So I'm building my new 406" sb and it will come with Dart SHP heads. The heads have the "new standard" 3/8" opening for a temperature sender. That's a problem because my old temperature sender was 1/2".

I'm not willing to enlarge the 3/8" head opening to 1/2", so my options are limited. I can buy a turned down sender from Willcox for $80, I can switch out my gauge for something new and new sender, apply resistors to get a gauge to read right, or maybe something else.

In my view, none of those are great answers as the best option IMHO is the 1/2" turned down to 3/8", but with the new versions of these senders (TU5, etc), many are finding out that they don't always read right in Ohms at various temperatures.

My alternate solution is different. It attempts to use the original gauge with almost *any* temp sender.

To do this, one first needs to determine the high/low and intermediate Ohm readings of the sender. That's easily enough done with some paper, a pot of water and a multimeter. Plot those values on a sheet of paper. This is the easy part.

The next part is piecing together some electronics to accomplish mapping the unknown sender Ohm values to a range of resistor values for the temp gauge. Since the temp gauge relies on the sender to be a variable resistor, my solution would have to reproduce that same reading to the gauge. While I'm no EE, I'm dangerous enough to get into trouble, and what I found is that a transistor (FET, MOSFET, etc) can be used to do the same thing.

So here's a rough schematic of what I'm proposing:



While the schematic is for an oil pressure sensor and gauge, the same schematic for the water temp should almost be exactly the same.

My "Voltage Divider Circuit" is working fine, however I'm not sure about the Temp Gauge side. I've done a lot of research and this setup seems to be common. My questions are what should the value of R2 be and what type of transistor should I be using NPN or PNP? Should I use a MOSFET instead? There are probably hundreds of transistors of varying types, how to I decide which one to use?

On a bit of a technical note, while testing my actual oil pressure gauge, 5V worked fine and with the sensor line disconnected, there was about 30ma of current, and with the sensor line grounded, there was a about 130-140ma of current. I wasn't sure how much of that current was going through the sensor.


If you're an EE and can help move this project along, I would like to hear from you.

 

So you may ask, how does this solution provide any real benefits?

The answer to that question is in the mapping function of your sensor values to what the variable resistor value should be that your gauge expects.

So take for example what is known about our original sensors in that they were generally in this range of Ohm readings for a particular temperature (taken from LectricLimited site):




and now maybe using a completely different sensor with different Ohm values for the same temp readings.

What was 80 Ohms at 212F is now maybe 100 with your new sensor. Well if you used that sensor in your car, your gauge would show a much lower temperature than what it truly is.

However, computer code comes to our rescue. I can create an array that says map the 100 Ohm sensor reading to my transistor to recreate the 80 Ohm variable resistor.

This is done using the Arduino editor. This picture shows my code to read any sensor and determine the Ohm readings using a voltage divider:



That means I can use any variable resistor sensor I want, and if I know it's range of Ohm values, I can map those to something I can use to vary my transistor value so that it replicates as close as possible the original GM sensor.

 

I assume the Lilypad nomenclature in your schematic is a placeholder for the real 3/8" 5VDC powered temperature sensor? What you are describing I think is to digitize that output in the Arduino and then provide a controlled resistance to ground to fake out your original gauge. I didn't try to figure out what year you have but being chrome bumpered that is two wire system that provides a variable resistance to ground. BTW, the resistance plots nicely on semilog paper. So your real problem is that is an analog circuit and what should that look like?

I had the same problem with a new motor and ended up tuning it with two two wire temperature sensors in parallel, one in each head, and a series resistor to get dead on at 212 degrees and less than 10 degrees off anywhere else.

No Arduino in mine but I do applaud your creativity with the Arduino. I've got one hooked to an electric fuel pump on my bench top now and plan to PWM modulate it to regulate fuel pressure (but that is another story).

Anyway I don't really understand your schematic without seeing the hookup to the temperature gauge and just one temperature sensor, and that not being a Lilypad, whatever that is.

 

Ignatz, you are correct, my Lilypad are just crude placeholders for the sensor:



and my gauge pins:



Fritzing didn't have correct placeholders for them so I used something "close enough".

My plan is to use the D8 output pin to modify the voltage to the base/gate of the transistor, thus creating a "variable resistor" which is what the temp sensor really is.

Right now I'm just playing with my oil pressure gauge which is a 3 pin, +12v source, ground, and signal. The signal line will go to either the collector/source or emitter/drain I guess depending on if I go with a PNP or NPN?

I've verified that my voltage divider is working well to determine the Ohm resistance value of my sender, so that part is done.

Now I just need to work out the other side, the "signal" side.

I found this wiring diagram for a "temp gauge" from LectricLimited. It's probably a generic gauge, but it has 4 posts like my 71 vette does, so I'm somewhat assuming mine works just like it.



With my idea, you could use any water, oil, fuel sensor you wanted and, through code mappings, could make them *all* work with our original gauges. I think that's a very cool idea.

Though some may scoff at my idea, that's their problem. It seems to me this is a very basic problem and the wiring schematic for this shouldn't be that complex, but I was asking for help as I write code for a living, I don't make circuits.

 

If you go this route use an NPN transistor. You shouldn't need anything big, but make sure it isn't a switching transistor. Seems like you would see similar results with a variable resistor in place of the ceramic resistor.

 

I guess I don't really need to know the specifics of why I should use a NPN, but it's good to know.

