I just want to make sure that I am understanding your concept.. So we would upload all the data, sort of like two EXCEL data columns, one with mV and the other one with the corresponding CO2 level, so that when the micro-controller gets a mV reading from the sensor, it will reference to the data to find the correct CO2 ppm?
Yes. More-or-less. Look at the examples on Tracy's page. It's complicated, because he has interpolation in the code as well. But that's the BASIC idea (hah hah).
Again, I think it's time to sit tight and wait for Tracy to help out again, but it wouldn't hurt to look over that page carefully. I am out of my league here - I understand the basic concept, but I've never done this stuff.
I just want to make sure that I am understanding your concept.. So we would upload all the data, sort of like two EXCEL data columns, one with mV and the other one with the corresponding CO2 level, so that when the micro-controller gets a mV reading from the sensor, it will reference to the data to find the correct CO2 ppm?
Dylan,
·· I was hoping that we (you Rocketeers!) could simply add·new programming code to the current program to convert·the millivoltage to CO2 ppm. I understand that this is why we to·calculate the 'constant' Dr. Allen and Sylvie mentioned several times before.·Think of it like this: Pretend the CO2 sensor reads out·688 mV when it's exposed to 320 ppm of CO2. What number would you multiply 688 by in order to get 320? In this simple, make-believe example, X would be the constant.
688 * X = 320
We would insert that constant into the computer program to make it calculate 688 * X to read out 320 (ppm) instead of 688 (mV). But it's more complex than that because, as Sylvie just said above, as CO2 concentration changes the relationship isn't linear. That means it·doesn't form a simple,·straight line. Look at the very first graph that Dr. Allen sent you during the first days of the project (on page 2 of the forum I think.) Look at the graph·line around 200-500 ppm or so.·It isn't straight the whole way.
A second thing we (you Rocekteers!) need to figure out is how to calibrate the CO2 sensor. Someone mentioned putting it in a "pure helium" environment (I don't fully understand why), another person suggested a pure CO2 environment. We need to·put the CO2 sensor into a known·concentration of CO2 (1,000 ppm for exaxmple) so we know what the millivoltage reading is for 1,000 ppm, etc. of CO2. Does that make sense?
Sylvie and Tracy, am I on track here...? Is my explanation theoretically correct? What task(s) do you suggest the Rocketeers·work on·during this coming weekend's 4-1/2 practice to make the most of their limited time together? Thanks and ciao!
Keep up the good work, Dylan! You have a great future waiting for you.
Again, I think it's time to sit tight and wait for Tracy to help out again, but it wouldn't hurt to look over that page carefully. I am out of my league here - I understand the basic concept, but I've never done this stuff.
Oh, I get it! A computer programming joke... the BASIC idea. You clever guy!··
Yes, I'm as eager as you are to hear from Tracy.·The answer is·in the constant and the code, I think. And getting a baseline CO2 reading. How would you go about doing that Paul? Put dry ice in a fume hood and then measure millivoltage? I don't understand the thing about putting the sensor in a "pure helium" environment to get baseline/calibration millivoltage. Why helium? Did I miss the simple explanation?
Yes, I'm as eager as you are to hear from Tracy.·The answer is·in the constant and the code, I think. And getting a baseline CO2 reading. How would you go about doing that Paul? Put dry ice in a fume hood and then measure millivoltage? I don't understand the thing about putting the sensor in a "pure helium" environment to get baseline/calibration millivoltage. Why helium? Did I miss the simple explanation?
I was just speculating about a way to get a calibration point with 0 ppm CO2 (that is, to expose the sensor to a gas that contains no CO2, and none of the other gases the datasheet says it is sensitive to). I'm not sure you're going to need to do anything at all with calibration points anyway, but if you do, that seemed to me to be one way to get a known CO2 concentration to work from.
Sorry for not being on the forum lately. I just didn't really understand how to do the first method, and the second method (as Mr. Kibler said) would be hard to do because the relationship between CO2 concentrations isn't linear. Dylan helped me to understand the first method in a more simpler way (thanks Dylan!). But the actual formula to convert it still eludes me. Like Sylvie said, I think this will all become more clear when Dr. Allen explains it, since he has a lot to teach us about this concept.
I'm back and am happy to see that you are thinking things through carefully about how to do the data processing to convert mV to ppm.
But first, do we have any data to process?
Compare my response graph with yours. At least the most recent one I could find. Are there others?
mine:
yours:
You can find those and associated comments earlier in this thread on page 22 and page 23 respectively.
You say you applied CO2 at 1:00 and then again at 1:30 and at 1:35. The CO2 should cause the mV to decrease. See the clear signal in my graph? If there is a signal in your graph, it would take no end of signal processing to extract it. The rule of thumb is, if your brain can't see it, don't expect the BASIC Stamp to find it!!!
I am puzzled by the low level of the mV reading in your graph. (about 0.45V versus 1.4V that I saw) That makes me suspicious that the heater is not getting 6.0 +/- 0.1 Volt. That would explain the lack of response to CO2. I know you have a multimeter. Please measure both the heater voltage and the output from TP1.
-- Is the heater getting 6.0 Volts?
-- Is the mV reading at TP1 the same on both the multimeter and on the STAMP debug display?
-- Do you have more than one CO2 sensor (in case one is damaged)?
