Engine and Transmission Tech

tips, technology, tools and techniques related to vehicle driveline components

Go to Thread: PreviousNext
Go to: Forum ListMessage ListNew TopicLog In


Moderator
Curtis Jacobson
Portland Oregon
(4414 posts)

Registered:
10/12/2007 02:16AM

Main British Car:
71 MGBGT, Buick 215

authors avatar
Stewart-Warner Temperature Sensor Specs & Alternatives
Posted by: Moderator
Date: January 18, 2013 02:37PM

I've wasted entirely too much time thinking about coolant temperature sensors this week. I'm keen to move on to more interesting topics, but I thought I'd first quickly post here. Perhaps someone might benefit from my struggles.

I've been using a Stewart Warner (S-W) "Deluxe" coolant temperature gauge and the sender that came with it since the early nineties. S-W provides three different adapters so that the same sender can be used in four different sized holes. It's the left-most sender in the photo below and as you can see it's a peculiar thing! Its threaded portion starts about 1/2" further down the barrel compared to any other temp sender I've seen. The sender's threads are ordinary 1/8" MPT. (M for male. MPT is simply the male version of NPT. The 1/8" size designation might also be confusing... I think of it as an inside diameter of a pipe with outside diameter of about 3/8".) To the left of the sender you can see the S-W 1/2" NPT adapter I've been using for years. That's the adapter you'd use too if you have a BOPR engine with either Buick or Edelbrock manifold.

StewartWarnerTempSensor-Alternatives.jpg

I'm installing a Rover EFI intake manifold, and I'd really rather not use the 1/2" NPT adapter. There actually IS a 1/2" NPT hole in my new manifold but it's at the very highest point where the sender would be out of the water current and coincidentally where wiring would be more conspicuous. My Rover manifold has an 1/8" NPT port too, which would put the sender in the water current and in a more convenient spot for wire routing. My reservation about putting the S-W sender in that hole is it would sit so damned high. Why did they thread the wrong part of the barrel!

What are those other senders in the photo? An old Rover (Lucas?) one-wire gauge sender is shown at lower right. An old Rover (Lucas) two wire temperature sender is shown at upper right - I'll use that one for EFI, but not for the dashboard gauge. The shiny sender in the middle is new... more about that below.

I thought a suitably-threaded replacement might be as near as my local auto parts store, so I pulled out my NAPA-Echlin catalog. I was stunned at the number of senders shown in it. I was also stunned at how vaguely and incompletely they're described. As I said already, the S-W sender is peculiar... but from the Echlin catalog I'm pretty confident S-W grabbed it "off the shelf" - it certainly looks like Echlin part number TS6646 which corresponds to Chrysler/Dodge/Plymouth part# 4220334, as used in 1982/83 cars. (I'm telling you this in case you want to buy a spare locally, probably cheaper than from S-W.) The NAPA catalog says it should have an output of 91-109 ohms at 220F, although I'm not equipped to verify that.

The one-wire Rover sender is Echlin part# TS6147 (a.k.a. "AMR1425"). The catalog says it was used on 93-95 Land Rovers, but no electrical properties are listed. That's true of most of the senders in the catalog.

I'm sure there are easily over a hundred senders in the book, and I looked at every one. I crossed-out every sender with the wrong threads, wrong specs, weird wire connections, etc. One or two good prospects turned out to be no-longer-available. NAPA's website provided some of the specs that are missing from their book, which helped a little to rule out some possibilities.

I paused at this point and tried to get advice from S-W, NAPA, Advance Auto, and Auto-Meter. The first three were completely unresponsive. Auto-Meter confirmed they don't offer a suitable part.

Echlin part# TS6690 (Ford# F32Z10884A) seemed my most likely match, so I ordered one. (~$10.) It's the shiny one above. FWIW, I actually prefer its spade type connection. The book says it was used on 93-97 Fords, but further research showed it was used on the Probe, which had a Mazda engine. In fact, the same sender was also used on many Mazda models. The NAPA book reports 375 ohms at 100F. (Who cares about that? I want my gauge to be accurate at the righthand side of the needle's sweep!)

I made a simple test rig... I threaded the three MPT senders into a piece of scrap steel and sat it across the lip of a 2Qt sauce pan filled nearly to the brim with water. I heated the water gradually while monitoring temp with a kitchen thermometer, and recorded gauge readings and resistance values at 20F intervals from 100-200F (plus 190 and boiling). Oh yeah: I powered the gauge with my Radio Shack 13.8V DC power supply, which incidentally I've found very useful over the years.

Long story short... the S-W sender/gauge combo are more accurate than I expected. The Rover and Ford senders are both off, but they're almost identical over the range I tested (so the Ford sender would be a cheap replacement for Range Rover owners.) The resistance value curves of the Rover/Ford senders stay parallel to the curve for the S-W sender for range I could test for, about 70 ohms low, which means you could conceivably put a 70 ohm resistor in your wiring harness to make a Rover/Ford sender work nicely with your S-W gauge.