As far as the switching type of transistor, yes I was aware that there are "switch" (on or off) transistors and the ones I'm looking at which are variable, but thanks for clarifying.

As for the water temp gauge resistor between the top/bottom posts, I have yet to play around with those but I would agree that it's one more variable in my solution. However, to me, it's just easier to say use XXX ohm resistor across those posts, and then I'll just adjust my "variable resistor" transistor to mimic an original/NOS GM temp sender as it come from the factory to get my accuracy in my gauge.

 

OK, now I see the voltage divider, a fixed resistance in series with the variable resistance of the sensor.

As far as the output, you can use an NPN with a resistor tied to the base and the collector tied to the gauge. The D/A voltage out of the Arduino and the value of the base resistor will determine the base current. That drives the transistor collector current and in turn that moves the gauge. You can just put a pot to ground and see where the gauge goes for various settings. That will tell you what the collector current should be.

A very common transistor is a 2N2222. All these devices have published characteristics which will get you close to the base resistor value. Then it is just a path to ground from your D/A through the resistor and a diode drop.

You'll want to see what the supply side voltage is for your gauge with the engine running, probably more like 14 volts.

 

The only reason I suggested the NPN is that your diagram has an NPN transistor and there really is no benefit to going PNP. Experiment 3 is basically what you are attempting.

 

I just finished creating my working demo of what I tried to articulate above.

I have some rather crude code to calc the resistance of my SB 1/2" NPT sensor, and then I set the transistor to make my gauge read whatever I want.

I'm just waiting for Vimeo to process my video and then I'll post a link.

*I* think it's pretty cool for about 8 hours of work. Is it perfect? Hell no. Will it solve world hunger? Hell no. Is it overly complicated for something so simple? Hell yes.

I'll post the link here, but it says there's 40 minutes remaining before it's ready.


I just watched the video to the end and my Samsung S5 phone video editor added a condensed version of the video at the end with sound, oops. Maybe I'll fix that tomorrow, maybe not.

 

Pretty cool. I assume you know not to put teflon tape on the sensor when it's installed. Have you thought about where you're going to mount this?

 

Fun project, but Wilcox already made an adapter board for adjusting the gauge to the new temperature sensors.
Of course with the uP you would be able to make the gauge perfectly linear.

 

Seems like your going the long way on this... This resistor should solve your issue with any temp sender.

http://willcoxcorvette.com/product_i...ducts_id=28320

 

That whole Teflon tape thing is pretty much a myth: http://www.lbfun.com/warehouse/tech_...eCF-Thread.pdf. I suppose it's possible to put 20 layers on and not tighten the nut enough to cut through the Teflon so it won't make ground, but it's unlikely.

As for my little project, no idea if I'll actually use it. As I've been well told by people here and "other" vette sites that my project is silly stupid, there are much simpler ways to accomplish this via pipe adapters, resistors on the sender, etc. I get it, it's not everyone's cup of tea.


I've seen lots of guys run aftermarket gauges because they had to, to interface with different senders than original. All I'm saying is that with this setup, I can make an original oil/fuel/water gauge work with nearly anything.

 

Yep, I've seen and admired your work on this, watched the video even.

I just had one question and that is, does this variable resistor make it accurate at the 100, firstmark (155?), 210, secondmark (what temp is that?) and the 250? Are all automotive temp senders non-linear, I'm curious.

What if I had a linear instead of a non-linear sender, then we've still got a problem right?

 

This forum has quite a few people who've done lots of things "their way" so I think you are mostly in the right place. The skill set you've got here is clearly not for everybody, but so what. It is the 21st century after all and new cars are way better thanks to computers, why not ours. Lot of spare capacity on that little board and I know you can find more functionality.

I've got my own little Arduino project going on a fooling around basis and am getting ready to install it during the winter downtime. I mostly don't care about other people's opinions but I do value their actual experience.

 

Will the system actually work in the car when subject to real world conditions such as ignition noise and voltage spikes?

My '64 has a Mallory box, solid core wires, and non-resistor plugs and I can't run many of the electrical devices (e.g., my g-analyst) due to electrical interference even when it runs on its own battery.

 

Your Mallory box and other goodies will not affect my resistor.

 

OTOH, your resistor will not work in my '64. At least you don't list that application.

 

Yes. you are correct... We started working on that solution last week.

I wasn't aware you had a 64 since this was a C3 thread. One of the issues with this so far has been inconclusive data or data variances between reproduction sending units and this variance is just a matter of more sender testing. We have another meeting planned on this on Weds of the coming week.

To answer the question above...

I've never seen a linear dash unit and I'm not sure why anyone would wish to use a linear sender with a non-linear gauge, doesn't make sense. Like I said above, seems like you are taking the long way to a solution. Your work looks nice and I can appreciate having a bug in your head and trying to figure it out, this is how this entire resistor project started for me... I just got side tracked with other projects which caused some long delays.

Willcox

 

Great thread Mooser. It sure would be nice if someone out-there would mold us an A-pillar gauge pod. I would like to mount this in that area as well. I'm currently planning to use the clock as a permanent location but would rather have it on the A-pillar alongside a vacuum gauge. A creative person could probably make a small fortune developing this little item.... Richard454 has also recommended this gauge and he is pretty thorough when it comes to his recommendations. With Mooser and Richard recommending it - I'm going to order mine today as well.