Dan M Posted February 27, 2009 Posted February 27, 2009 Well, I haven't talked to Wittner yet but will. In the meantime, let me argue the case that you all have it backasswards. Here goes: Break open your stock sensor. What is the small metal bit made of? It looks like copper to me. What does copper do really well? It transfers heat really, really well. It's also expensive. It would never be speced unless the spec-er wnated maximum heat transfer, for cost reasons. Then, after spec-ing expensive copper, what do they make the rest of the holder out of? Plastic. What does plastic do pretty well? Many things, but among them is that it is a poor conductor of heat. So, they encase a great conductor of heat, which is screwed directly into the heat source and then encase the conductor in an insulator. Why would someone do that? Let's see if your mental faculties are firing tonight . . . Holy shit. I'm offline for 11 days and 7 pages are added to this thread. I was thinking the activity was waning but evidently I had that thought bassackwards. So far as sensor material and operation, I need no education on that Greg. I've been testing thermistors for 30ish years. I've heated them, cooled them tested resistance and voltage drops. I've graphed their operation from cold start to full operating temperature. I'm quite certain I know how they work. My contention is whoever spec'd the sensor also spec'd the holder. Why not screw it directly into the head? That seems to be the most direct contact, eh? That would be even cheaper if your theory that less money was the driver here. Why on a bike that has expensive components like Brembo gold line brakes would they save a buck on a sensor holder? I still contend the plastic was spec'd for a reason. Sometimes engineers make mistakes, and have to shift on the fly especially when dealing with new (to them) systems. Why would you say Greg that they spec'd a vacuum controlled fuel pressure regulator then didn't hook it up to vacuum? It is because it is not appropriate to the vacuum pulses of a twin. That fact didn't stop it from making it on to the bikes. They just had to alter it and keep it open to atmosphere. That temp sensor is designed to be in a water jacket. The ECU is a simple one. I'm thinking the extreme and rapid temp changes of an air cooled cylinder head are a bit much for it (the ECU) to handle. Insulating it with plastic mimics it being in coolant. It slows variation. Makes for a much simpler map.
Guzzi2Go Posted February 27, 2009 Posted February 27, 2009 ... It slows variation. Wouldn't less frequent sampling have the same effect?
luhbo Posted February 27, 2009 Posted February 27, 2009 Only for forums. In RL this would never work, there are just too many parties involved. I guess the ecu is continuosly polling the analog inputs, anything else would make problems only. Hubert
Guest ratchethack Posted February 27, 2009 Posted February 27, 2009 I've been testing thermistors for 30ish years. I've heated them, cooled them tested resistance and voltage drops. I've graphed their operation from cold start to full operating temperature. I'm quite certain I know how they work. My contention is whoever spec'd the sensor also spec'd the holder. Why not screw it directly into the head? That seems to be the most direct contact, eh? That would be even cheaper if your theory that less money was the driver here. Why on a bike that has expensive components like Brembo gold line brakes would they save a buck on a sensor holder? I still contend the plastic was spec'd for a reason. . . . That temp sensor is designed to be in a water jacket. The ECU is a simple one. I'm thinking the extreme and rapid temp changes of an air cooled cylinder head are a bit much for it (the ECU) to handle. Insulating it with plastic mimics it being in coolant. It slows variation. Makes for a much simpler map. Welcome back to the fray, Dan. I'm just wrapping up a little write-up on my latest excursion on this, in which I arrived at some of the same conclusions you've noted above, and will post it in a bit. I'd very much appreciate your well-qualified perspective in response, soon as I get a ROUND TUIT.
Guzzi2Go Posted February 27, 2009 Posted February 27, 2009 Only for forums. In RL this would never work, there are just too many parties involved. I guess the ecu is continuosly polling the analog inputs, anything else would make problems only. Hubert Well, less frequent is still continuous. Anyway, the question was not what ECU does, but rather if the goal was to minimise impact of transients on temperature measurement, wouldn't there be other, simpler and more reliable methods then designing a special receptacle? An example of such a method would be infrequent sampling.
