| Author | Message | ||
     Finedaddy1 |
I applaud your efforts. Well done.  | ||
| Timebandit |
if you still have the FH012AA reg, I wouldn't hesitate to unplug the cables to swap out regulators and take the bike for one ride with the shunt reg; I'm sure that the shunt regulator won't kill your DIY stator in an hour, and the delta-T between series regulators and shunt regulators is the real number that we're after. But if you don't want to go there, that's understandable. the problem is that until someone performs a controlled experiment, we don't really know anything useful. (Message edited by timebandit on April 11, 2012) | ||
| Timebandit |
BTW, where are you measuring Tbatt? 285*F is just insanely hot ... boiling. | ||
| Timebandit |
"A graph comparing shunt v.s. series regulators would be the ultimate summary statement for us non-electrical engineer types." If Greg isn't interested in doing a regulator swap, one could obtain equivalent data by testing the series regulator under differing load conditions. (Not trying to pile more work on Greg.) What we're seeing right now is that with the series regulator, Tstator plateaus at about 195*F above ambient air with the standard load on the bike, and doesn't change all that much during the ride. It would be interesting to know how much load / current demand the "standard load" turns out to be. Another way to plot his data would be to normalize Tstator referenced to the ambient air reading and plot that: Tdelta = Tstator - Tambient Greg, do you have an excel data file that you could post? If so, I'll do the plot. Going a step further, if someone had a reading of the vreg's current output under the "standard load" conditions, that ordered pair (I, Tdelta) would give us one data point on the plot that Danny's asked about. To draw the whole plot, you'd have to repeat the test under varying load conditions and plot each plateau Tdelta value as f(I). The obvious question that comes to mind is how would the results look if someone with a series vreg were able to repeat the experiment, measuring Tdelta while varying Iregulator, to plot the results while increasing electrical loads on the bike. That would result in the desired plot of delta-T as f(I) without ever having to change the regulator. Obviously, the plot's endpoint at the full rated output would yield an equivalent estimation of the shunt regulation case, even if the maximal current number needed to be extrapolated. No shunt regulator required. | ||
| Reepicheep |
That's VBatt Timebandit, look on the scale on the right. (took me a bit to figure it out also) | ||
| Timebandit |
thanks. | ||
| Hildstrom |
Finedaddy1: Thanks. Timebandit: "we don't really know anything useful"... Really? Nice. I thought knowing that my setup will survive those ambient air temperatures was very useful. "Greg, do you have an excel data file that you could post?" There is a link to the data file on my page as well as the gnuplot script I used to plot it. I'm not sure I'd assume a linear relationship between ambient air temperature and stator temperature. The deltas would probably increase at higher air temperatures. I'm not going to switch regulators or add heated accessories. I am willing to repeat my hot idling test while I'm monitoring the values. I have a clamp-on current meter that I can use to manually log some corresponding current data. After the stator reaches a steady state temperature, I think these measurements would be easy to do: * two fans, low beams * two fans, high beams * two fans, no beams At idle rpm with that much current draw, the series regulator should draw the same current as a shunt regulator because there would be no excess voltage to switch off or shunt. Then you could attempt to quantify the worst-case temperature rise as a function of current and I can stop the test if it approaches damage levels. | ||
| Timebandit |
"we don't really know anything useful"... Really? Nice." i'm not trying to belittle your effort at all! but the first set of data says what it says, and it doesn't say anything more. so we have to be cautious about what conclusions we try to draw from it. i think that the purpose of your posts hasn't been with the intent to provide a set of data for drawing conclusions. rather, i think that it was provided to demonstrate that you've built a fantastic tool for data collection. please correct me if i'm wrong on that. what we're looking at in this data set is a very interesting observation, the output of a very useful tool, and some very helpful plotting scripts. you've built a fantastic tool that could be used to provide the answers we've all been looking for for a long time. but as much as we are all impressed by the cool-factor, we have to keep our feet on the ground -- the tool looks GREAT, but this set of data doesn't tell us all that much about the stator problem. it is a single data set that was collected without the experimental control needed to isolate the effects of any one variable. in that respect, it's incredibly useful at demonstrating the tool, but the truth is that first set of data doesn't tell us all that much. again, i am not belittling your work on the tool, ... i am only trying to contribute by providing