picaxe with LM355Z(temperature sensor)

[DeadNet]

Has anyone succeeded in creating a thermometer using the LM355z
temperature sensor? I am keen to use it rather than the popular
ds18b20 because of cost and the calibration facilities it provides. I
just have not seen any circuit diagrams around.
I am new to electronics and the picaxe.
Any thoughts/pointers?

 

[The Real Andy]

On Fri, 03 Feb 2006 20:36:12 -0000, Jasen Betts <jasen@free.net.nz>
wrote:

On 2006-02-03, David L. Jones <altzone@gmail.com> wrote:

The LM335 is a simple zener shunt temp sensor, just a series resistor
is all you need, check the datasheet. The voltage is proportional to
temperature at 10mV/degK

The LM35 has a 10mV/degC which might be a bit more sensible for you.

does that mean if goes negative if you freeze it ?

It does if you have the output tied to a negative rail with a
resistor. But i guess one could be clever and float the device up a
tad with a resistor on the ground pin, but then one would need to be
carful regarding the tempco of the resistor etc.

If you dont mind throwing an op-amp into the equasion, any PN juction
will measure temperature. A BC327 or similar jelly bean transistor is
good for about 2mV/degC from memory, and with some clean electronics
you can get about 0.1degC accuracy without difficulty. Adn IIRC they
silicon devices are quite linear, but dont quote me on that, it has
been a good 10+ years since i did analog stuff.

 

[bruce varley]

The Real Andy <will_get_back_to_you_on_This@> wrote in message
news:2h29u1hof49p3lsgvf48ua8m3i6f4h08ul@4ax.com...

On Fri, 03 Feb 2006 20:36:12 -0000, Jasen Betts <jasen@free.net.nz
wrote:

On 2006-02-03, David L. Jones <altzone@gmail.com> wrote:

The LM335 is a simple zener shunt temp sensor, just a series resistor
is all you need, check the datasheet. The voltage is proportional to
temperature at 10mV/degK

The LM35 has a 10mV/degC which might be a bit more sensible for you.

does that mean if goes negative if you freeze it ?

It does if you have the output tied to a negative rail with a
resistor. But i guess one could be clever and float the device up a
tad with a resistor on the ground pin, but then one would need to be
carful regarding the tempco of the resistor etc.

If you dont mind throwing an op-amp into the equasion, any PN juction
will measure temperature. A BC327 or similar jelly bean transistor is
good for about 2mV/degC from memory, and with some clean electronics
you can get about 0.1degC accuracy without difficulty. Adn IIRC they
silicon devices are quite linear, but dont quote me on that, it has
been a good 10+ years since i did analog stuff.

Having applied both to the same application, I'd recommend you go with the
LM35 unless there's a specified reason such as cost, it's higher output and
better specified. I amplify the LM35 output with an opamp, with a
noninverting gain circuit, so that the voltage swing across my temperature
range of interest is somewhere around the ADC span. If you use an amp that
can use a single ended 5V supply the overhead isn't large..

IIRC (also a heck of a long time ago), with a PN junction there was
noticeable nonlinearity across a 0 - 70C span.

 

[The Real Andy]

On Sun, 5 Feb 2006 21:19:30 +0800, "bruce varley"
<bxvarley@weqstnet.com.au> wrote:

The Real Andy <will_get_back_to_you_on_This@> wrote in message
news:2h29u1hof49p3lsgvf48ua8m3i6f4h08ul@4ax.com...
On Fri, 03 Feb 2006 20:36:12 -0000, Jasen Betts <jasen@free.net.nz
wrote:

On 2006-02-03, David L. Jones <altzone@gmail.com> wrote:

The LM335 is a simple zener shunt temp sensor, just a series resistor
is all you need, check the datasheet. The voltage is proportional to
temperature at 10mV/degK

The LM35 has a 10mV/degC which might be a bit more sensible for you.

does that mean if goes negative if you freeze it ?

It does if you have the output tied to a negative rail with a
resistor. But i guess one could be clever and float the device up a
tad with a resistor on the ground pin, but then one would need to be
carful regarding the tempco of the resistor etc.

If you dont mind throwing an op-amp into the equasion, any PN juction
will measure temperature. A BC327 or similar jelly bean transistor is
good for about 2mV/degC from memory, and with some clean electronics
you can get about 0.1degC accuracy without difficulty. Adn IIRC they
silicon devices are quite linear, but dont quote me on that, it has
been a good 10+ years since i did analog stuff.

Having applied both to the same application, I'd recommend you go with the
LM35 unless there's a specified reason such as cost, it's higher output and
better specified. I amplify the LM35 output with an opamp, with a
noninverting gain circuit, so that the voltage swing across my temperature
range of interest is somewhere around the ADC span. If you use an amp that
can use a single ended 5V supply the overhead isn't large..

DEfinately the easiest way.

IIRC (also a heck of a long time ago), with a PN junction there was
noticeable nonlinearity across a 0 - 70C span.

If i went out to the shed and did some hunting I could probably find
some of my old circuits and cal algorithms for PN junctions. Problem
is these days I dont have a 5.25" floppy drive

 

[Richard Freeman]

"DeadNet" <clancoma@yahoo.com> wrote in message
news:1138942852.140268.219530@g47g2000cwa.googlegroups.com...

Has anyone succeeded in creating a thermometer using the LM355z
temperature sensor? I am keen to use it rather than the popular
ds18b20 because of cost and the calibration facilities it provides. I
just have not seen any circuit diagrams around.
I am new to electronics and the picaxe.
Any thoughts/pointers?

Yes I have a Pic based (note not a Pic Axe however the hardware is the same)
Data logger with (amongst other options) a front end for an lm335. The
interface for the LM335 uses a CA3130 set up as a differential Amplifier
running off a supply 0.6V below the Pic chip and up to 6 V - could email you
the schematic if you are intereseted and can give a (suitably disguised of
course) email address.

Regards
Richard Freeman