Post by thread:[OT]: Basic theory of a fuel gauge

Yeah, I know -- this is really off-topic. And I apologise, But there are a lot of very knowledgeable people on this list, and perhaps one of them will be kind enough to give me a general answer. A couple of years ago, I had a new outboard motor installed on my old power boat. The installation included an instrument panel, and said panel included a fuel gauge. It was never hooked up, however -- merely a standard thing that was included. The other day, I had to have a leak repaired in the fuel tank, and I asked the guy to put in a new sending unit and hook it up to the fuel gauge for me. I picked the boat up today, and the fuel gauge does not budge, although the tank is about 3/4 full. (The guy who put in the sending unit said he would be glad to put it in, but that he didn't know anything about fuel gauges, and if it didn't work it would be my problem). How would I go about trouble shooting the fuel gauge? The sending unit itself is merely a float inside the tank, with a wire going to the gauge, isn't it? I would assume that as the float moves up or down, there is a varying resistance. And I would assume that the fuel gauge itself is merely an amp meter with the current varying as the resistance of the sending unit changes. I further assume that one side of the meter would go to ground, while the other side would be hot (going through the resistance). So the whole thing should be able to be checked, using merely a multimeter. Are these assumptions correct? Or am I missing something vital? Thanks for any tips, John -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> A couple of years ago, I had a new outboard motor installed on my > old power boat. The installation included an instrument panel, and > said panel included a fuel gauge. It was never hooked up, however > -- merely a standard thing that was included. > > The other day, I had to have a leak repaired in the fuel tank, and I > asked the guy to put in a new sending unit and hook it up to the fuel > gauge for me. I picked the boat up today, and the fuel gauge does > not budge, although the tank is about 3/4 full. > > How would I go about trouble shooting the fuel gauge? The sending > unit itself is merely a float inside the tank, with a wire going to > the gauge, isn't it? I would assume that as the float moves up or > down, there is a varying resistance. And I would assume that the > fuel gauge itself is merely an amp meter with the current varying as > the resistance of the sending unit changes. That's the general theory of operation. > I further assume that one side of the meter would go to ground, while > the other side would be hot (going through the resistance). So the > whole thing should be able to be checked, using merely a multimeter. I think one side of the meter goes to hot (through a fuse). Other side of the meter goes to the float sender. Easier to wire since you only need 1 wire from gauge to sender. Other side of sender goes to ground (usually via metal fuel tank). This also minimizes probability that a wiring short would allow full voltage/current to heat up the sender (submerged in gasoline). Is the fuel gauge hooked up backwards? Can you get a part number off of the fuel gauge and find a datasheet for it? Have you hooked up a dummy load (i.e. potentiometer) in place of the sender and seen if the fuel gauge will move as you adjust the pot? Lee Jones -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Your basic assumption is correct. But one side of the meter is hot and the other side of the meter goes to ground through the varying resistance of the sending unit. Usually most meters are calibrated to a particular sending units resistance since the overall resistance value of the sending units potentiometer will vary according to each manufacturer. if you hook up an ohmmeter to the lead coming from the sending unit you should see the resistance vary as the float is raised and lowered. dave. In a message dated 7/8/03 12:13:18 PM Pacific Daylight Time, spam_OUTjnall01TakeThisOuTALLTEL.NET writes: > And I would assume that the fuel gauge itself is > merely an amp meter with the current varying as the resistance of the > sending unit changes. I further assume that one side of the meter would go > to ground, while the other side would be hot (going through the > resistance). So the whole thing should be able to be checked, using merely > a multimeter. > -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

At 03:32 PM 7/8/2003 -0400, Dave wrote: > >Your basic assumption is correct. But one side of the meter is hot and the >other side of the meter goes to ground through the varying resistance of the >sending unit. Usually most meters are calibrated to a particular sending units >resistance since the overall resistance value of the sending units >potentiometer >will vary according to each manufacturer. OK, that may be a problem, with the calibration. There may be some sort of adjustment on the meter for that -- will have to check. > >if you hook up an ohmmeter to the lead coming from the sending unit you >should see the resistance vary as the float is raised and lowered. Well, it would be kind of hard to vary the float level, since it is sealed inside the fuel tank. I think that the suggestion by Lee Jones to substitute a wire to ground with a pot in series would probably be the easiest way to check it out. At any rate, now that I know that my basic assumptions on how it functions were pretty much on target, with some minor adjustments in what goes where, I think that I can probably proceed from there. Thanks, guys!! John -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

John, Here's a "quick and dirty" troubleshooting method for an automotive fuel gauge that isn't working (no need for a pot unless calibration is the issue and you know what the resistance of the sender should be)... At the sender, disconnect the wire coming from the gauge. Connect the terminal at the end of the wire to ground (making sure that the ignition switch is on so the gauge has power). After giving it some time for the gauge reading to stabilize (maybe several minutes), disconnect the wire from ground and again allow some time for the gauge to react. If the gauge and wiring are good, you should have seen the gauge indicate both "full" and "empty" depending on the connection (or lack of) to ground. If that happened, the sender is either inop, not grounded, or not well matched to the gauge. If no reaction was observed when you performed the above test, go to the gauge to make sure it has power and that the wire you were just using isn't open or shorted to something. If all the external connections are good, you probably have a faulty gauge. I can't speak for every gauge out there, but typically they are designed for a sender that approaches zero resistance in one of the extreme positions. So directly grounding the wire from the gauge to the sender isn't a problem. Randy http://www.glitchbuster.com {Original Message removed}