a rational interpretation of the data you've provided. what i see is a fantastic tool that could be used to collect very meaningful data in the future. for that, you deserve quite a bit of praise. getting back to practical application: the big question has always been: does buying a series regulator help us? if so, how much? we all think that it helps us, based on theory, but we still really don't know the answer any more than before you did all this work, because we don't have a reference temperature yet. to determine whether upgrading the rotor or vreg is going to be helpful, we need to know something about the stock bike's setup -- how hot the stator gets before someone drills on their rotor, or spends a lot of time and money to buy a shunt regulator and install it. it would be far more useful to know how hot the stator gets in the stock configuration, and then compare that number to how hot the stator gets after doing the things that you've done. you've built a great tool, but the one set of data you've collected so far isn't enough to tell us what we really want to know -- how much temperature reduction you've gained vs. the stock setup by doing what you've done with the rotor and stator. this is sort of the problem with taking these kinds of numbers -- as you know, there's a pretty big hurdle to be overcome in getting to the data -- to get any truly useful information, you have to disassemble the bike at least TWO times to compare two different sets of conditions: 1. install data probes, reassemble, gather data for stock setup; 2. make changes to the regulator, reassemble, gather data for modified setup. to do the regulator vs. oiling rotor test that I was talking about, i'd have to disassemble the bike FIVE times: 1. install data probes, reassemble, gather baseline data for stock rotor & stock vreg setup; 2. switch regulators, reassemble, gather experimental data set for stock rotor & substitute vreg; 3. switch rotors, reassemble, gather data for substitute rotor & substitute vreg; 4. switch vreg, reassemble, gather data for substitute rotor & stock vreg; 5. switch things to the way i want to keep them. obviously, there's a reason I haven't published the data yet -- collecting the data is going to take a lot of work. then there;s that inconvenient truth that if you want to isolate the effects of rotor cooling vs. a change in regulation paradigms, you have to design a valid experiment to isolate the effects of those variables. nobody ever said it would be easy to create an experimental design that will provide for the isolation of variables and the drawing of conclusive answers. unfortunately, if that isn't done with scientific rigor, it places serious limitations on how useful any set of casually obtained data is going to be. what we have right now with your data collection tool is a great foundation for where to go in the future. you've done some great work in designing a data capture system. you've collected one set of very interesting data that's piqued everyone's interest. but to keep it in perspective, we're looking at one very interesting data point that doesn't tell us much because you need two data points to draw a line. to provide comparisons, you need several sets of data points to draw different lines for each condition, and then compare and isolate the variables. you've done great work. now there's useful tool that could be used to answer questions that people have been speculating about on this forum for years. unfortunately, we still don't have the answers, but now we've got a tool that could be used to obtain them if it's used properly. thanks for all of your hard work. | ||
| Hildstrom |
Timebandit: You are interested in deltaT of various modifications. I am interested in maxT of my modifications. It's that simple. The fact that I am not performing an ideal experiment to obtain the numbers you want does not mean conclusions cannot be drawn or the data are not useful. I'm after two bits of information, nothing more: My specific rewind/rotor/regulator combo: 1) is sufficient 2) is not sufficient My regulator placement: 1) is sufficient 2) is not sufficient "i think that the purpose of your posts hasn't been with the intent to provide a set of data for drawing conclusions. rather, i think that it was provided to demonstrate that you've built a fantastic tool for data collection. please correct me if i'm wrong on that." Yes, you are wrong on that. The paragraph immediately following logdata01.gif on my page, currently the last paragraph, should have explained the information I seek, the need to log at higher ambient temperatures, and the fact that I did not build a data logger just to demonstrate that I built one. | ||
| Dannybuell |
Hildstrom ~ I enjoy your posts and appreciate your thoughtful interpretations for us non EE types. | ||
| Timebandit |
