| Author |
Message |
    
Hildstrom
| | Posted on Thursday, June 21, 2012 - 07:50 pm: | 
|
I connected channel 1 of my Rigol DS1052E across two stator leads to see what happens to the stator lead voltage at idle and when the CE-605 SB series regulator is actively regulating at higher rpm. No dual-probe differential input is necessary since the motorcycle was totally isolated from ground. The corresponding data files are on my page. Here are two shots at idle, 200F coolant temperature, both fans running, and 12.5V battery voltage:
Here is a shot at 4000 rpm, 185F coolant temperature, both fans running, and 14V battery voltage:
You can clearly see where the CE-605 SB is open-circuiting the stator leads, which causes the voltage to jump up quite a bit above battery charging voltage. |
Posplayr
| | Posted on Thursday, June 21, 2012 - 08:28 pm: |
|
Hildstrom,
Is that a Wye- wound or Delta wound stator?
Bubba,
I dare say you will not find any motorcycle (equipped with a PM 3 phase generator) manual that will not specify measuring open circuit leg to leg stator voltages using an AC voltmeter.
The voltage danger of measuring a SERIES r/r is no worse than what is in every manual already.
As usual you are trying to blowing voluminous amounts of smoke up everybody's @$$.
Pos |
Nightsky
| | Posted on Thursday, June 21, 2012 - 09:00 pm: |
|
Hildstrom's stator scope shots have rounded tops and sloping zero crossings like ours. |
Timebandit
| | Posted on Thursday, June 21, 2012 - 09:38 pm: |
|
quote:I dare say you will not find any motorcycle (equipped with a PM 3 phase generator) manual that will not specify measuring open circuit leg to leg stator voltages using an AC voltmeter.
You have a flair for making up false and distracting statements to confuse people, and pointing out things that are irrelevant. I never said anything about how to measure voltages with an AC voltmeter.
What I said was that your oscilloscope instrumentation methods SUCK. The result of your poor analytical methods is that your data is noisy, frequency response is poor, and your images are so crappy that you can't recognize regulator activity where it is obviously taking place.
Several pages back, I posted a series of photos showing how the regulator was active during the charging period. In response to this, you claimed that I was an incompetent idiot that could not understand something as simple as sticking a screwdriver across a battery. You ridiculed me as if I were a total moron.
In response, three different people have used oscilloscopes to post images that confirm my assertions and prove that you were wrong all along.
Now who's the one blowing smoke up peoples' asses? I made a prediction that there was regulator activity during the peaks, when you said there wasn't. Experimental data obtained by three different people confirm my assertion that in real world data the tops are rounded, the zero crossings are sloped, and there is high frequency regulator activity/switching during the battery charging portion of the trace. This is exactly the opposite of what you claim should happen -- you claim that the tops and baselines are always flat and that no regulation takes place during the peaks. The data doesn't lie.
By now it should be obvious to everyone why you continue to refuse to answer Nightsky's questions about why all of the real-world scope tracings look the same, and why your none of your data looks anything like the real-world data.
(Message edited by timebandit on June 21, 2012) |
Timebandit
| | Posted on Thursday, June 21, 2012 - 09:39 pm: |
|
Sparky, I would be interested in doing an unbiased regulator shootout if I could get a bunch of samples. |
Nightsky
| | Posted on Thursday, June 21, 2012 - 11:03 pm: |
|
Hildstrom,
Thanks for the scope shots.
12.5 is the engine off no load battery voltage. I get ~13.6 while idling the Buell 2009 with Ducati. This tells me at 1300 RPM, you're not charging.
It's a bit dangerous to ground the scope to one stator leg. With what we are now calling a "series" regulator, the vreg open circuits the stator from the load on at least one leg. There can be lethal voltages between scope ground and bike ground exposed at higher RPM.
(Message edited by Nightsky on June 21, 2012) |
Hildstrom
| | Posted on Thursday, June 21, 2012 - 11:05 pm: |
|
Posplayr: The Buell 1125 stators are delta. I could clearly recognize where regulator activity was taking place in your traces.
Timebandit: "and there is high frequency regulator activity/switching during the battery charging portion of the trace." I sill disagree with you on that point. If you look at the second image I posted, which is the zoomed-in idle shot, the tiny spike at the beginning of that half-cycle is the series regulator activity/switching from open to closed. The noise during the rest of that charging half-cycle is probably due to ECM, fan PWM, etc and I suggest that rectification is the only thing occurring after closing the switch. The 4000 rpm shot clearly shows that series regulator activity/switching is synchronous with the stator output half cycles; it occurs at the same frequency as the stator AC voltage, not at an order of magnitude higher frequency.
