View original topic: Temperature Control (electric) - Manual Transaxle gearoil cooler Page: 1, 2, 3, 4, 5  Next
Sodo Mon Apr 04, 2016 11:49 am

I'd like to open a topic specific to temperature control system for Transaxle cooling projects.

Due to a string of traumatizing transaxle failures I perhaps over-reacted and now I'm deep into the project of cooling/filtering/delivering cooled oil to my poor over-engined transaxle. 173 HP Subaru EJ25.

Starting out, for reference here Samba links to transaxle cooling discussions (just my threads). There are a few others linked from these.

- Drilling for 4th gear and mainshaft ball brg oiling with some good electrical controls discussion on p5
- Drilling for the Pinion oiling
- Transaxle Failures every 10-20k miles
- Homebrew fittings for transaxle cooling add-on (no drilling)

and the debacle of the new rebuild with a "possibly contaminated differential" just for the horror (Parental Advisory).

I'm using the Weddle pump and radiator, which are spendy, but I think high quality and suitable. However Weddle's thermoswitch is not accurate enough to control the system the way I want it to work. Weddle states that it turns the system on at 150*F, a sensible number, as there's little reason to pump thick, cool gear oil. However in testing, the first switch I rec'd turned on at 115*F and off a little less. I called and they said the range is 125*-150* and that thermoswitches in that range are difficult to find. Weddle replaced it immediately, but the new replacement turns on at 130*F which is OK I suppose------> but off at 75*F. So,,,, once it turns on it's never gonna shut off. Plus this thing is HUGE, requires a HUGE hole in the transaxle. Which I drilled and threaded (3/4"-16) and have to live with now.

There are other automotive thermoswitches all over the internet, which close at coolant temps (appx 180*F).

Here is a $20 unit 12VDC Temperature Controller 30-120 'C Thermo Switch 15A Normally Open Relay. Its sensor simply bolts down under a screwhead, probably about 5mm screw. But its not an outdoor unit. Not sure how accurate these are, or reliable under vibration, but the screw-down sensor is much less disruptive than drilling a huge 3/4" hole in your bucket. The Weddle switch being in direct contact with the oil temperature (not outside case temp) is attractive though.

Here's my wiring diagram. I want it to run fully automatic, always. With an override switch so I can turn it on whenever I wanna screw with it. Or maybe early when approaching a long climb.

======== NOTE ====== No longer using this schematic or method. It does work, but look down the page to the new system schematic onpost 19.

Normally the indicator LED turns on when the temperature switch closes and energizes the relay, starting the pump. A manual switch bypasses the thermoswitch and connects ign. signal to terminal86 to energize the relay so the pump can be switched on manually.

Interested to hear other thoughts on how to best control this system. Certainly a simple manual switch will work for some drivers and doesn't require much discussion. Manual is a great starting point for sure. This thread could provide good insight how to add controls later without drilling the gearbox.

crazyvwvanman Mon Apr 04, 2016 12:45 pm

If the 25 amp fuse blows the pump can't run. The so called "pump on" led would still come on when the temp switch or manual switch was on. You could think the pump is running when it isn't and can't.


Sodo Mon Apr 04, 2016 12:53 pm

Yes 12v+ from Ign would light the "pump-on" LED if the 25A was blown.

The pump draws 5 - 9 amps sustained, so there's a bit of headroom in 25. Temperature guage rising beyond normal could inform you that system is not operating.

crazyvwvanman Mon Apr 04, 2016 1:10 pm

That's fine, but it is just a "switch on" led, not a "pump on" led then.


Sodo Mon Apr 04, 2016 2:38 pm

Whenever the 25A fuse is "not blown",,,,,,(99.9% of the time?),,,, the LED will indicate that "relay terminal86 is energized". IF that's what you mean by "just a switch-on" then agreed. The LED lights to inform you that the thermoswitch has closed. Iif you notice it at the moment it turns on that's the switching temp.

Assumption that LED-on = relay-on = pump-on. But it's not a big stretch to assume that the systems are functioning. But you're right it IS an assumption.

