Jump to content

As of July 17, 2015, the LabJack forums here at forums.labjack.com are shut down. New registrations, topics, and replies are disabled. All forums are in a read-only state for archive purposes.

Please visit our current forums at labjack.com/forums to view and make new posts. To post on the current forums, use your labjack.com login account. Your old LabJack forums login credentials have been retired. There are no longer separate logins for labjack.com and LabJack forums.


Photo

Noise when more than 1 signal is connected (Easy stream example)


  • Please log in to reply
9 replies to this topic

#1 user1979

user1979
  • Members
  • 62 posts

Posted 16 January 2013 - 09:46 AM

Hello, I am using the U6 Easy stream example with Labview and a U6 Labjack module. My purpose is to read V1, V2, and V3 with the AI0, AI1, and AI2 channels. When connecting +V1 to AI0 and the ground of the circuit to the GND Pin I dont get any noise. When I add +V2 to AI1 I get something like 20mV noise (wether I read AI0 on its own or AI0&AI1). I get the same behaviour if I do this with the other Analog inputs and also if use only V2&V3 or V1&V3. V1, V2 and V3 have no noise when only each of them is connected on its own. V1,V2,and V3 are signals coming from 3 temperature sensors (attached file). Do you have an idea what is the problem? Thanks a lot, User

Attached Files



#2 LabJack Support

LabJack Support
  • Admin
  • 8677 posts

Posted 16 January 2013 - 10:02 AM

One test is to see if it is a general problem, or unique to these sensors in your setup. Try some other signals in your stream to see if they are affected: -Tied to GND. -1.5V AA battery or similar. -Different temperature sensor such as LM34 (labjack.com/support/app-notes/temperature-sensors). Do you see the problem in command/response mode (try LJLogUD), or just in stream mode (can test with LJStreamUD in addition to the VIs)?

#3 user1979

user1979
  • Members
  • 62 posts

Posted 17 January 2013 - 04:40 AM

Thanks for your help.

One test is to see if it is a general problem, or unique to these sensors in your setup. Try some other signals in your stream to see if they are affected:

-Tied to GND.
---------->The problem doesn't occur
-1.5V AA battery or similar.
---------->The problem doesn't occur
-Different temperature sensor such as LM34 (labjack.com/support/app-notes/temperature-sensors).
---------->I have tried with 2 pressure sensors (attached file n1). The reading is correct (attached file n2)

Do you see the problem in command/response mode (try LJLogUD), or just in stream mode (can test with LJStreamUD in addition to the VIs)?


---------->When using LJLogUD, I get attached file n3 : both signals seem correct. When disconecting one of them the other stays the same. From what I understand in the program it only measures one point where as the Easy stream example takes 1000 Samples on each of the lines.

Attached file n4 shows my problem. As the matter of fact it is not a noise problem but the dropping of the voltage you can see on the graph (mainly on channel 1(red curve)). When I disconnect sensor 1 on channel 0, the red curve gets correct (attached file n5) and the data changes from 0.97V (sensor1 connected on channel0 and sensor 2 on channel 1) to 0.93V (when only sensor2 is connected)

User

Attached Thumbnails

  • 2.PNG
  • 3.PNG
  • 4.PNG
  • 5.PNG

Attached Files

  • Attached File  1.pdf   77.36KB   722 downloads


#4 user1979

user1979
  • Members
  • 62 posts

Posted 17 January 2013 - 06:48 AM

I have made some more tests. - When I have this setting the measure using Easy stream example works fine when only 1 channel is scanned (no noise no offset) : each of the 2 sensors have their own 5V reference component (but same +12V) Both sensors are connected to the U6 When disconnecting one of the sensors the value doesn't change at all (one channel scanned) (When scanning 2 channels I get more noise and an offset the second channel) - When I have this setting the measure using Easy stream example I get some noise : Same 5V reference component for both sensors Both sensors are connected to the U6 If 1 channel is scanned I get some noise (no offset) but the average seems correct (If both channels are scanned I get noise and offset on the second channel) One solution could be to a 5V reference component for each sensor and launch the the Easy stream example, set to one channel, for each channel. But I would've liked to be able to use the example for all the cahnnels at the same time.