(Message edited by hildstrom on June 21, 2012) |
Hildstrom
| | Posted on Thursday, June 21, 2012 - 11:18 pm: |
|
Nightsky: No problem. I get 13-14V at idle too, but not forever. When you get ~13.6, what is your coolant temperature and how long have your fans been running at full blast? |
Timebandit
| | Posted on Thursday, June 21, 2012 - 11:22 pm: |
|
Greg, your data looks pretty clean. I'm thinking that some of that jitter on the traces represents common-mode noise could go away if differential sampling were used.
The risk of grounding the scope to a stator leg is that the scope's chassis now floats up and down at the full voltage of the stator. You definitely don't want the scope chassis to have an electrical potential that floats up and down with the oscillating AC voltage of one of the stator legs. At 4000 RPM the scope is going to have 50 VAC on it when the regulator opens. Things could get ugly if you reached to touch a knob on the scope. If the knob is not electrically isolated from the chassis, that pot will have 50 VAC on it's case -- touch the chassis and your finger has just been stuck into a light socket. Better hope that you're wearing a good pair of rubber soled shoes so that you don't provide a low resistance path to ground through your body; better hope and that you're not leaning on the bike frame, motor, or handlebar with the other hand -- you'll get the full voltage/current output of the stator across your chest. When the regulator switches to open circuit the voltages become lethal.
Bzzzzzzzzzz!
Any doubts? Hook up a voltmeter between your scope ground and the bike's chassis and watch what happens when the reg goes open circuit.
Bzzzzzzzzzz! |
Timebandit
| | Posted on Friday, June 22, 2012 - 01:31 am: |
|
"I suggest that rectification is the only thing occurring after closing the switch."
That's a pretty simplistic visualization of the problem.
The problem is that you're looking at this situation while you're wearing single-phase blinders. You see a single-phase time domain tracing, and you're stuck thinking in a single-phase mindset. You only see the trace that's on the scope in front of you. You fail to remember that there are two identical traces taking place that are offset +/- 120 degrees in time, and you can't see them because your instrumentation doesn't have enough channels to show them all concurrently. You need to train your mind to think in three phases while you look at one phase, otherwise you'll never be able to understand why your conceptualization is wrong.
It's important to remember that 3-phase power is a constant power source. That means that at all times, the battery is being driven by one of the three phases. When the single-phase trace that you're looking at begins charging the battery, the other phase that's 120-degrees ahead of it has just finished charging the battery and has begun the 0-volt period. Then, when the single phase trace that you're focusing all of your attention on begins the 0-volt period, you need to realize that the phase that's lagging 120-degrees behind it is beginning the battery charging period. At all times, one of the three phases is going to be in the battery charge interval. That means that at all times, there is regulator activity. 2/3 of the time it's occurring in the phase that you're not looking at and you just fail to think about that!
No, the HF ripples are definitely not PWM fan impulses. They occur at all times, regardless of coolant temperature, regardless of how long the motor has been running. They're visible right after startup when the coolant is cold and the fans are still off. They're also not fuel injection pulses, because they not synchronized to fueling. It seems as if you haven't had the time to think your assertions all the way through, and reason your way through why they can't be right.
Think about what types of regulator activity those HF pulses represent, and why some type of activity has to be taking place during the battery charge interval. How does the regulator determine the charge state of the battery? The answer to what that HF activity represents should be obvious.
To fully appreciate this, you need to train your brain to see things happening concurrently in three phases instead of just paying attention to what you see on the screen in front of you. Assessment of the battery charge state is something that has to be done at all times, because at all times, one of the three phases is actively charging the battery.
I've spent way too much time explaining concepts that seem straightforward enough that I don't think we should have ever become bogged down by them. I just can't see this topic drawing on for another 7 pages. The progress of clearing up these misconceptions is just taking too darned long. I'm losing interest. |
Hildstrom
| | Posted on Friday, June 22, 2012 - 11:54 am: |
|
Timebandit: Yes, the extremely dangerous situation I created has been mentioned and it was also mentioned before I did it. No doubts. I'm still alive and kicking; no shocks. I was as careful doing this as when I touched my voltmeter probes to my disconnected stator leads or when I wired a 240V circuit in a live panel.