I designed fuel indication systems for automotive in the past but not boats but I believe they are similar. There are two basic designs, bi-metal and air-core. The bi-metal is the cheapest and may be what you have. The sender in the tank changes resistance with the fuel level and changes the current going through the gauge bi-metal. The varying current causes varying heat that bends the bimetal. The needle is attached to the bi-metal and moves accordingly. You can usually tell if it's a bi-metal by the angle of display. If it's 45 degrees or less then it's probably bi-metal. A bi-metal system also requires a regulator of some kind to control the nominal heat, typically another bi-metal that alternates opening/closing to produce a square wave whose effective voltage can be adjusted by tightening or loosening the preload on the bi-metal element. The regulator is usually what fails in these systems. Shorting or opening the sender will move the needle but it will take a little while and will only work if the regulator hasn't already failed. The other system, air-core, is similar to the meters used in an analog VOM. It has a permanent magnet attached to the pointer shaft and then two coils wound 90 degrees to each other around the magnet. The varying resistance of the sender changes the current through the coils causing a change in the magnetic field and thus the position of the pointer (the pointer is the resultant vector). These type of gauges require no regulation and react rather quickly when the sender is grounded or open. These systems are usually have 90 degrees of movement or display. In any of these, the ground is always suspect. The tank is metal but the sender is usually mounted on a rubber seal. This insulates the sender base. If the ground is not connected to the sender base, the ground path is broken. Grounds also have the nasty habit of corroding. Hopefully my long winded description will lead you to the proper solution. --- John Nall <.....jnall01KILLspam@spam@ALLTEL.NET> wrote: {Quote hidden}> Yeah, I know -- this is really off-topic. And I apologise, But > there are > a lot of very knowledgeable people on this list, and perhaps one of > them > will be kind enough to give me a general answer. > > A couple of years ago, I had a new outboard motor installed on my old > power > boat. The installation included an instrument panel, and said panel > included a fuel gauge. It was never hooked up, however -- merely a > standard thing that was included. > > The other day, I had to have a leak repaired in the fuel tank, and I > asked > the guy to put in a new sending unit and hook it up to the fuel gauge > for > me. I picked the boat up today, and the fuel gauge does not budge, > although the tank is about 3/4 full. (The guy who put in the sending > unit > said he would be glad to put it in, but that he didn't know anything > about > fuel gauges, and if it didn't work it would be my problem). > > How would I go about trouble shooting the fuel gauge? The sending > unit > itself is merely a float inside the tank, with a wire going to the > gauge, > isn't it? I would assume that as the float moves up or down, there > is a > varying resistance. And I would assume that the fuel gauge itself is > merely an amp meter with the current varying as the resistance of the > sending unit changes. I further assume that one side of the meter > would go > to ground, while the other side would be hot (going through the > resistance). So the whole thing should be able to be checked, using > merely > a multimeter. > > Are these assumptions correct? Or am I missing something vital? > > Thanks for any tips, > John > > -- > http://www.piclist.com hint: The PICList is archived three different > ways. See http://www.piclist.com/#archives for details. ===== Chuck Hellebuyck Electronic Products chuckKILLspamelproducts.com **** Home of the BasicBoard Development Platform********* Great for learning, teaching and developing electronic, programmable robotics and other embedded applications. Starter packages from $149.95. http://www.elproducts.com __________________________________ Do you Yahoo!? SBC Yahoo! DSL - Now only $29.95 per month! http://sbc.yahoo.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

But be careful, troubleshooting anything dealing with the fuel system. Its extremely dangerous. --Bob At 12:29 PM 7/8/2003 -0700, you wrote: {Quote hidden}> > A couple of years ago, I had a new outboard motor installed on my > > old power boat. The installation included an instrument panel, and > > said panel included a fuel gauge. It was never hooked up, however > > -- merely a standard thing that was included. > > > > The other day, I had to have a leak repaired in the fuel tank, and I > > asked the guy to put in a new sending unit and hook it up to the fuel > > gauge for me. I picked the boat up today, and the fuel gauge does > > not budge, although the tank is about 3/4 full. > > > > How would I go about trouble shooting the fuel gauge? The sending > > unit itself is merely a float inside the tank, with a wire going to > > the gauge, isn't it? I would assume that as the float moves up or > > down, there is a varying resistance. And I would assume that the > > fuel gauge itself is merely an amp meter with the current varying as > > the resistance of the sending unit changes. > >That's the general theory of operation. > > > I further assume that one side of the meter would go to ground, while > > the other side would be hot (going through the resistance). So the > > whole thing should be able to be checked, using merely a multimeter. > >I think one side of the meter goes to hot (through a fuse). Other >side of the meter goes to the float sender. Easier to wire since you >only need 1 wire from gauge to sender. Other side of sender goes to >ground (usually via metal fuel tank). This also minimizes probability >that a wiring short would allow full voltage/current to heat up the >sender (submerged in gasoline). > >Is the fuel gauge hooked up backwards? Can you get a part number off >of the fuel gauge and find a datasheet for it? > >Have you hooked up a dummy load (i.e. potentiometer) in place of the >sender and seen if the fuel gauge will move as you adjust the pot? > > Lee Jones > >-- >http://www.piclist.com hint: The PICList is archived three different >ways. See http://www.piclist.com/#archives for details. --------------- NOTICE 1. This account can accept email & attachments up to 10M in size. 2. Federal Monitors: At request of client, some attachments are encrypted. Please DO NOT delay traffic; please reply with credentials for password. -------------- -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

