This is the first time I could get on the forum since getting back from my trip, and there was a lot I had to catch up on! I think I understand most of what was (and is) happening. Likewise, I haven't studied logarithms yet but I've been slightly exposed to them in testing. Sylvie posted a couple pages back that "Perhaps pure helium would be an available non-reactive gas (remember, the voltage is generated by chemical reactions), and one that won't cause a fire/explosion when you turn on the heater element." I beleive this is why a pure helium environment could work, but are there any other gases we could use for the same purpose?
This is the first time I could get on the forum since getting back from my trip, and there was a lot I had to catch up on! I think I understand most of what was (and is) happening. Likewise, I haven't studied logarithms yet but I've been slightly exposed to them in testing. Sylvie posted a couple pages back that "Perhaps pure helium would be an available non-reactive gas (remember, the voltage is generated by chemical reactions), and one that won't cause a fire/explosion when you turn on the heater element." I beleive this is why a pure helium environment could work, but are there any other gases we could use for the same purpose?
Justin
Welcome back and thanks for clarifying that. I believe helium would probably be the easiest gas to obtain, as it is available at Toys R Us, party stores, etc. I'm not sure what other gases are easy to obtain and suitable for this purpose. Suggestions, anyone?
I'm so sorry I haven't been active on the forums, I was on vacation all week and this is the first time I've even been on a computer all week (basicly since our meeting last Sunday). I went out deep sea fishing yesterday which is why I didn't go on then (by the way, great weather... we caught 200lbs of cod, haddock, and cusk, along with 2 baby sharks, and 2 sharks that ate our fish as we realed them in. One being 10-12 ft long, and one being about 7. Also, we caught a 20lb wolf fish, if you've never seen one before go on google and look one up, they are pretty scary looking fish.) Anyways, I'm with my dad now, and we are about to go to Chili's for lunch, but I promise as soon as I get back to my house I'll read up on post and get in the homework. Again, so sorry I haven't already gotten them in, I've just been extremely busy with vacation.
Be back shortly,
-Mike
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ "Learn to obey before you command."
I believe someone mentioned a pure helium environment to get a zero point for CO2 levels / millivoltage. However, I am not sure why pure helium was chosen. Could someone perhaps explain this to me? It has me curious ...
=====================================
Andrew,
It was Sylvie who mentioned a pure helium environment. He explained that if the environment was pure (100%) helium, then no other gases (like CO2) should be present. That's why I asked about putting the ASP-2 in a vacuum chamber (sometimes called a bell jar.) If all the air (gases) is sucked out there shouldn't be any CO2, or any other gases for that matter. It should give the same result as a pure helium environment but without the problem of containing the helium. Right...? I'm merely speculating.
=====================================
Sylvie,
"Houston, we have a problem..." I soldered the capacitor to the corresponding ADC pins and now I'm getting "spikes" in the data. The ASP outputs data OK but every 10th (or so) data point is either 60-75 units high or low. The range in the data is much wider than before and the graph line is not nearly as 'clean' (straight.)
Would sodlering the capacitor on to the ADC chip have caused this. I'm getting the same result with every chip I soldered (and I didn't think to keep an un-soldered chip as a control...!) More chips are on the way, along with two more CO2 sensor modules. One module was fries-- you can easily see the dark "burn" spot on the PCB-- and one sensor isn't working. We're down to our last CO2 sensor!
Dear Team,
So I read up, got very confused, and then tried the first homework problem. I got done the first part ((y2-y1)/(x2-x1)) and got -4/1000 and it is -0.004. then I did the easy 1/-0.004= -250. Then I got confused with the same thing Dylan did. I read even further down and got even more confused, and then it got a little better and I ended up with 250 * mV + 74250... homework question one done · sorry it was so overdue in the first place though
On to homework question 2... Did the math, got it all written down and completed, and ended with this: 0.01 * mV + 323... is this correct?
Now homework number 3... Is it because even though all of the points are on the same lines the line is curved and there's two lines? Again, correct me if I'm wrong.
I'm REALLY·sorry it took so long to get in, but I thought overdue homework was somewhat better then no homework at all.
Be back tomorrow,
-Mike
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ "Learn to obey before you command."
I'm a bit behind, I think, in what's happening, but a couple of comments...
I don't know what's going on with the parts, that they're turning up bad. I hope you're not destroying the expensive CO2 sensors as you're learning. You really only have to solder one wire to it, I think (the one to TP1). Is that right? It should be really simple.
Chips can be destroyed by leaving the soldering iron on the pins for too long (or, of course, by putting the soldering iron onto the body of the chip). I've never destroyed one, though, even when I was a completely inept novice solderer. I'm not sure what you're doing, and I wonder if maybe they're still just fine, but not wired up correctly.
As for collecting data points, I assume that Tracy has some plan for that, and I think you should wait for him to speak up again about that.
