Known dead fuel sensor gave up it's carcass for science.
This is a carbie unit, for whatever that's worth.
If there's anyone 'sparky' enough to understand how this type of sensor works, feel free to chime in.
Nothing obvious to me as to why this unit quit working, other than it's been submerged in gasoline for a decade.
The slightly ravaged condition is due to removing the potting.
(http://farm6.static.flickr.com/5175/5449377608_beeda975df_b.jpg) (http://www.flickr.com/photos/26339726@N03/5449377608/)
(http://farm6.static.flickr.com/5300/5449377316_c8f1e27100_b.jpg) (http://www.flickr.com/photos/26339726@N03/5449377316/)
Somekind of variable resistance thingy attached to float.
Believe it develops/wears bad connection where mechanic ends meet electrics,
resistance becomes irregular or fixed - stops working.
Same as in cars fr 50-60s I learned fr my father who worked as a car electrician then.
Many basics still the same, batteries, charging etc, still learning from him.
More like a variable capacitance, looks to me. Many electrolytic capacitors (the little can gizmo) are known to fail. Bad batches abound. I saved a 65" Mitsu TV with $11 of replacement capacitors in the power board. If the shiny end of the can is slightly bulged out, it gave up the smoke. The other parts, resistors and transistors should be reliable. Pity.
My old '59 alfa has a variable resistance fuel sender. Wrapped many turns of wire on that sucker.
also an analog filter to limit response to rapid changes (e.g. waves moving around in the tank).
Yep, bad capacitor. If you were so inclined, you could probably just solder a new cap on there and it would work fine. Doesn't look like there's really anything else on there that would go bad, at least not with the minimal current moving through that thing.
All of the components were potted inside the plastic sleeve (that I cut in half), so it was basically impossible to access anything without ruining it.
What I did figure out, or at least strongly suspect...Is that you can't check one of these for function with a VOM set to 'continuity'.
I'll know better when I get my hands on the brand-new replacement,
Yes, it looks like it needs juice before you can get anything out of it.
Speeddog ..excellent thread , i have had two of these units fail over the years on M750 and S4 and tried many things to clean and ressurrect them so it intersting to see an Autopsy and understand the ''cause of death'' would it not be possible to change capacitor and use some kind of epoxy to reseal the unit as they are damned expensive . i kept the last dud one off the S4 so have a spare to play with .
Perhaps an electronics specialist could fix it, but it would be very hard to get the components out of the epoxy potting without ruining them.
Capacitor gets my vote. Electrolytics have a finite life span. With the exception of bad vacuum tubes, they have been the #1 culprit in every guitar amp I've repaired over the years.
Yup, sure looks like capacitor.
I just guessed (sorry! :-[) it was the same as the one in my 907,
which my father repaired & is resistance based.
Hi,
I have what looks a newer version of this sensor from a 2001 Monster M750 which I've dissembled and am in the process of trying to repair:
(https://lh5.googleusercontent.com/-FieEdEUfMTk/T2-gexzT5JI/AAAAAAAAAOY/bu9G_VQVjNY/s1541/IMG_2469.JPG)
(https://lh3.googleusercontent.com/-JrheXFwna4I/UU4sa0-TpoI/AAAAAAAAAdA/eU8fYEgU5uM/s1425/IMG_2596r.jpg)
I've made a circuit diagram that maybe somebody here can troubleshoot. these are the symptoms I had and what I've tried so far:
The fuel level low light didn't work and was permanently off. I disassembled the sensor, dug out the board that had been fully potted.
I replaced the 2B transistor as I pulled it off when cleaning the board, the two electrolytic capacitors which smelled of gasoline,
and the LM393 comparator chip.
Now the low fuel light is always on and will not go off regardless of the fuel level switch state.
The markings came off a couple of the components, hence the ?'s
(https://lh6.googleusercontent.com/-J1cP0uI25dw/UU_OPZf4OnI/AAAAAAAAAfY/n3fWE010pjI/s1424/Circuit+Diagram.jpg)
If/when it's working I'm planning to relocate the circuit board external to the tank to help preserve it.
To me that sounds like the diode is buggered. (the leftmost component in your picture).
check the diode isn't leaking current in the other direction. should only have flow one way (i.e. very little resistance in one direction and almost infinite in the other)
and whats worse Ducati want ridiculous money for them ,, in UK£96 ( $150),,, shameful profiteering !!!!!!!!! for something that costs $10 to make
good job reverse engineering the circuit diagram, but i have a doubt about how the transistor is configured.
Its been a while since i've analyzed a diagram but here goes.
The theory is the fuel sensor/switch sends a low level signal to the input of an operational amplifier(lm339 op amp).