I've already wired the van with only 2 wires (as the schematic). I probably should have run extra wires just in case of good suggestions. I have some really nice 2-wire Sure-Seal connectors so "2 wires" was compelling. 8) Another wire could indicate there's system pressure.

Anyone have discussion/ideas regarding control systems for the temperature? Something better than the Weddle thermoswitch....

SteveMc Mon Apr 04, 2016 9:06 pm

Programmable temperature controller.;A=112617

Sodo Mon Apr 04, 2016 9:28 pm

SteveMC I might have to use one of those. Kinda cool, it performs the temp guage function too.

Sodo Mon Apr 04, 2016 9:29 pm

Here are the control system posts (repeated) from where they're buried in the other thread (pages 5 and 6). Sorry I hope this isn't boring y'all to death but if this cooling thing proves itself someday, the control system deserves it's own searchable (and on-topic-able) subject.

crazyvwvanman wrote: You have come this far so why not run a couple more wires that could make a difference? I would want the indicator led to be powered by the pump side of the relay so I would know that the pump was actually getting power, not just that the relay was being asked to turn on by one of the switches. Also, as already suggested I would wire the relay to self latch when triggered by either the temp switch or a momentary switch and then self reset to off when key was turned of off. Finally I would think about a second momentary switch for the ground wire of the relay, an NC switch, so the driver could turn off the relay and it would stay off unless the temp switch turned it on again or the driver turned it on with the first switch. The system would always automatically turn on when needed and reset to off with the key off but the driver could still trigger both the on and the off states. This adds 2 wires from the relay to the dash, one for the pump power led indicator and one for the relay ground at the dash through a second dash switch.

How many amps does the pump pull? (5-9Amps)


SteveMc wrote: Can't sleep so I thought I'd doodle based on the suggestions above. I think it works but seems kind of clunky. There's likely a more streamlined way to do this. Temperature auto starts no matter what. Driver can turn on as they desire and off if the temperature is below setpoint. Ignition off unlatches the relay. Flow switch lights the light. Relays powered from 'X' branch to remove load when starting engine. And I think that's three wires between dash and tranny.

boroko wrote: Here is an example of what I had in mind:

Either the momentary ON switch or the any activation of the TEMP switch would latch the pump, fan, and indicator on. My thinking was that if you were underway and got to the point that the temp were high enough to kick it on, you would probably want it to stay on for the duration of the trip until the van was shut off again. Doing it that way maintains the automatic function in case anyone else is driving, or you are just having such a good time that you forget to turn it on. If, say you came out of a hard climb out of the mountains and the heat dropped to the point you wanted to turn it off, once the temp switch stayed open, you could momentarily push the OFF button and it would shut down until the next time the TEMP switch closed, or you pushed the ON button. Seemed to me the least amount of fuss and parts, while still accomplishing the goal.

A couple of side notes:
The diode can be almost any part with enough voltage rating (almost all would meet the 12v in a van), and carry the current of the relay coil. A 1N4001 is good for 1A. Any 1N400x will do nicely. You can probably find one in some junk electronics laying around.

Sizing the fuse is obvious. It may be easier to fuse the fan and pump separately, but a fuse at the supply also protects the wiring and is a good idea.

If you wanted run it without the fan, a regular switch could be added downstream of the relay to open it's path while still running the pump.

The indicator: if you choose to use a discrete LED, something close to a 1k ohm resistor would be needed in series to limit the current, and the little flat side on the base of the LED is the negative side.

Hope that clears up some of the questions.

Sodo Sat Apr 09, 2016 12:47 pm

Moving on to the temperature control component.... I'm not getting a lot of satisfaction from the Weddle switch although it is the SIMPLEST. If it would simply work. Turn ON, then OFF somewhere near 150*F that would be the BEST. Also if it was 1/4NPT that would be great for you other fellers. But I already drilled and tapped that HUGE 3/4" hole in my trans. :shock:

Now trying something else, a micro-processor controlled unit. I bought the INKBIRD controller $16 on AMAZON. I do not know if they are reliable. Note that the temp controller also serves as a digital trans temp gauge.

