Permanent fix (in progress): Fuel/temp gauge spiking (pics, not dialup friendly)

Spook50

Get ready
 
Joined
Feb 16, 2005
Messages
5,377
Location
Spokane, WA
Update 3 Nov 2013:
The Kickstarter is now live! See here for more details.

---------------------------------------

Update 6 Dec 2013:
Unfortunately the Kickstarter didn't make it. I'll still continue with development of this and still intend to sell it once I have a tested and proven design and can produce a quality part with consistent results. It'll just take alot longer than it would have if the Kickstarter had been successful, and the boards won't be as snazzy/pretty as they would be if I could afford to have them professionally manufactured. The biggest limiting factor is our financial standing, which is pretty tight and will be for a while. But like I said I'll keep working on it as I can and will continue to update this thread as I make progress.

---------------------------------------

Update 22 Dec 2014
Well what's been holding this project back at this point is our finances. After having our son in May, we've been having to tighten our belts a great deal. We're still making ends meet but it's at the cost of a lot of projects and other money-involving plans we've had to change.
Since the Kickstarter wasn't successful I've started a GoFundMe campaign that hopefully will allow me to continue to work on this, if not as quickly as a successful Kickstarter would have been. The project is still alive. Just very much stalled. Hopefully I can breathe some new life into it and actually get some regulators out there for people having the wonky gauge issue.

Here's a link to the GFM site:
---------------------------------------

Update 6 Feb 2015

I got an order in for a set of prototype PCBs from Smart Prototyping. I'll have enough to make three prototypes (actually I have ten boards coming, but I only have three ICs on hand at this point). My guesstimate on lead time is about three weeks and I put in the order on Monday the 2nd. Huge thanks to those who've contributed to my GoFundMe campaign. I wouldn't even have been able to take this step without your help.

---------------------------------------

Update 30 Aug 2019

I'll be ordering the next batch of prototype PCBs finally from Germany this weekend (turnaround time is longer than I'd like, but they do good work at a reasonable price for small quantities). Next step is ensuring the holes line up with the posts on my gauge and if all is good, populate a few boards and send out a few test units. Feels good to finally be working on this again now that I finally have my work bench set up again and can sit in my shop and work during the evenings once I have all the components together.

---------------------------------------

It's pretty much common knowledge that as our 60s & 62s age, the contacts in the mechanical voltage regulator for the fuel & temp gauges can begin to stick and cause the needles to spike, or in bad cases max out and damage the armature that controls the needles from too much current.

I've been working on a permanent fix for this that eliminates the mechanical regulator from the circuit and replaces it with a bolt-in solid state regulator. I figured this thread would be cool for the "gee whiz" of watching the development stages that'll lead up to a final product that I plan on selling through the Marketplace (Storefront, specifically). Funding is very tight so I've got a Kickstarter submitted for approval that (if successful) will fund the remainder of development, equipment I'll need for assembly/testing and an initial production run once I have a PCB design finalized. But since that's not even set in stone yet, here's a bit of an overview up to where I'm at now.

First step was to figure out what I was working with exactly. Our gauges operate by accepting a source voltage that passes through the gauge, through the wire harness, through the sender (which in its most basic form is just a variable resistor) and then to ground. I got out the FSM and scrawled up a rough schematic of the circuit for the fuel and temp gauges based on the resistance values specified.



When I take the input voltage(s) and the maximum values of sender resistance I can use Ohm's Law to determine exactly how much current will be pulled from the regulator so that I can ensure I build one that will support it with a little bit of headroom. Unfortunately I've discovered that the values for gauge and sender resistance are far different from what is specified in the FSM. The gauges are described to be about 50Ω. When testing the gauges with my multimeter, both the fuel and the temp gauges read exactly 25Ω resistance, same thing when I used one of the better DMMs at school too. I asked another Mudder to test his gauges and he got the same results, so that verified that there was nothing wrong with my gauges.


