Permanent fix (in progress): Fuel/temp gauge spiking (pics, not dialup friendly) (1 Viewer)

[Spook50]

Another update

Unfortunately not as much progress on development as I had been hoping to make in the last couple weeks. With the wife ill I've had to take on taking care of her, the two-foot terrorist (our 2 y/o daughter) and preparing for this quarter's finals. I did manage to get a charge on my truck's battery and hook up the gauge assembly for testing a few days ago, and it's working great. Next step will be to remove the regulator, connect it to my bench test bed and verify the voltage it's set to. My test units I send out for real world testing will be sent pre set to the same voltage mine ended up at, but they'll still have trimmers to allow for voltage adjustments.

In place, the regulator is working great in my truck. I need to at least temporarily fix my dash plastic though before I can drive the truck around to get some longer term, real world results. But so far everything's looking good and once I can order the supplies I can start putting together the switcher circuits that I'll be using for the production regulators. Until I either have a hot air rework station or a reflow oven though, I'll be stuck with thru-hole components for those prototypes. I had initially intended to buy a rework station so that I could assemble the production units, but if I can get my hands on an infrared toaster oven and set up an Arduino to control it, it'll save money on that aspect and let me order a larger production batch of PCBs, along with extra components to make more regulator assemblies.

 

[Spook50]

Damn

Just got back inside from messing around with my regulator and further tweaking the output voltage, and it looks like I'll have to rethink my 12V+ and ground connections. The voltage gauge ends up reading low. My battery is a little depleted (should be around 10.5V or somewhere around there) since I've been messing with it most of last night and today, but my voltage gauge was showing 8V and bottomed out when I cranked the engine. No problems cranking though, so that told me the battery still had some juice in it. I'll have to get a new 9V for my multimeter and check it for sure while the engine's running, but if I'm right I'll have to work on a new layout.

Hopefully I'm wrong :whoops:

Oh well. That's the whole purpose of testing them completely before I sell anything.

 

[Spudman]

I'll support you Spook, I hope your wrong as well, but hey the way this is going so far you're probably right

 

[Spook50]

I'll support you Spook, I hope your wrong as well, but hey the way this is going so far you're probably right

I should find out tomorrow when I make a run to grab a battery for my multimeter. In case my assumption does prove correct, I've already drawn up a new schematic using the same parts as prototype 3 that will take advantage of the posts directly connected to the fuel gauge.

Honestly if that design works well, it'll make mounting the regulator assembly much easier too, so there is an added bonus to that.

Pulling an all-nighter to finish my thesis on ethanol production and subsidies tonight, so I've really got cars on the brain

 

[Spook50]

Well I didn't really learn anything today. I got my multimeter up and running, but when I fired up the truck so that I could take a reading from the battery, the voltmeter worked flawlessly, along with the fuel and temp gauges. Very perculiar to say the least. And my alternator was charging at a consistent 14.5V, so everything's free and clear there. I'm sure something about having the regulator connected the way it was contributed to it though. I just wish I knew exactly what so I could duplicate it and know for sure.

Regardless, to avoid the possibility of the same thing happening with someone else's voltmeter with my regulator installed, I redesigned the overall footprint and connection layout for it. This will still be very strong and sturdy, and much easier to install since there won't be any need to secure washers underneath the board where the voltmeter screws pass through. I found that to be more than a little annoying when I installed prototype 3. If prototype 4 is successful, the next step will be building and testing a switcher circuit.

Meanwhile, here's a quick shot of the new layout. First chance I get I'll put it together and test it.



Quick edit: In-vehicle test of the third prototype.