Greg Field Posted February 27, 2009 Posted February 27, 2009 Holy shit. I'm offline for 11 days and 7 pages are added to this thread. I was thinking the activity was waning but evidently I had that thought bassackwards. Note the use of the word "all." It is not singular. This was not aimed directly at you. So far as sensor material and operation, I need no education on that Greg. I've been testing thermistors for 30ish years. I've heated them, cooled them tested resistance and voltage drops. I've graphed their operation from cold start to full operating temperature. I'm quite certain I know how they work. My contention is whoever spec'd the sensor also spec'd the holder. Why not screw it directly into the head? That seems to be the most direct contact, eh? That would be even cheaper if your theory that less money was the driver here. Why on a bike that has expensive components like Brembo gold line brakes would they save a buck on a sensor holder? I still contend the plastic was spec'd for a reason. Sometimes engineers make mistakes, and have to shift on the fly especially when dealing with new (to them) systems. Why would you say Greg that they spec'd a vacuum controlled fuel pressure regulator then didn't hook it up to vacuum? It is because it is not appropriate to the vacuum pulses of a twin. That fact didn't stop it from making it on to the bikes. They just had to alter it and keep it open to atmosphere. That temp sensor is designed to be in a water jacket. The ECU is a simple one. I'm thinking the extreme and rapid temp changes of an air cooled cylinder head are a bit much for it (the ECU) to handle. Insulating it with plastic mimics it being in coolant. It slows variation. Makes for a much simpler map. I never said you needed an education. I implied that perhaps many of you are over-thinking this whole thing. Why not plug the sensor directly into the head? Are you serious? See how long that probe is? If you sunk that into the head, the piston'd launch it into the next county. Yes, the plastic would be for insulation. By acknowledging that, you are assuming rational thought on the part of its designer. Continue that, and you have to assume rational thought on the choice of copper as the medium designed in to transfer heat to the probe tip, too, right? If this rational designer wanted to damp or attenuate the temp reading, why would he not spec a material that transfers heat slowly, such as cast iron, eh? A gap between the tip and the copper could easily be a production error, rather than a lapse of rational thought. Look, I'll ask Dr. John. It's just that there's a very narrow time window in which it's OK to call him and everyday I find myself too busy during that window.
Dan M Posted February 27, 2009 Posted February 27, 2009 I never said you needed an education. I implied that perhaps many of you are over-thinking this whole thing. Why not plug the sensor directly into the head? Are you serious? See how long that probe is? If you sunk that into the head, the piston'd launch it into the next county. Yes, the plastic would be for insulation. By acknowledging that, you are assuming rational thought on the part of its designer. Continue that, and you have to assume rational thought on the choice of copper as the medium designed in to transfer heat to the probe tip, too, right? If this rational designer wanted to damp or attenuate the temp reading, why would he not spec a material that transfers heat slowly, such as cast iron, eh? A gap between the tip and the copper could easily be a production error, rather than a lapse of rational thought. Look, I'll ask Dr. John. It's just that there's a very narrow time window in which it's OK to call him and everyday I find myself too busy during that window. First off, I'm not trying to pick a fight here. When somebody tells me I have something assbackwards I feel compelled to respond, explain or say thanks for the correction. Certainly a shorter sensor could have been spec'd to screw directly into the head. That part is easy, they come in all shapes and sizes. Cast iron would hold temp and slow things too much. No? You're sampling temp of an alloy head. Cast iron would really hold heat. I don't think they were trying to alter readings, just smooth them some. This has all been speculation and I suppose it is why we have a tech board. Input from DrJ if he was involved in this end of the design will be welcome. We all want to know why. The truth is out there...