=================
I had a couple of days of rocket launching this weekend. Yesterday wasn't so hot. I had a chance to test a board I designed to work with a certain expensive altimeter...and as far as I can tell, it doesn't work. This was the second revision, so I'm a bit disappointed, and confused. I also flew my 4" diameter rocket on an I motor, using an altimeter, and the shock of the drogue ejection charge snapped the zip tie holding the battery in place, so I didn't get a main ejection, and I had a (minor) crash. No damage, and today I set it up again, this time on a Cesaroni J357, and had a perfectly flight.
So I set up my homemade onboard GPS transmitter (using XBees), and flew it on an Aerotech·J500, and it worked perfectly. I got good data through a flight to about 2400 feet, capturing the GPS data on my laptop (see attached images - the magenta line is the rocket, and the blue one is the track from my handheld GPS), and also displaying them in real time on a little receiver/LCD screen that I built using a Parallax Propeller Protoboard. From that,·I took the landing coordinates and entered them into my handheld Garmin GPS, and used that to walk almost directly out to where the rocket had landed. Very satisfying. To have come this far in three years from the "getting an LED to blink" days is really a testament to Parallax. I REALLY love learning.
·· Sounds like you had a good weekend of rocketry.·It must be rewarding to see a project that you designed and built yourself work so well. The GPS graph is really cool. So it allows you to predict where the rocket will land instead of just telling you where it landed? How does that work? That would come in really handy when you're launching in cornfield country, etc. Why are the different parts of the altitude graph different colors? How badly did you prang your rocket when the main ejection didn't fire? Hard to believe that the G-forces from ejection snapped a zip tie! I've learned from (many) previous·crashes not to hold·altimeter batteries in place with zip ties and rubber bands!
=============================
We're getting 'stray' spikes in the dataand I'm not sure why. We ran the ASP-2·on July 31st and the data·and graph lines were rock-solid. Then I moved· wires and parts around and, even though they should be interchangeable, we're getting stray spikes in millivoltage and CO2 ppm (millivoltage.) I suspect loose wires again, or a "bad" part. Whatever the case we have two more CO2 $en$or$ coming, and 10 more ADC chips. The program is working fine but the range of data is higher that before. But we'll get there. I do believe we fried on of the sensors; there's a dark spot on the back side of the PCB.
Bottom line question: What causes random number spike/ 'static'·(in both mV and CO2 ppm)? The sensor will read out 8-10 data points (e.g. 254, 252, 251, 254, 253, 250, 252, 249, 251... and then 168 or 337, then back the the 250's.)
NOTE: When I shut off (comment out) the command that turns the heat switch on and off·(HIGH 4) there is much less variability in the TP1 millivoltage. That seems diagnostic.
OK, the same questions lingers: What causes a wide range of millivoltage data? What causes 'spikes' (both high and low) in the data every 10th data point or so? I switched components over to another clone and right now there is virtually no variation-- no 'spikes'' in either the millivoltage or in the CO2 ppm (millivoltage.) In five minuts or so I'll apply CO2 to the sensor to see if the data plunges and then recovers as nicely as it did in Tracy's graph.
Could it be something as simple as having the wires plugged into a breadboard as opposed to soldering them to a perf board? Would "loose wires" cause data spikes and a wide(r) range of data?
We didn't have to wait 5 minutes. After a 2400 second (40 minute) warm-up the ASP-2 was ready for CO2. When we applied it-- with a 5-second breath-- the data plunged and recovered beautifully...! We'll post the graph and data.
So how do we ensure there is minimal variation in the millivoltage data and that it plunges and recovers this nicely EVERY time...?!
If we wanted to see if it plunges and recovers every time...could we just do multiple tests and see if one (or more) trials differs from the others? If we do multiple tests, the data should be consistent every time right?
···That's·a good point and a good question:· "If we do multiple tests, the data should be consistent every time right?" Yes!
That's exactly what we've been doing for·the past·few weeks: conducting extensive, time-consuming·multiple trials. What we're trying to do is isolate the variable,or variables. That means we're trying to figure out exactly what is causing the erratic data. The problem (and the fun!) is the fact that there are many variables, each of which·might cause erratic data. Each variable, as we've discovered, can also cause it's own unique problem(s). Some things (like crossed wires) can cause NO data to appear; another variable might cause a certain number (2650) to appear for all the data. Still another varible might shut the whole system down. That is the nature of our project and our work.
With Dr. Allen and Sylvie's generous help we've figured out many of the variables and "fixed" many of the problems along the way. For example, the program works now! That was a big accomplishment·as we worked our way through·several variables along the way. Now we're left with trying to find out why some of the data has a wider range than other data. This causes a spikey, "static-y" graph line that is harder to interpret than one that isn't. Compare the attached graphs #1 and #2 to graph #3 and you will see what I mean. The question is, what variable(s) are causing the "static", which results in the "spikey" graphs (#1 and #2.)
Justin, what do YOU think could be causing the "spikey" graphs? Hypothesize three likely and logical causes-- not science fiction guesses--·so we can try them out. You make a good point and asked a·good question, Justin! I'm glad to see you on the forum again. I'll look for your, and your teammates', replies and hypotheses. I'm also eager to hear what Sylvie and Dr. Allen think it might be.