"That's a pretty simplistic visualization of the problem." "think in three phases"
The argument that the noise during the battery charging portion is actually switching and rectification artifacts from the other two phases is the best yet. Your previous descriptions sounded like you were describing some sort of much higher-frequency switching of the phase being observed, which is what I was disagreeing with.
"I've spent way too much time explaining concepts that seem straightforward enough that I don't think we should have ever become bogged down by them. I just can't see this topic drawing on for another 7 pages. The progress of clearing up these misconceptions is just taking too darned long. I'm losing interest."
Oh, you must mean misconceptions like:
* the harness provides more duty cycle reduction than a series regulator
* stator current reduction at higher rpm can be overlooked
* series regulation can only be defined by the ratio of T-on to T-total; T-off should not be mentioned and it cannot be used to define series regulation
* a series regulator is actually shunt regulating between open-circuit events
* stator output voltage should look like a clipped sine wave
Yes there have been misconceptions posted by others, myself included, but you have also contributed to the misconceptions and you should try to be a bit more humble. Your posts are not always 100% correct, 100% textbook clear, or 100% objective and neither are mine. Don't get me wrong, you do have lots of valuable contributions and I appreciate them, but this forum is as much a place for discussion as it is for lecture.
(Message edited by hildstrom on June 22, 2012) |
Posplayr
| | Posted on Friday, June 22, 2012 - 11:56 am: |
|
Nightsky,
There is no danger in putting the scope clips directly across the stator leads. Of course you should not under any circumstances touch( with bare hands) a pair of stator leads when the stator is open circuited and you are running the bike. It will give you a jolt.
I have included a couple of pages out of the THS730A manual specifically showing voltage measurements with either the built in voltmeter/data logger or the scope probe. Both are very high input impedance and so no significant current will flow to the meter in either case.
I have also attached a section of a document discussing scope grounding. For most all cases method #3 which is to clip the scope probe ground clip near the point of the measurement is generally the best and is adequate in almost all but the most critical measurements.
If there is any confusion, I never suggested grounding your scope to the stator AC and simply connecting the scope probe ground clip certainly does not do that. I only suggested grounding the motorcycle frame to the power ground (NOT Neutral) as I was not sure how your scope was grounded.
You can read about scopes connections and grounds here.
http://www.keepandshare.com/doc/4184938/scope-meas urements-pdf-june-22-2012-8-06-am-912k?da=y
After analyzing the DELTA v.s WYE stator differences , I'm suspecting that there are at least two factors as to why my SOME of my plots seem to be flattop while the other three plots are not. I say SOME because there are example where this is not always true which you seem to have overlooked.
#1 Most of the plots that look flat top are on a 20V per division scale as opposed to a 5V per division scale.
I previously posted a link to this file which you may have overlooked. It shows plots on a 5V per division scale with voltages taken lead to ground. as opposed to lead to lead. There is more rounding evident in these plots.
http://www.keepandshare.com/doc/4143240/scr-mosfet -shunt-compare-pdf-june-14-2012-12-34-pm-905k?da=y
#2 When you short a Wye stator leg to leg, there are two windings in series between the short. When you short a Delta wound stator leg to leg there is only one winding between the short. With everything else being equal, you can then expect that that it will be harder to short a single winding that it would be to short two windings. Basically the equivalent resistance of the winding is twice as big compared to the equivalent resistance of the R/R short so the voltage drop of the R/R short will be approximately 1/2. This is at least a partial explanation, but regardless I suspect the difference is in the Wye v.s. Delta would stator.
I offered to send you a FH0012AA MOSFET R/R to test that out.
Pos |
Posplayr
| | Posted on Friday, June 22, 2012 - 12:14 pm: |
|
Nightsky,
There is a bantering idiot here on the board (who shall go unnamed) that continues to suggest that you have an issue with the the power plots I posted.
http://www.thegsresources.com/_forum/picture.php?a lbumid=1998&pictureid=10791
This is a plot estimating the total power loss due to heat in a WYE stator comparing a MOSFET SHUNT FH012AA regulator to a Compufire SERIES R/R.
The RMS current was measured using the previously described THS703A with current clamp using the internal measurement function to compute a realtime RMS current. I then adjusted the per winding current to compute the equivalent total RMS current in the winding (which in the case of the SERIES R/R matches the DC output current).