Checking the temp gauge


Testing the fuel gauge. You can see my MM reading 24.8 in the pic. It would rapidly bounce between 24.8Ω and 25.0Ω, but using the school's DMM (much more accurate) I was able to take finer readings and got 4,995mΩ, so easily close enough to round to 25Ω.
EDIT: This picture is wrong. I put the leads back on just for the sake of getting a pic. The red lead should've been connected to the post immediately above where it is now, and the black should've been on the post immediately below where it's currently connected. My multimeter is showing the value from when I first read the gauge and hit the hold button.

Now had I been smart I would have assumed that the same would be the case for the senders and ohmed out the ground leads for both gauges one of the several times I've already had the dash apart. Next time I tear into it I'll be doing that with a cold engine and an empty fuel tank so that I can record the maximum resistance of both senders and build a better test bed.

Next was to get a ballpark idea of what input voltage these gauges required, so they got connected directly to the DC power supply and I was able to slowly increase the voltage until the needles read at their upper limit. You can see by the indicators on the PS in the background that with no resistance other than the gauges themselves, 5.0V was the magic number for each.




This is to show the workings of our gauges. As current passes through the thin wire, it heats up and bends the armature, which causes the needle to move. No current = "cold" armature = bottomed out gauge.

Next step was to figure out how accurate they would be with a 5V feed and resistances to match what was specified in the FSM. So I threw together a basic test rig to mimic the circuit for the linear voltage regulator that I was going to use for my first regulator.




Current came from the power supply at 14VDC, through the voltage regulator, through the gauges, back through a couple trimmer pots I put in place to simulate the resistance from the senders and back to the negative lead from the power supply. What you see on the DMM is the current measurement between the regulator and the gauges at 297mA. I was going for a simulation of the complete circuit. One thing I noticed while testing was that the regulator got HOT HOT HOT, reaching the point of thermal shutdown. I needed to make sure there would be a heatsink in place if I stuck with linear regulators.

Save for the discrepancy between the FSM's measurement and my actual measurement, everything else seemed kosher. The gauge needles settled in the appropriate spots for the corresponding settings on the pots. Using a small thru-hole breadboard, I soldered a 5VDC linear regulator, three wires and a couple smoothing capacitors into a rough prototype to test in my 62.





Notice the cumbersome heatsink attached to the regulator. Linear regulators are for the most part inefficient in that they will continue to draw a certain amount of current (depending on the output voltage and maximum output current), but any unused current is dispelled as heat. A simpler way to figure it is that the greater the difference between input and output voltages, the more heat a linear regulator will generate. The benefit is that their output voltage is much more stable than that of a switching regulator, which can make them desirable for sensitive electronics (depending on the application). Not being sure just how sensitive the armatures of the gauges are without the inductor in place (black capacitor-looking part inside the fuel gauge), I opted to err on the safe side as a "just in case" measure.

So that I would know that the fuel gauge needle will max out with the correct source voltage, I made sure my tank was topped off. I buttoned my dash back together, started the truck, and the fuel gauge needle only rose just slightly over halfway. After driving my truck and getting the engine good and warm, the temp gauge showed just about half of what it normally does.

So, design 1 was a bust. The heatsink also generated another problem in that it was very easy to snag and accidentally yank when reaching behind the instrument panel. When pulling the panel to remove it, I snagged a section of wiring harness on it and with very little force broke the regulator off of the main assembly. At least the solder joints weren't what broke, so I guess that was a plus :meh: I also had a feeling that even with that heatsink in place, the heat some of our trucks see in warm climates would prevent it from conducting enough heat away from the regulator.

Prototype 2 used a 6.3VDC regulator (still linear), but the same essential design. I also used a dual heatsink setup on this that allowed for less than half the total thermal resistance of the previous heatsink, plus did not protrude nearly as far back as the other.



In the heat of summer here, the inside of my dash measured 163*F at the hottest I measured, which is less than 100* difference from the thermal shutdown point of the regulators I was experimenting with. Not a very large window to dissipate much heat from the regulator, and even less when it's operating and putting out its own heat. But surprisingly, it survived without shutting down. By how narrow of a margin though, I'm not sure. You'll also notice that I didn't bother with ring terminals on this one since it is only a prototype. No sense in wasting consumables when you don't have to.

