 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Microcontroller and PC projects : Temp LM335 voltage fluctuations
  Page 2 of 7   |
|||||
| Author | Message | ||||
pcaffalldavis Senior Member  Joined: 17/10/2011 Location: United StatesPosts: 187 |
Certainly twofingers, I had to stop the program to get the Maximite voltages but here they are: I have Outdoor sensor on Pin(2) The Boiler sensor is in Pin(3) Outdoor Multimeter on Breadboard pin to MM Pin(2): reads 2.71 volts Control C/Print Pin(2): Maximite display reports 2.72959 Boiler Voltage divided Circuit Multimeter on Breadboard pin to MM Pin(3): reads 2.76 to 2.77 volts Control C/Print Pin(3): Maximite display reports 2.79243 I guess there is a discrepancy that shows either my multi-meter is reading about .02 volts low, or maybe the Maximite is reading .02 volts high? We're all here 'cause we're not all there. |
||||
twofingers Guru  Joined: 02/06/2014 Location: GermanyPosts: 1593 |
Are the readings constant? The discrepancy between the multimeter and the MM is not relevant at this moment. Ignore it! causality ≠correlation ≠coincidence |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1593 |
I would say: Just connect the shield and the ground (GND) of the micro.  causality ≠correlation ≠coincidence |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
The meter shows a bit of fluctuation on the Boiler/divided circuit; 2.76-2.77, but the meter shows a steady 2.71 volts on the outdoor circuit pin(2). I'm not displaying pins as voltages on the Maximite display. Should I do that so I can see the fluctuation while it is running the program? We're all here 'cause we're not all there. |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1593 |
No need for that! We only need to see if the values fluctuate and compare this with the multimeter values. If we see one value (Maximite) is steady and the other (multimeter) is not then we can say more. causality ≠correlation ≠coincidence |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
Okay, I puled the can cap on the Outdoor pin-board point where the sensor and Maximite pin(2) connects. I put a 1uF ceramic cap there instead, and across to ground. I didn't see any polarity for the ceramic cap, and boy is it small! After a reboot I saw Outdoor temp readings of O C to -2C in under one minute. We're all here 'cause we're not all there. |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1593 |
There is no polarity for the ceramic caps!  The Outdoor sensor is now in house? With 0°C ?  You know that breadboards contacts are not very reliable? Better is to solder. How about the multimeter readings? Are they steady? causality ≠correlation ≠coincidence |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
No, the outdoor is still outdoors. I wasn't watching the outdoor sensor voltage on the multi-meter connected to the bread-board Outdoor sensor pin when I saw the fluctuating Maximite Outdoor readings. I'll go do that now. And, yes, it is just at freezing this afternoon. I will say yahoo, 'cause that is an improvement! We're all here 'cause we're not all there. |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1593 |
It's nearly the same temperatur in Germany!   causality ≠correlation ≠coincidence |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
Testing update I just ran an 8 minute test. I like using the Outdoor sensor as it has the period High & Low constantly displayed on my screen. I also get the current/last loop Outdoor temp. I'm only sending the Maximite out to get sensor results every 50 program loops. This is about every 8 seconds. In the 8 minute boot test the pin(2) voltage was monitored using a multi-meter. I never saw it deviate from 2.74 volts the entire time. This is the breadboard pin I put the 1uF ceramic cap on with the other cap leg to ground. In 8 minutes The Low was .4 C and the High was 4.7 C For those in F land that was a Low of 32.7 and the High was 40.4 I've had the wood-fired Boiler program running the past six years, and for comparison, it reported a steady 36 F or 2.5 C the entire 8 minutes. I saw one example in the 8 minute test on this CGStick1 Maximite running 3.0 where two samples, 8 seconds apart, were 2.3 C, and 4.1 C. That is almost 2 C difference, 8 seconds apart. In F that is about 4 degrees. I'm just having a difficult time accepting that much variance in 8 seconds. Maybe it is okay? And maybe I just need to use a temperature average, but for the purposes of documenting the daily highs and lows (which I also do for the weather service) I just am having trouble bringing myself to believe these extremes are accurate. I suppose I could write a routine that removes extreme data points and keeps them out of the averages, but what a machination to go through. Or I suppose I could switch the immediate data points which are used to display the daily high and low temps as they are happening with a collected average high and a collected average low instead. These past few years I've liked using immediate temperature data points because when a current temp data point is the same as the daily high or low, it displays the number in reverse text. This draws attention to the screen as a sort of alert that we are currently setting a new high or low at that moment. Sorry if this is all too much. I just wish I could get these little buggers to stay within a more precise range like the ones on the Wood-fired boiler do. For the life of me I don't know what the difference is. I built/am building both, but they don't act the same. We're all here 'cause we're not all there. |
||||
TassyJim Guru Joined: 07/08/2011 Location: AustraliaPosts: 6283 |
If the LM335 voltage is steady but the MM gives a varying reading, I would look closely at the 3.3V on the analogue Vin pin. I would also update the Maximite firmware to a more recent version. 18 degrees C at 10am this morning. We are taking the dogs for a walk on the beach. Better than wrestling grizzlies. Jim VK7JH MMedit |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1593 |
Peter, if I understand you right, the multimeter shows steady a value of 2.74V? If you get this voltage direct from the LM335 - not in the middle of the voltage divider - then this 2.74V means you have 2740mV/10 °K = 274-273.15 = 0.85°C on your Outdoor sensor. Okay, if so then we are a step further.  This means your sensor is okay and your cable too! I think we should examinate the Maximite part of your setup. Can you replace the Maximite? Have you a backup Maximite or Micromite? ... Its late now in Germany. If you like we can continue tomorrow. Michael causality ≠correlation ≠coincidence |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