By squaring this current and multiply by the measured leg to leg resistance we can estimate the heat lost in the stator. Even if the exact power lost in heat is in error(due to the assumed stator resistance which surely changes for the worst with temperature), the ratio of the estimates will still be accurate (assuming stator temperatures are the same).
http://en.wikipedia.org/wiki/Copper_loss
You should not confuse this with 3 phase power measurements which require and independent measure of voltage and current to (potentially in each phase) so that the power factor is accounted for in total power delivered. I am only estimating copper losses not total power.
Pos |
Hildstrom
| | Posted on Friday, June 22, 2012 - 12:27 pm: |
|
Posplayr: If you connect a probe ground directly to a stator lead, touch one finger to that same probe ground conductor at the scope, and touch another finger to the motorcycle chassis at the same time, you could get shocked during shunt regulator open-circuit events. If your scope is mostly plastic and rubber and you are careful, I'd say the risk is low, but it is not zero. It would not be too hard to have one hand touching metal at the handle bars while manipulating the throttle and accidentally brush your other hand up against channel 2's exposed connector ground at the scope.
(Message edited by hildstrom on June 22, 2012) |
Posplayr
| | Posted on Friday, June 22, 2012 - 12:49 pm: |
|
Hildstrom,
You are correct; I would not suggest touching either stator lead with the engine on. If you look at the pictures I posted everything is clipped on and there are no hands near the connections.
You would really have to try to get shocked if one hand is on the throttle. That would required resting you hand on the frame while holding a stator wire.
Beyond controlling the throttle watching the tach I have to monitor the DSO to capture frames on the scope. Of course if you don't have a DSO then you are also having to operating a camera.
Generally you don't have enough hands to hold stator wires and do the other required operations, so clipping to the stator wires for measurements is preferable in almost all cases.
Pos |
Timebandit
| | Posted on Friday, June 22, 2012 - 04:21 pm: |
|
quote:Oh, you must mean misconceptions like:
* the harness provides more duty cycle reduction than a series regulator
* stator current reduction at higher rpm can be overlooked
* series regulation can only be defined by the ratio of T-on to T-total; T-off should not be mentioned and it cannot be used to define series regulation
* a series regulator is actually shunt regulating between open-circuit events
* stator output voltage should look like a clipped sine wave
evidently, these misconceptions still havien't been cleared up. Now I get to beat my head against the wall to explain them yet another time.
* the harness provides more duty cycle reduction than a series regulator
To date, all of the real data that has been collected shows that the harness provides a maximum duty cycle reduction to 33% and the other product produces a maximum duty cycle reduction to 50%. 33% amounts to more duty cycle reduction than 50%. That's what the data shows. We don't have any actual data that says anything different.
You've suggested a theory that over time, the advantage may be nullified by ECM activity. It's an interesting theory. But it's only a theory. We don't have the ECM code to look at which would verify or refute your theory. And you have collected no data to support this theory, so you don't have the ability to prove whether it is right or wrong. Your idea remains an interesting unproven conceptualization. Until someone actually collects data that proves your theory to be correct, it remains nothing more than speculation. To think otherwise is the misconception. I'm waiting for real world data.
* stator current reduction at higher rpm can be overlooked
I don't think I ever said it can be "overlooked," but I agree with EBR and HD that it is irrelevant. For some reason people want to believe that current continues to rise as a function of RPM. Voltage does that, but current doesn't. It only climbs from idle to 4000 RPM, where it reaches a plateau that is caused by the limitations of the B-field of the rotor. Current does not rise as a function of RPM above 4000. It stays flat. (Charts that demonstrate this have already been published on this site. I don't think it's necessary to go dig them up.)
H-D and EBR have decided that the at-risk RPM for the bike are not at high RPM (above 4000) where the bike has effective cooling and current is not increasing; the at-risk area is at low RPM (below 4000) where the bike does not have effective cooling and current is continually increasing. That is why the harness operates in the RPM band that it does. Above 4000 RPM the bike has enough cooling, and the primary risk area for the stator is below 4000 RPM. This is not a misconception.
Your own data shows that even with your CE series regulator, which reputedly "solves" the stator problem, the bike continues to be at risk at low RPM. Note that in the areas of your data that I've surrounded with red boxes, the bike is operating at low RPM (where the harness would operate), and that even after substituting your CE regulator for the harness, the bike remains at risk because the stator temperature continues to rise.
