LED replacement idiot lights...

[jrt]

[al_roethlisberger]

Well, I received and installed the LEDs from LEDtronics:

 

http://www.led.net/datasheets/Pages/miniat...roducts/12a.htm

 

They are a perfect fit and replacement for the OEM incandescent bulbs, and quite a bit brighter. Although I also bought some white LEDs from them, I ended up installing the correctly colored LEDs, as it made for a more true and intense color than the white. And in some cases, the vibration has rubbed small places through the color in the idiot light lenses, and white would show through. With the colored LEDs, this was not an issue.

 

However, interestingly, the only LED that has proven to be a problem is the fuel idiot light... the light that inspired this whole project

 

 

It turns out that even with the tank full(and the incandescent bulb dark) there is about 1V on the socket for the low fuel idiot light, and that is apparently more than enough to fully light the W203CO3K-14V Orange LED. With the tank nearly empty, this light has about 2V across the terminals.

 

So, with the LED, it's apparent high-efficiency allows it to light full on, even with the tank full <_>

 

 

So here's a question, I wonder how I can make the LED only light(see voltage) when there's say..... greater than ~1.75V? I would assume there would be a transistor?? I could attach inline to fix this, or maybe a different voltage LED from the LEDtronics catalog?

 

Jason??

 

al

[Guest Gary Cheek]

The fuel light may have a bit more going on than you realize . I haven't really dug into it yet but it appears the resistance of the indicator filament may be a functional ,critical part of the circuit . If it is what I am thinking they may be using a pair of thermistors , the differential between them when covered with fuel or uncovered will unbalnce the bridge and start the light .

In any event there may be quite a bit more invloved to drive an LED .

[al_roethlisberger]

Yeah, I thought about that .... but dunno

 

Anyone else have any "bright"(no pun intended) ideas as well

 

...otherwise, I guess I can just put the incandescent back in, but alas, that is where I really wanted the brightness and a more binary lighting.... as opposed to a gradual burn from dim to full-on.

 

al

[jrt]

No good deed goes unpunished, eh?

My first thought is that the circuit is constantly powered. Low voltage at a full tank of gas slowly goes to higher voltage at low fuel. A thermistor is one way to do it, a float valve is kind of old fashioned. Due to the binary nature of LEDs (on or off), as long as the voltage is higher than some threshold, then it will be on. I think we know what that threshold is . You could try to put a resistor in series with the LED- that'll drop the voltage. I'll look into how it works when a get a copy of the wiring diagram in front of me.

Cheers,

J

[Guest Gary Cheek]

Actually I think the DC resistance of the bulb is a crtitical part of the circuit . An LED behaves differently than a plain resistance . It has a fairly constant voltage drop across a wider voltage range .

It will take a closer look to know for sure but I am pretty certain the new level indicator is not a simple switch .

[jrt]

You're absolutely correct about the behaviour of an LED vs a filament bulb, but let's figure out what the detector (indicator?) is before anything else. The wiring diagram is a simple circuit, but does not say how the 'low fuel level' indicator works. I can't find anything in the manual (the CD one). I'm not really inclined to pull the thing off my bike- sorry Al-, but I'll keep looking.

It occurs to me that the real solution to this would be a cascade of LED's that would more closely mimic the behaviour of a filament bulb. That's a little too bulky for what you want, though.

J

[docc]

In prior discussions the conjecture was that the sensor is a Hall effect device activated by a magnet in a float. Hence the tendency to have some intermittent 'glow' before the fuel level triggers a bright 'on' condition at the bulb.

 

Using this theory I removed my tank and petcock and cleaned the tank out really well. Lot's of crap came out of it. Afterwards the low fuel sender appeared to work more reliably. ( Notably, this was after a tipover which I thought must have dislodged all the miling and sealing debris in recesses of the tank).

 

It would be great to understand what the sender really is . . .

[al_roethlisberger]

Thanks for the insight guys, I'll await your further theories

 

The 2003+ senders(that are part of the in-tank pump assembly) definitely do not use a float though. It is a long wire, with what looks like capacitor at the end.

 

Here is a photo:

[al_roethlisberger]

......well, just let me know what you find/think Jason

 

Until then, I'll just leave my dash apart, as I hate taking the fragile plastic apart multiple times.

 

al

[jrt]

After reading about fuel level indicators (booorrrinnngg), I'm thinking more and more that Gary is right. I bet that it's either a Hall effect (I doubt this one) or a thermistor (my bet). I'm thinking it's not a Hall effect because of the very low voltages (and moving parts and cost) involved with that setup- but I could be wrong. I am NOT an engineer and I suspect Gary and some others know more about this stuff than I do. A thermistor makes more sense to me because it is a variable resistor based on temperature. A constant voltage/current is passed through the thermistor and a lamp in series, and when it is cold ( thermistor under fuel) it has a high resistance (no current to lamp), when it heats up (not under fuel) resistance drops, and the light comes on. If this is what is used, then it's going to be more troublesome to set up an LED as an indicator, but it can be done. I'm lazy, so here's a web page that gives some schematics and explains everything very well. Apparently a Toyota uses a thermistor setup so that would be the one to follow, only the Guzzi would be on-off, not a continuum.

Here's the link:

http://www.westol.com/~beaurega/gas.htm

 

Sorry Al. You could try a simple benchtop set of experiments that would allow you to work it out. Do you have an extra fuel level sending unit that you could play with? Can Charlotte do soldering yet?

J

[Guest Gary Cheek]

IF it is indeed a resistance baesd design , a simple op amp comparator using a 741 IC could easily drive the LED for you . A little ohmeter investigation should get the ball rolling .

Sounds more like a Winter project . We only ride the 850T here so maybe this winter ? We need to take gearbox apart again anyhow so what the hey . Stay tuned .......

 

The turn on ,turn off points are programmable with the comparator circuit too !

[al_roethlisberger]

Thanks guys!

 

I'll have to look into this for sure.

 

I wonder if my local Frys has this op-amp? ..have to go check.

 

al

[docc]

So today I try the acid test again. The lamp is imperceptibly dim at 145-150 miles. At 152 miles it comes on bright for about a mile. Again at 154. At 156 it comes on bright again and sputters out of fuel.

 

I brace myself with the bike in gear and do three or four very deep tip-slosh maneuvers. This takes me another 8 miles to a fuel station.

 

The tank then takes 4.7 USG.

 

I'll pull the tank again and clean it out of the petcock hole. I'll try again to fit the manual petcock TX sent me.

 

Because of the return line to the right ( dead) side of the tank , I've decided it's best to run the throttle as wide open as possible when running low on fuel. This way more fuel is used to project the bike down the road and less is, theoretically ,trapped in the right side.

 

"Er, honest Officer, 100 mph is the only way I could get to the gas station."

[Guest Gary Cheek]

Got it ! I have worked out a nice LED driver that gives a sharp turn on . As soon as I get the thing packaged for easy installation we''ll see if some dealer such as MPH is interested in distribution .