 

[snowakii]

IR toaster oven.. less than $50 and you have your reflow oven.

preheat to 100 degrees for a short time. then crank it up to 450 degrees (both top and bottom heat sources on) and stay put watching the progress. less than 3 minutes and you will see the solder paste become shiny before your eyes.

seriously a cheap toaster oven that has ir output. hot air gun works too. with the hot air gun a little preheat on a cheap hot plate... just set to 100 degrees. The problem is with the pwb substrate you are likely to buy. the preheat is low enough not to cause problem and is more a prep step for the solder paste and relieve stress in the parts if they've been out of their original package for very long. then put the coals to it so the solder can take the that before the board starts to smoke. kind of like tinning wires, too low of temperature iron and all you do is melt insulation. hotter iron and the solder melts before the insulation can.

 

[Spook50]

Yeah I'll definitely have to give the UV method a shot and find me an IR oven to get some experience working with SMD parts.

I figured out why my voltmeter was cutting out with my last regulator design. I had forgotten that the current flow into and out of the voltmeter is through the screw shanks and the contact between the screw heads and copper pads on the original PCB. Without having holes plated all the way through the only way for the circuit to be complete was if the washers I was using as spacers were in contact with the screw shanks. If it wasn't just a quickly done homemade board and had plated thru holes (which I had been planning to have once I have them manufactured anyway) the voltmeter would've consistently worked without issue. However, I'm not sure if pulling current from the voltmeter's circuit for the long term might have any negative effects, so my new design pulls power from the circuit meant to power the gauges.

Yet another hand drawn board. It's ugly, but it works. Nevermind the hole without a pad on it. I goofed with the drill press



It doesn't look nearly as elegant as my third prototype, but installation is much easier with this layout. The one voltmeter screw being used to help support the board (really just to reduce vibration that could put extra stress on it) has a pad on both sides of the board with a piece of wire soldered into place to basically simulate having a plated thru hole in place. Though I suspect I'll have this issue with the posts I'm using as mounting points too, so I'll have to keep that in mind when I get it installed for testing.


Found a selection of copper washers at the local Ace, which should help prevent any extra resistance in the circuit over using steel washers as spacers.


Hopefully tomorrow I'll get the chance to do an in-vehicle test of the regulator.

 

[Spook50]

Well that was stupid of me. I forgot to draw the ground trace on my #4 board

Looks like I'll be cutting another one and transferring components from the boogered up one to the new one.

Now to go back out to my shop and suck kerosene fumes for a while.

Then I'll turn on the heater....

 

[Spook50]

Hokay! Managed to escape to the shop today for a bit and did up my 5th revision, which is a bit more compact than #4 and will give me a good long term real-world impression of how well a solid state in general will perform in the long term. Still a linear (had to cannibalize components from #4 for this one), but the next order from Digikey will be parts to build and test a switcher so that I can hopefully get a few steps closer to an actual production unit.

Since the Kickstarter didn't make it, I'm currently planning to use thru-hole components (don't have the money to invest in the initial expense of producing SMD units) until I can build up enough funds to upgrade to SMD production equipment.

Meanwhile, here's a quick shot of #5 on the test rig I slapped together for dialing in exact voltage and for testing units before shipping. Yes, that's a 12V lawn mower battery. Not perfect but it does work for this purpose
So far it looks like about 7.6-7.7VDC is the money shot for output to the gauges. Long term testing will help confirm or contradict my findings.

 

[Spike Strip]

Updates ?

 

[Granite Grinder]

Cool project, I thought the spiking gauges was random, but my '88 had the problem, and my new to me '89 has the same problem, only worse. Sign me up

 

[Spudman]

I'm also sitting patiently waiting for this mod as I've just about finished restoring my 60, keep us updated spook, you're doing an excellent job

 

[Spook50]

I gotta check Mud more often. I've been driving the 62 to work for the last couple weeks and the gauges are working beautifully. The temp gauge is much more accurate, the fuel gauge works great and there are no spikes at all. Because the needles are actuated by a metal band that moves when it heats up, the gauges do take a minute or so to rise up to an accurate reading, unlike gauges with stepper motors which instantly pop to an accurate reading.

Long story short: a (IMO very minor) caveat, but I think we've got a winner. How well it performs (what with being a linear regulator) in warmer weather will be the real stress test for it.