My hypothesis #1: Could the "static"-- the spikey data-- be caused by loose wires on the BOE or breadboard?
My hypothesis #2: Could thicker,·solid-copper-core wiring (from the breadboard to the BOE)·have more resistance than thinner, stranded wiring, causing the spikey data?
I agree with Justin on this one on his methods, but couldn't we just allow the 5 minute warm up period every time before testing, then start the data collecting? Maybe create another program that runs just the heater for five minutes without any other actions occurring, then start data collecting on to the Flash drive? But I am confused, did you mean 2400 second, as in 2.4 seconds, (As if you wrote it in miliseconds), or actually that you had to wait almost a hour before testing? If you really did have to wait that long, then my theory could be wrong...
After looking through the graphs, (I must written my post after you posted), I noticed the yellow line. This seems to of been able to take the widely varying data from the y axis, and concentrated it into a thinner, more accurate line. I do not think I am making myself clear enough, but I cannot explain it any other way. Could these, spikes, be caused by the accuracy of the CO2 sensor? I was just thinking that maybe the CO2 sensor is not meant to sense this accurate of an amount of CO2. But then again, as I was writing that I realized, didn't we talk about increasing or decreasing the resolution to increase accuracy?
1) I notice that on graphs #1 and #2 (above), the ones with spikey graph lines, that millivoltage DECREASES to around 500 mV.
2) I also notice that on graph #3, the one with·a·"clean" (not spikey)·graph line, that the millivoltage INCREASES to over 3,000 mV.
To me this seems diagnostic, but I'm not sure exactly what it means. Could it be that, on the graphs with lower millivoltage, the voltage is being used to make the spikes and so it is lower...? How do we fix it?
...couldn't we just allow the 5 minute warm up period every time before testing, then start the data collecting?·...did you mean 2400 second, as in 2.4 seconds...
Hi Dylan,
·· Dr. Allen and I use a 30-40 minute (1800-2400 second) heater warm-up period before adjusting the potentiometer to 1,000 mV. This allows the heater/sensor time to·burn off·other gases and stabilize·(see Dr. Allen's suggestion in previous posts.)
After looking through the graphs... I noticed the yellow line. This seems to of been able to take the widely varying data from the y axis, and concentrated it into a thinner, more accurate line.
Dylan,
· The yellow line was added AFTER the sensor collected the data. I did it on Microsoft Excel when I took the data off the flash drive. The yellow graph line is called a "moving average trendline." It's purpose is to make the data/trend easier to see and interpret. You hit the nail on the head. Good thinking!
So what do YOU think is causing the static/ spikes in the graph lines? What are your three hypotheses? Hard to believe that we're off to Nevada tolaunch the ASP-2·in 5-1/2weeks! Our next practice is August 14th from 1:00 -4:30 PM at my house.
Mr. Kibler
Post Edited (Mark in NH) : 8/3/2010 1:44:08 AM GMT
I believe that the reason why they are increasing and decreasing, one to around 3000 mV and the other to 500mV, is because one graph was made out of data that had our false constant in the program and the other one did not. So that would mean that the one without the constant would increase around 3000 mV and the one with the constant to decreased to around 500 ppm. Because remember, as mV decrease ppm increase. And as mV increase (3000) ppm decrease (500).
*Had to edit this post to make it more understandable*
Post Edited (Dylan Landry) : 8/3/2010 1:46:07 AM GMT
...·they are increasing and decreasing ...because one graph was made out of data that had our false constant in the program and the other one did not.
Dylan,
·· Download the most recent version of the program (or any version) and check to see if it has 'our' constant in it. I don't believe it does.
Think more critically and more thoroughly···· before posing your next hypothesis.·Be sure·that·it's logical and·testable.·You only have TWO hypotheses left now...!
Mr. Kibler
Post Edited (Mark in NH) : 8/3/2010 1:52:48 AM GMT
I have a question before I post my hypothesis, for trials 3 and 2, they seem to be at least reacting to CO2, for the the first graph, was CO2 applied during testing? And if so, at what times?
I fully agree with Dylan's first hypothesis that a wrong constant could be the culprit. If that's not the case, then my next hypothesis would be to check the wiring (since wires can become loose very easily as we have found out). How are the wires attached? Are they soldered? That could be another problem if they are. My thrid hypothesis would be that maybe the CO2 sensor is fried? Did you use the same sensor for all 3 graphs?
I have a question before I post my hypothesis, for trials 3 and 2, they seem to be at least reacting to CO2, for the the first graph, was CO2 applied during testing? And if so, at what times?
Dylan,
·· Don't be hasty in forming your next two hypotheses:be thoughtful. What we're (really!) trying to do is use all of our resources-- that means everyone on the project·team-- to help figure out why the data is "spikey." Then we can test the hypotheses and fix the problem.. but only if the hypotheses are thoughtful,logical, and testable. I hope team members don't just·throw out random "ideas" to "get the homework done." That's not what we're about. We're scientists, trying to solve a very real scientific problem. We want our ASP-2 to work, right?