...You say you applied CO2... [noparse][[/noparse]it] should cause the mV to decrease...
Tracy,
·· Welcome back! I hope you had a good time in Chicago and Michigan. I hope it wasn't as hot there as it has been here in southeasyern Ohio... 10 days of 90+ degrees, with humidity to match the temperature.
You read my mind and I'm not sure why it didn't occur to me at the time (in Maine two weeks ago.) It came to me about two days ago as I was driving down the road. Of course! The millivoltage should decrease when CO2 is applied, not increase. This has me puzzled now. and so I'll run another series of trials to see·if I can·get a "dip" in the graph like we see in yours. I'll also upload the data (*the forum doesn't like Excel.)·I'll also check to see that the heater is on; last time it was hotter than summer but I'll check the voltage just to be sure.
The Rocketeers are really sticking with it, a positive thing because much of this is over their heads (and mine, and Sylvie's, so he says. I think he's downplaying his brilliance.) The meet this weekend for 4-1/2 hours at Andrew's house and I'm curious what will result from their meeting. Some excellent learning is going on and not all of it can be quantified. But they're sticking with it; that's the important thing.
You can delete your own duplicate posts, if you click on the little X at the upper right hand corner of its window, next to the editing pencil.
According to the MG811 CO2 sensor documentation, the raw sensor output would be around 320 mV at ambient CO2 concentrations, and that amplified times 6 2/3 by the Parallax interface circuit would bring it up to around 2.1 Volts where it goes to TP1 and to our analog-to-digital converter. However, at ambient, I am seeing around 1.4 Volts and you are seeing around 0.45 Volts. There may be considerable variation among MG811 sensors in what that resting level is, and also perhaps in the sensitivity factor. But 0.45 Volts is way way too low.
·· I just learned something new. I deleted the triplicate posts, etc. (where did the 1's and 0's go, Sylvie? Do they still exist at all...?)
I've got 'em over here. I try to recycle when I can. Of course 0's are cheap, but you can never have enough 1's.
I see I'm getting more unwarranted credit ("brilliance"? Nope. I'm an intermediate beginner).
Okay, they need to put the CO2 sensor on sale, so I can play around with this as well and see what kind of signals I get.
Re. Excel and the forum, you can always put the graph on the screen in Excel, then use the Windows Snipping Tool to grab the graph and put it into a .jpg file. That's what I did above.
Now, I'm surprised to see 1500 mV values in Tracy's graph for what I assume is regular air. All of the graphs on the sensor's datasheet go no higher than 400 mV. Granted, they're not displaying the kind of CO2 content you expect to see in air, but I wouldn't have expected it to go that high. Is that the raw reading from the sensor, or is it boosted·by that fancy "buffer" circuitry? I believe it's the latter, and that we've already talked about that earlier in the thread. I'll go back and look for it now.
Ah yes, there it is on page 15, where Tracy writes "Any Voltage that you measure on the test point, TP1, will be 7.67 times the actual output of the CO2 sensor". So that 1500 mV reading really represents a sensor reading of about 195 mV ( = 1500/7.67) from the sensor, leaving me even more confused, as I'd already concluded that the voltage from the sensor probably never goes below 250 mV.
I'm less brilliant than I was even a few minutes ago, and I'm working my way backwards quickly.
Post Edited (sylvie369) : 7/22/2010 12:51:05 AM GMT
·However, at ambient, I am seeing around 1.4 Volts and you are seeing around 0.45 Volts. There may be considerable variation among MG811 sensors in what that resting level is, and also perhaps in the sensitivity factor. But 0.45 Volts is way way too low.
Dr.Allen,
Is 1.4 volts considered the average voltage at ambient? How do we bump up our voltage from being 0.45 volts to around your voltage (1.4 volts) if not exactly that?
I'm not sure why the output from your sensor is 0.45 Volt instead of something higher. I am suspicious that there is still something wrong with the wiring of the circuit, that somehow it is not getting enough power to its heater for example. I asked whomever has one of the sensor setups and a multimeter to test of the heater voltage, or at least the power supply voltage for the sensor, and also to test to be sure that the voltage on TP1 matches what you are reading with the PBASIC program. It is also possible that the sensor is defective or has been toasted somewhere along the way.
Not only is 0.45 Volt lower than expected, there is no real evidence for a response to CO2. The signal is very noisy, bouncing up and down, but look at it in the graph posted earlier. Do you see any evidence of a response to CO2 at 1pm or at 1:30 pm? If not, no amount of boost to bump up the voltage to 1.4 is going to help. There will be natural variability from one sensor to another of this type, but we have not tested enough of them successfully yet to fathom the extent of that variability. How many CO2 sensors do we have? I have one here. Sylvie needs one.
Sylvie is right about the boost inside the Parallax circuit. It amplifies the raw output of the sensor by a factor of 7.67. So ambient air should give about 2 to 3 volts output from TP1.
Is your graph's data composed of reading from TP1? If so, then Sylvie's post makes sense!
Sylvie said...
Ah yes, there it is on page 15, where Tracy writes "Any Voltage that you measure on the test point, TP1, will be 7.67 times the actual output of the CO2 sensor". So that 1500 mV reading really represents a sensor reading of about 195 mV ( = 1500/7.67) from the sensor...