Guest ratchethack Posted February 27, 2009 Posted February 27, 2009 UPDATE: Here we have a bit of a “wrap-up and write-up” for the little head temp sensor extravaganza I launched a month or so back. You might have to be somewhat bored, have all the Guzzi maintenance up to date, and be fairly inclined to do semi-fiddly stuff with Guzzi’s in the first place to indulge in this kinda behavior yourself. Being as loopy as a bag o’ squirrels with a workshop full o’ tools on hand in Winter ain’t a prerequisite. But it couldn’t hurt. Since it’s warmer here now, and it’s stopped raining (again), I’m pleased to’ve done this in bad weather (yes, bad weather is a relative thing ), or I more’n likely wouldn’t have started on it in the first place. But in retrospect, the results were more positive than expected, and I’ve advanced my understanding considerably. Based on the results and what I learned in the process, I’d do it again -- even in good weather -- in a heartbeat. Sincere thanks again to John A. for the inspiration a year and a half ago. WARNING: There are no designer sound bites to be found here. So if you’re not sincerely interested in reading what for some, no doubt, would be beyond insufferabe in the direction of too much to read, and would demand an excruciatingly painful effort from at least a select few -- I’m begging you -- please consider this my cordial invitation to adjourn immediately to the Hooter’s thread, where your time might be much better spent. If you’re compelled to reply to this post, I welcome you to do so with utmost sincerity. But for God’s sake, DO NOT reply without reading it, or “cherry pick” it for stuff you disagree with by pulling things out of context and/or leaping to false conclusions just to indulge some irrepressible compulsion to fling baseless accusations of dementia and lunacy about, without any comprehension of wot you're responding to. This is only common Forum courtesy, and nearly always more likely to save you the embarrassment of stepping in something nasty that you can’t get yourself out of. That’s the deal, take it or leave it. I hope you’ll consider this a fair enough exchange. If not, ‘bye now. EXEC SUMMARY I was into my experiments and analysis on this awhile before I’d concluded that the extent of the inherent error in the OE head temp sensor/holder went far beyond, and in a significantly different direction entirely, from what might first be evident by casual observation -- yes, even by measurement with diagnostic tools. Thanks also to Pete and others for their info. It was an invaluable starting point. Per previous posts, I’ve wired up a linear taper 1K Ω variable resistor in series with the head temp sensor, and have tested this satisfactorily with thermo-paste in the plastic holder, with a heat sink under the body of the sensor. I’ve got somewhere around 50 miles of mixed-mode riding (in traffic and on the slab up to ~90 mph) on it at this point. With the thermo-paste in there, the bike would not run even remotely acceptably without either the heat sink or the variable resistor (one or the other, both being significantly more satisfactory than either alone) with my current map, and equally poorly with the brass holder, regardless of what I put in it, or empty. I’ve arrived at the above strictly as an academic exercise. Per multiple previous posts, I HAD NO REAL “PROBLEM” TO SOLVE here, unless very occasional slight hot over-lean symptoms at and off idle and very occasional pinging (detonation) after hard riding in hot weather be considered "problems", which I'd never considered either to be. If I had, I’d have re-mapped 6 years ago. But since my “library” PC III map has always delivered far better than “acceptable” performance under all conditions of operation without meddling with the OE plastic sensor/holder a-tall, I ain’t fixin’ wot I figure ain’t broke by replacing the map. Now that I’ve thoroughly tested my slight head temp sensor/holder modifications, it’s likely to stay “as is” for the forseeable. Cost of the hardware (variable resistor, knob, and heat sink) was $2 USD. WHAT, EXACTLY, DOES THIS ACCOMPLISH? All I’ve done here is restore the very nearly flawless running my Guzzi exhibits without thermo-paste in the OE plastic sensor holder, and get it to run the same (or better) WITH thermo-paste by fooling the ECU into believing the motor is running cooler than it actually is, only when needed "on the fly", still using my original “library” PC III map with OE sensor/holder error encoded. Capice? OBLIGATORY GREAT WALLOPING DISCLAIMER: (Which experience tells me just might actually throw a bit of a fence around the sheer magnitude of expected leaping to false conclusions in group formation, and keep the usual ration of spectacularly false interpretations to a somewhat manageable level.) This was a $2 experiment. If you believe value is always directly proportional to cost, and that this is a hard and fast fundamental rule o' nature, there's next to no value to be found for you here. There was no re-mapping, no brake dyno, no exhaust gas analyzers, Tune Boys, Axone diagnostic tools, and no