|
In Chamber Testing
I have been working with pressure sensors to try and understand what is going on inside the rotor chamber.
What I am finding is that the best peak pressures seem to be located around 45° ATDC. But notice the variation of cycles. We think of a more even flow of power pulses. This log is taken while accelerating through 5000 rpm. The range above this is still too retarded. http://i287.photobucket.com/albums/l...relocation.jpg |
The test equipment is made by TFX Engine Technology.
You can familiarize yourself with the program by going to http://www.tfxengine.com/hardware3.html The test mule is a RE 13b running 12 psi (10 psi spring) TO4s, HKS cast divided 50/50 meth water pre and post turbo. HKS Twinpower Power FC |
More interpretation.
The red arrows indicate an engine misfire (in this case the mixture was too rich) . The blue arrows depict a too retarded cycle at about 60° with subsequent lower pressure. The green arrow show a cycle closer to 45° and higher pressure. I am amazed at the scatter. http://i287.photobucket.com/albums/l...location-1.jpg |
Just roughly looking at this chart I would say we should add a degree to section 2 and two degrees to section 3.
http://i287.photobucket.com/albums/l...location-2.jpg |
http://i287.photobucket.com/albums/l...tionrating.jpg
This is what the burn rate looks like. (One of the best examples of the 210 samples taken during this log.) The red line is the pressure of rotor compression without ignition. The blue line is ignited pressure. The purple line is the temperature of the burning mixture. Notice it is still 1400°C when the exhaust port opens. The pressure at exhaust port opening (EOP) is 91 psi. A detonation rating of 65/88/0 shows up for this #64 cycle. Since it has 0 frequency I am assured by the Mentor that it is benign. (But whenever the peak pressure is advanced below 45° it gets some of these readings.) The HP and torque readings are off in this sample because I am using the wrong displacement calculations. Barry |
Wow, I dunno how I missed this, but thanx for sharing such wonderful data!
Do you mind me asking where the location of the pressure and temperature probes were? -Ted |
Can you please show me where and how the sensors (pressure) are fixed into the rotor housings.
Where is the EGT sensor located? what type is it. What calculation (model is used) to guess the temperature at places aside from where a physical hard measure is taken? How are you measuring the AFR exactly. I thought about using their stuff but they could not answer any of my more detailed questions on their systems so I left it alone, interested to see what you have found and if you can shed any light on the above and also what cross checks you use to validate the outputs from their equipment. |
First time seeing this as well. Nicely done. Care to comment on the last output and advancing the timing to move the peak pressure closer to TDC? Or do you feel that doing so would lead to a much higher chance of detonation?
I've always wanted to play around and learn more about what is going on in the chamber itself. I feel a better understanding will lead to more power/reliabilty. Seeing what's actually going on rather than reading and interpreting what the engine is spitting out. |
If I had to guess the pressure sensors are directly bolted in to the combustion chamber via a spark plug hole as I doubt he's using MEMS. Temperature of the combustion chamber could be taken almost directly by placing a temp sensor on the housing immediately outside of the combustion housing (ie in the coolant flow). This would allow the modeling of the temperature fairly easily, though would be rather troublesome to do.
If what I said is true then I wonder how the pressure transducers affect your readings. |
Quote:
This will almost always affect the spark ignition due to changing the location of the spark (plug). It might be a minor thing, but this might throw the data off just enough... -Ted |
wish i could see the pictures at work...
|
Quote:
|
Yes if memory serves, Barry's sensors are the sensor/spark plug type.
This is awesome data, and could lead to some very neat results. Its just a pity that all of this data is essentially going to be setting up a bench mark being that no one's really looked at this - in the capacity of the end user - before. Barry, have you done any comparisons yet between pressures with and without water? Can't recall if I brought that up at deals gap..... :) |
Interesting thread, subscribed
|
Awesome!
Thanks for sharing Barry! Next time I see you, I need to pick your brain again! |
I agree that your low pressure pulses are due to improper fuel mix at your single point of ignition (since you are using the one spark plug hole for a pressure transducer). The flame front velocity is greatly reduced in improper mix, so the flame front propogation is too slow from a weak initation event.