With 6.3V, my needles were almost where they should be. With a topped off tank, my fuel gauge still only read between 3/4 and full though, so I still have some fine tuning to do yet. I'm waiting on a handful of components needed to build prototype 3, then I'll post more updates as I continue with developing this project.

BTW this is my first "from scratch" project planned, tested and built (and hopefully sold) by me since I started my engineering course, so it'll be pretty cool once I end up with a marketable project :D
 
Last edited:

Spook50

Get ready
 
Joined
Feb 16, 2005
Messages
5,377
Location
Spokane, WA
Good to see some positive feedback for this project. My box from Digi-Key was just delivered (it's like Christmas when I get a Digi-Key or Mouser box) so with any luck I'll be able to get prototype 3 assembled and installed by mid next week. Hopefully the new ideas I have for it will work out the way I'm expecting.
 

Spook50

Get ready
 
Joined
Feb 16, 2005
Messages
5,377
Location
Spokane, WA
Might as well start a list.

First!
That'll be the plan. Once I get a Kickstarter approved and launched I'll use as much of those funds as possible to do as large an initial production run as I can. I'm hoping I'll be far enough along by the time it closes that I'll be ready to order a prototype batch (usually about three or four) of SMT-ready boards and can allocate the majority of it to a marketable production batch.
 

Spook50

Get ready
 
Joined
Feb 16, 2005
Messages
5,377
Location
Spokane, WA
Update

I decided on a design change that I'll be using for my third prototype, so assembling that has been postponed a few days while I clear out a spot that I can set up some PCB etching supplies. I also extended my production date goals as a just-in-case to allow for any unforeseen problems that might come up between now and shipping. As of now my goal is to be shipping units on (or if luck holds out) or before 15 January.

The Kickstarter also got approved, so I'll be launching that in a couple days (just allowing a couple days in case I think of anything else to add or change before it goes live) and will edit my original post when I do.

Meantime, here's a screengrab of my current Rev3 schematic that I drew up when I was planning on using thru-hole proto boards for testing. The circuit is staying pretty much the same, but I'll etch my own PCBs for prototyping to allow for easier integration of changes as I go, and keep the overall units as small as possible. The smaller footprint for these, the better.

 

gregnash

Anal Retentive Analyst
SILVER Star
Joined
Nov 4, 2011
Messages
9,684
Location
Carson City, NV
Awesome dude!!! I will definitely be in line for one of these. Can't get enough of supporting our MUD guys! I just need to be able to afford more stuff!!
 
Joined
Oct 13, 2007
Messages
193
Location
So Cal.
Spook, I like what you're doing here, but may I make a suggestion? Why are you using a voltage regulator with such a low out put voltage? I am not looked into the mechanical regulator on these rigs but the lower your meter source voltage the more the regulator has to drop. duh. I know. But you have already seen that you need a heat sink. with the 6.3 volt output the regulator has to dissipate ~ 1.4 W. If you were to up the regulator voltage to say 10 V even that power would be reduced to ~ .5 W. Of course this presume the mechanical regulator output is up there too.

Would simplify your product both in parts count and complexity. Both good things.
 

Spook50

Get ready
 
Joined
Feb 16, 2005
Messages
5,377
Location
Spokane, WA
Spook, I like what you're doing here, but may I make a suggestion? Why are you using a voltage regulator with such a low out put voltage? I am not looked into the mechanical regulator on these rigs but the lower your meter source voltage the more the regulator has to drop. duh. I know. But you have already seen that you need a heat sink. with the 6.3 volt output the regulator has to dissipate ~ 1.4 W. If you were to up the regulator voltage to say 10 V even that power would be reduced to ~ .5 W. Of course this presume the mechanical regulator output is up there too.