wind tunnel. If you believe that the absence of any or all of these invalidates anything or everything I observed and noted below from my extensive testing on the road, you're entitled to your opinion, but I beg you differ, and I b'lieve the results speak more credibly than well enough for themselves, but that's just me. Though I believe I've uncovered new, and also confirmed existing principles of operation of the sensor/holder COMMON TO ALL V11's, THAT HAVE NOT BEEN WELL UNDERSTOOD hereabouts, there’s no claim of any magic “cure all” here. So by all means, let no one be so silly as to mistake this post for a recommendation to do wot I’ve done here! Each and every motor, intake and exhaust config, map, and every combination thereof is different than the next, so what worked well for me here and led me in the direction my observations took me based on my "seat o' the pants dyno" MAY BE just as likely NOT to work as well on the next Guzzi. I seem to’ve had similar positive results to those of John A., and I’ve been experimenting here, that’s all. Many of my observations here ARE NOT CLAIMS OF FACT, and may be subject to error and/or revision. On the other hand, some may recognize similarities to their own situation and be encouraged to try something I've demonstrated here. Per previous posts, I’ve fully tested both the OE sensor/plastic holder and the brass holder under a variety of modifications, with and without thermo-paste, filled solid with lead and various gradations of air gaps, and with and without the heat sink. If I had a line on a different sensor or sensor/holder altogether that might be more suitable than the above, I’d have tested it. I don’t know of any, so I’ve gone in the only known direction that my testing could lead me. Without any question or equivocation, that would be directly back to the OE sensor/plastic holder setup, and I reckon it’ll stay installed on my Guzzi the way I’ve doctored it up here, unless or until I find and test something that I believe works better. THE SEMI-FIDDLY PART Ergonomics: I was obliged to find and fit a small, sturdy, weather-resistant enclosure for the variable resistor, and locate it close to the left-hand switch cluster in a secure, unobtrusive place, where the adjustment knob is both easily accessible without having to take my eyes off the road to look for it, without taking my hand off the grip, and at the same time visible enough so I can determine its position at a glance, and where it’s not a too terribly awful eyesore. It took a little extra unanticipated effort. Otherwise, all of this was as simple, easy, and straightforward as might be expected. Semi-fiddly to fab up, but in the end, not too horribly awful, very good ergo's -- and it works like a Champ CRITICAL CONSIDERATIONS OF THE OE TEMP SENSOR/HOLDER ERROR Despite what I’ve now confirmed to be a common misconception on this thread, the error inherent in the heat read via the OE sensor/holder setup IS NOT SIMPLY LIMITED to its inability to read high enough temperature due to the air gap. It’s being used to both monitor and control FI here – something it was apparently not designed to do. The sensor was engineered to be immersed in continually circulating liquid -- NOT in air, NOT in a dab of thermo-compound at the tip, and NOT entirely immersed in a static, uncirculating mass of thermo-compound, oil, anti-seize, or what-have-you. As noted by many previously, the thing appears to’ve been intended for use as a simple temperature monitoring instrument sensor and/or fan control sensor in a cooling jacket and/or oil gallery -- NOT as a FI monitoring and control sensor – and almost without question, NOT for FI monitoring and control of an air cooled motor. Properly designed, and specifically purpose-built and developed, seems to me a FI control sensor for an air cooled motor would be considerably different in both form and function, one application to the next (notably, far less massive for a more accurate dynamic read). WHAT NEW AND SIGNIFICANTLY DIFFERENT OPERATING PRINCIPLE WAS CONFIRMED? ANSWER: THERMAL INERTIA OF THE OE SENSOR BODY Regardless of original design intent, and quite possibly MOST significantly, the plastic sensor/holder, as delivered from Mandello, (and the brass holder combo also, for that matter) can’t read temp changes FAST ENOUGH to keep up with engine dynamics and mode of operation – regardless of air gap or thermo-paste in either one. As mentioned previously, the sizeable mass of the sensor body acts as a heat reservoir with significant THERMAL INERTIA. This is a condition that WOULD NOT EXIST if the entire temp probe were immersed in continually circulating liquid, as the sensor was designed to be used. The THERMAL INERTIA problem (that is, the sensor body not responding quickly enough to temp changes) is NOT RESOLVED by either of 2 conditions: 1. an air space, as I believe was intended for both plastic and brass holders by Guzzi (.015" and .25", respectively), or 2. thermo-compound, lead, or other heat conductor between holder base and sensor tip. After adding thermo-compound, I found that a heat sink with