You /might/ have some luck putting it in an oven at about 100 F to 130 F for a few hours, then letting it cool slowly in the oven. The theory being that moisture will evaporate and eventually leave the meter due to the steam pressure. The air that replaces it as it cools should contain less moisture than what came out. If you have it apart, tape a little packet of recently dried silica gel inside, and then use some caulk or other sealer to re-seal it. It all depends on how much water is in, how it got in, and whether it can get back out on its own. -Adam John Nall wrote: {Quote hidden}> Thanks for all the advice -- it has been extremely helpful! > > A follow up question: There is a great deal of moisture in the fuel > gauge > -- so much that it is condensed on the inside of the glass, thus > making the > gauge difficult to read even if it did work. I'm not at all sure this > little puppy comes apart -- anyone have a suggestion on how to get rid of > the moisture? Hair dryer? > > John > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email listservspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body > > > -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email @spam@listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

Chuck, I know all about the "old way" of making fuel gauges... On Tue, 8 Jul 2003 18:19:07 -0700, Chuck Hellebuyck wrote: >...< > The sender in the tank changes resistance with the fuel level and changes the current > going through the gauge bi-metal. The varying current causes varying > heat that bends the bimetal. ...but my current car has me foxed! It's a Rover 600-series, and the fuel gauge shows the contents all the time - even when the ignition is switched off (in fact even with the battery disconnected) so it must be using some sort of "intelligent" actuation, rather than using a current-meter type of thing. Any ideas how they're doing this? Perhaps a servo? Looking at the wiring diagram the sender is the usual variable resistor to Earth, so the gauge itself seems to be where the difference is. Cheers, Howard Winter St.Albans, England -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

I have a Nissan Sylvia (SX240 for those in the US), the fuel gauge does this by virtue or a liquid that changes properties depending on an electric current through it. No current the liquid becomes extremely viscous and therefore the needle cannot move. Bung a current through it and the needle can move to wherever it is required. Now so I can invent the constant temperature butter dish, what I need is a liquid that changes its heat conducting properties depending on either the amount of current or direction of current. Oops perhaps I should apply for a patent pending on that one. Colin On Wed, 16 Jul 2003 11:00:34 +0100, Howard Winter wrote: Any ideas how they're doing this? Perhaps a servo -- cdb, KILLspambodgy1KILLspamoptusnet.com.au on 16.07.2003 I have always been a few Dendrites short of an Axon and believe me it shows. Light travels faster than sound. That's why some people appear bright until they speak! -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

----- Original Message ----- From: "Jeremy Darling" <RemoveMEjdarlingTakeThisOuTASKPIONEER.COM> <snip> > I would be willing to bet that if you pull the battery > and then add fuel to the vehicle that the gauge would > not move. Many car companies have added this feature > to their gauges. Basically their is a tooth or some > other form of a holder that doesn't let the needle move > when their is no power. When power is applied then > the holding apparatus moves back and the needle can flow > freely. Or at least that's how the gauges in my old > Saturn and 4Runner worked. I'd be willing to bet that's how it works on my '92 Miata. As you say, even filling the tank with the ignition off prevents the needle from moving. One other complicating factor, however. After refilling the tank and turning on the ignition, the needle has a huge rise-time. It takes about 2 minutes for it to go from empty to full. I'm inclined to think that it may actually have that dampening fluid someone else mentioned, unless it's just electronically dampened, to prevent a jumping needle on bumpy roads. - Robert -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> I'd be willing to bet that's how it works on my '92 Miata. As you say, even > filling the tank with the ignition off prevents the needle from moving. One > other complicating factor, however. After refilling the tank and turning on > the ignition, the needle has a huge rise-time. It takes about 2 minutes for > it to go from empty to full. I'm inclined to think that it may actually have > that dampening fluid someone else mentioned, unless it's just electronically > dampened, to prevent a jumping needle on bumpy roads. Robert, You would be correct in thinking that. In the old days the tanks were baffled to make the measurement more accurate. Today they slow down the polling of the sensor and then average it with the last few results to create a more realistic measurement. -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

part 1 1612 bytes content-type:text/plain; charset=iso-8859-1 (decoded 7bit) Not too long ago I disassembled a spare fuel gauge for my experimental airplane to see how the gauge dampens sloshing in the tank. The needle is spring loaded to return it slowly to empty once power is removed. This spring acts as a low-pass filter and it works quite nicely. Regards, Bob Jeremy Darling <spamBeGonejdarlingspamBeGoneASKPIONEER.COM> wrote .. {Quote hidden}> > I'd be willing to bet that's how it works on my '92 Miata. As you say, > even > > filling the tank with the ignition off prevents the needle from moving. > One > > other complicating factor, however. After refilling the tank and turning > on > > the ignition, the needle has a huge rise-time. It takes about 2 minutes > for > > it to go from empty to full. I'm inclined to think that it may actually > have > > that dampening fluid someone else mentioned, unless it's just electronically > > dampened, to prevent a jumping needle on bumpy roads. > > Robert, > You would be correct in thinking that. In the old days the tanks were > baffled to make the measurement more accurate. Today they slow down the > polling of the sensor and then average it with the last few results to > create a more realistic measurement. > > -- > http://www.piclist.com hint: PICList Posts must start with ONE topic: > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads part 2 2 bytes -