If you did take your reading from a Multimeter attached to TP 1 this could prove our problem. Also, for Mr. Kibler's graph, I see an increase at both the times of 1:00:00 and 1:30:00. Since it is suppose Decrease in mV when more CO2 is applied, this could be due to maybe backwards wiring or a small program error. I do not want to take anymore time troubleshooting at this stage. The reason is that I believe that no one wants to end up doing this during our practice time. For those times could best be applied for building.
Also, If we would have to upload multiple constants due to it not being linear, would we need to upgrade yet again to a new stamp?
Or could we possible make another program which could insert the mA into a formula, and it would be able to calculate the constant on the fly, then multiply the mA by it to get ppm? Or would this by chance delay readings?
Do you have a BS2e? That is one that has 8 program slots. I am sure you have a lot of things to build, and building is always more fun that troubleshooting! Troubleshooting can be extremely time consuming and requires patience and analytical thinking, and tools, and understanding of the underlying principles. It is great when you isolate the source of the trouble, but painful when it just goes on and on with no resolution. Then, is there a backup plan?
What all do you need to build at your Saturday meeting? Is everything else ready for liftoff?
In my way of thinking, the conversion from mV to ppm is secondary. It would be nice to have PBASIC do it in real time, nevertheless, you can easily do that in EXCEL after the flight. It is much more important to be sure the sensor is responding as it should, and that a valid raw reading is going into the datalogger file!!!
Do you have a BS2e? That is one that has 8 program slots. I am sure you have a lot of things to build, and building is always more fun that troubleshooting! Troubleshooting can be extremely time consuming and requires patience and analytical thinking, and tools, and understanding of the underlying principles. It is great when you isolate the source of the trouble, but painful when it just goes on and on with no resolution. Then, is there a backup plan?
What all do you need to build at your Saturday meeting? Is everything else ready for liftoff?
Tracy,
· Saturday's pratice is at Andrew's house and all the Rocketeers will be there... except me. I won't return from Ohio until Wednesday or Thursday of next week. This complicates things and bottlenecks us a bit; that's an unfortunate reality.
Andrew has everything the team needs to·get the ASP-2 working on Saturday and it has a BS2e stamp on it.·Andrew also has a CO2 sensor (we have three total) and everything else he needs to "clone" the ASP-2 prototype that I have.·By now he should have in working and running in tandem with the one I have here with me.
Like you and Sylvie, I suspect that there's still a problem with the wiring. I'll go over it with a fine-tooth comb when I return (but darn it! I thought I had it dialed in befire I left Maine for Ohio.)
·· OK, I'm back. I found a power supply, with 4% battery supply left. Also, I nearly fried the 'NASA' computer Andrew, like I did when I was at Marshall Space Flight Center. Remember when·I put it·in the double-insulated black carrying case without making sure it powered down?·I did the same thing today and it was in the case, in the van, in the·90 degree heat for 1-1/2 hours... Close call!
Like I was saying Tracy, I suspect that it's still an issue with the wiring. But Christopher and I spent literally·hours-- days in fact-- with it before I left Maine, cross-checking all your diagrams, the programming, etc.·I thought we had it rigth before I left for Ohio, but like I said a few posts back, it occured to me a few days ago that the numbers should not have incraased when we added CO2. It's back to square one with the wiring and it reminds me of your comment to Dylan:
"Troubleshooting can be extremely time consuming and requires patience and analytical thinking, and tools, and understanding of the underlying principles. It is great when you isolate the source of the trouble, but painful when it just goes on and on with no resolution. Then, is there a backup plan?
Not fully understanding that interplay between all the underlying principles is part of the problem on my end. And yes, it is (has been!) painful to go on and on without resolution. But we'll stick with it like diligent scientists. We'll figure it out, isolate the variable, and make it work. That is my commitment and hope.
Our backup plan? Launch last year's ASP (ugh!), which I'd really rather not do. We owe it to those who have helped make the trip to Nevada in September possible-- but most importantly to ourselves-- to figure it out and make it work. That's what aerospace engineers, and so that is what we will do.
Andrew, please conatct me by e-mail and let me know what the operational status is of our ASP-2 clone, 'Phipdippides'. I hope he plans to make an appearance at this weekend's practice!
Tracy Allen said...
Dylan,
Do you have a BS2e? That is one that has 8 program slots. I am sure you have a lot of things to build, and building is always more fun that troubleshooting!
Dr. Allen,
I'm not sure what Dylan's situation is, but both Mr. Kibler and I have BS2e's on our ASP-2 platforms. I don't believe we have too much building left to do. Mounting the A/D converter and CO2 sensor comes to mind, but I'm not quite sure how we will tackle that -- perhaps a good topic for this weekend's team meeting?
Tracy Allen said...
Troubleshooting can be extremely time consuming and requires patience and analytical thinking, and tools, and understanding of the underlying principles. It is great when you isolate the source of the trouble, but painful when it just goes on and on with no resolution. Then, is there a backup plan?
I can relate to what you said about troubleshooting. As you know, we've definitely been there and done that before. However, it certainly feels great once you "cross the finish line," so to speak.
At this upcoming team meeting, is it possible that we are gonna split up with some of us working to mount the A/D converter and some of us working to mount the CO2 sensor? Still, we would need to discuss it.
At this upcoming team meeting, is it possible that we are gonna split up with some of us working to mount the A/D converter and some of us working to mount the CO2 sensor? Still, we would need to discuss it.
