A. The junction at the starter is the clearest choice that stands out looking at the wiring diagrams. There are probably others that are acceptable too.

 

[johnjoe]

Alright, I'm going to give the starter a try. I'll post the resolves once I'm done. I greatly appreciate the help.

 

[oddball]

Please re-read the OP. The main power feed to the fan should **NOT** go to an existing fuse or relay. I tied into the main lug and installed a new fuse. The fan pulls 43A on full speed, which you have it set up to do.

Relay 6 is the horn, so I hope that's not how it's wired, although it'd be hilarious.

Anyway, your wiring size is probably insufficient. Need to use 6ga, which would have been nearly impossible to tap into an existing relay. I can only assume that whatever you're tying into is overloading and killing the engine. There's already a nice 4ga run from the main power junction to the fuse box.

 

[johnjoe]

It doesn't matter. You're trying to take more power from a circuit than it can deliver. As a result, you're killing power to the engine electronics. This is why the engine stalls when the fan starts. You have to find and use a circuit that can spare an extra 30A or 40A. The junction at the starter is the clearest choice that stands out looking at the wiring diagrams. There are probably others that are acceptable too.

Alright I got it to work, but I'm hearing the way I might of wired it I got it ready for failure

 

[johnjoe]

Please re-read the OP. The main power feed to the fan should **NOT** go to an existing fuse or relay. I tied into the main lug and installed a new fuse. The fan pulls 43A on full speed, which you have it set up to do.

Relay 6 is the horn, so I hope that's not how it's wired, although it'd be hilarious.

Anyway, your wiring size is probably insufficient. Need to use 6ga, which would have been nearly impossible to tap into an existing relay. I can only assume that whatever you're tying into is overloading and killing the engine. There's already a nice 4ga run from the main power junction to the fuse box.

You seem as if you know a lot about electrical work and other stuff, I on the other hand only know power and no power. I connected the main power source to a 80A fuse holder (Maxi the same one you have in your pic) and ran it to the starter. Connected the ground wire same place you did, left the black and white wire just free hanging and I connected the green wire to fuse #12 in the fuse box. One thing I know I must change all my 8ga wires to 6ga cause the wires get hot