Look at the·dips in each graph line and you can see how far into·each trial (in minutes) CO2 was added. Graph #3, the "good" graph, actually has the time printed on the graph in red letters (41, 45 and 48 minutes.)
In trial #2 CO2 was added about 44 minutes into the trial (the X-axis doesn't show it.) In trial #1 CO2 was added at 1 hr, 7 minutes (see the X-axis label.)
I fully agree with Dylan's first hypothesis that a wrong constant could be the culprit. If that's not the case, then my next hypothesis would be to check the wiring (since wires can become loose very easily as we have found out). How are the wires attached? Are they soldered? That could be another problem if they are. My thrid hypothesis would be that maybe the CO2 sensor is fried? Did you use the same sensor for all 3 graphs?
Justin
Guys, ··We haven't added·the constant to the program yet (have we?) so I don't believe it's the cause of the problem.
FYI, the same CO2 sensor was used to record the data that made all three of the attached graphs. I don't believe it's "fried." If it was, how could it have given·"good" data for an (excellent) graph/trial #3?
This was a solid hypothesis Justin. To test it we should be able to·simply change the sensor (the variable) and keep everything else the same (the constants.) Maybe I'll try this in the morning "just in case!" I still hypothesize that it's loose wiring, or a different diameter and type of wire (so no one can use my hypotheses!)
I fully agree with Dylan's first hypothesis that a wrong constant could be the culprit. If that's not the case, then my next hypothesis would be to check the wiring (since wires can become loose very easily as we have found out). How are the wires attached? Are they soldered? That could be another problem if they are. My thrid hypothesis would be that maybe the CO2 sensor is fried? Did you use the same sensor for all 3 graphs?
Justin
Justin,
It's a good theory, but since the constant ... well ... stays "constant" I don't believe it would cause variations in the graph. I believe this would be something more hardware related, but I can not exactly think of what it would be. My first and most logical hypothesis would be interference. However, is even a little bit of interference able to create this much variation in our data? Does anyone else consider this as a possibility?
This graph that Mark posted on page 27 looks very nice. Something was working well! The line is smooth and is resting above 2 Volts, and the response to CO2 is very clear. That is what we need to see more of!
Spikes in data. Sometimes those can come from 1) having long loops of wire, or 2) "ground loops" that come from poor or misplaced connections at the ends, or 3) inadequate power supply filtering, and 4) noise intrinsic in the sensor. Okay, four hypotheses, but number 4 is a special case. You have to realize that the BASIC Stamp during its operation draws significant pulses of current.
In the case of (1) , you can think of the wires to the ADC and sensor as an antenna that picks up external noise, and the solution is to shorten or shield or tighten up the loops.
In the case of (2) you can think of every path on the circuit board as a resistor, and where current flows through a resistor there is also a voltage by V = I*R. One thing to look for especially is where the ground wire to the sensor attaches to the MCP3202 ADC. It should attach directly to the Vss pin of the ADC, not to some remote point on the ASP-2. Be sure the connections are good.
As to case (3), maybe an additional capacitor across the power supply for the MCP3202 might help. Try something like a 100 µF 16 Volt electrolytic capacitor You already have a 0.1 µF capacitor there, right? Leave that there and put the large one in parallel with it. You could also try a 0.1µF capacitor connected between Vss and the TP1 input to the MCP3202, to help filter residual noise.
I'd look at 2 first, then 3 because it is easy, and then 1, and ultimately characterize 4, the intrinsic noise.
We're back from camping. It did not happen quite as expected. All the campgrounds in the Yosemite high country were full, despite the fact that we arrived well before noon on Friday. Never before! So we went down the eastern slope of the Sierra to Mono Lake, which is a salt lake even saltier than the ocean or the one in Utah. It is famous for its pillars of calcium carbonate. Then on Saturday we went back up to the high country and took several nice hikes, including one up to Mono pass. It is at 10500 feet, above tree line, replete with meadows, small lakes and wildflowers. It was not cold at all. It is much colder now that we are home with the coastal fog.
What is "noise intrinsic in the sensor"? How could this affect the spikes in the data? Could wrong wiring cause the spikes in the data? I know we have about checked the wiring over ten times but it could still be a possibility I think.
I agree with some of Dr. Allen's hypothesis. Since the CO2 sensor is not meant to output data, just an alarm, TP1 could also have some current going through it that we are picking up that is not necessarily coming from the sensor's output depending on CO2 levels, maybe some power required for other components or the heater.
I agree that interference could be a possibility, but what would be causing it?
Justin
Justin,
·· This is the point I make above, restated. And interence isn't a possibility: it's a reality. Look at the graphs from both·Trial #1 and Trial #2. Compare them to Trial #3. How are they different?·
>
The graph line in Trial #1 and #2 is very "up and down." There's variability in the data and the data falls over a wider range than in Trial #3. This is cause by interference.
So I restate my questions:
1) What do·YOU think is causing the variability (interference) in the data?