part 1 1699 bytes content-type:text/plain; charset=iso-8859-1 (decoded 7bit) Andrew, I have been thinking about the tanks sloshing around as being noisy signal, and that is one of the reasons why I've been contemplating the median filter. Do you have a better suggestion? I tried an analog low-pass RC filter, and didn't have good luck with it, because of the low resistance of the senders (30-240 ohm Stewart-Warner float types) and not to mention it made me nervous to have large caps tied to the fuel senders, in an airplane, no less. Regards, Bob Andrew Warren <EraseMEaiwCYPRESS.COM> wrote .. {Quote hidden}> Bob Japundza <RemoveMEPICLISTEraseMEEraseMEmitvma.mit.edu> wrote: > > > I just started back on my median filter code for the new fuel > > gauges I will be installing in the RV (using a pic, of course...) > > Bob: > > Why a median filter? Median filters are excellent for pulling square > pulses out of very noisy signals, but not so good for what I imagine > to be your filtering requirement (specifically, low-pass filtering to > keep the fuel gauge from bouncing as fuel sloshes around in the > tank). > > -Andrew > > === Andrew Warren -- RemoveMEaiwspam_OUTKILLspamcypress.com > === Principal Design Engineer > === Cypress Semiconductor Corporation > === > === Opinions expressed above do not > === necessarily represent those of > === Cypress Semiconductor Corporation > > -- > http://www.piclist.com hint: PICList Posts must start with ONE topic: > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads part 2 2 bytes -

Bob Japundza <RemoveMEPICLISTTakeThisOuTspammitvma.mit.edu> wrote: > I have been thinking about the tanks sloshing around as being > noisy signal, and that is one of the reasons why I've been > contemplating the median filter. Do you have a better suggestion? Seems to me that a straight low-pass filter would work well. > I tried an analog low-pass RC filter, and didn't have good luck > with it, because of the low resistance of the senders (30-240 ohm > Stewart-Warner float types) and not to mention it made me nervous > to have large caps tied to the fuel senders, in an airplane, no > less. Try a digital low-pass filter. Given a series of input samples in the range 0-255, this code (all 7 words of it) should perform a 256-sample moving-window average and give you the filtered result in AVEHI. ; Run this on every input sample. Enter with the new input ; sample in NEW. Exits with filtered result in AVEHI, and ; NEW unchanged. ; ; Might want to initialize AVEHI to something reasonable at ; program start; if it were me, I'd probably just initialize ; AVEHI to the first input sample. ; ; (WIDTH-1)*AVE + NEW ; AVE = ------------------- ; WIDTH AVEHI EQU [ANY FILE REGISTER] ;Holds the average [0-255]. AVELO EQU [ANY FILE REGISTER] ;Holds the fractional part ;of the average. NEW EQU [ANY FILE REGISTER] ;Holds the new sample. FILTER256: MOVF AVEHI,W SUBWF NEW,W SKPC DECF AVEHI ADDWF AVELO SKPNC INCF AVEHI -Andy === Andrew Warren -- EraseMEaiwspamspamBeGonecypress.com === Principal Design Engineer === Cypress Semiconductor Corporation === === Opinions expressed above do not === necessarily represent those of === Cypress Semiconductor Corporation -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

part 1 2018 bytes content-type:text/plain; charset=iso-8859-1 (decoded 7bit) Yep..the wing tanks have ribs in them and the lightening holes are not cut out, except for some holes to allow fuel to pass in between the ribs, or 'bays' as they're referred to. Its designed to help keep down sloshing, but there is still quite a bit of sloshing even when nearly full. There is one rib (#2 from the root) which has a trap door to allow fuel to easily flow into the first bay but not as easily to flow out, so that in an uncoordinated turn while low on fuel the pickup tube doesn't unport, which would obviously cause the engine to quit. One problem with the RV's are the float-type fuel senders are not very accurate above 1/2 full, due to the dihedral of the wing and the location of the senders being attached to the root rib. I've been thinking about putting a second sender on the other end, and using a pic to extrapolate fuel level with a table lookup based on the a/d results of sampling both senders...There are some other alternatives, such as capacitance-typ! e senders, but they are inaccurate due to different fuel densities and temperature changes... Regards, Bob Picdude <RemoveMEpicdudeKILLspamNARWANI.ORG> wrote .. {Quote hidden}> On Wednesday 16 July 2003 17:26, Bob Japundza scribbled: > > I have been thinking about the tanks sloshing around as being noisy signal, > > and that is one of the reasons why I've been contemplating the median > > filter. Do you have a better suggestion? > > Not sure if you already have a tank in there, but if not, you can add internal > baffles, of probably use a fuel cell instead.... but check the regs on > that. > > BTW, RV6 ..... sweet! > > Cheers, > -Neil. > > -- > http://www.piclist.com hint: PICList Posts must start with ONE topic: > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics part 2 2 bytes -

part 1 2673 bytes content-type:text/plain; charset=iso-8859-1 (decoded 7bit) Cool! I appreciate the code. I will give it a try. Very elegant and simple. Regards, Bob Andrew Warren <aiwSTOPspamspam_OUTCYPRESS.COM> wrote .. {Quote hidden}> Bob Japundza <spamBeGonePICLISTSTOPspamEraseMEmitvma.mit.edu> wrote: > > > I have been thinking about the tanks sloshing around as being > > noisy signal, and that is one of the reasons why I've been > > contemplating the median filter. Do you have a better suggestion? > > Seems to me that a straight low-pass filter would work well. > > > I tried an analog low-pass RC filter, and didn't have good luck > > with it, because of the low resistance of the senders (30-240 ohm > > Stewart-Warner float types) and not to mention it made me nervous > > to have large caps tied to the fuel senders, in an airplane, no > > less. > > Try a digital low-pass filter. Given a series of input samples > in the range 0-255, this code (all 7 words of it) should perform > a 256-sample moving-window average and give you the filtered > result in AVEHI. > > ; Run this on every input sample. Enter with the new input > ; sample in NEW. Exits with filtered result in AVEHI, and > ; NEW unchanged. > ; > ; Might want to initialize AVEHI to something reasonable at > ; program start; if it were me, I'd probably just initialize > ; AVEHI to the first input sample. > ; > ; (WIDTH-1)*AVE + NEW > ; AVE = ------------------- > ; WIDTH > > AVEHI EQU [ANY FILE REGISTER] ;Holds the average [0-255]. > AVELO EQU [ANY FILE REGISTER] ;Holds the fractional part > ;of the average. > NEW EQU [ANY FILE REGISTER] ;Holds the new sample. > > FILTER256: > > MOVF AVEHI,W > SUBWF NEW,W > SKPC > DECF AVEHI > ADDWF AVELO > SKPNC > INCF AVEHI > > -Andy > > === Andrew Warren -- KILLspamaiwspamBeGonecypress.com > === Principal Design Engineer > === Cypress Semiconductor Corporation > === > === Opinions expressed above do not > === necessarily represent those of > === Cypress Semiconductor Corporation > > -- > http://www.piclist.com hint: PICList Posts must start with ONE topic: > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics part 2 2 bytes -