In all of the red-boxed areas RPM is low. During low RPM the stator temperature continues to climb. The accumulation of heat in the stator is only terminated by a switch to higher RPM, which is indicated by the green boxes. This is the RPM band in which the aftermarket regulators claim to offer the most protection. It is also the RPM band in which the stator current with the stock setup stops increasing because of the B-field limits of the rotor. In these areas, the temperature has stabilized to the point that it is no longer rising. Sometimes it actually decreases.
Your own data shows that the CE regulator allows stator temperature to rise continuously as long as low RPM is maintained. Where's the protection below 4000 RPM where the bike is at the greatest risk? I don't see any.
The great misconception is that the series regulator protects the bike better than the harness in the high risk areas.
POS data shows that the CF doesn't even start to series regulate until RPM reaches the band in which the Buell current has already stopped increasing, and cooling has started. It's not helping where help is needed the most.
Your data shows that the CE regulator acts similarly -- it doesn't help at the low RPM / high-risk band where temperature and current are increasing, and cooling is low.
The harness is designed to work in the areas on which thermal load on the bike is most detrimental -- low speeds that lack cooling. The oiling rotor is even better.
While we're talking about the CE setup, I'm going to side track for a minute -- I agree with Nightsky's observation that your charging system data table on your web page shows that your custom rewind stator and the CE regulator are not charging at idle. Your idle RPM measurements show 12.6 volts which is the resting battery potential, and that no charging is taking place. You asked a question about how long the OEM setup will charge at idle. The answer: Forever. I can leave the bike idling for an hour and the voltage will never drop below 13.9. If your system is "bailing out" to 12.6 at idle then your system isn't performing as well. That would suggest that you have two issues to look into: Failure to protect at low RPM and failure to charge at idle. At least the Buell system will charge the bike while it's sitting at a stoplight. Your own data says that the CE won't.
* series regulation can only be defined by the ratio of T-on to T-total;
I'll let the Wikipedia definition of duty cycle handle this one:
http://en.wikipedia.org/wiki/Duty_cycle
* a series regulator is actually shunt regulating between open-circuit events
Not an accurate quote. I never said that about "a series regulator", thereby implying all series regulators. Your paraphrase of my comment derives an entirely different meaning from what I was saying. What I said was that Pos' data showed that using a black box analysis, his data was indistinguishable from the operation of a shunt regulator that's controlled by a series pass transistor. Using a black box model in which you're not allowed to peek at the electronics inside -- which is exactly the condition that was specified for the analysis -- I reached the correct conclusion.
* stator output voltage should look like a clipped sine wave
I have to give credit to Nightsky for bringing up the idea that Pos' claims that the traces should be rectangular were totally wrong. Nightsky is the guy who came up with the assertion that the sine waves should be observable on top of the regulator traces, and that the square waves that were suggested were totally wrong.
"Clipped" is an accurate description. Maybe you're just thinking about clipping sine waves in amplitude, rather than phase angle. The clipping reference is entirely correct. That's how asynchronous regulation works -- by clipping the sine waves produced by the stator. Nightsky and I have both published photos that document this artifact if you're willing to page back to look at them.
Like I said, explaining this stuff is becoming way too time consuming. The problem is that when someone doesn't understand something that's been clearly stated, it's not worth the time to bring everyone up to speed. I can't afford to give up an hour to type a 6000 character post every time that somebody holds onto a misconception. Some people are just going to have to be left behind. |
Hildstrom
| | Posted on Friday, June 22, 2012 - 06:13 pm: |
|
Timebandit:
http://www.badweatherbikers.com/buell/messages/290 431/671285.html?1334011053
"If the engine is in that band for an extended period of time it uses the relay to open one leg of the 3 phase charging system reducing charging output and heat generated. After a period of time it will turn the leg back on returning the charging system to full output." To have duty cycle, there must be a cycle.
I never said the CE-605 SB alone solves the stator problem. Again, my air-cooling rotor plays a part and my data are not representative of a stock bike. My stator is not at risk. Even after a long ride and prolonged idling, I have yet to record a stator temperature over 300F, which is well below the 464F rating of my stator's wire insulation and the 500F rating of my epoxy. Your last red box contains a low-rpm low-speed stator temperature plateau, but I probably should have idled longer in the second red box to capture the idle plateau. I'll idle for 20 minutes next time to see if I can capture it. I never said a series regulator protects better than stock at low rpm. I did say that lower temperature while driving around at higher rpm and when starting to idle can only be beneficial compared to the harness and stock regulator.