This voltage is compared to a predetermined voltage set by a voltage divider(couple of resistors). When the level switch's voltage reaches or surpasses the predetermined voltage the output of the op amp changes states. btw as you can see there are 2 op amps per lm339. Normally the outputs of these op amps are not used to directly drive hardware (relays,solenoids, lights) so this is where the transistor comes in.There are 2 types of transistors PnP and NpN. A transistor has 3 legs, Base, Collector and an Emitter. In this case you can think of them as a switch where current will flow from the collector to the emitter. The way they basically work is you apply a positive voltage to the base(PnP type) or a negative voltage (NpN type) to get the current to flow from the collector to the emitter.
The output of the op amp drives the base of the transistor. The transistor will now be used to drive the light(flow current through the collector to the emitter and then the light). Now this is where the diagram confuses me. One leg of the transistor(i'm assuming collector) is connected to the supply 12v, so far so good. The base of the transistor should be connected to the output of the op amp and its not. This is the part of your diagram you should maybe re-check.
A shorted diode will have the light on all the time so an easy check would be to just lift it out but remember that they have a polarity and must be put back with the same orientation.Your Transistor and capacitors also need to be oriented correctly. When soldering a transistor be very, very carefull not to apply to much heat, this will destroy them.
QuoteWhat I did figure out, or at least strongly suspect...Is that you can't check one of these for function with a VOM set to 'continuity'.
poking around the circuit board with a VOM in Resistance measurement mode is a bad idea. The vom sends a voltage through its probes(in resistance measurement mode) in order for it to calculate a resistance to display.This voltage can easily fry your op amps or transistor.
Sorry for rambling on.......
How about maybe, when I replaced the LM393 I installed it backwards and somehow reversed the supply polarity. Check out the attached diagram with the LM393 and supply polarity reversed, does it make sense then ?
(https://lh6.googleusercontent.com/-9tqxuzOlpm0/UVI5qRKYeYI/AAAAAAAAAhc/KcxIjzKutQc/s1100/Reversed.jpg)
It looks better to me. If you call the bottom horizontal line ground, cut the emitter loose from it and tie it to the side of the lamp with the ground sign you might have a working circuit.
Looks like the first comparator (output on pin 1) will go low when the fuel switch is engaged. R2/R8 form a reference point at 12V / 2 = 6V. Noting that the LM393 is an open-collector output, this means that when its output goes low, it forms a discharge path for C2 (through R4) at a time constant of 1/(2.2k * 100u) = 4.5 seconds. The charge path for C2 is through R5 (I assume that's 345k, which gives a time constant of 35 seconds. The capacitor output is monitored through R6 (100k) by the second comparator, against the same 6V reference from R2/R8. The idea behind R6 is so that the leakage (bias) current of the LM393 input does not discharge C2. This comparator drives the transistor, which lights up the fuel lamp. I suppose when this happens, the LM393 loses power briefly, but the capacitors help the circuit "retain state" when the transistor turns off and the power is restored.
All I can say is "crazy effing Italians!" Adding a separate input for +12V would have made it a lot more sane, at least to my eyes.
Quote from: Bill in OKC on March 27, 2013, 05:10:41 PM
cut the emitter loose from it and tie it to the side of the lamp with the ground sign
Would that not make a short circuit when the transistor switches ?
Quote from: tuxicle on March 27, 2013, 08:11:28 PM
Looks like the first comparator (output on pin 1) will go low when the fuel switch is engaged. R2/R8 form a reference point at 12V / 2 = 6V. Noting that the LM393 is an open-collector output, this means that when its output goes low, it forms a discharge path for C2 (through R4) at a time constant of 1/(2.2k * 100u) = 4.5 seconds. The charge path for C2 is through R5 (I assume that's 345k, which gives a time constant of 35 seconds. The capacitor output is monitored through R6 (100k) by the second comparator, against the same 6V reference from R2/R8. The idea behind R6 is so that the leakage (bias) current of the LM393 input does not discharge C2. This comparator drives the transistor, which lights up the fuel lamp. I suppose when this happens, the LM393 loses power briefly, but the capacitors help the circuit "retain state" when the transistor turns off and the power is restored.
All I can say is "crazy effing Italians!" Adding a separate input for +12V would have made it a lot more sane, at least to my eyes.
Can you please elaborate on the separate +12V ?
Quote from: gjscott on March 26, 2013, 05:32:08 PM
How about maybe, when I replaced the LM393 I installed it backwards and somehow reversed the supply polarity. Check out the attached diagram with the LM393 and supply polarity reversed, does it make sense then ?
(https://lh6.googleusercontent.com/-9tqxuzOlpm0/UVI5qRKYeYI/AAAAAAAAAhc/KcxIjzKutQc/s1100/Reversed.jpg)
So I've reworked the circuit to look like this and the gas light is still permanently on.
There is 12V going in and 9.5V coming out to the lamp regardless of switch state. Any ideas or things for me to check ?
Quote from: gjscott on March 27, 2013, 10:43:12 PM
Would that not make a short circuit when the transistor switches ?