Justin
Justin,
This is not one of my top priorities, as I stated previously. We can definitely discuss it and work with what resources we can, but we probably can't physically build anything at the upcoming meeting since Mr. Kibler has the ASP-2 with him in Ohio. I only have a mockup. Also, I don't believe this is an area where we want to "split up," since the two components are closely related and will likely be mounted to the same place. Nonetheless, we have plenty of other important tasks to complete at the next meeting.
Andrew, Justin, and Rocketeers across America, · I agree with Justin. I think this weekend's practice time will be well spent working with Phidippides and trying to get him to do what the ASP-2 and my clone are supposed to do. Andrew·has all the necessary components and virtually all the information you need has been posted to the forum. This includes wiring schematics, the program code, pictures of wiring, explanations, etc.
If you·make this·your sole focus·you'll be amazed how fast 4-1/2 hours will go by! Trust me. It will give you a (very)·full taste of what the project is·really about... the problems, the problem solving skills, the analytic thinking that Dr. Allen mentioned, and ultimately the thrill of figuring out how to make the ASP-2 work. This is what I would like to see you accomplish because I want·you to have pride and ownership for the completed project. Be sure to schedule August's practice when you meet this weekend. I propose·August 14th or 15th at Dunbarton Space Flight Center (my house) from 1:00 - 4:30 PM. After that, because·school will be starting and Chritsopher will be headed off to college,·we will probably only meet twice before we head into the desert to launch the ASP-2. Our primary mission goal is to launch the ASP-2 fully operational, e.g. with the CO2 sensor and all·of last year's·systems fully functional.·Please define, discuss, and then list your mission goals sometime during·this weekend's practice.
Andrew, plan ahead for practice and have everything ready that your team·will need. Please keep me posted·on what you accomplish and what questions still need to be solved. I'm excited to hear all about it!
Good luck, and aim high,
Mr. Kibler
(PS ...and if you think it's hot in New Hampshire, see the attached pictures of where I've been "camping" the last three weeks during Ohio's summer heat. The fossil hunting is excellent here!)
Andrew (ARLISS) said...
Justin Abbott said...
Andrew,
At this upcoming team meeting, is it possible that we are gonna split up with some of us working to mount the A/D converter and some of us working to mount the CO2 sensor? Still, we would need to discuss it.
Justin
Justin,
This is not one of my top priorities, as I stated previously. We can definitely discuss it and work with what resources we can, but we probably can't physically build anything at the upcoming meeting since Mr. Kibler has the ASP-2 with him in Ohio. I only have a mockup. Also, I don't believe this is an area where we want to "split up," since the two components are closely related and will likely be mounted to the same place. Nonetheless, we have plenty of other important tasks to complete at the next meeting.
We are so sorry for the overdue work, but we met today at Andrew's house and we have really accomplished a lot (including the very overdue work). We are having an extended practice from 12:00-4:30. We think we MIGHT have figured out the constant today. Although we might not have the correct constant, we now have a much better understanding of how to get it. Justin looked at the sensitivity curve graph and took the first CO2 point (400 ppm CO2=324 mV) and divided them. Therefore, to get 1 ppm of CO2 he divided 400 by 400 and 324 by 400 (this because what you do to one side you have to do to the other). We then found out 1 ppm of CO2 equals 0.81 mV. Then we switched the equation around. This time we divided 324 by 324 and 400 by 324 to find out how much CO2 ppm there was at 1 mV. We then figured out the magic number (we think) -- 1 mV equals 1.234567901 ppm of CO2. Does this seem right?
~ The Rocketeers
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ "Learn to obey before you command."
·· From the sound of the e-mail I got from Andrew about today's practice, it sounds like you got a lot accomplished. Good work!··
I understand that·you got his clone of the ASP-2 working and producing CO2 data. Is that correct? If so,·then we're in a great position for our next practice (August 14th?) when I return! I'd love to see the data and the graphs that you're getting so we can compare them to Dr. Allen's and mine.
·· After thinking things through I deciphered that I started the·CO2 sensor·at 1000 mV rather than adjusting it to 1000 mV after the heater had warmed up for 2000-2500 seconds as in your graph. Would this cause the discrepency?
When I run the next trial I'll adjust the potientiometer to 500 mV at start-up, then wait 2500 seconds for warm-up, then adjust to 1000 mV. What are your thoughts?
Based on·an e-mail I just got·from Andrew it sounds like the team had a really productive practice today (see Mike's post above.) Andrew said they "got his ASP-2 clone running perfectly" and I'm eager to hear exactly what that means...! I'm very, very pleased with the team; they are· by far the most involved yet. Please be patient and·stick with us! With your help, we'll figure it out.
Hi Mark,
The CO2 signal (blue line above), is not affected by the potentiometer (red line). The potentiometer sets the level for the alarm LED to turn on. It might appear from the timing in the graph that somehow the potentiometer setting affected the CO2 reading, but that was simply a coincidence. I happened to be there with a screwdriver in hand and decided to fiddle with the potentiometer so that I'd be able to make the alarm LED come on when I breathed on the sensor Subsequently I did breath on the sensor and at one point with much effort and huffing and puffing I succeeded in making the blue line drop below the red line -- LED on.