T-on = T-total - T-off
T-total = Ton + T-off
You can define either in terms of T-off and the other variable, but you said I was wrong to point out T-off. Yet T-off was your only concern when discussing harness duty cycle.
"Notice that at 4000 RPM the device continues to act primarily as a shunt regulator, but that the device is acting to shut-off the connection to the load more often than it was at 2000 RPM." You edited your previous posts after I quoted this the first time and pointed out the shunt (short-circuit) and series (open-circuit) events. Now you say it is not an accurate quote and I paraphrased incorrectly.
Now you're claiming that Nightsky's traces, your traces, and my traces actually look like clipped sine waves. I'll agree they are not square waves and not sine waves.
I don't really think I'm clinging to misconceptions here. |
Timebandit
| | Posted on Friday, June 22, 2012 - 10:51 pm: |
|
Yawn. You are still clinging to arguments over the same old misconceptions. And I'm becoming quite bored on the subject. Right now, you're trying to find one hole that you can poke into anything I've said, just so that you can claim that you've done it.
reading curve_carver's email from EBR doesn't tell me anything that I don't already know about the harness. I reverse-engineered the original set of data, back in the days when nobody here -- you included – actually understood what the harness was doing. i was the guy who first explained that it single-phased the system when nobody here had a clue what single-phasing meant. it's not as if posting the observations of latecomers will prove at all enlightening to the people who were first to explain how the harness operates. It's also interesting to note that the email from EBR did not contradict any of my observations. in fact, it verified the observations that i had originally made.
duty cycle is what it is. my use of the term has never been incorrect. I can't see any point in posting images of Venn diagrams to explain third grade math. People either understand subsets or they don't. I won't spend any more time on that.
on the subject of post editing, I did edit those posts to move the images to photobucket without changing their content. the plan was to embed a lot more educational photos in this thread, and quite frankly I don't like having to re-size every image to meet BadWeb's arbitrary file size and image dimension restrictions, and the required compression levels to make them fit as uploads – the board is a pain in the ass that way. It's easier to just upload images to photobucket, which doesn't have such severe restrictions on images. Photobucket also makes it a lot easier to post files and to view them as a slideshow. I think that a slideshow makes understanding the series of slides easier. When the images got moved to photobucket, they kept all of the text that was embedded in them. It was duplicative of what was written in the posts. They still say the same things that they always said. The file timestamps on the images should still be the same as the time that I originally generated them, if photobucket preserves that information.
What's interesting is that in going back to those old posts, you were complaining (wrongly I might add) that “regulation doesn't take place” in the images I had referenced. The truth, as you acknowledged a couple of posts up, is that you failed to think in three phases, and responded with the wrong answer based solely upon an errant conceptualization that was restricted to single-phase activity. A couple of posts back it looked like the light bulb finally went on. But now you're regressing back to making arguments that are based upon the single-phase mindset. There is no point in arguing over these points or rehashing them. At this point you either see the big picture or you don't. I think it's pretty foolish to continue to claim that I'm wrong when the problem is that some peoples' scope of comprehension seems to be limited to what can be observed on a single-phase tracing. I don't see any point in continuing to rehash these points. They're correct. Failure to comprehend isn't my problem. I won't spend any more time on remediation in an effort to make everyone understand, so I'm moving along.
Your claim that I referenced sine waves in your data is a totally false statement on your part. I never said that your images looked like clipped sine waves. Check the facts on that. All of the relevant posts are on this page. It was Nightsky who said that the three sets of data that were produced by Him, me, and you all looked alike:
http://www.badweatherbikers.com/cgibin/discus/show .cgi?tpc=290431&post=2233884#POST2233884
quote:Hildstrom's stator scope shots have rounded tops and sloping zero crossings like ours.
Show me where I said your data looked like sine waves. Look closely. I never said it. Your claim is not accurate and your accusation is false.
As far as your stator data goes, I think it's premature to make a claim that your data shows that your stator is not at risk. Your data fails to prove that. In every case your data clearly shows that the temperature-time index for your stator was continuing to increase – indicating heat accumulation – and the heat-accumulation proceeds until you changed your riding behavior to stop the increase. In other words, changing rider behavior before failure occurs is a necessary condition of your system in order to keep the stator from being at-risk. It's important to keep that fact in perspective. I agree with your assessment that prolonged idling tests are required to capture all of the relevant observations. I don't think that your idle periods are anywhere near long enough. To keep this in perspective, you need to understand that your idling tests don't come anywhere near to being close to the hot-weather idling tests that Buell performed. They'd take the bikes to the Arizona desert, and park them inside of a small boxed-in area in the desert sun for the entire day, and let them idle in the hotbox in the desert sun for hours on end. A 20 minute data set doesn't come close to giving comparable reliability data.
