Page 1 of 1

Wheatstone bridge voltage drop: ME who doesnt know enough EE

Posted: Sun Jul 01, 2007 6:56 pm
by iae2101
I'm trying to figure out the resistance of a thermistor using a wheatstone bridge by measuring the Vout and entering that into a program i created that tells me the resistance. im using three 220ohm resistors and one 3kohm@25C thermistor. I set my voltage supply to 3.3V but as soon as i connect the circuit the voltage of the supply drops to 1.7 to 1.8 volts. does anyone have any suggestions to prevent this voltage drop such that my input voltage will stay fixed at 3.3V? Thanks!
-ian

Posted: Sun Jul 01, 2007 10:30 pm
by rshayes
There may be a wiring error. A bridge with four 220 ohm reisitors would look like a 220 ohm load to the power supply. At 3.3 volta, this is about 15 milliamp, which most supplies should br able to supply.

Re: Wheatstone bridge voltage drop: ME who doesnt know enoug

Posted: Mon Jul 02, 2007 1:06 pm
by Bob Scott
iae2101 wrote:I'm trying to figure out the resistance of a thermistor using a wheatstone bridge
Ian, is there a reason for using a wheatstone bridge instead of a DMM? Is it for educational purposes?
Isn't the wheatsone bridge from an earlier age of buggy whips and whale bone corsets?

Bob

Posted: Mon Jul 02, 2007 9:59 pm
by iae2101
my goal is to measure the temperature using an ADC. im building the temperature circuit that must use the LEAST amount of power and has the lowest cost possible. It will be powered off of batteries between 3V to 4.5 Volts. I have talked to friends who suggested using op amps, but considering they have to be powered, the bridge just seemed easier. Do you happen to have any suggestions? thanks!
-ian

Posted: Mon Jul 02, 2007 10:44 pm
by Chris Smith
Op amps can draw extremely low currents.

National Semi has a whole section of just temperature measurements,.. op amps,...Instrument amps,... that draw close to nothing.

100 micro amps draw with higher and lower figures available.

J/K Junction sensors, IC temp chips, and even Thermistors.

Posted: Tue Jul 03, 2007 1:11 am
by rshayes
The simplest arrangement would be a voltage divider using the thermistor as one resistor and a resistor as the other leg. If the A/D converter has an internal voltage reference, a small amount of current may be available to supply the voltage divider from this reference.

If the same reference voltage is used for the voltage divider and the A/D converter, the A/D converter will indicate the ratio between the thermistor and the resistor. This means that the reference voltage does not have to be accurately calibrated. Some A/D converters can use the power supply as the voltage reference.

The main problem is that the measured voltage will not be a linear function of temperature (this is a characteristic of the thermistor itself). This may require a microprocessor to read a look up table and possibly interpolate between entries to get the actual temperature.

There are assemblies using multiple thermistors that can give a linear voltage with temperature. These are somewhat more expensive that a single thermistor.

Posted: Sun Jul 08, 2007 9:26 am
by iae2101
I just checked out National Semi and saw they had some useful temperature devices. The only problem with it is in order to use them we would need several components that i think we could do without. The chip we are using has 4 single point ADC inputs and a few digital I/O. I was trying to avoid using additional ADCs on other components if it wasnt necessary. Also, the ADC has a fixed range of 0 to 1.2V.

So here is my senerio:
I have a wheatstone bridge working in the ranges i want it to with the output between 0 and 1.2V for my temperature range, its tested and works. Im powering it off of the device that im trying to integrate it to. with a Vcc of 3.3V. The bridge it powered between Vcc and ground. Im reading a voltage with my DMM off the bridge, Unfortanately, the ADC is a single point so im not reading the correct bits off the ADC because the Vout of bridge has a common ground (of 0) to the device, hence the ADC is not reading the voltage across the bridge, its measuring between ground and Vout+ of the bridge. Is there anyway to use the device to power my circuit but to have the ADC read the correct voltage?
Thanks again for all the help and advice!
-Ian

Posted: Sun Jul 08, 2007 1:46 pm
by Robert Reed
Have you thoght about inserting an Op-Amp with differential input and differential output between bridge and ADC ?

Posted: Sun Jul 08, 2007 1:50 pm
by Engineer1138
I suspect that you don't really want to go there, but you can use a diode as a temperature sensor. Each unit will require calibration, but that would be the lowest total cost solution.

What are your cost and accuracy targets? A National Semiconductor LM35 doesn't cost much and will give a stable 10mV/degree output. It's probably the easiest sensor there is on the market and requires no external components.

Posted: Sun Jul 08, 2007 3:04 pm
by Bob Scott
iae2101 wrote:my goal is to measure the temperature using an ADC. im building the temperature circuit that must use the LEAST amount of power and has the lowest cost possible. It will be powered off of batteries between 3V to 4.5 Volts. I have talked to friends who suggested using op amps, but considering they have to be powered, the bridge just seemed easier. Do you happen to have any suggestions? thanks!
-ian
You did not specify the temperature range or whether you need information in degrees F, C or Kelvin.

As was mentioned, National has its LM34, LM35, LM335, etc. Analog devices (www.analog.com) also has a WIDE range of temperature sensors that should meet any of your requirements, with analog or digital outputs.

ASIDE: I managed to get some AD590 samples from Analog, but for our application we ended up using the National LM35. The AD 590 reads degrees K, and outputs a tiny current, requiring use of very high precision load resistor to maintain accuracy.

Is "output" usable as a verb? :shock:

Bob

Posted: Sun Jul 08, 2007 8:23 pm
by philba
Bob Scott wrote: Is "output" usable as a verb? :shock:

Bob
hey, you're not a geek until you verb nouns and noun verbs. Though prepositioning adjectives is only for the truly geeked.

Posted: Sun Jul 08, 2007 9:24 pm
by Robert Reed
OUT PUT - Noun
PUT OUT - Verb

Posted: Sun Jul 08, 2007 9:33 pm
by iae2101
my temp range is approx -20C to 50C. i found this which i think will work perfectly, http://www.analog.com/en/prod/0,,TMP36,00.html . no WSB, and i dont think it should cause any grounding inssues.
But does anyone happen to know of any simular devices to this that measure (ambient) light intensity or moisture? Thanks again!
-Ian

Posted: Mon Jul 09, 2007 9:54 am
by Engineer1138
Hmmm. Temperature, light level, soil moisture. This sounds an awful lot like something I was asked about designing last year... Is it solar powered?

You can use CdS cells, photodiodes, or even LEDs to measure light levels. For moisture, you can measure soil resistance, but the probe must be driven with AC for it to work reliably.