Honda ATV Forum banner

DIY Cooling Fan Gizmo

25K views 226 replies 14 participants last post by  fishfiles 
#1 ·
Let's attempt to fix the air-cooled Honda ATV excessive oil temp problem that has plagued us all for these last 30+ years, ehh?

This project will be a joint effort by @wheelsquad and myself. We are gonna attempt to build a better FCU gizmo for his Rancher 400AT. We are beginning this project with a new oil temp sensor (#37750-HC4-751) in hand that wheelsquad purchased and then graciously sent to me... so's I could get my meters busy on it.

Our ideas consist of making a gizmo that will be fully adjustable from end to end. We'd like to precisely control the cooling fan kick on temp and precisely control the cooling fan kick off temp, independently. That means we'll need to provide a wide range of hysteresis adjustability in our design. The circuit will need to be very stable as well, not be overly sensititive yet retain very fine-grained tuning abilities. The circuit basis must be expandable too, as we'd like to be able to put the over-temp idiot-light feature to better use. Our resulting gizmo design and build (if we are successful) should provide a suitable FCU replacement for all Honda ATVs that have a factory oil temp sensor in the sump.

1st step today was taking some baseline resistance measurements at various oil sensor temperatures. The oil sensor was suspended in a small pan of vegetable oil and slowly heated. Each temp point where measurements were taken were held steady for five minutes to allow sensor resistance to completely stabilize.

As you can see in the pics, we will need a circuit design that works primarily within a resistance range of under 3.5k ohms. The oil temp sensor resistance (vs temp) is not linear at all either, and it responds to temperature change lazily. We'll have to tolerate between 100-150 ohms of that laziness between rising and falling oil temp trends.

I learned that the thermo-couple and the infrared gun are running neck & neck, both display the same accuracies all of the way up to 240 degrees F. So from now on I'll just use the thermo-couple for all future temp measurements.

Pic dump...
 

Attachments

See less See more
5
#2 ·
This is a simple circuit that I put together many years ago which utilizes an LM741 Op-Amp reconfigured to serve as a voltage comparator. This schematic was originally drawn with reversed LM741 pins for heating purposes. It was used with a 10k ohms thermistor, which controlled a heater and circulating fan for my homebuilt egg incubator/hatchery. In this drawing the circuit has been corrected to be used for cooling purposes.



As you've seen in the pics above in the 1st post, we will need to work with oil temp sensor resistances less than 3.5k ohms. So its likely that our basic voltage comparator circuit will change right away. I'm thinking that those 10k resistors on the left-hand side of the schematic might get swapped out for 5k resistors... we'll soon see whether I'm right or not.

Half of our problems may be solved by choosing this simple LM741 Op-Amp and transistor circuit. Our remaining needs/ideas can be added on to this... more about those ideas later.

I'll mock this up on a perf board this week an' see what can be learned. Any suggestions? Requests? Critiques? :)
 

Attachments

#14 ·
I'll mock this up on a perf board this week an' see what can be learned. Any suggestions? Requests? Critiques? :)
Don't forget the fail-safe feature where the fan will actually run when the sensor goes bad.... vs fan never coming on, how Honda has it.
 
#3 ·
I am no ele wizard at all, correct me if I am wrong here !, but isn't veggy oil thinner than atv motor oil ?, wont this temp reading be off a tad once you dip that temp sensor into atv oil because its thicker ?.
 
#5 ·
Part of my problem while testing my sensors was measuring the lower resistance values with water above 200F. My sensors seem to read just a bit too high, so I don’t know if they would trip the fan until after the critical temp threshold has been crossed—i was it to trip before.

Retro I don’t know how the Op-Amp works and barely remember how to read a circuit.

I see a diode in there and many resistors. Do you have rheostats in there or is that handled by the OpAmp?

I often thot one could trick the fan relay into thinking the sensor circuit resistance was lower by wiring a smaller value resistor (or variable resistor) in parallel with the sensor. However a parallel circuit seems unlikely due to construction of the sensor itself.

So are you left with manipulating a series circuit?

Great job BTW i enjoy innovations like this.
 

Attachments

#8 ·
Ladies and gentlemen, be careful what you ask for when saying "hey retro would it be possible to _______?"

Fun for me too, and learning a lot also.
@shadetree, great questions. And yes the RATE of heat transfer can vary between liquids, but retro is only taking measurements after the temperature has completely stabilized. So the resistance reading at said temperature will be exactly the same regardless of the liquid it is in, provided it has stabilized.