I'm done quibbling over points that people fail to comprehend. The presentation was clear, and at this point people fall into one of two groups: those who understand and those who don't. I'm moving forward. If people don't understand what's been said, I'm willing to leave them behind. It's not unexpected that much of this information is going to go over most peoples' heads, and I can't see any point in arguing with people who have just enough understanding of the topic to be argumentative, but not enough understanding to see the big picture.
My closing point will be to say that anyone can read an engineering textbook that discusses voltage regulator design, assimilate what's in the lesson, and then proceed to operate under the false belief that they understand how state of the art regulators actually work. The facts are that the modern, state of the art designs that are covered by patents and are implemented in industry go well beyond the simple teaching methods that you'll find in a textbook. If anyone really wants to understand how state of the art regulators work, they aren't going to get the information from the people who have read textbooks. The only way to get the data that's required to understand state of the art regulator design is to read the most current patents, to disassemble the state of the art products to determine how they work, and to then begin to actually design and build another regulator prototype that utilizes the type of technology that you can't find in a textbook. Doing that gives you much more insight into regulator design and operation. That's what I've been doing in my spare time, and it sounds like that's what Nighstky has been doing too. |
Nightsky
| | Posted on Saturday, June 23, 2012 - 12:09 pm: |
|
Hildstrom's plots show stator temp rising at low RPM and low speed with the CE-605 SB.
At 87 minutes, high speed driving drops stator temp below where it had climbed at idle.
We don't know maximum stator temp because that requires testing the bike sitting idling for an hour, but it certainly would be higher than any point on this plot.
At time 55 minutes, Hildstrom maintains speed but apparently shifts down. Speed same, RPM up. Notice this is the inflection point where stator temp starts climbing.
The data indicate there are two opposing forces: air flow from speed cooling the bike and stator, RPM generating heat.
Above 70MPH, airflow cooling overcomes RPM heat generation.
Idling at 0MPH, there is no airflow. RPM heat generation is less too, but something is bigger than nothing, and stator/bike temp rises with the bike sitting still.
We don't know the crossover point, but the indication is stop and go urban traffic heats the stator.
Country riding limits stator temp.
Even with a "series" regulator. |
Nightsky
| | Posted on Saturday, June 23, 2012 - 12:19 pm: |
|
Electrocution hazard exists when you ground your scope to a stator leg. Touch the metal scope frame with one hand, any metal on the bike with the other, and if the relay happens to be open, you may go into cardiac arrest as electricity surges across your chest.
Bzzzzzzt, thump.
Enough said on the subject. |
Nightsky
| | Posted on Saturday, June 23, 2012 - 12:28 pm: |
|
"I offered to send you a FH0012AA MOSFET R/R to test that out."
I'll take you up on that.
Electrically, you should not be able to differentiate between wye and delta if they are wound to be electrically equivalent externally.
Power calculations made externally do not depend on type.
Also, in addition to copper loss, there is also core loss, so I^2R is only part of the loss. |
Timebandit
| | Posted on Saturday, June 23, 2012 - 05:31 pm: |
|
One way to look at that last data plot is to say, "The stator never goes over 300*F. It's all good."
If you look closely, the data actually tells you quite a few interesting things about the bike:
1. From 0-30 minutes there is a rapid rise in Ts (stator) during low-speed riding (below 4000 RPM). Heat is accumulating in the stator while cooling is not sufficient because speed is too slow. The other data that we've seen in this thread shows that series regulators like the CF begin with very small reductions in duty cycle at low RPM, and don't begin to do much to effect duty cycle until RPM climbs above 4000.
The series regulators don't look like they offer much protection in this low RPM zone from 0-30 minutes. This is an area in which the H-D harness update does maximally decrease stator duty cycle by maximally reducing the stator's output via single-phasing. Of course, not even the harness update is going to do much for you when you don't have adequate airflow for cooling. It's obvious to me that the oiling rotor is the only thing that is going to help you under the conditions defined in this area of the chart. The data shows that no regulator can help you here -- not the CE, not the CF, not the Ducati+harness.