part 1 2958 bytes content-type:text/plain; charset=iso-8859-1 (decoded 7bit) Andy, just one question. How would you change the code to handle a 10-bit a/d sample? Probably not necessary in the case where I'm determining fuel level (1/256 is plenty of resolution), but I'm curious if it can be done yet kept simple. I suppose I could really make the fuel gauge rock-steady by using this algorithm along with sampling around 250ms or so... Regards, Bob Andrew Warren <EraseMEaiwEraseMECYPRESS.COM> wrote .. {Quote hidden}> Bob Japundza <@spam@PICLIST@spam@spam_OUTmitvma.mit.edu> wrote: > > > I have been thinking about the tanks sloshing around as being > > noisy signal, and that is one of the reasons why I've been > > contemplating the median filter. Do you have a better suggestion? > > Seems to me that a straight low-pass filter would work well. > > > I tried an analog low-pass RC filter, and didn't have good luck > > with it, because of the low resistance of the senders (30-240 ohm > > Stewart-Warner float types) and not to mention it made me nervous > > to have large caps tied to the fuel senders, in an airplane, no > > less. > > Try a digital low-pass filter. Given a series of input samples > in the range 0-255, this code (all 7 words of it) should perform > a 256-sample moving-window average and give you the filtered > result in AVEHI. > > ; Run this on every input sample. Enter with the new input > ; sample in NEW. Exits with filtered result in AVEHI, and > ; NEW unchanged. > ; > ; Might want to initialize AVEHI to something reasonable at > ; program start; if it were me, I'd probably just initialize > ; AVEHI to the first input sample. > ; > ; (WIDTH-1)*AVE + NEW > ; AVE = ------------------- > ; WIDTH > > AVEHI EQU [ANY FILE REGISTER] ;Holds the average [0-255]. > AVELO EQU [ANY FILE REGISTER] ;Holds the fractional part > ;of the average. > NEW EQU [ANY FILE REGISTER] ;Holds the new sample. > > FILTER256: > > MOVF AVEHI,W > SUBWF NEW,W > SKPC > DECF AVEHI > ADDWF AVELO > SKPNC > INCF AVEHI > > -Andy > > === Andrew Warren -- spamBeGoneaiwKILLspamcypress.com > === Principal Design Engineer > === Cypress Semiconductor Corporation > === > === Opinions expressed above do not > === necessarily represent those of > === Cypress Semiconductor Corporation > > -- > http://www.piclist.com hint: PICList Posts must start with ONE topic: > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics part 2 2 bytes -

> -----Original Message----- > From: Jeremy Darling [SMTP:.....jdarlingspam_OUTASKPIONEER.COM] > Sent: Wednesday, July 16, 2003 2:15 PM > To: TakeThisOuTPICLIST.....TakeThisOuTMITVMA.MIT.EDU > Subject: Re: [OT]: Basic theory of a fuel gauge > > I would be willing to bet that if you pull the battery > and then add fuel to the vehicle that the gauge would > not move. > Err..probably not seeing as the gauge is electricaly driven! You would be correct in thinking that. In the old days the tanks were baffled to make the measurement more accurate. Today they slow down the polling of the sensor and then average it with the last few results to create a more realistic measurement. In the old days most fuel gauges used a hot wire movement, whose very slow response time effectively integrated the senders output. Mike ======================================================================= This e-mail is intended for the person it is addressed to only. The information contained in it may be confidential and/or protected by law. If you are not the intended recipient of this message, you must not make any use of this information, or copy or show it to any person. Please contact us immediately to tell us that you have received this e-mail, and return the original to us. Any use, forwarding, printing or copying of this message is strictly prohibited. No part of this message can be considered a request for goods or services. ======================================================================= Any questions about Bookham's E-Mail service should be directed to TakeThisOuTpostmasterKILLspamspambookham.com. -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