"I'm after two bits of information, nothing more: My specific rewind/rotor/regulator combo: 1) is sufficient 2) is not sufficient My regulator placement: 1) is sufficient 2) is not sufficient " My mistake. I thought you were interested in the same questions that everyone else in the 1125 community is interested in regarding the stator life problem. I didn't realize that you were only interested in "two bits of information, nothing more." Knowing that, I'd still like to compliment you on developing what looks like a very useful tool for anyone who is interested in studying the bigger picture. Maybe you've paved the road for other people to investigate those other questions that you're not interested in. Since you're not interested in anything beyond your own two bits, I'll stop talking about my interest in the bigger picture and confine my comments in your thread to helping you assess your upgrade. 1. Thermal Stability: you already know that Kapton polyimide wire is rated at 450*F and that you've selected the right wire for the application. To go beyond 450*F you don't really have any other options in stator wire. We also know that at 450*F other systems in the bike are subject to catastrophic failures, which renders the stator wire temperature question moot. In the real world, we know that the bike will never see 450*F, which suggests that measuring Tmax for your setup is an interesting, if unnecessary quest. In reality, if you wind with polyimide wire, measuring Tmax has no purpose other than an academic one -- if you approach or exceed the 450*F limit of the wire, there's nothing that you can do about it. 2. Stator Output: Looking at DC voltages tells us virtually nothing about stator output, because the voltage regulator takes a wide range of AC stator voltages from about 20-120 VAC and makes them all look like 13-14 VDC at the battery. Confirming that you've got 13-14 VDC at the battery does tell you two things: a. that the stator's AC output is greater than the minimum AC output required to establish 13-14VDC at the regulator output. In other words, stator output is indeterminate in the range of 20-120 VAC. that;s a pretty wide range. b. the voltage regulator works if given adequate AC input. To understand stator performance, and to make sure that you don't have some dead coils, it would be helpful to look at AC voltages at the stator. That will tell you important things that your current measurement paradigm is not looking at. That's why measuring AC voltages is the basis for the H-D stator troubleshooting tests in the electrical diagnostics manual. We also know that if you duplicate the OEM wire gauge, the winding pattern, and the number of turns, then induction motor math tells us that the rated output of the stator cannot be different. Even if some variable, such as the number of turns, wire gauge, etc. were subtly different, someone versed in induction motor theory understands that the rated output of the stator/rotor charging system will not change because it is defined by the rotor and not by the stator; Here's why: Faraday's Law tells us that the AC output voltage of the stator is proportional to the rate of change of magnetic flux, i.e. rotor RPM. Ampere's law tells us that the stator's B field has an upper limit that is defined by the permanent magnet rotor, which is fixed. The stator core material has a magnetic permeability and cross sectional area that by design can handle the rotor. Because these variables were not changed, the output current of the stator remains limited by the B field of the permanent magnet in the rotor, not by your stator windings. In other words, the ROTOR will limit the stator's plateau output. This makes taking stator measurements an interesting but unnecessary exercise. 3. Stator testing: If you duplicate the OEM winding pattern, turns ratio, and wire gauge, all that you need to do to test the stator is to accurately measure inductance (or resistance) in each coil, and confirm non-continuity with ground. In other words, you just have to verify that your windings are good and not defective. That's a pretty simple measurement that obviates the need for an experiment if you have a either an inductance meter or a 4-wire kelvin ohmmeter; obtaining the answer could be done without going to the trouble of building the data acquisition equipment. As I recall you said that you don't have precision meters, which is probably why you decided to build the recorder. 4. Regulator placement: A simple hand-held IR thermometer could give you a pretty good answer for Treg. Of course, that approach totally lacks the cool-factor provided by in-flight recording. 5. Epoxy: The biggest question that lingers in my mind is the choice of epoxy for the stator windings. I'm interested in hearing how well the cheap epoxy fares in the long term vs. the expensive special-purpose epoxy that's designed for the specific application. Will you be monitoring oil samples? Thinking the problem through, I think that you could have gotten all of the answers that you needed to assess the success of your rewind without going to the trouble to build the in-flight recorder, but I understand why you did it. And I'm really glad that you did it. You did a good job, and I'd like to thank you again for telling us how you did it. Maybe some other people will use your in-flight recorder design to answer more of those questions we're interested in. | ||
| Hildstrom |