2 How do we fix it?
See if you can come up with three logical hypotheses that are different from any that have been posted. That way we'll have lots if variables to consider.
··· It sounds like you had a good camping trip in the Yosemite high country. The pictures are great... and such a·beautiful clear-blue sky! I didn't realize that you were hiking up so high, to·over 10,000 feet. It wasn't·cold up there? I'll bet the stars were stunning! It sounds like you did a lot of hiking, too. How far did you hike? If the campground was full did you camp out in the "wilderness"?
On the topic of stars, an... ASTRONOMICAL NEWS FLASH! ·· There may be an aurora borealis-- Northern Lights-- tonight!
====================================
Thanks for the (four!) hypotheses. We'll shorthen the long "antenna" wires and solder the Vss wire directly to the ADC chip. We'll also add the capacitors you suggested to filter out 'noise' ('noise' is the same as interference Justin, sometimes called·"static." It·causes the spikey graph lines.)·We also ordered two new CO2 sensor modules, two·extra sensors,·10 more ADC chips, and a handful of resistors so we can solder the ADC set-up onto perf board. Having never done it before we've already managed to go through several resistors and a few capacitors and ADC chips. They're·now permanently soldered to perf board·set-ups·that didn't work. Like our graph lines, soldering is a (learning) curve, sometimes exponential, but ultimately it levels out... or so we hope!
Stay tuned for details and be sure to check out the Northern Lights tonight. Thanks for sharing the camping details and pictures. "Tell us a scary camping story...!" (Did you see or hear Sasquatch?!)
The graph looks very nice... the line is smooth and is resting above 2 Volts...
Tracy,
·· So ultimately the resting point for a "good" graph should be above 2,000 mV? Is that one litmus test? Sorry if you mentioned this before and I missed it. I just want to be surer than sure
Comments
This is the first time I could get on the forum since getting back from my trip, and there was a lot I had to catch up on! I think I understand most of what was (and is) happening. Likewise, I haven't studied logarithms yet but I've been slightly exposed to them in testing. Sylvie posted a couple pages back that "Perhaps pure helium would be an available non-reactive gas (remember, the voltage is generated by chemical reactions), and one that won't cause a fire/explosion when you turn on the heater element." I beleive this is why a pure helium environment could work, but are there any other gases we could use for the same purpose?
Justin
Welcome back and thanks for clarifying that. I believe helium would probably be the easiest gas to obtain, as it is available at Toys R Us, party stores, etc. I'm not sure what other gases are easy to obtain and suitable for this purpose. Suggestions, anyone?
Thanks,
Andrew
I'm so sorry I haven't been active on the forums, I was on vacation all week and this is the first time I've even been on a computer all week (basicly since our meeting last Sunday). I went out deep sea fishing yesterday which is why I didn't go on then (by the way, great weather... we caught 200lbs of cod, haddock, and cusk, along with 2 baby sharks, and 2 sharks that ate our fish as we realed them in. One being 10-12 ft long, and one being about 7. Also, we caught a 20lb wolf fish, if you've never seen one before go on google and look one up, they are pretty scary looking fish.) Anyways, I'm with my dad now, and we are about to go to Chili's for lunch, but I promise as soon as I get back to my house I'll read up on post and get in the homework. Again, so sorry I haven't already gotten them in, I've just been extremely busy with vacation.
Be back shortly,
-Mike
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"Learn to obey before you command."
-Solon
Mr. Kibler,
I believe someone mentioned a pure helium environment to get a zero point for CO2 levels / millivoltage. However, I am not sure why pure helium was chosen. Could someone perhaps explain this to me? It has me curious ...
=====================================
Andrew,
It was Sylvie who mentioned a pure helium environment. He explained that if the environment was pure (100%) helium, then no other gases (like CO2) should be present. That's why I asked about putting the ASP-2 in a vacuum chamber (sometimes called a bell jar.) If all the air (gases) is sucked out there shouldn't be any CO2, or any other gases for that matter. It should give the same result as a pure helium environment but without the problem of containing the helium. Right...? I'm merely speculating.
=====================================
Sylvie,
"Houston, we have a problem..." I soldered the capacitor to the corresponding ADC pins and now I'm getting "spikes" in the data. The ASP outputs data OK but every 10th (or so) data point is either 60-75 units high or low. The range in the data is much wider than before and the graph line is not nearly as 'clean' (straight.)
Would sodlering the capacitor on to the ADC chip have caused this. I'm getting the same result with every chip I soldered (and I didn't think to keep an un-soldered chip as a control...!) More chips are on the way, along with two more CO2 sensor modules. One module was fries-- you can easily see the dark "burn" spot on the PCB-- and one sensor isn't working. We're down to our last CO2 sensor!
"Pass the chips, please!"
Mark
So I read up, got very confused, and then tried the first homework problem. I got done the first part ((y2-y1)/(x2-x1)) and got -4/1000 and it is -0.004. then I did the easy 1/-0.004= -250. Then I got confused with the same thing Dylan did. I read even further down and got even more confused, and then it got a little better and I ended up with 250 * mV + 74250... homework question one done
·
On to homework question 2... Did the math, got it all written down and completed, and ended with this: 0.01 * mV + 323... is this correct?