To make 10bit A/d into 8-bit, just shift right twice. That's it. --Bob At 11:33 PM 7/16/2003 -0500, you wrote: {Quote hidden}>Andy, just one question. How would you change the code to handle a 10-bit >a/d sample? Probably not necessary in the case where I'm determining fuel >level (1/256 is plenty of resolution), but I'm curious if it can be done >yet kept simple. I suppose I could really make the fuel gauge rock-steady >by using this algorithm along with sampling around 250ms or so... > >Regards, Bob > >Andrew Warren <.....aiwRemoveMECYPRESS.COM> wrote .. > > Bob Japundza <RemoveMEPICLISTspamBeGonemitvma.mit.edu> wrote: > > > > > I have been thinking about the tanks sloshing around as being > > > noisy signal, and that is one of the reasons why I've been > > > contemplating the median filter. Do you have a better suggestion? > > > > Seems to me that a straight low-pass filter would work well. > > > > > I tried an analog low-pass RC filter, and didn't have good luck > > > with it, because of the low resistance of the senders (30-240 ohm > > > Stewart-Warner float types) and not to mention it made me nervous > > > to have large caps tied to the fuel senders, in an airplane, no > > > less. > > > > Try a digital low-pass filter. Given a series of input samples > > in the range 0-255, this code (all 7 words of it) should perform > > a 256-sample moving-window average and give you the filtered > > result in AVEHI. > > > > ; Run this on every input sample. Enter with the new input > > ; sample in NEW. Exits with filtered result in AVEHI, and > > ; NEW unchanged. > > ; > > ; Might want to initialize AVEHI to something reasonable at > > ; program start; if it were me, I'd probably just initialize > > ; AVEHI to the first input sample. > > ; > > ; (WIDTH-1)*AVE + NEW > > ; AVE = ------------------- > > ; WIDTH > > > > AVEHI EQU [ANY FILE REGISTER] ;Holds the average [0-255]. > > AVELO EQU [ANY FILE REGISTER] ;Holds the fractional part > > ;of the average. > > NEW EQU [ANY FILE REGISTER] ;Holds the new sample. > > > > FILTER256: > > > > MOVF AVEHI,W > > SUBWF NEW,W > > SKPC > > DECF AVEHI > > ADDWF AVELO > > SKPNC > > INCF AVEHI > > > > -Andy > > > > === Andrew Warren -- spamBeGoneaiw@spam@spam_OUTcypress.com > > === Principal Design Engineer > > === Cypress Semiconductor Corporation > > === > > === Opinions expressed above do not > > === necessarily represent those of > > === Cypress Semiconductor Corporation > > > > -- > > http://www.piclist.com hint: PICList Posts must start with ONE topic: > > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > >-- >http://www.piclist.com hint: The list server can filter out subtopics >(like ads or off topics) for you. See http://www.piclist.com/#topics --------------- NOTICE 1. This account can accept email & attachments up to 10M in size. 2. Federal Monitors: At request of client, some attachments are encrypted. Please DO NOT delay traffic; please reply with credentials for password. -------------- -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Well hate to argue, but I've worked and built many car's and truck's And while the gauge was positivly fed it is typically fed off of a simple solution that is composed of magnet wire wrapped around some sort of (depending on age) cardboard center. Then an arm moves up and down via a float in the tank. This was common in the 50's and 60's. In the 70's they started using baffles in the tank as it was found that the sender could (and will) change too rapidly in quick accelerations, breaking, and tight turns. Mid 70's to late 80's it was common to find a solution that was closer to what you are describing, but usually the tank still had some sort of baffling system incorporated into it (look at your tank, their are these dents in it that are very uniform) the new baffling system also worked well to hold the tank straps at uniform positions. Late 80's until today they use many different types of senders (thou most are still based on the same idea) and typically the buffering occurs either in the onboard computer (witch tells the gauge where it should be) or in the gauge itself. Want to see a perfect example tear apart the gas tank on a VW bug some time. Even today if you tear down a tank their will be a float connected to a plastic cylinder open the cylinder and you find a simple wire wrap solution. Since gas can't be ignited in its liquid state it is completely acceptable to submerge this (usually along with the fuel pump). {Original Message removed}

> -----Original Message----- > From: Jeremy Darling [SMTP:TakeThisOuTjdarlingspamASKPIONEER.COM] > Sent: Thursday, July 17, 2003 2:21 PM > To: PICLISTEraseMEMITVMA.MIT.EDU > Subject: Re: [OT]: Basic theory of a fuel gauge > > Well hate to argue, but I've worked and built many car's and truck's > And while the gauge was positivly fed it is typically fed off of a > simple solution that is composed of magnet wire wrapped around some > sort of (depending on age) cardboard center. > The difference being I am used to tearing apart components made by Lucas (Prince of darkness) over here in the UK. It's very likely the US manufacturers took a different approach. Having said that I have seen some older French and italian gauages that used the same hot wire method. Most lot of them use a coil of resistance wire wrapped around a bi-metalic strip rather than the classic hot wire type guage. {Quote hidden}> Then an arm moves up > and down via a float in the tank. This was common in the 50's and > 60's. In the 70's they started using baffles in the tank as it was > found that the sender could (and will) change too rapidly in quick > accelerations, breaking, and tight turns. Mid 70's to late 80's it > was common to find a solution that was closer to what you are > describing, but usually the tank still had some sort of baffling > system incorporated into it (look at your tank, their are these > dents in it that are very uniform) the new baffling system also > worked well to hold the tank straps at uniform positions. Late > 80's until today they use many different types of senders (thou > most are still based on the same idea) and typically the buffering > occurs either in the onboard computer (witch tells the gauge > where it should be) or in the gauge itself. Want to see a perfect > example tear apart the gas tank on a VW bug some time. Even today > if you tear down a tank their will be a float connected to a > plastic cylinder open the cylinder and you find a simple wire wrap > solution. > Most modern senders don't use a wound resistance wire per se, but have a printed element on a ceramic substrate, a technique also used in some flap type air flow meters and throttle position sensors. Mike ======================================================================= This e-mail is intended for the person it is addressed to only. The information contained in it may be confidential and/or protected by law. If you are not the intended recipient of this message, you must not make any use of this information, or copy or show it to any person. Please contact us immediately to tell us that you have received this e-mail, and return the original to us. Any use, forwarding, printing or copying of this message is strictly prohibited. No part of this message can be considered a request for goods or services. ======================================================================= Any questions about Bookham's E-Mail service should be directed to RemoveMEpostmasterEraseMEspam_OUTbookham.com. -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