Timebandit: Word count != results. Please show us the progress and results of your experimentation in a new thread. "I might have been able to hook you up with some embedded temperature sensors that you could have placed inside of the windings on a couple of key poles, which would have allowed you to take direct measurements of stator temperatures at a couple of key poles, like 1 (at 6:00) and 3e (2:00). That would have been a great opportunity to get direct stator core temps." "unnecessary quest", "no purpose" Well, which is it? "I'd be interested in knowing how hot the unit gets in real world driving." "A simple hand-held IR thermometer could give you a pretty good answer for Treg. Of course, that approach totally lacks the cool-factor provided by in-flight recording." Well, which is it? | ||
| Timebandit |
wow, you respond to analytical, constructive criticism with resentment. that's not a good recipe for open collaboration. there is no need to be defensive. i've repeatedly given you compliments for your work, and I've politely provided information to help people who are not specialists in this field to understand what is going on. armed with that sort of knowledge, it's possible to eliminate a bunch of unnecessary and time-consuming steps. if you're willing to accept constructive criticism, i can help you. i can't be of any help if you are unable to accept valid suggestions and well-intended critique without becoming defensive and attempting to turn this into some sort of confrontation. if you prefer to morph this conversation into something that involves confrontation, then the collaborative effort will have to end. there has been no inconsistency; there are simple facts that remain as true now as they were before: the OEM stator in the bike's stock configuration shows asymmetric thermal damage. (a large collection of photos that establish this observation are posted in another thread.) that thermal damage is thought to be a function of asymmetric oil distribution and asymmetric coil cooling in the stator. at the risk of stating the obvious, if anyone is interested in prolonging stator life, one can look at the asymmetric thermal damage that occurs to the OEM stator in it's native state and use it as a guide. by measuring the temperature difference between the area of highest thermal stress and the area of lowest thermal stress, one can quantify how much temperature reduction (cooling) is necessary to prevent the observed signs of augmented heat-induced damage. that is useful information. at the risk of stating something even more obvious, you need TWO temperature measurements to make a comparison. TWO temperature measurement conditions provide a quantifiable answer. ONE temperature measurement condition does not. your experiment is hampered by the fact that you only collected temperature readings under one experimental condition. the data provided by your experiment would be far more helpful if you obtained a reference temperature in the pre-modification state. having two numbers would demonstrate the thermal improvement that's been provided by the regulator swap. not only would that be a very useful bit of data, it's a pretty easy measurement to take, as you have all of the pieces to the puzzle sitting on your workbench. All that you need to do is to collect the data. to get that answer -- or any useful answer about stator life -- you need to measure temperature in TWO different experimental conditions -- the temperature condition that causes failures, and the temperature condition that prevents failure. unless you know the temperature of the stator that causes failure modes, knowing the temperature in some other mode provides you no basis for reference, rendering the measurement indeterminate in predictive value. | ||
| Timebandit |
at the risk of being redundant -- i compliment you on the work you've done so far, and i encourage your ongoing efforts. | ||
| Hildstrom |
Maybe my last post was unwarranted, but some of the things in your posts just rubbed me the wrong way. I apologize if I misinterpreted your writing. Thanks for the kind words. I don't need two experimental conditions because I know the approximate failure temperature of my rewound stator. I agree that I would need the baseline temperature if my stator's insulation and epoxy were unknown. Even if I wanted to, I can't measure a good baseline because I drilled air cooling holes in my rotor. (Message edited by hildstrom on April 13, 2012) | ||
| Timebandit |
thanks for understanding. i feel like we're on better ground now. Looking at your page, it isn't exactly clear to the casual reader just what the ProtoShield is and what it does. In the text, you referred to it without first explaining what the ProtoShield is. Doing a little research I was able to find the answer, but if other people are confused about this (I think they will be if they dig into the actual build) then it might be worth updating the page to explain the interface board in a little more depth. I commend your use of open source tools like gnuplot on a project like this one. It's possible that if you're willing to fully document the build of the sampling unit, then other people who are interested in participating in this sort of project might be able to follow in your footsteps. In regard to our stator temp reduction project, because the test conditions are