No, the light bulb will be the load. The way it is now only the fuel level switch turns the light on with R8 inline - the rest of the circuit does nothing as it is. That might be what you are seeing during testing. You only have one ground point for the entire circuit - everything grounds through the lamp. You will need the bottom horizontal line to be ground to set up your voltage dividers, ground the negative side of your capacitors and ground pin 4 of the IC - the comparators will not work without that. Put the lamp between the emitter and ground.
Quote from: gjscott on March 27, 2013, 10:44:12 PM
Can you please elaborate on the separate +12V ?
Well on further thought, I was wrong. I thought the circuit would oscillate (as soon as the transistor turns on, it shorts out the supply to the comparators). I forgot that the lamp has a finite resistance, so the circuit will still have power even when the transistor's on.
The circuit supply common doesn't make sense to me. I know us italians want to be different but this borders on stupid design for such a simple control.
Please check the supply's gnd or negative and make sure that diagram really is that way.
Quote from: scaramanga on March 30, 2013, 12:04:24 PM
The circuit supply common doesn't make sense to me. I know us italians want to be different but this borders on stupid design for such a simple control.
Please check the supply's gnd or negative and make sure that diagram really is that way.
Which version doesn't make sense, the original or the reversed one?
Her are some pics of the board without some components so so can see the layout:
(https://lh3.googleusercontent.com/-rFFiZnfcEpc/UVcw4UvzxeI/AAAAAAAAAh8/0Zi0u83YZVI/s1541/IMG_2010r.jpg)
(https://lh5.googleusercontent.com/-OBlY7Ab-XLQ/UU5EyWN-WTI/AAAAAAAAAeQ/b_Yirhw1R9k/s1378/IMG_2593r.jpg)
reversed, sorry.
you see those 3 holes to the right of the black wire(second photo), they communicate to the other side and connect to the trace underneath. things are getting clearer...now get back to work :)
Quoteand whats worse Ducati want ridiculous money for them ,, in UK£96 ( $150),,, shameful profiteering !!!!!!!!! for something that costs $10 to make
small price to pay for a safely engineered electric device that sits "IN" a fuel cell just below your face. ;D
Quote from: scaramanga on March 30, 2013, 06:16:41 PMsmall price to pay for a safely engineered electric device that sits "IN" a fuel cell just below your face. ;D
No it is too much to pay for a dumba$$ engineered device, when a mechanical float switch would perform the same function. This another example of a solution to a problem that did not exist. ;D
Quote from: Dry Martini on March 31, 2013, 11:08:13 AM
No it is too much to pay for a dumba$$ engineered device, when a mechanical float switch would perform the same function. This another example of a solution to a problem that did not exist. ;D
I'm guessing you don't remember the days of non electronic floats that would change greatly with approach and decent on hills and low fuel lights that would be constantly blinking once you used some fuel.
Quote from: howie on March 31, 2013, 11:34:14 AM
I'm guessing you don't remember the days of non electronic floats that would change greatly with approach and decent on hills and low fuel lights that would be constantly blinking once you used some fuel.
Yes I do, however, those drawbacks can be dealt with.
Quote from: Dry Martini on March 31, 2013, 11:08:13 AM
No it is too much to pay for a dumba$$ engineered device, when a mechanical float switch would perform the same function. This another example of a solution to a problem that did not exist. ;D
A float system without some kind of electronic dampener would be crappy too. They should use a thermistor transducer like the Japanese bikes do.
Quote from: gjscott on March 31, 2013, 11:56:03 AM
They should use a thermistor transducer like the Japanese bikes do.
Whatever. I bet it does not cost $150.
I can see at least one difference between the circuit card traces and your schematic. The 8pin IC has pin 8 (bottom right leg) going to +V and pin 4 (top left leg) going to ground. I am assuming the white wire is + and the black wire is -. It looks like the transistor is an open collector with the bulb tied to +V. Your schematic should have the bulb between +V and the collector.
I'm very surprised that they don't use sensors(capacitive proximity switches) on the outside of the new plastic tanks. Very easy to do, safer and would last forever.
Quote from: Speeddog on February 15, 2011, 02:30:38 PM
Known dead fuel sensor gave up it's carcass for science.
This is a carbie unit, for whatever that's worth.
....
Hi, I am a new member, joined this board mostly because of this thread.
I'm having trouble with the fuel sensor on my 2002 ST4s, it reads OK when the tank is full but goes randomly to empty after about 70 miles.
I wondered if anyone on here had a good understanding of the way these things work?
E.g. what are the signals on the two wires?
what are the levels for full/empty?
Is it simply an analogue current drawn proportional to fuel level or is it more complex?
I am slightly hampered with workshop space at the moment, but I am an electronics designer by trade, I'm also reluctant to throw money at replacing unreliable parts with the same.
I also can't just pull the bike apart and work it out from scratch, as I commute on it most days, so any headstart I can blag is going to be welcome!
Ideally I'd also like the gauge to tell me more usefully how little is left in the tank, not go to zero when there's still about a gallon left!
I seem to always get jobs where I have a choice of pushing the tank range or filling up every day.
Cheers,
Chris