part 1 4585 bytes content-type:text/plain; charset=iso-8859-1 (decoded 7bit) I understand that, but I would like to maintain 10-bit or greater resolution through the averaging algorithm. Regards, Bob Bob Axtell <@spam@cr_axtellRemoveMEEraseMEYAHOO.COM> wrote .. {Quote hidden}> To make 10bit A/d into 8-bit, just shift right twice. That's it. > > --Bob > > > At 11:33 PM 7/16/2003 -0500, you wrote: > >Andy, just one question. How would you change the code to handle a 10-bit > >a/d sample? Probably not necessary in the case where I'm determining > fuel > >level (1/256 is plenty of resolution), but I'm curious if it can be done > >yet kept simple. I suppose I could really make the fuel gauge rock-steady > >by using this algorithm along with sampling around 250ms or so... > > > >Regards, Bob > > > >Andrew Warren <EraseMEaiw@spam@CYPRESS.COM> wrote .. > > > Bob Japundza <@spam@PICLISTspam_OUT.....mitvma.mit.edu> wrote: > > > > > > > I have been thinking about the tanks sloshing around as being > > > > noisy signal, and that is one of the reasons why I've been > > > > contemplating the median filter. Do you have a better suggestion? > > > > > > Seems to me that a straight low-pass filter would work well. > > > > > > > I tried an analog low-pass RC filter, and didn't have good luck > > > > with it, because of the low resistance of the senders (30-240 ohm > > > > Stewart-Warner float types) and not to mention it made me nervous > > > > to have large caps tied to the fuel senders, in an airplane, no > > > > less. > > > > > > Try a digital low-pass filter. Given a series of input samples > > > in the range 0-255, this code (all 7 words of it) should perform > > > a 256-sample moving-window average and give you the filtered > > > result in AVEHI. > > > > > > ; Run this on every input sample. Enter with the new input > > > ; sample in NEW. Exits with filtered result in AVEHI, and > > > ; NEW unchanged. > > > ; > > > ; Might want to initialize AVEHI to something reasonable at > > > ; program start; if it were me, I'd probably just initialize > > > ; AVEHI to the first input sample. > > > ; > > > ; (WIDTH-1)*AVE + NEW > > > ; AVE = ------------------- > > > ; WIDTH > > > > > > AVEHI EQU [ANY FILE REGISTER] ;Holds the average [0-255]. > > > AVELO EQU [ANY FILE REGISTER] ;Holds the fractional part > > > ;of the average. > > > NEW EQU [ANY FILE REGISTER] ;Holds the new sample. > > > > > > FILTER256: > > > > > > MOVF AVEHI,W > > > SUBWF NEW,W > > > SKPC > > > DECF AVEHI > > > ADDWF AVELO > > > SKPNC > > > INCF AVEHI > > > > > > -Andy > > > > > > === Andrew Warren -- spamBeGoneaiwEraseMEcypress.com > > > === Principal Design Engineer > > > === Cypress Semiconductor Corporation > > > === > > > === Opinions expressed above do not > > > === necessarily represent those of > > > === Cypress Semiconductor Corporation > > > > > > -- > > > http://www.piclist.com hint: PICList Posts must start with ONE topic: > > > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > > >-- > >http://www.piclist.com hint: The list server can filter out subtopics > >(like ads or off topics) for you. See http://www.piclist.com/#topics > > --------------- > NOTICE > > 1. This account can accept email & attachments up to 10M in size. > 2. Federal Monitors: At request of client, some attachments are encrypted. > Please DO NOT delay traffic; please reply with credentials for password. > -------------- > > -- > http://www.piclist.com hint: The list server can filter out subtopics > (like ads or off topics) for you. See http://www.piclist.com/#topics -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics part 2 2 bytes -