So, you don't have to follow any special procedure with the potentiometer setting. The potentiometer reading is there mostly as a check that the MCP3202 ADC is working as it should. Now we just need a good rocketeer CO2 graph, and then move on to the math!!
Good luck with the fossil hunting, and bearing the heat!
Mike, and Justin, good work thinking about the elusive constant. You picked the point at 400ppm, 324mV. Why that one? If you pick the point at 6000ppm, 274mV, you would get a very different answer, right? But you are definitely on the right track to find the slope, rise over run.
When you say the change is so many mV per ppm, that is the slope of the line on the graph. Or, to put it another way, that is how much it changes in its y coordinate, divided by how much it changes in its x coordinate. Here:
Lets look at a small piece of the graph. You have one point at (x1,y1) = (400,324). Another nearby is (x2,y2) = (500,318). Here is how to calculate the slope of the line...
or if you invert that (like you did in your try), you come out with
slope2 = 1 / -0.06 = -16.6667 ppm / mV
Do you know the equation for a line? We want an equation for the line that connects the point (400,324) to the point (500,318). We just found the slope. Here is an easy way to write the equation of the line segment:
ppm - 500 = -16.6667 * (mV - 318)
Now, plug in the values from the known point, (400,324) and (500,318), and see if they satisfy the equation. That is, the same number on both sides. Like 2 = 2, NOT 2 =1. If so, they are on the line. You can juggle that around by moving the 500 to the right hand side and combining terms.
ppm = -16.6667 * mV + 5800
Is that right? Plug in mV=318 and mV=324, and see if you get ppm=500 and ppm=400 as the answers.
Can you plug in 277mV and get anywhere near the answer of 5000ppm that you can read off the graph? Why?
Now, here is the homework. Do the same thing, but now for the values when ppm=4000, mV=281, and the close by point, ppm=5000, mV=277. Do you get the same slope and same equation as above?
Comments
Again, I think it's time to sit tight and wait for Tracy to help out again, but it wouldn't hurt to look over that page carefully. I am out of my league here - I understand the basic concept, but I've never done this stuff.
·· I was hoping that we (you Rocketeers!) could simply add·new programming code to the current program to convert·the millivoltage to CO2 ppm. I understand that this is why we to·calculate the 'constant' Dr. Allen and Sylvie mentioned several times before.·Think of it like this: Pretend the CO2 sensor reads out·688 mV when it's exposed to 320 ppm of CO2. What number would you multiply 688 by in order to get 320? In this simple, make-believe example, X would be the constant.
688 * X = 320
We would insert that constant into the computer program to make it calculate 688 * X to read out 320 (ppm) instead of 688 (mV). But it's more complex than that because, as Sylvie just said above, as CO2 concentration changes the relationship isn't linear. That means it·doesn't form a simple,·straight line. Look at the very first graph that Dr. Allen sent you during the first days of the project (on page 2 of the forum I think.) Look at the graph·line around 200-500 ppm or so.·It isn't straight the whole way.
A second thing we (you Rocekteers!) need to figure out is how to calibrate the CO2 sensor. Someone mentioned putting it in a "pure helium" environment (I don't fully understand why), another person suggested a pure CO2 environment. We need to·put the CO2 sensor into a known·concentration of CO2 (1,000 ppm for exaxmple) so we know what the millivoltage reading is for 1,000 ppm, etc. of CO2. Does that make sense?
Sylvie and Tracy, am I on track here...? Is my explanation theoretically correct? What task(s) do you suggest the Rocketeers·work on·during this coming weekend's 4-1/2 practice to make the most of their limited time together? Thanks and ciao!
Keep up the good work, Dylan! You have a great future waiting for you.
From Ohio,
Mr. Kibler
Yes, I'm as eager as you are to hear from Tracy.·The answer is·in the constant and the code, I think. And getting a baseline CO2 reading. How would you go about doing that Paul? Put dry ice in a fume hood and then measure millivoltage? I don't understand the thing about putting the sensor in a "pure helium" environment to get baseline/calibration millivoltage. Why helium? Did I miss the simple explanation?
Mark
Sorry for not being on the forum lately. I just didn't really understand how to do the first method, and the second method (as Mr. Kibler said) would be hard to do because the relationship between CO2 concentrations isn't linear. Dylan helped me to understand the first method in a more simpler way (thanks Dylan!). But the actual formula to convert it still eludes me. Like Sylvie said, I think this will all become more clear when Dr. Allen explains it, since he has a lot to teach us about this concept.
Justin
I'm back and am happy to see that you are thinking things through carefully about how to do the data processing to convert mV to ppm.
But first, do we have any data to process?
Compare my response graph with yours. At least the most recent one I could find. Are there others?
mine:
yours:
You can find those and associated comments earlier in this thread on page 22 and page 23 respectively.
You say you applied CO2 at 1:00 and then again at 1:30 and at 1:35. The CO2 should cause the mV to decrease. See the clear signal in my graph? If there is a signal in your graph, it would take no end of signal processing to extract it. The rule of thumb is, if your brain can't see it, don't expect the BASIC Stamp to find it!!!
I am puzzled by the low level of the mV reading in your graph. (about 0.45V versus 1.4V that I saw) That makes me suspicious that the heater is not getting 6.0 +/- 0.1 Volt. That would explain the lack of response to CO2. I know you have a multimeter. Please measure both the heater voltage and the output from TP1.