To resemble genuine non DIY temperature sensors, I decided to make a globe-end "thermowell".

3/4-16 "plug", drilled and "thermowell pipe" inserted then welded.

Need to pot the sensor into the thermowell. There are several "heat transfer compounds" available. It seems like you'd want the tem sensor to react quickly to the oil temp, so filling the 'well' with this compound, then epoxy-ing (potting) the sensor in, is next. The 'well' will be screwed UPwards into the trans, not like a 'well' .... (holding water like buckets don't). If the potting is not reliable, it's possible the 'heat treat compound' could leak out. Not sure what to do there yet. Some compounds are silicone based which might cause trouble with the 'potting'.

Fitting nickel plated, then polished.

========== EDIT rather than a new post ======

OK after a discussion with Samba member Boroko (control systems tech/design) decided to pot the sensor into the housing with JBWeld. Reason being that the system is not monitoring temperature swings in "seconds" but "minutes". Bo felt that the JBWeld would conduct the temp in less than a minute, and I can observe its behavior by testing it in a pot of boiling water or ice water. Will add a heat-shrink tubing strain relief and pot that in too.

Anyway, this is where I'm at on my project. I've been messing with the temp control unit, pretty easy to set. Waiting for JBWeld to cure so I can test the sensor. Looking forward to the day I can put oil in it (all the holes are plugged)!

Brian Anthony Sat Apr 09, 2016 4:24 pm

You could use this type of switch at your suction port on the trans. ACN sells it in 180 and 190 degrees and #6, 8, 10, & 10 AN fittings.

Don't know if 150 or 160 degree thermostats are available. I use the 190 degree version for an engine oil cooler fan.

chase4food Sat Apr 09, 2016 4:35 pm

Brian Anthony wrote: You could use this type of switch at your suction port on the trans. ACN sells it in 180 and 190 degrees and #6, 8, 10, & 10 AN fittings.

Don't know if 150 or 160 degree thermostats are available. I use the 190 degree version for an engine oil cooler fan.

I would prefer this kind of bi-metal switch to switch the relay for this kind of function. Simple and few things to fail. These switches are available in all sorts of temperature ranges and on/off hysteresis. Only problem is buying ones and twos can be difficult.

Sodo Sat Apr 09, 2016 5:17 pm

Brian I'm not sure if 150* is the right temp (for oil cooler projects) to focus on. But it's the temp that everyone seems to like.

180,190 is not too hot for oil but once the oil temp has reached 180 you can bet some components in the trans (4th & mainshaft ballbearing) are quite a bit hotter (up high out of the oil). Not sure inline is a good place to put the controller because when the pump is not running no oil is flowing, so the controller can't do it's job. That controller would work OK in a full-flow (pump always running) situation, to control a fan, for example.

Sodo Sun Apr 10, 2016 9:20 am

Sensor is done. I put some heatshrink on as a strain relief and then potted the strain area with black "Shoo Goo" E6000. Onward to testing, I think this is going to work out great.

Stovetop testing with a candy thermometer.

Works as expected. I can't think of any use for the "heating" cycle, but connected a test lamp to it just to watch it work. The "cooling cycle" will energize the pump relay. The relay in this INKBIRD controller is rated to 10A but I've seen my Weddle pump draw 9A sustained (but this was while 90w was cold 50*) so I'm going to use the INKBIRD to energize a 40A Bosch type relay to start the Weddle pump.

In this example, the "cool test lamp" is the Weddle pump. At 150+*F, the controller has turned on the Weddle pump. Current plan is to run the radiator fan and pump in parallel (Pump+Fan on & off concurrent).

The controller is set for 150*F but it does not turn on the pump right at 150*, it has a delay to prevent cycling around the setpoint. At 150*F the "cool" light starts blinking, while it's doing the 'compressor delay'. At about 154*F the "cool" light went solid and the "pump bulb" lit. However the pump shuts down precisely at 149*F. This is going to be kinda fun.

rmcd Wed Apr 20, 2016 10:34 am

I've got some MGL Avionics gauges that support cylinder head temperature (CHT) senders. I've spoken with a rep and the CHT washer-like senders will read low to super high temps from 100 plus. They can also be installed under a bolt head which seems like an easy and reliable assembly.