And yes the boiling/burning point of vegetable oil is lower than motor oil, but we should be well below those temperatures for this. If we were at those temps with the sensor installed I'd be having much bigger motor problems.
 
#13 ·
The cold ones sound interesting, Fouriers & Ficks sounds like an upscale NY dept. store to me tho? :)

I mocked up the gizmo today using 3.65k ohms resistors (can't find any 5k in my bins) on the left side, following my hunch. It works good with those 3.65k's in there, but I think it will work even better with 10k ohms resistors put back in there.

It is overly sensitive during calibration right now IMO, using those 3.65k ohms resistors in there. I'm gonna swap the 500k ohms Hysteresis trimpot out for a 1 Meg ohm trimpot too, to help relieve some of the over-sensitivity on that function of the circuit. So I'll test again tomorrow and hopefully, get started on adding a coupla' mo' ideas to it.

No pics taken today, sorry. I ran out of time. I'll get a coupla' pics of the mockup tomorrow.
 
#15 · (Edited)
Yeah, matter of fact I was thinking about that feature today so I'm glad that you mentioned it. We have the idiot light in the display that will likely never, ever come on. So, if we can design the gizmo so that the fan starts if/when the sensor fails high out of range... AND we can start the fan if the oil sensor wire ever gets unplugged/ripped in half by a stick or something...? Should we tie that over-temp LED into that half of the device, so that the idiot light comes on whenever the sensor fails high, or it gets unplugged?

We can take advantage of that light somewhere else besides an overheat condition I feel, because the primary function of the gizmo should already eliminate any overheating by its sensitivity and precision.

Of course, the oil cooler could get plugged up with mud or the cooling fan motor could fail or blow a fuse, causing an overheat. Unless we add a thermister onto the cooler to monitor cooler temps and add a hall-effect tachometer circuit to the fan motor... :)

Think about it... lets bounce some ideas.

So far I think we are gonna have to monitor the ambient air temperature, because sub-zero air temps would false-trigger our sensor failing high feature. Complexities are imminent. :)

EDIT:
Actually we might be able to turn that over-temp LED on if the motor overheated too, so don't rule that possibility out just yet.
 
#18 ·
We have the idiot light in the display that will likely never, ever come on. So, if we can design the gizmo so that the fan starts if/when the sensor fails high out of range... AND we can start the fan if the oil sensor wire ever gets unplugged/ripped in half by a stick or something...?
Yep exactly what happened to mine, sensor wire came unplugged, my oil cooler may have been restricted some, it was really hot out and I was in slow going in old pasture bogs, but since the default fail from mode from being unplugged was to not run the fan, and oil temp light also doesn't come on then, mine got REALLY hot before I realized it.
 
  • Like
Reactions: retro
#16 ·
I promised a pic of the mockup so here it is. As you can see I am using a Red LED (its lit in the pic) to indicate to me when the transistor turns on/off, rather than tripping a fan relay at this point. I'm also using a 5k ohms pot in place of the oil temp sensor to speed up testing. Since we've already learned our operating resistance range from the veggie oil tests we can quickly dial in our targets using the pot.

I didn't get to spend any time on it today cause universe made other plans. I'll try again tomorrow. I am putting together a short list of parts that we might need, probably order those up this weekend.
 

Attachments

#20 ·
@retro, I'll try and think of ideas to best utilize the idiot light. But I'm thinking I like your idea most for having it come on if there is an error in the fan or fan circuit itself. Since it should be obvious with the gizmo in place of a temp sensor or wire issue with the fan running continuously.
 
  • Like
Reactions: _Wilson_ and retro
#21 ·
Ya, monitoring the fan motor operation itself might be impractical since there is no tachometer built into those.

So if we can make it so the fan relay energizes by default (and turns the light on to warn ya that there's an error?) if the sensor wire gets unplugged, we probably solve one of the more important faults that Honda failed to account for. I think we can do that pretty easily right now... Do you want it to do anything else?

So far we have a very basic but far superior fan control unit design that is very inexpensive, very sensitive and fully adjustable for:
1) Operating oil temp range extremes. We control both the ceiling and the floor, fan kick-on/kick-off temp points, independently of each other using trimpots.
2) We control the overall width of the operating temp range. AND we can shift that width up/down. As the oil sensor ages and decays, we can follow that decay with a trimpot, extending the life of the system while retaining sensitivity. All Hondas can benefit from that feature, especially the really old ones.