the plastic holder is an effective PASSIVE measure that helps to overcome this (see below), and that using added resistance to the sensor output as needed is an effective ACTIVE measure to modify the worst of the undesirable sensor output. Using both yielded the best results on the road with my PC III map. A FEW MORE THINGS I OBSERVED, LEARNED, AND/OR CONFIRMED The inherent error in the OE sensor is incorporated into any map that was created with the OE sensor setup in place, and the ECU will accordingly attempt to compensate for the sensor error as directed by each and every map made to do exactly that. As I’ve demonstrated consistently and posted previously, if you’re running a map that was created with the OE sensor installed (as I do), and then attempt to “correct” that inherent sensor error by simply adding thermo-paste and/or use of the brass holder without re-mapping, there’s considerable downside risk of “confusing” the ECU into making the bike run WORSE, to the point of being unrideable at full operating temperatures, (particularly in warm weather), with varying degrees of symptoms including stumbling, balking, surging, coughing, bucking, popping, snuffing out altogether, etc., at idle, off idle, and at trailing throttle, where the ECU is feeding over-lean pulsewidth signals to the FI. This is the “double astigmatism” syndrome I noted earlier, whereby the ECU is now “confused” by a MORE ACCURATE READ via the old map, which still “expects” the sloppy output it was mapped to interpret -- and now doesn’t see. In other words, inherent error exists not only in the OE sensor/plastic holder setup, but in any map created with it in place. So by all means, if you’re inclined to “correct” the sensor by adding thermo-compound, this may make very good sense, and it MAY BE very effective -- but unless you want to consider “cheating” like Yours Truly here, best ALSO consider the need for doing this only BEFORE re-mapping to complete the correction, otherwise the remaining potential for inherent error in the map, while possibly not absolutely “a lock”, can nevertheless be significant enough to result in NEW unwanted symptoms, as it did in my case. OK, BUT WHY THE HEAT SINK? Heat flow of a similar example, and as installed Based on my observations, as noted previously, this provides an escape route for dumping excess heat from the sensor body at all times, directly and "passively" overcoming the THERMAL INERTIA of the sensor body on the heat input side of the thermistor. It shortens up lag times in heat flow changes at the sensor probe tip caused by the sensor body acting as a relatively dead-end heat reservoir at all operating temperatures. As long as the sensor tip is thermally connected to the holder base (and therefore the head) with thermo-paste, temperature reads ARE NOT adversely distorted by the heat sink, since the sensor tip (where the resistance signal of the thermistor is actually read) is by far more directly (and correctly) influenced by the constantly changing heat source of the cylinder head than it is by the heat sink at the other end of the sensor body -- as long as the heat of the sensor body is not significantly higher than the head when the head cools down. By logical analysis and the laws of thermodynamics, by lowering the heat content of the sensor body with the heat sink, a more dynamically changing heat flow through the tip occurs, which provides a more accurate read. I believe the results as tested on the road demonstrate and fully support this analysis. I believe the finning machined into the brass holder was an attempt to leverage the same beneficial heat sink principle to a lesser degree. It doesn’t work well due to the high heat conductive property of brass, which, though brass has very good properties as a heat sink, and is of course stronger than the plastic holder, and is relatively immune to ham-fisted wrenching, it conducts prohibitively high heat flow directly into the dead-end heat reservoir of the sensor body, negatively overriding any benefit of the heat sink effect, and hammering the ability of the sensor to respond quickly. It makes responses to changes in temperature even slower by maintaining a high heat content in the sensor body. The high thermal inertia property of the sensor body causes false reads at the sensor tip, both higher and lower than at the head itself, depending on conditions, map, motor config, etc. This is just me, but I seriously doubt if cost of materials ever had the slightest bit to do with Guzzi spec’ing one vs. the other -- large lot buys and bean counter edicts or not. The plastic holder simply does a superior job overall, since it's an insulator rather than a conductor of heat, and it doesn't make anywhere nearly as direct a thermal connection with the sensor body, and it conducts relatively more of the source heat flow to the sensor tip than to the sensor body -- even over the OE .015" air gap. Hence the justification for the factory upgrade from brass to plastic. Adding the heat sink simply makes the plastic sensor/holder setup more responsive to engine temp dynamics with the thermo-paste installed – at least in my case, as the the results on the road demonstrate. WHY THE VARIABLE RESISTOR? 