What are you using for ignition? The spark energy appears to be too low. What is your ignition timing? Peak cycle efficiency for a piston engine is typically with peak combustion pressure occuring at 12 degrees ATDC, so I would look for 18 degrees on a rotary. It seems that your 45 degree peak pressure is running a terribly retarded ignition. Have you tried adjusting your fuel mix to improve your combustion? That is cool that you found some high speed trasducers.I have been wanting to get some for a while now, we use cylinder pressure transducers at work to maximize horsepower and efficiency, as well as emission predictions. Area under the pressure curve is your friend (indicated torque). Speaking of that, it is very low. I see where you mention that it is wrong. You should be seeing somewhere around 210 Lb-ft for two rotors, 105 lb-ft for a single rotor based on my output dyno findings. Your blowdown (pressure at exhaust open) will decrease if you get your ignition timing corrected. Still this blowdown number argues for a later exhaust port opening... You should see combustion initiation (blue and red diverge) slightly before TDC. (Not too much obviously). Thank you so much for posting this awesome information! I really like seeing data like this. Your peak cylinder pressure is low, but very good for 45 degrees (although that is only 30 on a piston engine). I am guessing that this is a 9.0:1 Compressiion ratio engine... I am guessing that the sensor is in the leading plug hole, as you couldn't read pressure to the end of the power stoke in the trailing plug hole. I would love to drill an additional hole at the bottom of the housing to monitor so that I could observe effects of leading/trailing split. |
I asked them for a cycle to properly map a wankel rotary 1080 deg.
Have they done this Barry? without it the software is useless in my opinion. From memory when I asked this they stopped returning my E-Mails. Can you map out one cycle at 0 to 1080 deg @ say 5000rpm for me? given where the sensor is this wont be possible sadly, to properly instrument this you would need 3 different pressure sensors equidistant machined into the rotor housing surface and all three sensors would need to be collated into each other to form one map. Me personally I don't care much for mathematical models, it would take allot to run proper sensors all over the engine (internal and other wise to get all the real information of what is happening). Still It would be good fun to play with, let us know more when you get to test it mate. |
The awesomeness of pressure transducer curves is not the math, but that ability to compare curve to curve what improvement was seen by the changes that you just made.
The numbers might not be right, but you really only need to see the 270 degree power stroke to compare the power that you are making at each new setting. The other three should be basically the same at different AFRs and ignition timing settings. The pressure transducer does not substitute for the dyno, it only allows deeper insight into what the cycle is doing and what adjustments can be safely made. You are correct about the entire cycle requiring three transducers, but I really don't think that anyone cares about the intake and exhaust stroke, and the end of the compression stroke (all that really matters) is able to be observed. (You might be able to observe the pressure on the intake stroke in the oil injection hole). |
This is what I use > http://www.picoauto.com/applications/trium.html
I long ago stopped posting important "current stuff" on the interweb as too many people copied it!. Anyway I went with picoscope and use it for pretty much everything (actuators and sensors), the amount of stuff you can do with it is only limited by your imagination. Use by lots of OEM big manufacturers too, as In above link. |
Why are you guys assuming Barry is only using one spark plug per rotor?
http://www.tfxengine.com/images/SparkplugSensors.gif |
Quote:
|
Do those combination spark plug + pressure transducer mimic the typical (NGK) spark plugs we normally run?
Mazda went with the surface gap spark plug design for a reason. Do they have a spark plug that is surface gap also? I think we all know that shrouded versus unshrouded spark plug electrodes do change the way the engine responds? -Ted |
Quote:
The sensor is a modified Optrand sensor from TFX. It is housed in a special spark plug. I have two PLX EGTs but the program is calculating the port opening EGTs. http://i287.photobucket.com/albums/l...6-2663_IMG.jpg This is my first attempt. It burned up because of the Rotary’s high temps. Barry My second attempt (with advise from my mentor) was this .025” orifice to protect the sensor face. http://i287.photobucket.com/albums/l...s/IMG_8430.jpg |
Quote:
The specific multi function super response sensor I use on the Pico is like $1+k only measures to 500psi though or 3.4Mpa can do running tests with it just not high powered ones, I am looking for another sensor that is as accurate and fast (>1% FS) and lower cost lol. My next engine I will machine the rotor housings to fit three sensors permanently equidistant into the engine and map the whole cycle, see how excited I get haha. Can tell you from running a car permanently VBOX equipped after a year or more of looking at reams of information you do get slightly over it :beatdeadhorse5: Then again I have been doing crap like this since 1991 (beginning of mech eng study), combine that with the internet age now where every single customer/peer is a mech eng enthusiast/expert cause of wiki and forums lol and it gets on your nerves needing to explain shit over and over in a never ending attempt to educate some cunts. |
Barry how much are TFX charging now for a system? and the generic plug sensor btw (like the one you have pictured)?