-- Is the heater getting 6.0 Volts?
-- Is the mV reading at TP1 the same on both the multimeter and on the STAMP debug display?
-- Do you have more than one CO2 sensor (in case one is damaged)?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Post Edited (Tracy Allen) : 7/21/2010 4:21:17 PM GMT
·· Welcome back! I hope you had a good time in Chicago and Michigan. I hope it wasn't as hot there as it has been here in southeasyern Ohio... 10 days of 90+ degrees, with humidity to match the temperature.
You read my mind and I'm not sure why it didn't occur to me at the time (in Maine two weeks ago.) It came to me about two days ago as I was driving down the road. Of course! The millivoltage should decrease when CO2 is applied, not increase. This has me puzzled now. and so I'll run another series of trials to see·if I can·get a "dip" in the graph like we see in yours. I'll also upload the data (*the forum doesn't like Excel.)·I'll also check to see that the heater is on; last time it was hotter than summer but I'll check the voltage just to be sure.
The Rocketeers are really sticking with it, a positive thing because much of this is over their heads (and mine, and Sylvie's, so he says. I think he's downplaying his brilliance.) The meet this weekend for 4-1/2 hours at Andrew's house and I'm curious what will result from their meeting. Some excellent learning is going on and not all of it can be quantified. But they're sticking with it; that's the important thing.
Mark
You can delete your own duplicate posts, if you click on the little X at the upper right hand corner of its window, next to the editing pencil.
According to the MG811 CO2 sensor documentation, the raw sensor output would be around 320 mV at ambient CO2 concentrations, and that amplified times 6 2/3 by the Parallax interface circuit would bring it up to around 2.1 Volts where it goes to TP1 and to our analog-to-digital converter. However, at ambient, I am seeing around 1.4 Volts and you are seeing around 0.45 Volts. There may be considerable variation among MG811 sensors in what that resting level is, and also perhaps in the sensitivity factor. But 0.45 Volts is way way too low.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Post Edited (Tracy Allen) : 7/21/2010 8:14:00 PM GMT
·· I just learned something new. I deleted the triplicate posts, etc. (where did the 1's and 0's go, Sylvie? Do they still exist at all...?)
Mark
I understand it, most of it anyways. I'll go to you or Dylan if I need something clearified in the future... Sorry so late on things.
See you all soon,
Mike
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"Learn to obey before you command."
-Solon
I see I'm getting more unwarranted credit ("brilliance"? Nope. I'm an intermediate beginner).
Okay, they need to put the CO2 sensor on sale, so I can play around with this as well and see what kind of signals I get.
Re. Excel and the forum, you can always put the graph on the screen in Excel, then use the Windows Snipping Tool to grab the graph and put it into a .jpg file. That's what I did above.
Now, I'm surprised to see 1500 mV values in Tracy's graph for what I assume is regular air. All of the graphs on the sensor's datasheet go no higher than 400 mV. Granted, they're not displaying the kind of CO2 content you expect to see in air, but I wouldn't have expected it to go that high. Is that the raw reading from the sensor, or is it boosted·by that fancy "buffer" circuitry? I believe it's the latter, and that we've already talked about that earlier in the thread. I'll go back and look for it now.
Ah yes, there it is on page 15, where Tracy writes "Any Voltage that you measure on the test point, TP1, will be 7.67 times the actual output of the CO2 sensor". So that 1500 mV reading really represents a sensor reading of about 195 mV ( = 1500/7.67) from the sensor, leaving me even more confused, as I'd already concluded that the voltage from the sensor probably never goes below 250 mV.
I'm less brilliant than I was even a few minutes ago, and I'm working my way backwards quickly.
Post Edited (sylvie369) : 7/22/2010 12:51:05 AM GMT
Is 1.4 volts considered the average voltage at ambient? How do we bump up our voltage from being 0.45 volts to around your voltage (1.4 volts) if not exactly that?
Thank you all for your time and effort,
Sean
I'm not sure why the output from your sensor is 0.45 Volt instead of something higher. I am suspicious that there is still something wrong with the wiring of the circuit, that somehow it is not getting enough power to its heater for example. I asked whomever has one of the sensor setups and a multimeter to test of the heater voltage, or at least the power supply voltage for the sensor, and also to test to be sure that the voltage on TP1 matches what you are reading with the PBASIC program. It is also possible that the sensor is defective or has been toasted somewhere along the way.
Not only is 0.45 Volt lower than expected, there is no real evidence for a response to CO2. The signal is very noisy, bouncing up and down, but look at it in the graph posted earlier. Do you see any evidence of a response to CO2 at 1pm or at 1:30 pm? If not, no amount of boost to bump up the voltage to 1.4 is going to help. There will be natural variability from one sensor to another of this type, but we have not tested enough of them successfully yet to fathom the extent of that variability. How many CO2 sensors do we have? I have one here. Sylvie needs one.
Sylvie is right about the boost inside the Parallax circuit. It amplifies the raw output of the sensor by a factor of 7.67. So ambient air should give about 2 to 3 volts output from TP1.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Is your graph's data composed of reading from TP1? If so, then Sylvie's post makes sense!