Nuther question for ya on how to mount the fan motor relay(s)... should we use external relay(s) (Bosch type, 12v automotive?) to allow for easier replacement, or should we mount relay(s) inside the enclosure permanently?
 
#22 ·
How often do relays like that go bad? I'd still lean toward all in one enclosed unit for ease of mounting and connecting.

Also, is the way you are planning this one to bypass the entire existing fan control module? Or just to add to its existing function? I have no opinion on what is better cause I know so little about the existing circuit, just curious.
 
#23 ·
This will be an independent replacement for the existing fan control, so the original ECM circuit will be abandoned. We'll reuse the OEM oil temp sensor wire & plug is all.

Relays don't fail very often, so combining everything into one enclosure will probably be the cleanest and simplest. The enclosure will have a removable lid anyway, so replacing a relay should be easy if one ever failed. I'll use high-amperage (30/40A Bosch style) relays, they're common, cheap and reliable.

The front fender will have to be removed to install it. So far it looks like there might be 5 or 6 wires to connect up. I'll calibrate it before it ships... which reminds me, what oil temp do you want the fan to start? And stop? Do you use synthetic oil in your Rancher?
 
#25 ·
This will be an independent replacement for the existing fan control, so the original ECM circuit will be abandoned. We'll reuse the OEM oil temp sensor wire & plug is all.
Ok, I'll have to look at my wiring harness then, cause I'm pretty sure the oil temp sensor wire goes into the same harness as the 3 shift position switches on the 400AT. Which is towards the rear of the motor. Pretty sure you plan on connecting into that oil temp sensor after more towards the front? Not a problem, just good to plan out.

Relays don't fail very often, so combining everything into one enclosure will probably be the cleanest and simplest. The enclosure will have a removable lid anyway, so replacing a relay should be easy if one ever failed. I'll use high-amperage (30/40A Bosch style) relays, they're common, cheap and reliable.
?

The front fender will have to be removed to install it. So far it looks like there might be 5 or 6 wires to connect up. I'll calibrate it before it ships... which reminds me, what oil temp do you want the fan to start? And stop? Do you use synthetic oil in your Rancher?
Ok with 5 or 6 wires I think we've surpassed what I'll recruit the old man for. I'll recruit my A team, preferable someone more comfortable with soldering wires.

I didn't write it down at the time, but last summer after fully cleaning out my oil cooler, on a warm (but not hot) day I warmed up the machine good using it, then let it idle with fan on and I think near base of the cylinder, stabilized about 210-215F with my laser. Then turned off the fan and it stabilized about 240F. So hopefully that measurement was representative of oil temp?? If so maybe switch on temp at 230F and then switch off once it gets back below 215F? This is my first go-round with this sort of thing, but seems like a good starting point? That's why we are building in adjustability, right? Lol
 
  • Like
Reactions: retro
#24 ·
My head, OMG, my head! It's gonna explode!!!! Man Retro, you're something else son! Your so flipping smart you make my head hurt!!!!! ? I've always sucked at electrical stuff and this is amazing to me and I'm sure others on here. My mind doesn't work like yours I guess and I know it's mostly just understanding how it works and how to manipulate things like that, but still, just crazy. And ol WheelSquad hanging right in there like a hair in a grilled cheese with ya, pretty darned impressive!!!! Kudos my friend! I'm going to have to give ya a buzz so you can explain this gizmo on my new ride, I have a couple questions. Keep it up and Honda will be putting ya on the payroll! LMAO!!!!!
 
#26 ·
hanging right in there like a hair in a grilled cheese
LOL'd for real when I first read that. Never heard that line before. Ha.

Yeah sounds like Honda could have easily implemented some of this cause they already have a fan control module, just add a touch more functionality. If only retro weren't already "retired"....
 
#27 ·
That is diffidently a better mouse trap.

Couldn't you just wired the fan to run all the tie while the ignition is on?
Why not a switch instead to turn it in when you want it on, if you didn't like
it on all the time.
Just a thought.
The fan on my 450es has come on yet, I have only ridden it for a few minutes, but did let it idle 5 minutes and still no fan.
I am going to put it on a switch tomorrow.
NO it wont be automatic but I can handle the responsibility.
MrC.
 