|
Quote:
TFX uses calculations for burn rate by watching the pressure rise. We enter degrees for port opening then the calculations give temp and pressure at opening. My AFRs are measured in the downpipe 18” from the turbo flange. I am getting some early detonation indications when the peak pressure is located less than 45º ATDC. Clint, "the mentor", says we can check the stock knock sensor for accuracy. Unfortunately my computer cannot run both programs (datalogit & TFX) concurrently. Some of the specs on the sensor: Non-linearity & Hysteresis Full Scale Output ±1% Combustion ±0.5% Non-Combustion ±0.25% Available Frequency Response 0.1 Hz to 10 kHz 0.1 Hz to 20 kHz 1.0 Hz to 30 kHz More at http://www.optrand.com/products.htm Barry |
Quote:
My understanding of this very preliminary data is that we will have to keep the lower squatter points greater than 45º ATDC (at higher RPM anyway). I think the main focus may become lessening the reversion from the exhaust port to decrease or minimize the squatter. Barry |
Quote:
I was simply pointing out the error in thinking that only one plug per rotor was being used. |
Quote:
I will discuss with Barry. |
Quote:
Who do you talk to there? you can E-Mail me if you like or PM to save commentary from others, thanks. |
Do we really need to turn this into a playground spat?
Can we drop the name calling? This thread is GOLD. Anyone who understands the data being presented knows that this kinda stuff is PRICELESS. I think this is the first time I've seen such data on ANY RX-7 type of forum. I've only read about stuff like this is SAE papers. To have a member actually messing around with such equipment is just unheard of. It would be sad if this thread gets closed. -Ted |
Quote:
Quote:
Quote:
Not using two plus per rotor would definitely invalidate the findings. My third iteration spark plug is a –10 heat range TFX model, with a platinum electrode. The only negative effect that it might have on combustion is that it is ¾” vs the Rotary designed reach of 7/8”. Things to be tested are: with and without water, water/meth, splits, negative splits in vacuum, advance curves, port shapes, you name it. Barry |
Quote:
|
Quote:
|
Quote:
My contact at TFX Engine Technology is Clint Gray. Their website has a better explanation of all the functions. http://www.tfxengine.com/index.html Their system goes for about 5k. The plug was $1300. All technical input is appreciated, Barry |
Quote:
Quote:
|
Sorry about the confusion, I thought the same thing that others did, this was a simple transducer installed in the leading spark plug hole.
From what I have seen, there is more available power by building that peak pressure earlier in the cycle. If you are seeing detonation with earlier pressure peaks, 45 degrees is what you get. (This really isn't too bad, at is equates to 30 degrees on a piston engine. As I recall, 12-15 degrees is the sweet spot for peak cycle efficiency (on a piston engine). The rotary may be different. Actually, now that I think about it, the rotary has a longer combustion chamber and probably requires longer for the flame to propagate. This may mean that the pressure will spike too quickly is it is initiated any sooner, while taking too long to propagate if initiated at this time. Anyway, I hope that you are planning to analyze the effects of leading/trailing split. Some claim that it makes a big difference, I found no change on the dyno at all for pretty much the entire test. In fact, I unplugged the trailing plugs and saw no change under 6,000 rpm. |
Barry do you know your %error on the calculations at all?
|
Quote:
There are a lot of concepts to interrelate when considering what is going on inside of a rotary engine. This is from a Mazda paper Rotary86v6a4, Fig. 14, showing flame propagation. I think this is probably Mazda Research at its best! If you haven't seen it before please take your time trying to understand it. http://i287.photobucket.com/albums/l...metravel-1.jpg Some things to note: Because the mixture is flowing the flame front hardly moves upstream at all. In fact the trailing portions of both flame patches are pushed backwards part of the time. The squish generation and trench shape further complicates this movement. When the leading and trailing flame fronts collide (at about 20º ATDC) that their speed diminishes. The knock region is from 30º - 45º ATDC and where the knock sensor is located. Barry |
Quote:
The timing trigger that I fabricated adjacent to Mazda’s timing wheel would be the area for greatest possible error. To check this a test run is then made where the engine ignition is cut at 6000 rpm and the throttle is opened fully. This double-checks TDC in relation to the logged actual compression hump. To my knowledge the rest are calculations. http://i287.photobucket.com/albums/l...s/IMG_8427.jpg |
| All times are GMT -5. The time now is 12:10 PM. |
Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
Hosted by www.GotPlacement.com