If you did take your reading from a Multimeter attached to TP 1 this could prove our problem. Also, for Mr. Kibler's graph, I see an increase at both the times of 1:00:00 and 1:30:00. Since it is suppose Decrease in mV when more CO2 is applied, this could be due to maybe backwards wiring or a small program error. I do not want to take anymore time troubleshooting at this stage. The reason is that I believe that no one wants to end up doing this during our practice time. For those times could best be applied for building.
Dylan Landry
Or could we possible make another program which could insert the mA into a formula, and it would be able to calculate the constant on the fly, then multiply the mA by it to get ppm? Or would this by chance delay readings?
Do you have a BS2e? That is one that has 8 program slots. I am sure you have a lot of things to build, and building is always more fun that troubleshooting! Troubleshooting can be extremely time consuming and requires patience and analytical thinking, and tools, and understanding of the underlying principles. It is great when you isolate the source of the trouble, but painful when it just goes on and on with no resolution. Then, is there a backup plan?
What all do you need to build at your Saturday meeting? Is everything else ready for liftoff?
In my way of thinking, the conversion from mV to ppm is secondary. It would be nice to have PBASIC do it in real time, nevertheless, you can easily do that in EXCEL after the flight. It is much more important to be sure the sensor is responding as it should, and that a valid raw reading is going into the datalogger file!!!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
· Saturday's pratice is at Andrew's house and all the Rocketeers will be there... except me. I won't return from Ohio until Wednesday or Thursday of next week. This complicates things and bottlenecks us a bit; that's an unfortunate reality.
Andrew has everything the team needs to·get the ASP-2 working on Saturday and it has a BS2e stamp on it.·Andrew also has a CO2 sensor (we have three total) and everything else he needs to "clone" the ASP-2 prototype that I have.·By now he should have in working and running in tandem with the one I have here with me.
Like you and Sylvie, I suspect that there's still a problem with the wiring. I'll go over it with a fine-tooth comb when I return (but darn it! I thought I had it dialed in befire I left Maine for Ohio.)
More later . battery dying,
Mark
·· OK, I'm back. I found a power supply, with 4% battery supply left. Also, I nearly fried the 'NASA' computer Andrew, like I did when I was at Marshall Space Flight Center. Remember when·I put it·in the double-insulated black carrying case without making sure it powered down?·I did the same thing today and it was in the case, in the van, in the·90 degree heat for 1-1/2 hours... Close call!
Like I was saying Tracy, I suspect that it's still an issue with the wiring. But Christopher and I spent literally·hours-- days in fact-- with it before I left Maine, cross-checking all your diagrams, the programming, etc.·I thought we had it rigth before I left for Ohio, but like I said a few posts back, it occured to me a few days ago that the numbers should not have incraased when we added CO2. It's back to square one with the wiring and it reminds me of your comment to Dylan:
"Troubleshooting can be extremely time consuming and requires patience and analytical thinking, and tools, and understanding of the underlying principles. It is great when you isolate the source of the trouble, but painful when it just goes on and on with no resolution. Then, is there a backup plan?
Not fully understanding that interplay between all the underlying principles is part of the problem on my end. And yes, it is (has been!) painful to go on and on without resolution. But we'll stick with it like diligent scientists. We'll figure it out, isolate the variable, and make it work. That is my commitment and hope.
Our backup plan? Launch last year's ASP (ugh!), which I'd really rather not do. We owe it to those who have helped make the trip to Nevada in September possible-- but most importantly to ourselves-- to figure it out and make it work. That's what aerospace engineers, and so that is what we will do.
Andrew, please conatct me by e-mail and let me know what the operational status is of our ASP-2 clone, 'Phipdippides'. I hope he plans to make an appearance at this weekend's practice!
Mr. (Mark) Kibler
Dr. Allen,
I'm not sure what Dylan's situation is, but both Mr. Kibler and I have BS2e's on our ASP-2 platforms. I don't believe we have too much building left to do. Mounting the A/D converter and CO2 sensor comes to mind, but I'm not quite sure how we will tackle that -- perhaps a good topic for this weekend's team meeting?
I can relate to what you said about troubleshooting. As you know, we've definitely been there and done that before. However, it certainly feels great once you "cross the finish line," so to speak.
Andrew
At this upcoming team meeting, is it possible that we are gonna split up with some of us working to mount the A/D converter and some of us working to mount the CO2 sensor? Still, we would need to discuss it.
Justin
Justin,
This is not one of my top priorities, as I stated previously. We can definitely discuss it and work with what resources we can, but we probably can't physically build anything at the upcoming meeting since Mr. Kibler has the ASP-2 with him in Ohio. I only have a mockup. Also, I don't believe this is an area where we want to "split up," since the two components are closely related and will likely be mounted to the same place. Nonetheless, we have plenty of other important tasks to complete at the next meeting.
Andrew
What do you have in your mockup? Is it everything that is on the ASP right now?
Sean
· I agree with Justin. I think this weekend's practice time will be well spent working with Phidippides and trying to get him to do what the ASP-2 and my clone are supposed to do. Andrew·has all the necessary components and virtually all the information you need has been posted to the forum. This includes wiring schematics, the program code, pictures of wiring, explanations, etc.
If you·make this·your sole focus·you'll be amazed how fast 4-1/2 hours will go by! Trust me. It will give you a (very)·full taste of what the project is·really about... the problems, the problem solving skills, the analytic thinking that Dr. Allen mentioned, and ultimately the thrill of figuring out how to make the ASP-2 work. This is what I would like to see you accomplish because I want·you to have pride and ownership for the completed project.