Now homework number 3... Is it because even though all of the points are on the same lines the line is curved and there's two lines? Again, correct me if I'm wrong.
I'm REALLY·sorry it took so long to get in, but I thought overdue homework was somewhat better then no homework at all.
Be back tomorrow,
-Mike
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"Learn to obey before you command."
-Solon
I don't know what's going on with the parts, that they're turning up bad. I hope you're not destroying the expensive CO2 sensors as you're learning. You really only have to solder one wire to it, I think (the one to TP1). Is that right? It should be really simple.
Chips can be destroyed by leaving the soldering iron on the pins for too long (or, of course, by putting the soldering iron onto the body of the chip). I've never destroyed one, though, even when I was a completely inept novice solderer. I'm not sure what you're doing, and I wonder if maybe they're still just fine, but not wired up correctly.
As for collecting data points, I assume that Tracy has some plan for that, and I think you should wait for him to speak up again about that.
=================
I had a couple of days of rocket launching this weekend. Yesterday wasn't so hot. I had a chance to test a board I designed to work with a certain expensive altimeter...and as far as I can tell, it doesn't work. This was the second revision, so I'm a bit disappointed, and confused. I also flew my 4" diameter rocket on an I motor, using an altimeter, and the shock of the drogue ejection charge snapped the zip tie holding the battery in place, so I didn't get a main ejection, and I had a (minor) crash. No damage, and today I set it up again, this time on a Cesaroni J357, and had a perfectly flight.
So I set up my homemade onboard GPS transmitter (using XBees), and flew it on an Aerotech·J500, and it worked perfectly. I got good data through a flight to about 2400 feet, capturing the GPS data on my laptop (see attached images - the magenta line is the rocket, and the blue one is the track from my handheld GPS), and also displaying them in real time on a little receiver/LCD screen that I built using a Parallax Propeller Protoboard. From that,·I took the landing coordinates and entered them into my handheld Garmin GPS, and used that to walk almost directly out to where the rocket had landed. Very satisfying. To have come this far in three years from the "getting an LED to blink" days is really a testament to Parallax. I REALLY love learning.
Post Edited (sylvie369) : 8/2/2010 3:26:42 AM GMT
·· Sounds like you had a good weekend of rocketry.·It must be rewarding to see a project that you designed and built yourself work so well. The GPS graph is really cool. So it allows you to predict where the rocket will land instead of just telling you where it landed? How does that work? That would come in really handy when you're launching in cornfield country, etc. Why are the different parts of the altitude graph different colors? How badly did you prang your rocket when the main ejection didn't fire? Hard to believe that the G-forces from ejection snapped a zip tie! I've learned from (many) previous·crashes not to hold·altimeter batteries in place with zip ties and rubber bands!
=============================
We're getting 'stray' spikes in the dataand I'm not sure why. We ran the ASP-2·on July 31st and the data·and graph lines were rock-solid. Then I moved· wires and parts around and, even though they should be interchangeable, we're getting stray spikes in millivoltage and CO2 ppm (millivoltage.) I suspect loose wires again, or a "bad" part. Whatever the case we have two more CO2 $en$or$ coming, and 10 more ADC chips. The program is working fine but the range of data is higher that before. But we'll get there. I do believe we fried on of the sensors; there's a dark spot on the back side of the PCB.
Bottom line question: What causes random number spike/ 'static'·(in both mV and CO2 ppm)? The sensor will read out 8-10 data points (e.g. 254, 252, 251, 254, 253, 250, 252, 249, 251... and then 168 or 337, then back the the 250's.)
NOTE: When I shut off (comment out) the command that turns the heat switch on and off·(HIGH 4) there is much less variability in the TP1 millivoltage. That seems diagnostic.
Good morning,
Mark
OK, the same questions lingers: What causes a wide range of millivoltage data? What causes 'spikes' (both high and low) in the data every 10th data point or so? I switched components over to another clone and right now there is virtually no variation-- no 'spikes'' in either the millivoltage or in the CO2 ppm (millivoltage.) In five minuts or so I'll apply CO2 to the sensor to see if the data plunges and then recovers as nicely as it did in Tracy's graph.
Could it be something as simple as having the wires plugged into a breadboard as opposed to soldering them to a perf board? Would "loose wires" cause data spikes and a wide(r) range of data?
Mark
We didn't have to wait 5 minutes. After a 2400 second (40 minute) warm-up the ASP-2 was ready for CO2. When we applied it-- with a 5-second breath-- the data plunged and recovered beautifully...! We'll post the graph and data.
So how do we ensure there is minimal variation in the millivoltage data and that it plunges and recovers this nicely EVERY time...?!
Mark and Chris
If we wanted to see if it plunges and recovers every time...could we just do multiple tests and see if one (or more) trials differs from the others? If we do multiple tests, the data should be consistent every time right?