Radio Shack's website lists a 68 ohm 2W resistor for $1.68. Part# NTE2W068-10 (a.k.a. 55050282)... That's the one I'll try. Shipping is ~$7 if you order from the Radio Shack website, but free if you purchase via eBay, here:
[www.ebay.com]

For the record, Stewart Warner sender output measured as follows:
100F = ~415 ohms
120F = ~320 ohms
140F = ~220 ohms
160F = ~184 ohms
180F = ~145 ohms
190F = ~125 ohms
200F = ~120 ohms

Note: S-W offers OTHER senders for gauges with different sweeps, etc. These measurements are for the gauge/sender photographed above. Apply this info at your own risk...



Edited 1 time(s). Last edit at 01/23/2013 01:40PM by Moderator.


lars49
Larry Barnes
Colorado Springs
(176 posts)

Registered:
06/11/2009 02:12PM

Main British Car:
1980 MGB GM LA1 3400 V6

authors avatar
Re: Stewart-Warner Temperature Sensor Specs & Alternatives
Posted by: lars49
Date: January 27, 2013 11:42AM

Too bad you can't get up to 210F -:)


Moderator
Curtis Jacobson
Portland Oregon
(4414 posts)

Registered:
10/12/2007 02:16AM

Main British Car:
71 MGBGT, Buick 215

authors avatar
Re: Stewart-Warner Temperature Sensor Specs & Alternatives
Posted by: Moderator
Date: January 27, 2013 12:38PM

Damned altitude. Maybe I could have rigged up a pressure cooker...

I forgot to mention that I tested two thermostats while I had the water boiling: the old thermostat from the Rover (probably OEM, marked made-in-UK) and the old thermostat from my Buick engine (aftermarket, marked made in USA, but I don't recall by whom). At boiling, the Rover thermostat was barely open whereas the "Buick" thermostat was open about 1/4". I'll discard the Rover thermostat and instead use the American one.


roverman
Art Gertz
Winchester, CA.
(3006 posts)

Registered:
04/24/2009 11:02AM

Main British Car:
74' Jensen Healy, 79 Huff. GT 1, 74 MGB Lotus 907,2L

Re: Stewart-Warner Temperature Sensor Specs & Alternatives
Posted by: roverman
Date: January 27, 2013 12:40PM

If I understand this correctly, SW does'nt have a shorter compatible sender ? Being a machinist, I would consider "re-threading" the SW sender. Best done on a lathe, gripped on the hex. Might need to turn the diameter down, so as not to overload the NPT die.This assumes there is enough room for the sender, to go in deeper. We know the sender needs to be in the "hottest" coolant, with reasonable flow,(no steam pocket or dead area). Being "one wire" means this a grounding sender and therefore, no teflon tape or other non conductive sealant, on the pipe threads ? Cheers, roverman.


Moderator
Curtis Jacobson
Portland Oregon
(4414 posts)

Registered:
10/12/2007 02:16AM

Main British Car:
71 MGBGT, Buick 215

authors avatar
Re: Stewart-Warner Temperature Sensor Specs & Alternatives
Posted by: Moderator
Date: January 27, 2013 01:49PM

Quote:
We know the sender needs to be in the "hottest" coolant, with reasonable flow,(no steam pocket or dead area).

I agree, but evidently Rover engineers don't (or didn't). Neither of their senders (shown above) protruded into the coolant flow at all! They're pretty much just measuring the temperature of the aluminum around them. To get the senders down into the flow like I'd like it, you'd need to mill down the bosses on the manifold by at least 1/2".

Quote:
Being "one wire" means this a grounding sender and therefore, no teflon tape or other non conductive sealant, on the pipe threads?

I'm not too concerned about this issue - I figure the teflon tape I applied overtop factory applied sealant will help fill the thread root - but that the working part of the threads will cut the sealant(s) and be metal-to-metal. I was concerned enough about getting a clean metal-to-metal connection that I used a tap to clean up the holes.

What a ground path! Brass sender to steel adapter to aluminum manifold to cad-plated steel bolt to aluminum cylinder head to black oxide head bolt to aluminum block to ?-plated ground strap to steel floorpan to copper ground cable to battery... it's a wonder one-wire senders work at all! They'll surely give different readings due to varying supply voltage (e.g. alternator output) too. EFI want/needs more reliable/accurate temp info, hence two-wire sender.


ex-tyke
Graham Creswick
Chatham, Ontario, Canada
(1088 posts)

Registered:
10/25/2007 11:17AM

Main British Car:
1976 MGB Ford 302

authors avatar
Re: Stewart-Warner Temperature Sensor Specs & Alternatives
Posted by: ex-tyke
Date: January 28, 2013 02:46PM

Quote:
What a ground path!
Aesthetics aside, you could always solder a ground wire to the hex to provide a 2-wire connection.
FWIW, I use the std MG sender with thread sealant - (I have to since the sender straight thread is not compatible with the tapered NPT housing thread) and haven't had grounding issues (although, I agree with Art's comment entirely).


Sorry, only registered users may post in this forum.