so labor intensive (requiring several assemblies/disassemblies) we had planned to have different people perform different legs of the experiment. I had some scheduling conflicts that got in my way of being able to work on my leg of the project right away, so I asked Tim to forward the measurement gear to another person who is participating in the project. That person hasn't done anything yet, and now that I'm ready to work on my part of the project, I don't have the gear. The parts seem to have fallen into a black hole. Our plan involved using commercial data acquisition equipment that was on-loan, and shipping it back and forth. If sharing it turns out to be a huge logistical headache, then perhaps it might be worthwhile to build some DIY data acquisition equipment as a backup. Your hardware seems like a good alternative. | ||
| Squish |
Very technical discussion here. Good information. I haven’t researched how exactly a series regulator works, but I understand the basic principle that the series regulator will operate to produce only the power needed to maintain a target voltage to meet the power demands of the bike, and thus if there if there is lower power demand from the bike’s electrical system, the series regulator will not demand full current from the stator like the shunt regulator. I probably should not ask this until I have researched in more detail how a series regulator works, but, anyhow, will the series regulator work with the harness mod? I ask since I guess the 2009 1125 with series regulator and without harness mod could end up in the condition were full power is being demanded by the bike’s electrical system, and the engine RPM is low and in the zone where there in insufficient oiling of the stator, and the 2009 1125 (with series regulator and without harness mod) is right back at root cause issue – max power demand / max current from stator, low RPM, and resultant high stator temps which lead to stator failure. Comments? | ||
| Posplayr |
at the risk of stating something even more obvious, you need TWO temperature measurements to make a comparison. TWO temperature measurement conditions provide a quantifiable answer. ONE temperature measurement condition does not. Hi temp powdercoated epoxy seems to burn at somewhere above 450 -500 degF. 400+ degF is used just to cure the powdercoat. Hildstrom's stator measurements show something like 270-280 degF so for most intents and purposes the comparison of know characteristics to a set of measured data provides all the information that is necessary to decide on the effectiveness of Hildstrom's mods. 280 degF will provide significantly improved longevity for the stator. } Having performed detailed measurements comparing a MOSFET R/R to a SERIES R/R , I would also have little motivation to measure a baseline which is know to burn stators and is therefore known to be defective (i.e. severely limited longevity)Confirming that the temp can be lowered to less than 300 degF is proof enough for me}}. I commend Hildstrom in his efforts and the detailed results he has provided. | ||
| Timebandit |
Posplayr, are you the guy who wrote this: http://www.posplayr.100megsfree3.com/FH012AA_Charg ing/RR_Tutorial.pdf | ||
| Reepicheep |
I believe I will have one of these inbound at the end of April... and it's in the back of my mind to see if it will work as a poor man's data logger. http://www.sparkfun.com/products/10244 I think it will out of the box, and if not, it should be easy to hack to make it do so. My thought is that for $100, its a fantastic little diagnostic tool no matter what. A common little easy to use scope that will work fine as a DVM as well for just about any problem on any vehicle. So people might want to buy them just to have them as a useful tool, even after they gather a few test runs for this particular problem. $100 aint cheap, but just shipping around the good stuff could be 1/3rd of that cost easy. And $100 for a good DVM isn't unreasonable, and its stupid cheap for a portable scope. (Real EE types dismiss these things because they don't have great bandwidth... that's true, but non EE types don't *need* great bandwith for a huge percentage of the problems they need to solve... this should be a great tool even if it only samples up to 1000 Hz). | ||
| Zac4mac |
Bill - that looks pretty cool. I might have to get one. I blew up my old Fluke 77-III futzing with a dead microwave... Zack | ||
| Posplayr |
TimeBandit, yes I was. The website I used to host all my public files has deleted them and I have not tried to locate or relocate all of the files. I had a lot of work there. That DSO looks like it would be well worth the cost. | ||
| Reepicheep |
You can get them for $79 as well on ebay and other places, but I think the SparkFun one has a nicer metal case. I'll post a review when I get mine. Only having a couple inputs would be a bit of a limitation, but I suppose we could either just run multiple tests under the same conditions and loop (probably good enough) or pick some universal reference (rear head temp or something). I think that DSO is also limited in how high a voltage it can take, so you probably still want to replace the basic fluke with another fluke... which you can hook to anything anywhere. It's the perfect vehicle mechanics tool though... | ||