Justin
···That's·a good point and a good question:· "If we do multiple tests, the data should be consistent every time right?" Yes!
That's exactly what we've been doing for·the past·few weeks: conducting extensive, time-consuming·multiple trials. What we're trying to do is isolate the variable,or variables. That means we're trying to figure out exactly what is causing the erratic data. The problem (and the fun!) is the fact that there are many variables, each of which·might cause erratic data. Each variable, as we've discovered, can also cause it's own unique problem(s). Some things (like crossed wires) can cause NO data to appear; another variable might cause a certain number (2650) to appear for all the data. Still another varible might shut the whole system down. That is the nature of our project and our work.
With Dr. Allen and Sylvie's generous help we've figured out many of the variables and "fixed" many of the problems along the way. For example, the program works now! That was a big accomplishment·as we worked our way through·several variables along the way. Now we're left with trying to find out why some of the data has a wider range than other data. This causes a spikey, "static-y" graph line that is harder to interpret than one that isn't. Compare the attached graphs #1 and #2 to graph #3 and you will see what I mean. The question is, what variable(s) are causing the "static", which results in the "spikey" graphs (#1 and #2.)
Justin, what do YOU think could be causing the "spikey" graphs? Hypothesize three likely and logical causes-- not science fiction guesses--·so we can try them out. You make a good point and asked a·good question, Justin! I'm glad to see you on the forum again. I'll look for your, and your teammates', replies and hypotheses. I'm also eager to hear what Sylvie and Dr. Allen think it might be.
My hypothesis #1: Could the "static"-- the spikey data-- be caused by loose wires on the BOE or breadboard?
My hypothesis #2: Could thicker,·solid-copper-core wiring (from the breadboard to the BOE)·have more resistance than thinner, stranded wiring, causing the spikey data?
Think, then reply,
Mr. Kibler
1) I notice that on graphs #1 and #2 (above), the ones with spikey graph lines, that millivoltage DECREASES to around 500 mV.
2) I also notice that on graph #3, the one with·a·"clean" (not spikey)·graph line, that the millivoltage INCREASES to over 3,000 mV.
To me this seems diagnostic, but I'm not sure exactly what it means. Could it be that, on the graphs with lower millivoltage, the voltage is being used to make the spikes and so it is lower...? How do we fix it?
Help, Sylvie!
Standing by,
Mr. Kibler
·· Dr. Allen and I use a 30-40 minute (1800-2400 second) heater warm-up period before adjusting the potentiometer to 1,000 mV. This allows the heater/sensor time to·burn off·other gases and stabilize·(see Dr. Allen's suggestion in previous posts.)
Mr. Kibler
· The yellow line was added AFTER the sensor collected the data. I did it on Microsoft Excel when I took the data off the flash drive. The yellow graph line is called a "moving average trendline." It's purpose is to make the data/trend easier to see and interpret. You hit the nail on the head. Good thinking!
So what do YOU think is causing the static/ spikes in the graph lines? What are your three hypotheses? Hard to believe that we're off to Nevada tolaunch the ASP-2·in 5-1/2weeks! Our next practice is August 14th from 1:00 -4:30 PM at my house.
Mr. Kibler
Post Edited (Mark in NH) : 8/3/2010 1:44:08 AM GMT
*Had to edit this post to make it more understandable*
Post Edited (Dylan Landry) : 8/3/2010 1:46:07 AM GMT
·· Download the most recent version of the program (or any version) and check to see if it has 'our' constant in it. I don't believe it does.
Think more critically and more thoroughly··
Mr. Kibler
Post Edited (Mark in NH) : 8/3/2010 1:52:48 AM GMT
I fully agree with Dylan's first hypothesis that a wrong constant could be the culprit. If that's not the case, then my next hypothesis would be to check the wiring (since wires can become loose very easily as we have found out). How are the wires attached? Are they soldered? That could be another problem if they are. My thrid hypothesis would be that maybe the CO2 sensor is fried? Did you use the same sensor for all 3 graphs?
Justin
·· Don't be hasty in forming your next two hypotheses: be thoughtful. What we're (really!) trying to do is use all of our resources-- that means everyone on the project·team-- to help figure out why the data is "spikey." Then we can test the hypotheses and fix the problem.. but only if the hypotheses are thoughtful,logical, and testable. I hope team members don't just·throw out random "ideas" to "get the homework done." That's not what we're about. We're scientists, trying to solve a very real scientific problem. We want our ASP-2 to work, right?
Look at the·dips in each graph line and you can see how far into·each trial (in minutes) CO2 was added. Graph #3, the "good" graph, actually has the time printed on the graph in red letters (41, 45 and 48 minutes.)
In trial #2 CO2 was added about 44 minutes into the trial (the X-axis doesn't show it.) In trial #1 CO2 was added at 1 hr, 7 minutes (see the X-axis label.)
Mr. Kibler
··We haven't added·the constant to the program yet (have we?) so I don't believe it's the cause of the problem.