How my gauges are reading with about 20 or so miles on a full tank. When it's full my fuel gauge needle is just barely touching the F, which is the reference point I used to get the voltage dialed in.

 

[JohninFla]

Very cool thread! I just read hru it, since the Superbowl isnt worth watching!

Looks like you got a winner there, it's doing the job.

AND red lights in your gauges!

 

[Granite Grinder]

Hi Spook, isn't your temp gauge reading a little high? Mine is usually pointed more towards the middle, other than when it's spiking. Just curious, love the red cluster

 

[Spook50]

Hi Spook, isn't your temp gauge reading a little high? Mine is usually pointed more towards the middle, other than when it's spiking. Just curious, love the red cluster

Mine did too with the mechanical regulator. When warmed up it would be just below the second thin line up. I honestly prefer it this way since if it does get any hotter than normal it won't be by much at all by the time the needle is in the red, which will lower my chance of causing any damage. Basically just letting me catch an overheat condition that much quicker. This is what it shows just before my electric fan kicks on. It dips to about a needle's width above that second thin line with the fan going full tilt boogie.

That said, it's still obviously winter, so I want to see how it'll read when the weather gets warmer. Worst case scenario is I have to design a dual output regulator, which would be a bit more complex and require severing the connection between the fuel and temp gauge voltage feeds (not to mention adding to the build cost). Before going down that road I'll experiment with dialing the voltage back bit by bit and seeing how it affects the fuel gauge's reading.

The red cluster was just swapping the light bulbs out for red LEDs from Super Bright. Did that a few years back and I LOVE how much better it feels on my eyes at night. My eyes are bad anyway and for some reason the green made me have to squint to see the needles clearly, but the red looks nice and sharp.

 

[Granite Grinder]

I so need to get a mechanical temp gauge, heck, even a decent electrical one. I don't trust anything the stock gauges say. Saying that, I still would like to have my stock gauges would be nice if they read somewhat correctly. Red light is perfect for night driving, I believe it doesn't dilate the pupil which makes it easier to see around you. Unfortunately, the idiots driving towards you with high beams or poorly adjusted lights will continue to blind you (me)

 

[Spook50]

Just wanted to let you guys all know that this project hasn't fallen through. Development & testing are just moving very slow. The linear-based regulator in my 62 is still working well, and hopefully soon I can get the supplies needed to build a switcher based circuit for testing. From the looks of it, the fuel gauge MAY be reading a little low. I'll know for sure once I get it down to about a quarter tank and go to refill it. If that's the case I'll use two circuits on the same PCB so that the temp gauge and fuel gauge can be adjusted separately.

Unfortunately my starter just gave up the ghost so until I can rebuild that I won't be able to run my gas tank down to a quarter capacity. I hate using the drain plug since it makes such a big mess, and I can't spare a drop of fuel at this point.

 

[Spudman]

Good work Spook, I'm still sitting patiently waiting, all my gauges are playing up now, fuel was at 3/4 a week ago, now reading empty, haven't been anywhere. Alt gauge reads low then comes good, did a battery test all good.

 

[Spook50]

Good work Spook, I'm still sitting patiently waiting, all my gauges are playing up now, fuel was at 3/4 a week ago, now reading empty, haven't been anywhere. Alt gauge reads low then comes good, did a battery test all good.

Your voltmeter issue could be a loose connection or a bad ground somewhere. If your temperature gauge shows the same behavior as your fuel gauge, this regulator will correct it. Unfortunately it's also possible for gauges themselves to go bad. As our trucks get older it's going to become more and more common though. You might want to try the emory paper trick with the contacts of the mechanical regulator in your fuel gauge to see if that makes a difference in its behavior. If it does then there's your culprit. If the issue persists, then there's another cause behind the problem that will still give you grief even with a solid state upgrade.

Just a thought: does your voltmeter show a good charge after revving your engine high, but low until that point? If yes, that's the sign that the alternator's regulator is going bad.