Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
10/12/10 06:09
Read: times


 
#179057 - RE: Resolution is not there...
Responding to: ???'s previous message
Thanks all for replies.

Here is what customer says "I need my meter reading to match (within +/- 0.1degrees) with a standard meter like fluke" :-( This is required using a thermistor... Thermistor has at least 5% tolerance and other effects like self-heating. But a thermistor is must for cost reasons. (as always customer wants a lowest cost rocket science)....

Per, Dallas one-wire sensors are still quite expensive than thermistors.
Michael,
Even my first reaction was that it is not possible to accommodate 1410 steps with 10bit ADC. But I wanted to confirm if I am missing something..
Since, cost is a major constraint here, changing resolution to 0.5 at least should be the way out.
Using a silicon diode/B-E junction also seems a good option. Have you used that sometime? In this case, constant current by using a precision voltage reference and a resistor should be enough?

Neil,
Will it help to increase resolution by linearizing?

Another option suggested somewhere was to use oversampling. Has anyone tried that?

List of 24 messages in thread
TopicAuthorDate
Measure temperature @0.1deg resolution using thermistor            01/01/70 00:00      
   Sorry - doesn't work            01/01/70 00:00      
      Thermistors aren't such bad...            01/01/70 00:00      
         Remember the original requirement of 0.1°C            01/01/70 00:00      
   Resolution is not there...            01/01/70 00:00      
      What about the Termistor            01/01/70 00:00      
      RE: Resolution is not there...            01/01/70 00:00      
         one-wire sensors are still quite expensive than thermistors            01/01/70 00:00      
            Thermistor Vs one-wire sensor cost            01/01/70 00:00      
               Your processor has a voltage reference good enough?            01/01/70 00:00      
                  Voltage reference            01/01/70 00:00      
                     That voltage reference has large error...            01/01/70 00:00      
                  Vref is Not Needed            01/01/70 00:00      
                     Depends            01/01/70 00:00      
                        Ah but,            01/01/70 00:00      
                           Still not trivial            01/01/70 00:00      
         The normal fluke doesn't give 0.1C correct readings            01/01/70 00:00      
         I think your customer ...            01/01/70 00:00      
            quite common problem with resolution/precision            01/01/70 00:00      
         linearing only solves one issue            01/01/70 00:00      
            Precision            01/01/70 00:00      
               hm            01/01/70 00:00      
               Not an opamp            01/01/70 00:00      
               Comparison with Fluke a big problem            01/01/70 00:00      

Back to Subject List