1K Ω variable resistor Lacking any known alternative to the far less than ideal sensor/holder combo itself, like the heat sink, this is merely another "work-around". Nothing more, nor less. Where the heat sink is a direct, PASSIVE measure on the heat input side of the thermistor, this is an indirect, ACTIVE way to counteract the worst shortcomings on the signal output side of the thermistor, caused by the wonky OE plastic sensor/holder. Per posts by John A. in this thread back in June of ’07, if you’re so inclined, and don’t want to re-map, but would like to be able to selectively rectify an over-lean condition, (which seems to be most prevalent) adding a variable resistor can be an effective way to accomplish this. John used a DPDT switch so he could bypass the temp sensor entirely and toggle back and forth. I took a different approach and obviated the use of a switch, since in my case, I have no need to provide lower resistance than the sensor provides at any condition of operation, and simply wanted to be able to selectively provide more resistance, using the sensor output as a base. The output of the NTC thermistor in ohms vs. temperature is on a logarithmic scale at the sensor (see output chart below). The interpretation of the sensor output by the ECU, and the A/F control output of the FI pulsewidth signal by the ECU, however, is based strictly on providing optimal A/F at the temperature it reads (and nothing else – no scale of output by the ECU here, logarithmic or any other) via the sensor output on the TempR line (again, see temp/resistance output chart and note below as posted numerous times previously). This means that when adding resistance by dialing up the linear taper variable resistor, one need not be concerned about too much or too little gain (as would be the case on a logarithmic output scale at the injectors) from a given rotation of the adjustment knob depending on what part of the adjustment rotation scale the knob happens to be. In other words, the linear taper provides added richening of the A/F (that richening still being directly dependent upon the temperature of the head, as read at the sensor tip) that does not go “out of proportion” at either end of the adjustment range. The variable resistor, in combo with the heat sink, allows me to run with thermo-paste in the OE plastic sensor by allowing me to enrich the A/F as needed, on the fly. This completely eliminates over-lean symptoms at full operating temp at idle and low RPM, typically by dialing up an additional ~300 Ω, and has the positive additional effect of dropping operating temp at the heads by eliminating lean-burn-generated higher temps. Once underway again, it allows me to flick the dial back to zero for more optimum fuel mileage. I’ve also found that dialing up an extra 500 Ω at fully warm startup fires it up quicker than without. ACKNOWLEDGEMENT OF PREVIOUS ERROR ON MY PART I had mistakenly posted the intent to use a 500K Ω variable resistor. My apologies. This was a bone-headed flub, due to my lack of recent hands-on with my DMM, not having much use of it in recent years. I had just checked the output of the sensor to make sure it was working correctly (it was), and had concluded that a 500 Ω variable resistor would be about right according to the output chart (below), but I had the DMM set on the wrong scale. So I wrote down the wrong value and had to exchange the one I bought. FWIW, they didn’t have a 500 Ω unit in stock, so I was obliged to go with a 1K Ω unit. After I got it installed and had tested it for awhle, I found the 1K Ω unit preferable, because it doesn’t take as much rotation of the knob to dial up 500 Ω, (the most I ever seem to need, as expected), and I find the ergo’s of using only partial dial rotation more to my liking on the road anyway for a quick flick with the index finger. BUT IT’S A KLUGE! IT’S CHEATING! IT’S BARBARIC! IT’S AN ABOMINATION! Yes, Binky. If you insist. For Yours Truly, it beat all the known alternatives, and quite handily at that. Make of it wot you will, the performance of adding the heat sink and variable resistor is for all intents and purposes a 100% functional success by my analysis, and to each his own. It allows use of the thermal goop in the sensor/holder without requiring a re-map, it doesn’t require any analysis tools, and it puts me in more direct control on an as-needed basis. At a cost of $2 USD, for Yours Truly, the "bang f'er the buck" here beats the living thermo-paste out o' lots of other alternatives. . . But o' course, that's just me. One might think of it this way: Going in the opposite direction in an attempt to overcome a persistent over-rich condition with an entirely different set of circumstances based on ambient climate, motor config, and map -- is wot I've done here any more of a kluge, cheating, barbarism, or abomination than wrapping a brass holder in duct tape in an attempt to get a higher heat read at the sensor tip than without? What this does, by my read, is at least