Thanks Tassy Jim, Do you mean put a meter on the Maximite 3.3 volt pin and monitor that while the program is running? I've never flashed a Maximite or a CGMMStick1 (if those even can be flashed) Who is a good source these days for a newer version of a MM? Beach sounds good, really good right now. I don't wrestle with um, just shoot um, but only if they think they own my house. Bears are like that though. They act like everything is theirs. I see why Russia had the bear as their symbol for years. We're all here 'cause we're not all there. |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
Thank you Michael twofingers, Yes, tomorrow will be another day. I don't have a voltage divider on the Outdoor or House sensor circuits, just the Boiler water temp circuit. I am currently trying this on a CGMMStick1 Maximite running 3.0. I also have a Maximite SD1 on the bench which seems to be exhibiting fluctuating readings too, but it takes a few seconds to connect up the pins from the breadboard so I'll collect similar test results later and post them while you sleep. I've never had a Micromite. No rush, ever. Sleep well Peter We're all here 'cause we're not all there. |
||||
| TassyJim Guru Joined: 07/08/2011 Location: AustraliaPosts: 6283 |
The latest firmware for the mono maximites (and all others) is on Geoff's website http://geoffg.net/maximite.html Scroll down to the bottom. The ZIP file includes the Windows bootloader program and instructions. It also has the correct manuals for the version of MMBasic. Make sure you get the one for the Monochrome Maximite. To put the CGMMSTICK into bootloader mode, you need to put a jumper on J3 while plugging it in. Details about the CGMMSTICK are on CG's website http://circuitgizmos.com/documentation/hardware-datasheets/cgmmstick1-technical-information/ The analog Vin (AVDD) is on pin 19 of the chip. You might be able to get a reading across C11 There is a 100 ohm resistor between the Analog and Digital 3.3V rails so reading the external 3.3 is not the best place to look at, but better than nothing. Internally, MMBasic takes 10 readings and discards the highest and lowest before averaging so high frequency noise should be filtered out but not mains hum etc. Jim VK7JH MMedit |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
Thank you Jim, I just took reading of CGMMStick1 at C11 while it was running my boiler prog. It fluctuates between 3.3 and 3.29 constantly. Might that be the problem? How steady should it read? Peter We're all here 'cause we're not all there. |
||||
| TassyJim Guru Joined: 07/08/2011 Location: AustraliaPosts: 6283 |
10mV is one degree C at the output of the LM335 At the reference, it would be similar as you are reading close to maximum voltage. You seem to be seeing a one degree error at most. The question is, how stable is your multimeter when probing little solder pads. Jim VK7JH MMedit |
||||
| pcaffalldavis Senior Member Joined: 17/10/2011 Location: United StatesPosts: 187 |
It is kind of difficult to hold a needle probe on that tiny solder pad for more than a few seconds. I wasn't on it long enough to see a large fluctuation on the temp sensor display. I would not have been able to look away anyway without losing my contact point. If I use clip-on tester leads on the 3.3 volt pin will that be good enough for a longer running test? We're all here 'cause we're not all there. |
||||
| isochronic Guru Joined: 21/01/2012 Location: AustraliaPosts: 689 |
Quick check of the multimeter - measure a cell voltage, say 1.456 v. Then reverse the leads and retest, should be -1.456 v. Any offset will show up as different numbers ie a non zero midpoint, and a offset needed. Then check the 3.3 v supplied to the sensor. -> steady 3.300 or close, multimeter probly is only 1 or 2 % accuracy anyway. Watch it a cro for noise. As a test, run the sensor itself off a cell/battery and see if the signal still fluctuates.Nrunning it off a pcb rail is going to be noisy and inaccurate. check current drawn by sensor. It should only be a few mA even when hot. Fluctuations should be very small. |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 5091 |
Dear pcaffalldavis, I hope you are not overwhelmed by all this advise. And to be honest, I am not sure I can give you any. But if you mention to see something like 40mV (4 degrees) variation on the ADC voltage there may be an other cause for the problem. And it may not be easy to locate. Let me try to explain. The Mono Maximite measure voltage referenced to the voltage present at the chip's AVdd (pin19) and AVss (guess pin 20). You have measured these voltage to be accurate (3.29V) and stable, but the sensor is attached to the IO connector. I am worried in your case about the voltage at the ground pin (AVss). The chip is mounted on a PCB, and the PCB connects the chips ground pins to the connector (even pin numbers) and the power supply (IC2 pin 1). The voltage across the PCB copper can present an error in the ADC reading. This voltage drop is depending on the current in the PCB copper. Is it possible (in your application) that the maximite powers relays, optocouplers, LED's, that may cause this variation in voltage drop (and thus cause the readings to be inaccurate). Most likely this comes into play if you have output pins that "pull low" i.e. a LED to +5V. Or by the VGA or PS2 keyboard, or a defective SD card. It can also be caused by an external ground loop (i.e. the +9V(12V) or USB power supply ground being connected to the sensor "-" in some way. I have a Mono Maximite myself, and noticed the thin copper traces for the power lines and ground. Maybe there is a relation. easiest to test is by connecting the maximite to ONLY the USB (providing power and serial communication), and the one problematic sensor. In that case ground loops cannot exist. You will have to use the terminal (tera term on windows) to check the values of the readings. I hope this does not confuse you more, or exceeds your skills. Just there to help. Regards, Volhout PicomiteVGA PETSCII ROBOTS |
||||
| Page 2 of 7 |
|||||
 Print this page |
 |
The Back Shed's forum code is written, and hosted, in Australia. | © JAQ Software 2025 |