#5 LabJack Support

LabJack Support
  • Admin
  • 8677 posts

Posted 17 January 2013 - 02:53 PM

I am wondering if the output of the sensors and/or the power supply to the sensors is weak and slow, and the sampling in stream mode causes a droop? Have you tried using an oscilloscope to look at the signal from AIN1 to GND to see if the scope shows the same thing? In your third image (image #1125 according to the URL when you click on it), it looks like AIN1 has negative spikes every 25 seconds, and AIN0 has smaller spikes. Is this from LJStreamUD? What sort of settings? Why are you using stream mode anyway? The time constant of these sensors is over 1 second, so no reason to collect data so fast. Have you tried just using VS for power supply, rather than your circuit with 12V/resistor/zener? Or if you want something more stable than VS, set DAC0 to 4.0V and use that as the power supply. I would try just using VS first, though.

#6 user1979

user1979
  • Members
  • 62 posts

Posted 18 January 2013 - 05:53 AM

I am using the easy stream example only because it was the example I got to first and also because usely I take 1000 samples of my signal at 1kHz to make an average, these are the only reasons why (infact I only need one final value per minute... but I would like to solve the problem though). I have tried with the 5V power pin of the Labjack and I have the save problem. Concerning the negative spikes, the print screen was taken when using the easy stream mode (Except if I have made a mistake the spike hapen roughly every 25 or 20 ms (40- 50 Hz) ( the setting is the default setting of the example). When adding a 10K resistance between AI0 and Ground and another 10K between AI1 and Ground, and using the 5V pin of the Labjeck I get very good measure ("aa" attached file). I have made an average of the 1000 samples on each line and both lines give the same value as the ones I get when I use a voltmeter and touch the pins of the Labjack When using these 10K resistances with my original circuit (but with two 5V references) I get attached file "bb". Which is by far the best I have had up to now when measuring 2 channels. The 70 mV offset I use to see when I was measuring 2 channels has disappeard since I have added those 2 resistances. The 1000 samples average gives the same value as the volmeter on both channels. So the only problem left is the small negative spikes (that I don't have with the 5V power pin of the labjack). I am willing to use the 5V reference solution to really have a stable power and to be close to the conditions the temperature sesnor is used in our products. Do you of an idea concerning the spikes? The offset comes back if I use only one 5V reference. Thansk a lot, User

Attached Thumbnails

  • aa.PNG
  • bb.PNG


#7 LabJack Support

LabJack Support
  • Admin
  • 8677 posts

Posted 18 January 2013 - 11:55 AM

My best guess is that the output of the sensors and/or the power supply to the sensors is too weak, and the sampling in stream mode causes a droop in one or both? If you want us to investigate it further, I suggest you send us some parts (resisters, reference zener, temp sensors) so we can replicate the spikes here at LabJack. Contact [email protected] and we will give you an RMA# if you want to send us something.

#8 user1979

user1979
  • Members
  • 62 posts

Posted 28 January 2013 - 01:54 AM

Thanks for offering testing the sensor but I won't have time to wait for the result. I will just measure each line seperatly since this isn't a problem in this particular project. As a matter of intrest, concerning the negative peaks that appear every 20-25 ms (or 20-25 Samples since I am recording 1000S at 1kHz) could that be linked to the fact that the samples may be recorded every 20-25 samples before changing inputs and thus generating an "adapting time" for the signal ?

#9 LabJack Support

LabJack Support
  • Admin
  • 8677 posts

Posted 28 January 2013 - 10:59 AM

Perhaps you get a spike every 25 samples, because the default for LJ_chSTREAM_SAMPLES_PER_PACKET is 25?

http://labjack.com/s...ers-guide/4.3.7

That means the U6 sends out a USB packet every 25 samples. When it sounds out this USB packet there is a spike of extra supply current, and this could cause the VS supply to droop. We know this does not affect measurements in general (per your results in post #3), so not sure why it might affect your particular signals.

#10 user1979

user1979
  • Members
  • 62 posts

Posted 28 January 2013 - 02:48 PM

ok. Thanks again for your help. Best regards, User


0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users