FYI, the same CO2 sensor was used to record the data that made all three of the attached graphs. I don't believe it's "fried." If it was, how could it have given·"good" data for an (excellent) graph/trial #3?
This was a solid hypothesis Justin. To test it we should be able to·simply change the sensor (the variable) and keep everything else the same (the constants.) Maybe I'll try this in the morning "just in case!" I still hypothesize that it's loose wiring, or a different diameter and type of wire (so no one can use my hypotheses!)
Mr. Kibler
Justin,
It's a good theory, but since the constant ... well ... stays "constant" I don't believe it would cause variations in the graph. I believe this would be something more hardware related, but I can not exactly think of what it would be. My first and most logical hypothesis would be interference. However, is even a little bit of interference able to create this much variation in our data? Does anyone else consider this as a possibility?
Andrew
Spikes in data. Sometimes those can come from 1) having long loops of wire, or 2) "ground loops" that come from poor or misplaced connections at the ends, or 3) inadequate power supply filtering, and 4) noise intrinsic in the sensor. Okay, four hypotheses, but number 4 is a special case. You have to realize that the BASIC Stamp during its operation draws significant pulses of current.
In the case of (1) , you can think of the wires to the ADC and sensor as an antenna that picks up external noise, and the solution is to shorten or shield or tighten up the loops.
In the case of (2) you can think of every path on the circuit board as a resistor, and where current flows through a resistor there is also a voltage by V = I*R. One thing to look for especially is where the ground wire to the sensor attaches to the MCP3202 ADC. It should attach directly to the Vss pin of the ADC, not to some remote point on the ASP-2. Be sure the connections are good.
As to case (3), maybe an additional capacitor across the power supply for the MCP3202 might help. Try something like a 100 µF 16 Volt electrolytic capacitor You already have a 0.1 µF capacitor there, right? Leave that there and put the large one in parallel with it. You could also try a 0.1µF capacitor connected between Vss and the TP1 input to the MCP3202, to help filter residual noise.
I'd look at 2 first, then 3 because it is easy, and then 1, and ultimately characterize 4, the intrinsic noise.
We're back from camping. It did not happen quite as expected. All the campgrounds in the Yosemite high country were full, despite the fact that we arrived well before noon on Friday. Never before! So we went down the eastern slope of the Sierra to Mono Lake, which is a salt lake even saltier than the ocean or the one in Utah. It is famous for its pillars of calcium carbonate. Then on Saturday we went back up to the high country and took several nice hikes, including one up to Mono pass. It is at 10500 feet, above tree line, replete with meadows, small lakes and wildflowers. It was not cold at all. It is much colder now that we are home with the coastal fog.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Post Edited (Tracy Allen) : 8/3/2010 3:38:12 PM GMT
What is "noise intrinsic in the sensor"? How could this affect the spikes in the data? Could wrong wiring cause the spikes in the data? I know we have about checked the wiring over ten times but it could still be a possibility I think.
Sean
I agree that interference could be a possibility, but what would be causing it?
Justin
·· This is the point I make above, restated. And interence isn't a possibility: it's a reality. Look at the graphs from both·Trial #1 and Trial #2. Compare them to Trial #3. How are they different?·
>
The graph line in Trial #1 and #2 is very "up and down." There's variability in the data and the data falls over a wider range than in Trial #3. This is cause by interference.
So I restate my questions:
1) What do·YOU think is causing the variability (interference) in the data?
2 How do we fix it?
See if you can come up with three logical hypotheses that are different from any that have been posted. That way we'll have lots if variables to consider.
Mr. Kibler
··· It sounds like you had a good camping trip in the Yosemite high country. The pictures are great... and such a·beautiful clear-blue sky! I didn't realize that you were hiking up so high, to·over 10,000 feet. It wasn't·cold up there? I'll bet the stars were stunning! It sounds like you did a lot of hiking, too. How far did you hike? If the campground was full did you camp out in the "wilderness"?
On the topic of stars, an... ASTRONOMICAL NEWS FLASH! ·
====================================
Thanks for the (four!) hypotheses. We'll shorthen the long "antenna" wires and solder the Vss wire directly to the ADC chip. We'll also add the capacitors you suggested to filter out 'noise' ('noise' is the same as interference Justin, sometimes called·"static." It·causes the spikey graph lines.)·We also ordered two new CO2 sensor modules, two·extra sensors,·10 more ADC chips, and a handful of resistors so we can solder the ADC set-up onto perf board. Having never done it before we've already managed to go through several resistors and a few capacitors and ADC chips. They're·now permanently soldered to perf board·set-ups·that didn't work. Like our graph lines, soldering is a (learning) curve, sometimes exponential, but ultimately it levels out... or so we hope!
Stay tuned for details and be sure to check out the Northern Lights tonight. Thanks for sharing the camping details and pictures. "Tell us a scary camping story...!" (Did you see or hear Sasquatch?!)
Mark
·· So ultimately the resting point for a "good" graph should be above 2,000 mV? Is that one litmus test? Sorry if you mentioned this before and I missed it. I just want to be surer than sure
Thank you,
Mark