as crude, and perhaps more likely to produce a raft of unwanted side-effects (of which there are none in my case). BUT WHY, EXACTLY, WOULD ANYONE DO THIS -- especially if their Guzzi already runs just fine without messing with it? ANSWER: By my analysis so far, first and foremost, it allows significantly cooler and smoother running at extended traffic lights once up to full operating temp, even in cool weather, but particularly in warmer weather. As noted above, it also provides noticeably quicker warm start-up. As I'd hoped from the start, over time, having the sensor tip thermally connected to the holder and head with thermo-paste (again, not possible otherwise with my map) will undoubtedly reveal more differences in operation on the road to come, due to a more consistent, accurate, dynamically responsive, and better synchronized heat read by the ECU than with the OE sensor setup and my current map – yes, even though the ECU is now “confused” by a more accurate temp read via the existing map, as noted above -- even better fuel mileage than my previous average (~40 mpg) being the expectation. This part remains to be seen, and time will tell. I’ll be running a few tanks of fuel through it to get a new MPG read. Will advise (Part IX). UPDATE: BETTER MILEAGE CONFIRMED over back country and hard, heavy-throttle, higher RPM mountain climb at higher temps of ~75-80°F (~24-27°C). As I'd hoped for, better than previous mixed-mode riding avg. mileage of ~40 MPG was achieved with the addition of thermo-paste, heat sink, and var. resistor, returning ~43 MPG. NOTE: As I’d flagged conspicuously up front and detailed above, many seem to've had experiences with their sensors ENTIRELY OPPOSITE to those of mine above. This doesn’t mean either experience is WRONG. I believe it just means we have different combinations of engine, intake, exhaust, and maps, therefore, we have entirely different head temp sensor/holder and FI control experiences – for what would seem to be fairly obvious reasons. Many -- but by no means all -- seem to be running overall richer maps (particularly at idle and low throttle openings) than Yours Truly. NOTE: (Part II): Any Q’s, observations, and especially any CHALLENGES to any of the above (or to my understanding of my observations) from anyone sincerely interested are as equally or even more sincerely welcome, and will be responded to in kind, time willing, and to the best of my ability. NOTE (Part III): As before, any reply posts demonstrating mindless ridicule, thudding ignorance, less than sincere interest (obvious evidence of not reading what one is responding to being the common trend), or any combination thereof, will be ignored. Well then. There you have it. The sun shines today and the open road beckons. Seductively, at that. . . ________ Temperature Sensor Output The following data table on the NTC thermistor (which appears in the Guzzi Service manual) was shamelessly lifted from Cliff’s link here: http://www.cajinnovations.com/MyECU/temperature.htm Temperature °C vs. Resistance ohms As noted in prior posts, this is the only relevant data to consider when adding a variable resistor to enrich A/F at operating temps. It maps the output of the thermistor under the only temp sensor feed to the ECU that is read at operating temperatures after startup. It’s the same data in Cliff’s previously posted TempR line. At 125°C, dialing up an additional 1K Ω “fools” the ECU into believing the motor is running at ~50°C, richening the A/F considerably. At 125°C, and dialing up an additional 100 Ω, the ECU is “fooled” into believing the motor is running at ~100°C, richening A/F relatively slightly. -40 100950 -30 53100 -20 29120 -10 16600 0 9750 10 5970 20 3750 25 3000 30 2420 40 1600 50 1080 60 750 70 525 80 380 90 275 100 205 110 155 125 100
Dan M Posted February 27, 2009 Posted February 27, 2009 That is a bit of time spent there, thanks. Interesting stuff. It also adds to parking lot mystique triggering "what the hell is that" looks from bystanders. You can come up with a crafty tag for it "That's my fuel ratio dial" Your work proves what an archaic lump the V11 is. Seems it can't be made to run properly at anything near stoich. Fooling the ecu into thinking it is anywhere from 10C to 30C cooler than actual is likely providing substantially more fuel. It would be interesting to sniff it with a four gas analyzer, play with the resister and watch CO climb. I wonder if a fixed resister of say about 150 ohms would be a good compromise adding enough fuel to aid running without completely wiping out economy. Keeping the dial in place does aid in drivability and starting with the option of dialing down to regain some economy. A little too much fiddling for me though. Fuel injection is supposed to be accurate all by itself. Starts to make the carb option look attractive. For your next trick add some sort of ignition timing adjusting mechanism to the other bar, and a dial type fuel regulator (with gauge) You'll be akin to Igor making adjustments on the fly in the lab while readying for the big experiment. Hmm, maybe a discovery has been made.