| Hildstrom |
I noticed some oily stuff coming out of the thermocouple wire where it penetrates the case. Two small zip ties tightened up the silicone tape enough to stop that; there was a big air gap in there. I should have put a dab of black silicone there when I assembled it too. Also, don't reuse the paper gasket; mine seems to be seeping around the bottom and around the bottom bolts. Squish: Thanks. Yes you can run the harness with a series regulator, but I am not. A series regulator works like the harness, but it can open-circuit (no current) all three phases, at all rpm, with no relay or moving parts. However, the harness is controlled by the ECU to do its thing during very specific conditions identified by the manufacturer. Timebandit: Thanks. The stuff on my page should be enough for someone to duplicate my effort. I added a part list. The code comments explain the pin connections. The only other required thing is the MAX6675 datasheet so you know how to attach it. I do not have a nice looking schematic yet though. Posplayer: Thanks; and thanks for doing that work before. Repicheep: The DSO Nano V2 should be cool. I almost bought one before I decided to buy my Rigol scope. The software is open-source and you may even be able to open it up to gain additional I/O. I'm not sure if it can function as a logger out of the box though. | ||
| Timebandit |
regarding the seal at the egress of the wires from the case -- one option that you've mentioned is adding a dab of RTV sealant. another idea that came from Tim Reeves was to poke a 14-gauge hollow needle through the grommet, pass the wires up through the needle, and remove the needle. this allows the rubber grommet to collapse around the wires forming a seal. i didn't have problems with gasket leaks. the guys at EBR tell me that they reuse the gaskets as long as they aren't damaged, so maybe you just need to check to assure the mating surfaces are clean and double check your fastener torques after riding. regarding the DSO -- bandwidth shouldn't be a problem for most bike applications, but the voltage input rating of 80V could be a problem. if you wanted to observe AC stator voltages during switching (ie: looking at back-EMF), you could easily overshoot the device's input rating under medium to high RPM conditions. i'd build a removable voltage divider into the connector cable. (Message edited by timebandit on April 17, 2012) | ||
| Hildstrom |
If anyone wants to go a similar route, here is the pseudo schematic of the logger I built. There are plenty of I/Os left for additional thermistors or thermocouples. The SPI lines to the MAX6675 could be uncrossed with a software change, but I am going to leave it as-is for compatibility with the one I built.  (Message edited by hildstrom on April 19, 2012) (Message edited by hildstrom on April 19, 2012) | ||
| Reepicheep |
Nice. That MAX6675 chip saves you from quite a few headaches. | ||
| Squish |
I googled and found this PDF on shunt versus series R/R - covers stator current and temps (link provided below). Great information. Have I seen this same article posted here on BWB? (I think I remember seeing this PDF file from a link on this message board). Anyway, wow, the shunt regulator is a poor design – if the article is correct, the shunt regulator simply shunts all current back through the stator windings when regulating, current increases to the point of saturation. There isn’t a lot of detail in the internal circuitry of the regulators, but the PDF communicates what 1125 owners need to know. http://www.google.com/url?sa=t&rct=j&q=shunt%20vs. %20series%20r%2Fr%20gs&source=web&cd=1&ved=0CCQQFj AA&url=http%3A%2F%2Fwww.vfrdiscussion.com%2Fforum% 2Findex.php%3Fapp%3Dcore%26module%3Dattach%26secti on%3Dattach%26attach_id%3D8336&ei=LQqRT_PCAYfhiAKp l622Aw&usg=AFQjCNEpK02iPjOUxomKnYlo6g22gKVvaQ | ||
| Reepicheep |
I don't know if "poor design" is the right choice of words... it's an economical and effective solution to a problem in most cases. In the case of the 1125, the jury is still out about how much difference it makes. At the point where the stator needs the thermal margin, I'm not sure a shunt regulator is doing much different than a series regulator. And a switching regulator could make the thermal problem *worse* at some RPM's (though in fairness, probably not the ones the 1125's suffer at). | ||
| Squish |
I was going to say “lame design”, but I tried to contain myself. With economical being the keyword, I should have said “economical design”. The problem with this design is nothing new. The notorious Honda turbo series of the early 80’s had a classic stator failure problem – the ’82 CX500 Turbo and the ’83 CX650 Turbo. Engines would run on the hot side, being a turbocharged motor. And it was well know that riding the machine with more than 14,500 miles on the stock stator was riding on borrowed time. These bikes also had shunt regulators. This was 3 decades ago. Interesting to note that the Buell 1125 stators are failing even earlier than the old Honda turbo bikes. (Message edited by squish on April 20, 2012) (Message edited by squish on April 20, 2012) |