Bob Japundza <PICLISTspamBeGonemitvma.mit.edu> wrote: > Andy, just one question. How would you change the code to handle > a 10-bit a/d sample? Probably not necessary in the case where I'm > determining fuel level (1/256 is plenty of resolution), but I'm > curious if it can be done yet kept simple. Bob: The code wouldn't be AS simple, but it certainly wouldn't be complicated; the fast that the moving window is 256 samples wide is the real key to making the implementation easy. > I suppose I could really make the fuel gauge rock-steady by using > this algorithm along with sampling around 250ms or so... Yes. A 256-sample window, with a sample taken 4 times per second, gives you a 64-second-wide window. If that's too wide, and you want to narrow the window without affecting the simplicity of the code, just call the filter routine multiple times per sample. -Andy === Andrew Warren -- RemoveMEaiw@spam@spamBeGonecypress.com === Principal Design Engineer === Cypress Semiconductor Corporation === === Opinions expressed above do not === necessarily represent those of === Cypress Semiconductor Corporation -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Yep.... some automotive instrument manufacturers have finally started using microcontrollers in "analog" gauges to compensate for the non-linearity of the (traditional) sensors. The analog part is the traditional needle, which is not handled by stepper motors. I've not specifically seen this being used for fuel-level gauges, but definitely for boost, oil-pressure, etc. My guess is that at some point, they will find it pointless to retain traditional meter movements for just a handful of gauges such as fuel-level. Cheers, -Neil. On Thursday 17 July 2003 15:13, Peter L. Peres scribbled: > > Any ideas how they're doing this? Perhaps a servo? > > It can be a stepper. Does it index (go to zero) sometimes ? Otherwise it > could be simply a differential instrument with deliberately high friction > bearings. With modern magnets that should be easy. When you say it works > w/o ignition, do you mean it shows what it showed when you switched off or > does it really show what's in the tank (assuming youare filling it). > > Peter -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Wednesday 16 July 2003 23:17, Bob Japundza scribbled: > Yep..the wing tanks have ribs in them and the lightening holes are not cut > out, except for some holes to allow fuel to pass in between the ribs, or > 'bays' as they're referred to. Its designed to help keep down sloshing, > but there is still quite a bit of sloshing even when nearly full. Absolutely. It's a major problem in a car, and I know from flying that it's an greater problem while being tossed about in the air. I *never* trust or even bother with fuel gauges in small aircraft. Only once when I first get into the aircraft, before even doing a walk-around check, I look at the fuel gauges to know if to call for a fuel truck. > There is > one rib (#2 from the root) which has a trap door to allow fuel to easily > flow into the first bay but not as easily to flow out, so that in an > uncoordinated turn while low on fuel the pickup tube doesn't unport, which > would obviously cause the engine to quit. Do you have any diagrams for this, I had started redesigning a new fuel tank for my (kit)car, but never finished it. Primary reasons were to have dual in-tank pumps, more fuel capacity, and a proper sump, but I know I'll need some good baffling to prevent running the pumps dry, which would be very, very bad. I had looked a bit at trap doors on oil pans, but felt uncomfortable with metal scraping metal (all aluminum) inside a fuel tank. Of course, since I have electric pumps in there, I shouldn't be such a pansy about it. :-) If you have any diagrams showing how the trap doors are implemented, I'd love to see them. > One problem with the RV's are > the float-type fuel senders are not very accurate above 1/2 full, due to > the dihedral of the wing and the location of the senders being attached to > the root rib. I've been thinking about putting a second sender on the > other end, and using a pic to extrapolate fuel level with a table lookup > based on the a/d results of sampling both senders... If you can get the full range, a PIC lookup-table can very easily compensate for non-linearity. If you run 2 senders, you'll have to do some in-flight calibrating, since averaging may always be low ... during sloshing, I can imagine situations where one sender would go up while the other would go down, and the average would be correct, but there are other situations where both could be lower than actual. I cannot see how both would ever get higher ... unless you're inverted :-) > There are some other > alternatives, such as capacitance-typ! e senders, but they are inaccurate > due to different fuel densities and temperature changes... Personally, I feel that using flow-sensors (~3% accuracy) would do an excellent job here. If EFI, you could also calculate the consumed fuel from pressure and injector-pulse widths. And if you use either of these, you could still use an in-tank float-type level sensor with a very very long averaging period to essentially "re-calibrate the fuel-flow" sensor over a longer period of time. BTW, a new/recent wave of "arm-less" fuel senders have hit the market. A bit more expensive, but I like the idea of not having the swing arm. Cheers, -Neil. -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

{Quote hidden}> -----Original Message----- > From: Picdude [SMTP:.....picdude@spam@EraseMENARWANI.ORG] > Sent: Friday, July 18, 2003 12:18 AM > To: .....PICLISTRemoveMEMITVMA.MIT.EDU > Subject: Re: [OT]: Basic theory of a fuel gauge > > On Thursday 17 July 2003 01:57, Michael Rigby-Jones scribbled: > > ... In the old days the tanks were > > baffled to make the measurement more accurate. ... > > Allow me to question this statement. Baffles, trap doors, and sumps are > primarily put in for one purpose ... to reduce sloshing so that the fuel > pump > does not run dry. This is not so critical on float-bowl / carb'ed > engines, > but very critical on high-pressure EFI systems, which is most everything > being produced today. The fact that baffles also help measurement is a > positive side-effect, which just means they put less effort into solving > this > problem. But I'm considerably sure that they don't baffle just for > measurement. > I'm afraid you have misquoted me, Jeremy Darling was the OP of that statement. Mike ======================================================================= This e-mail is intended for the person it is addressed to only. The information contained in it may be confidential and/or protected by law. If you are not the intended recipient of this message, you must not make any use of this information, or copy or show it to any person. Please contact us immediately to tell us that you have received this e-mail, and return the original to us. Any use, forwarding, printing or copying of this message is strictly prohibited. No part of this message can be considered a request for goods or services. ======================================================================= Any questions about Bookham's E-Mail service should be directed to .....postmasterSTOPspam@spam@bookham.com. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

On Friday 18 July 2003 03:39, Michael Rigby-Jones scribbled: {Quote hidden}> > -----Original Message----- > > From: Picdude [SMTP:picdudeEraseME@spam@NARWANI.ORG] > > Sent: Friday, July 18, 2003 12:18 AM > > To: RemoveMEPICLISTspamBeGoneMITVMA.MIT.EDU > > Subject: Re: [OT]: Basic theory of a fuel gauge > > > > On Thursday 17 July 2003 01:57, Michael Rigby-Jones scribbled: > > > ... In the old days the tanks were > > > baffled to make the measurement more accurate. ... > > > > Allow me to question this statement. Baffles, trap doors, and sumps are > > primarily put in for one purpose ... to reduce sloshing so that the fuel > > pump > > does not run dry. This is not so critical on float-bowl / carb'ed > > engines, > > but very critical on high-pressure EFI systems, which is most everything > > being produced today. The fact that baffles also help measurement is a > > positive side-effect, which just means they put less effort into solving > > this > > problem. But I'm considerably sure that they don't baffle just for > > measurement. > > I'm afraid you have misquoted me, Jeremy Darling was the OP of that > statement. > > Mike Ah.... okay. Sorry for the confusion. Cheers, -Neil. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.