#29 ·
Yes a few guys have made theirs to just run with key on, or a switch, which is what I currently have. The reason I went with a switch was because either my sensor went bad, and the way Honda designed it, once either goes bad not only does your fan lose the ability to run, but you also lose the idiot light telling you when it gets too hot. So I didn't bother replacing my temp sensor since you really have no indication for when it would ever go bad again. That's how this whole idea got started, wanting to have a working temp sensor, but only wanting to bother replacing it if I could have some indication to when it did go bad again. So I asked retro if that would be possible, and he says not only is that possible, but we can improve it in these other ways too since we're already going that far.

What would be icing on the cake for this project is me making the 6 hour one-way drive to retro and letting him install and test his gizmo on the machine itself....
 
  • Like
Reactions: _Wilson_ and retro
#30 ·
Well folks, this project may be taking a big turn this coming week, in a very surprising manner!

I was putting together a website parts order on mouser.com yesterday and decided to look into swapping the single channel LM741 Op-amp circuit out for a dual channel voltage comparator IC. And well, that dual channel comparator research led into research for programmable micro-controllers... which led to discovery of some pretty cool development code and tools on Github... which led to discovery of some pretty cool looking (functionally speaking) devices having that Github code supported, programmable uC onboard... which led me to place orders for two of the more featured and useful looking... of those devices!

It looks to me like Honda oil temp sensors might possibly work very well with those devices with minor modification. So this project will be on hold for a week or longer, until the 1st order arrives and I've spent some time on them with my thermocouple and soldering station.

I don't know anything about those devices yet (gotta study up), and I may have jumped in way over my head this time. So hang in there... either my mind has blown, or both yours and mine soon will be. :)
 
#34 ·
Going high tech are you. Did you take a look at the Arduino offerings? They have a few different models and tons of sample code available all over the net for them. I bought one a few years and put together a fancy controller for the lights on my wife's fish aquarium. Was a fun project and it has been rock solid.
 
#33 ·
Oh and you asked earlier and I never said.... But full synthetic yes, I'm a Rotella T6 guy... heck I even run that in my lawnmower, and wife's minivan!
 
  • Like
Reactions: retro
#35 ·
I haven't messed with an Arduino yet but they look pretty useful and fun. A few years ago I thought about making wireless game cameras based on a few broken Sony digital cams I had lying around and those tiny computer boards, so I might download photos and videos onto my laptop once per day without leaving my back yard. But that project never really got going because I could not find a source for used 900mhz WiFi boards that Tx/Rx with enough power to penetrate the forest over long distances. Years ago those powerful WiFi boards could be bought, but they disappeared when .govcorp reprovisioned the 900mhz wireless bands over to 4G cell phone providers. By the time I find something interesting to do with them, they'll probably be obsolete.
 
#37 ·
Awe crap... after all this time... 'an I figured that I must be the leading zero all along! :)

I got the 1st parts boards order in this week and been doing some testing on the stovetop. I've definitely stepped into the darkside big time, this time! But wait'll you folks see what these 'lil buggers can do 'an y'all might find a lil' sympathy for me!?! :)

1st thing I am reminded of was the export chinamans' penchant for scamming, when the 2nd parts order arrived the following day. All five boards in the 2nd order were fakes. :-(
Same 'ole, same 'ole game.... ehh? They're all being returned at the liars' expense for a full refund, but still wasted some unearned time disposing of that disgusting parasite.

Anyway, the µCs that have piqued my interest are the ST Microelectronics 8S003F3. They are not only cheap and common nowadays, but can be reprogrammed by the user to provide custom operation/features if desired.

In order to both meet & beat the goals of this simple fan control project though...? I doubt if we will ever need to write & flash any new code onto the STM8S, 'cause there are at least two decent solutions already for sale. Dirt cheap too! The only difficulty that I've experienced thus far have been sourcing boards populated with genuine STM8S chips on them. But of the lot of genuine STM8S boards that I have received so far, they're very impressive!

I've got several orders out and awaiting delivery for both; the W1209, and the W1219 temperature control boards. Those boards will be stripped down for fitting inside of tiny enclosures approximately the size of a pocket key fob. Only the displays and buttons will remain onboard, everything chintzy will be replaced with high quality components and relocated to a 2nd small power supply & relays enclosure.

The OEM Honda oil temp sensor has proven to be a very useless performer... so now that I am completely fed up with it, will be replaced with a custom two-wire, gutted & repotted, high quality thermistor homebuilt. Stay tuned for more details of these stupid ideas! :)
 

Attachments

#38 ·
Wow, what a can of worms that was opened. But retro is a Robin that hasn't eaten for a week and is gobbling it all up.