Would simplify your product both in parts count and complexity. Both good things.
For my first prototype, I wanted to start with a low output voltage and work my way up to the correct voltage for the gauges. I would've much rather erred on the safe side than start with a higher voltage (without knowing the limit that the gauges can take) and risk toasting my gauges. I could barely afford small orders of parts from Digi-Key, so I definitely couldn't afford to replace both my fuel and temp gauges. Much better to keep experimenting until I find the right voltage than start high and damage my gauges. As it turned out, my initial voltage was far lower than needed, so I stepped it up with a higher output regulator that I had in my parts box (the 6.3V that you mentioned) that brought the gauges much closer to being accurate. My third prototype I managed to get a bit of money (from selling a couple old parts from my '74 Plymouth that's now long gone) for a few parts, so now instead of using a fixed output regulator, I'm using one with an adjustable output. Made the circuit a little more complex than with the fixed regulator, but the adjustability saves me having to spend money on "throw-away" parts and saves me the time of waiting for shipping. Plus now I can use a trimmer pot to very slowly increase the unit's output voltage until I get my gauges accurate. The small extra bit of complexity doesn't concern me, since the cost increase in components is still minor enough that I won't need to increase the selling price I'm shooting for; and the SMT components I intend to use for the final product will save a great deal of real estate over using thru-hole parts. My next prototype will use a switching regulator to drop the heat output even more (though even as a worst case scenario I'd be looking at 3.75W heat output with a linear regulator now, but the more I can decrease the heat, the better). I wish I would've thought to order one back when I ordered the last batch of components, but I just wasn't familiar enough with them at the time and would rather have definite numbers figured out and then figure out how to integrate a switching regulator, than wait to figure out the regulator and then figure out numbers I need to abide by.
 
Joined
Oct 13, 2007
Messages
193
Location
So Cal.
Like I said I like what you are doing and did not mean to sound like I was criticizing. sorry if it sounded like I was. The switching regulator is a good idea for the heat concerns. These days you may even be able to find a tiny DC-DC converter to suit your design.

Digikey is convenient and I have bought parts from them often. I also use Allied, Mouser, and Arrow in case you are unfamiliar with them. Allied at times has clearance items due to odd lots. Oh. almost forgot, try findchips.com. They will list the prices from various vendors and most of the time the number in stock.

You're right about the smt saving board real estate and money. Might be able to buy SMT resistor for fractional pennies each f you get lucky with the value you need. I just got done making a hundred little boards for my son's high school band each with a dozen parts (transistors, resistors, and such) and the total for each board was about 70 cents all in.

Good luck and keep it up,

Steve
 

Spook50

Get ready
 
Joined
Feb 16, 2005
Messages
5,377
Location
Spokane, WA
Like I said I like what you are doing and did not mean to sound like I was criticizing. sorry if it sounded like I was. The switching regulator is a good idea for the heat concerns. These days you may even be able to find a tiny DC-DC converter to suit your design.

Digikey is convenient and I have bought parts from them often. I also use Allied, Mouser, and Arrow in case you are unfamiliar with them. Allied at times has clearance items due to odd lots. Oh. almost forgot, try findchips.com. They will list the prices from various vendors and most of the time the number in stock.

You're right about the smt saving board real estate and money. Might be able to buy SMT resistor for fractional pennies each f you get lucky with the value you need. I just got done making a hundred little boards for my son's high school band each with a dozen parts (transistors, resistors, and such) and the total for each board was about 70 cents all in.

Good luck and keep it up,

Steve
No worries. I welcome constructive criticism and suggestions that might help (or give me an idea) to make for a better product.

Digikey is usually my preferred supplier since I'm used to their layout and their prices are good. I've used Mouser and Newark before too. Haven't heard of Allied or Arrow though, so I'll have to give those a peek.

Tonight I'm planning on tracing out and etching my third prototype. Won't be home this afternoon (NTHS induction ceremony, booya!) but hopefully there will be a little daylight to take advantage of when I get back so I can pull my gauges out and do the work in my shop. Hopefully I bought enough copper clad PCB since this will be my first actual attempt at etching and I imagine I'll have a few goofs :D
 
Top Bottom