gstallons Posted February 27, 2009 Posted February 27, 2009 People can f**k with things and fool themselves into thinking it is better.The way to fix this driveability and performance issue properly is with a PC or similar device, a gas analyzer, a dyno and most importantly a good mechanic with the knowledge to use these tools to get the most out of the bike.
Guest ratchethack Posted February 27, 2009 Posted February 27, 2009 That is a bit of time spent there, thanks.Interesting stuff. Thanks for the feedback, Dan. Your work proves what an archaic lump the V11 is. Seems it can't be made to run properly at anything near stoich. Fooling the ecu into thinking it is anywhere from 10C to 30C cooler than actual is likely providing substantially more fuel. It would be interesting to sniff it with a four gas analyzer, play with the resister and watch CO climb. I wonder if a fixed resister of say about 150 ohms would be a good compromise adding enough fuel to aid running without completely wiping out economy. A crude fix for a crudely "misfit" temp sensor, I figure. :Keeping the dial in place does aid in drivability and starting with the option of dialing down to regain some economy. A little too much fiddling for me though. Yep. Not ever gonna be something everybody's just gotta have. Good thing I'm not selling anything here, eh? Fuel injection is supposed to be accurate all by itself. Starts to make the carb option look attractive. Indeed. Brings to mind the taper needle adjustment for the Amal Concentric on the left-hand handlebar of my old '68 BSA A65 Thunderbolt. There's no equivalent for the tickler button on the float bowl, though. Thankfully, Guzzi seems to've missed the mark by a pretty wide margin with the sensor/holder, and yet still wound up with something that can be made to operate efficiently enough by many alternative routes. Hmm, maybe a discovery has been made. T'were on my end -- several, at that.
Skeeve Posted February 27, 2009 Posted February 27, 2009 For your next trick add some sort of ignition timing adjusting mechanism to the other bar, and a dial type fuel regulator (with gauge) Forward, into the past! It'll be just like turning the clock back 75 years or so and having one twistgrip be the throttle and the other being the spark advance! Boo-ya! Heck, if we're really smart, we can get it right & swap the throttle to the left hand grip like the old Indians had it, to leave our right hand free for shooting [which was a BIG point the Indian reps made when pursuing police fleet sales... ] Wheee! Ride on!
Dan M Posted February 27, 2009 Posted February 27, 2009 Temperature Sensor Output This data table on the NTC thermistor was shamelessly lifted from Cliff’s link here: http://www.cajinnovations.com/MyECU/temperature.htm Temperature °C vs. Resistance ohms As noted in prior posts, this is the only relevant data to consider when adding a variable resistor to enrich A/F at operating temps. It maps the output of the thermistor under the only temp sensor feed to the ECU that is read at operating temperatures after startup. It’s the same data in Cliff’s previously posted TempR line. At 125°C, dialing up an additional 1K Ω “fools” the ECU into believing the motor is running at ~50°C, richening the A/F considerably. At 125°C, and dialing up an additional 100 Ω, the ECU is “fooled” into believing the motor is running at ~100C, richening A/F relatively slightly. -40 100950 -30 53100 -20 29120 -10 16600 0 9750 10 5970 20 3750 25 3000 30 2420 40 1600 50 1080 60 750 70 525 80 380 90 275 100 205 110 155 125 100 Just another quick thought here and I'll shut up. Looking at the spec'd sensor's values. It is ranged to be most sensitive in the 90C to 110C area. Interestingly, the exact operating temperature range of a modern liquid cooled motor. (thermostats typically open at 90 and cooling fans come on at about 110) The air cooled lump varies widely on either side of this range. Seems to lend some credit to someone's earlier theories on trying to mimic a coolant reading.
Recommended Posts