It's interesting and now that I think about it not surprising that you came to the conclusion that the oem Honda temp sensor is worthless for this type of control. Also potentially explains why it's all over the map with some people saying their fans never run, and others saying theirs run quite often.
 
  • Like
Reactions: retro
#39 ·
Yep, the temp sensor should have two wires for precise, linear control purposes; not just one, else voltage passing through it might measure all over the place due to motor/frame ground resistance path lengths/materials/inconsistencies. This one has proven to be quite terrible even while connected using short soldered leads, so I'm gonna gut some donor sensors and epoxy precision thermistors inside the donor brass for testing.

I plan to make two sensors in one go actually. One sensor containing a 4.7k ohms thermistor, in order to compare its' linear accuracy (inside our target temp control range) to the other sensor containing a 10k ohms thermistor. With both new two-wire sensors having very tight +/- 0.2 degrees C tolerance ratings. I may add a 3rd value precision thermistor to that test if I can find a suitable one on mouser.

I'm looking for a linear, accurate, pulled-up voltage response thru the sensor, vs. temperature. Sounds simple, but that might be difficult to nail down on the 1st try. The STM8S boards are definitely capable of tight control tho, so I'm going for it.
 
#40 ·
I ordered some good (TI) 9 volts DC output LDO voltage regulators today, along with some good 9V coil voltage relays to supply power for the boards (they have a 5 volt regulator on board, the 9v regulated supply should make them last a lot longer) & to switch on/off the Bosch style 12V/40A main fan relay.

Also ordered a selection of precision high temp (+/- 0.1 degrees C tolerance, -60 to +150 degrees C environment) thermistors... got some 30k, 20k, 10k & 6k ohms coming to add to my existing collection of 10k & 4.7k thermies, along with some SMD pull-up resistors in case I gotta swap resistors on the boards to maintain +5 volts supply to the sensor.

So... it'll be the end of the week before I can get back on the stovetop.
 
#41 ·
Well, I got all of the recent orders in today and the four W1209 boards (+ two more W1219s) are all genuine this time! So, after two weeks of trying I finally found legit suppliers! Man that game was getting old... I'm glad that crap is behind us now.

I really like these W1209s a lot! I've found a new thrill! LOL

They're small (44mm x 40mm), roomy, very well laid out boards and are much quicker & easier to operate than the W1219s are. The hardware, features and firmware are identical in both models, but these W1209s are a better implementation all around. I like the buttons layout the most on these... there is more space between the display and buttons and they are centered under the display which would make it easier to add a weatherproof strip of silicone buttons to an enclosure for these W1209 boards.

The mouser order arrived today too, as well as the two donor sensors that @shadetree donated to our project! Thanks again shadetree!

So, this weekend I'll be testing thermistors and trimming the pullup voltage on the boards until they're linear and accurate on the stovetop. Then I'll pot the new precision thermistors into the donor sensor bodies and verify those on the stovetop.

Next stage will be housings/mounting options to be hammered out. We can/probably should build this gizmo out using two separate enclosures. I found a small enclosure candidate (thin, pocket-size, about the size of a large key fob) that might be suitable for handlebar mounting the W1209 if ya want that option? A similar, larger enclosure might be fashioned if you prefer the W1219 board.

My idea for that so far consists of buying two tiny enclosures that I found online and then rob the silicone buttons from one enclosure (that one is too small to use, but has some real nice buttons!) to install into the other tiny enclosure.... I'd make a window for the display and a capillary for the LED... the board inside would be stripped down to about 12mm thickness. A china handlebar clamp from ebay could be used to mount it on the handlebar somewhere. Five 18 AWG or smaller wires would exit that tiny enclosure and connect into a 2nd, larger weatherproof enclosure mounted out of sight somewhere under the front fender. The 2nd enclosure would contain the voltage regulator & supporting discretes, the two relays, etc. on a perf board. Each enclosure + the sensor would have weatherproof plugs soldered into the ends of their wiring for ease of installation. In my mind this all seems easy and clean, but it will take a bit of fussing and time to complete. I'm all ears... this might be a great time to finalize our decisions.

Here are some pics of the W1209. The 1st pic shows LLL on the display cause the sensor is disconnected. Thats one of the more important features we wanted, right there. :) Next pic shows the size difference between the single display W1209 and the dual-display W1219. Third pic shows the red LED turned on which indicates that the fan relay is triggered.

If you'd like @wheelsquad I can send ya a couple of these boards, so that you can compare them yourself before we decide on the final product?
 

Attachments

This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top