Be sure to schedule August's practice when you meet this weekend. I propose·August 14th or 15th at Dunbarton Space Flight Center (my house) from 1:00 - 4:30 PM. After that, because·school will be starting and Chritsopher will be headed off to college,·we will probably only meet twice before we head into the desert to launch the ASP-2. Our primary mission goal is to launch the ASP-2 fully operational, e.g. with the CO2 sensor and all·of last year's·systems fully functional.·Please define, discuss, and then list your mission goals sometime during·this weekend's practice.
Andrew, plan ahead for practice and have everything ready that your team·will need. Please keep me posted·on what you accomplish and what questions still need to be solved. I'm excited to hear all about it!
Good luck, and aim high,
Mr. Kibler
(PS ...and if you think it's hot in New Hampshire, see the attached pictures of where I've been "camping" the last three weeks during Ohio's summer heat. The fossil hunting is excellent here!)
Sean,
That's correct. There are (theoretically) no hardware differences between the ASP-2 mockup and the real thing.
Andrew
We are so sorry for the overdue work, but we met today at Andrew's house and we have really accomplished a lot (including the very overdue work). We are having an extended practice from 12:00-4:30. We think we MIGHT have figured out the constant today. Although we might not have the correct constant, we now have a much better understanding of how to get it. Justin looked at the sensitivity curve graph and took the first CO2 point (400 ppm CO2=324 mV) and divided them. Therefore, to get 1 ppm of CO2 he divided 400 by 400 and 324 by 400 (this because what you do to one side you have to do to the other). We then found out 1 ppm of CO2 equals 0.81 mV. Then we switched the equation around. This time we divided 324 by 324 and 400 by 324 to find out how much CO2 ppm there was at 1 mV. We then figured out the magic number (we think) -- 1 mV equals 1.234567901 ppm of CO2. Does this seem right?
~ The Rocketeers
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"Learn to obey before you command."
-Solon
·· From the sound of the e-mail I got from Andrew about today's practice, it sounds like you got a lot accomplished. Good work!··
I understand that·you got his clone of the ASP-2 working and producing CO2 data. Is that correct? If so,·then we're in a great position for our next practice (August 14th?) when I return! I'd love to see the data and the graphs that you're getting so we can compare them to Dr. Allen's and mine.
Way to go team!
Mr. Kibler
·· After thinking things through I deciphered that I started the·CO2 sensor·at 1000 mV rather than adjusting it to 1000 mV after the heater had warmed up for 2000-2500 seconds as in your graph. Would this cause the discrepency?
When I run the next trial I'll adjust the potientiometer to 500 mV at start-up, then wait 2500 seconds for warm-up, then adjust to 1000 mV. What are your thoughts?
Based on·an e-mail I just got·from Andrew it sounds like the team had a really productive practice today (see Mike's post above.) Andrew said they "got his ASP-2 clone running perfectly" and I'm eager to hear exactly what that means...! I'm very, very pleased with the team; they are· by far the most involved yet. Please be patient and·stick with us! With your help, we'll figure it out.
Regards,
Mark
The CO2 signal (blue line above), is not affected by the potentiometer (red line). The potentiometer sets the level for the alarm LED to turn on. It might appear from the timing in the graph that somehow the potentiometer setting affected the CO2 reading, but that was simply a coincidence. I happened to be there with a screwdriver in hand and decided to fiddle with the potentiometer so that I'd be able to make the alarm LED come on when I breathed on the sensor Subsequently I did breath on the sensor and at one point with much effort and huffing and puffing I succeeded in making the blue line drop below the red line -- LED on.
So, you don't have to follow any special procedure with the potentiometer setting. The potentiometer reading is there mostly as a check that the MCP3202 ADC is working as it should. Now we just need a good rocketeer CO2 graph, and then move on to the math!!
Good luck with the fossil hunting, and bearing the heat!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
When you say the change is so many mV per ppm, that is the slope of the line on the graph. Or, to put it another way, that is how much it changes in its y coordinate, divided by how much it changes in its x coordinate. Here:
Lets look at a small piece of the graph. You have one point at (x1,y1) = (400,324). Another nearby is (x2,y2) = (500,318). Here is how to calculate the slope of the line...
or if you invert that (like you did in your try), you come out with
Do you know the equation for a line? We want an equation for the line that connects the point (400,324) to the point (500,318). We just found the slope. Here is an easy way to write the equation of the line segment:
Now, plug in the values from the known point, (400,324) and (500,318), and see if they satisfy the equation. That is, the same number on both sides. Like 2 = 2, NOT 2 =1. If so, they are on the line. You can juggle that around by moving the 500 to the right hand side and combining terms.
Is that right? Plug in mV=318 and mV=324, and see if you get ppm=500 and ppm=400 as the answers.
Can you plug in 277mV and get anywhere near the answer of 5000ppm that you can read off the graph? Why?
Now, here is the homework. Do the same thing, but now for the values when ppm=4000, mV=281, and the close by point, ppm=5000, mV=277. Do you get the same slope and same equation as above?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Post Edited (Tracy Allen) : 7/25/2010 4:55:43 AM GMT