2. Time 30-100 minutes. There is an overall decrease in the slope of the temperature-time function at highway speeds. The stator is continuing to accumulate heat based upon RPM, but the airflow at highway speed is enough to provide heat dissipation and cooling.
3. Time 40-52 minutes: blue/green curves show that CT remains stable. 55-64 minutes the blue/green curves show that RPM increases and CT increases. At constant V=80, increasing RPM from 5k to 6k results in increasing Ts from less than 230 to over 240. The obvious conclusion is that increasing RPM causes an increase in Ts, while cooling airflow remains constant. While keeping speed constant at an ambient temp of 100*F, increasing RPM serves to overwhelm the effects of air cooling. Riding at sustained high RPM is probably not a good idea. Running through a variation of RPM (as you would in a race) is going to be better.
4. 72-87 minutes: Decreasing to V=0 and RPM to idle results in increasing Ts and CT, even with the fans on. Heat accumulates in the stator and coolant; Ts climbs from 240-270 and Tc climbs from 180 to 194. Also note that at idle, Vbatt falls to 12VDC. No charging takes place at idle. You're draining the battery. You can't let that go on forever.
5. Time 87-95: Resuming speed to 60-80 mph causes both CT and Ts to fall. Forced air cooling helps.
6. Time 95-100+ minutes: CT remains stable while Ts exceeds previous maximum. Stator is retaining heat while the coolant temp is falling. Notice that Ts continued to climb throughout the test, and the only reason that it didn't climb higher was because the test was terminated. The latter sections of this graph give us a clue about Recovery vs. Non-Recovery after placing a thermal load on the engine. They tell us that stator temp continues to climb while coolant temp is falling. To correct this problem, you need to carry heat out of the stator, into the rest of the engine, where the oil and coolant radiators can act to release it. The obvious answer is that to get the heat out of the stator and into the oil, you need the oiling rotor. Drilling some air holes into the rotor isn't working.
The data tells us a lot more than "The stator temp never gets too high and all is good." Viewing the data with a critical eye tells us a lot about the mechanisms of heat transfer in the engine, what works to help get rid of heat, what doesn't work to get rid of heat, and what makes the heat problem worse.
The obvious conclusion is that a series regulator doesn't help at all during the low RPM heat accumulation areas. To avoid heat accumulation at low RPM, you need to do one of two things: install the oiling rotor, or stop riding the bike at low RPM. Low RPM heat accumulation is what's killing the stators. It should be pretty obvious that commuting and getting stuck in traffic is the worst thing that you could do on this bike while riding it fast on highways or backroads is a lot better for stator life. The thermal data clearly shows that this bike operates well when ridden fast as a sportbike, changing RPM, and that it's having heat issues when it's ridden as a commuter, especially at low RPM. |
Timebandit
| | Posted on Saturday, June 23, 2012 - 05:32 pm: |
|
re-posting the chart so you don't have to scroll as far:
|
Timebandit
| | Posted on Saturday, June 23, 2012 - 11:13 pm: |
|
Just to clarify an earlier point: In my previous post I said, "Low RPM is bad." You could make the argument that low RPM isn't the problem, that Low Speed is actually the problem. I think it's likely the combination of both.
LOW SPEED is bad because you don't have decent airflow over the radiators and oil coolers.
LOW RPM is bad because oil flow (oil pump function) and coolant flow (water pump function) are both dependent on RPM. That means that you have to have a "decent" but not "excessive" amount of RPM to generate enough water pressure to pump the coolant through the radiators, to generate enough oil pressure to pump the oil through the oil cooler, and to generate enough oil pressure to spray oil onto the stator (if you have the oiling rotor)... without making the RPM so high that the bike is generating excessive heat from the extra combustion events.
So I guess the best way to look at the problem would be to say that you need to make enough RPM to have good oil & coolant flow, but not so much that you're generating excessive heat. And you need to make enough MPH that you have good air flow across the oil & water coolers. Obviously, you're better off cruising on the open road than being stuck in traffic.
You can do this sort of experiment even if you don't have special instrumentation on your bike to make cool plots like Greg has done. Just eyeballing the changes in coolant temp at various speed/RPM combinations can give you a pretty good idea of what minimum RPM and speed are required to keep the engine from getting too hot. Anything that you can do to keep the temps down is going to be good for the bike, including the stator. |
|