My transmission is in place and I don't want to remove it to add this sender. I also don't want to glue the sender. Any idea which transmission bolt would be ideal for placing the sender? What size is the bolt?

Sodo Wed Apr 20, 2016 11:08 am

You wouldn't have to remove trans to add a sender, you could drill it with an annular cutter, then thread it with air pressure blowing the chips out the flutes. When there's a will there's a way. 8)

rmcd wrote: I've got some MGL Avionics gauges that support cylinder head temperature (CHT) senders. I've spoken with a rep and the CHT washer-like senders will read low to super high temps from 100 plus. They can also be installed under a bolt head which seems like an easy and reliable assembly.

My transmission is in place and I don't want to remove it to add this sender. I also don't want to glue the sender. Any idea which transmission bolt would be ideal for placing the sender? What size is the bolt?

If it's as simple as a ring under a bolt head I'd chose one that's below oil level and another that's up near the 4th gear / mainshaft ball bearing and switch between the two!

rmcd Wed Apr 20, 2016 12:36 pm

Thanks for the tips. I'd prefer not drilling until I get to the oil-cooler stage.

I'm a little worried about which bolt. I don't want to pull out the wrong bolt for fear that it will "drop" something on the inside of the tranny or mess up some kind of torque pattern.

Any advise on which bolt to chose given these criteria? A picture is worth 100 posts or something like that. : )


insyncro Wed Apr 20, 2016 12:54 pm

The rear nose cone bolts will be fine to hold the sender ring, but the number that you get will be way off.
One is used for a grounding strap attachment point.

Sodo Wed Apr 20, 2016 3:30 pm

There's nothing happening in the nose cone, and very little oil circulation. There would be a long delay, in other words your trans may be hot but the guage won't depict this for a long time, if ever. You want to put an external sensor near the action. Here are a couple possibilities that have easy access and a sensible placement. You could buy two sensors and unlug/replug your wires for each location (or move one sensor between 2 locations for testing).

The mainshaft ball bearing and 4th gear are poorly lubricated sources of heat and are in the middle of the case directly behind the upper bolt. It is above the oil level. The two on the left would sense the bearing heat soonest and are probably equal except the upper is difficult to access on a Syncro.

The lower bolt is below the oil level. Below the oil level will give you a good average temp.

Quite likely the case temp averages within 10 or 15 minutes anyway, just a guess. It would be very interesting to have several sensors around a transaxle with the instrumentation to display several at once. That's for Boroko I think. Boy if that guy were retired instead of working we'd have that info inside of a week. 8)

These are interesting temperature locations, I think. Perhaps there are better. None of the external bolts secure stuff inside, and you don't have to worry about torque patterns. You can choose any bolt. Well maybe not the CV joint bolts.. :shock:

Sodo Fri Apr 22, 2016 12:51 pm

OK here's my updated system schematic. I have all the components, now it's down to mounting them and getting the wiring tidied up. I tried to put all important info on the schematic, but probably missed some stuff.

NOTE: This schematic REV B was revised on Apr28, 2016, adding the oil pumping circuit.
Also forgot to mention that the INKBIRD controller has an on-off button.
If "ON" it retains that state. So if IGN is cycled off then back on, the controller comes back "ON". Perfect!

In abandoning the "old" system I have an extra wire going up front, perhaps it can be an LED indicating system pressure.

Here's what the momentary latching relay system contraption looks like. I'm not sure what to do with it, wrap it in tape?

There are circuitboards out there that might do the same job, for example this one for $7. But they don't tell you if it cancels or retains the STATE when you switch off the ign key. Need it to CANCEL whenever Ignition key is switched OFF.

SteveMc Fri Apr 22, 2016 6:46 pm

Fully insulated connectors may have been a safer choice. Also, relay sockets and small terminal strip could clean up all those jumpers and make it easier to follow and troubleshoot.

Powered by phpBB © 2001, 2005 phpBB Group