LM34 to A/D

[John Larkin]

John, just for everyone's benefit, what have you seen the LM34
do at higher voltages?

Latch up. Several things can make an LM34/35 latch up, one of them
being to pull the output down, as the datasheet suggests to allow
reading negative temperatures; sequencing matters. Latchup is less
likely if Vcc is below about 6 volts, but it can still happen,
especially if the output is pulled low statically or by a spike, like
from an ADC mux.

These gadgets also like to oscillate with the improper load. And the
internal pulldown is very weak, so mux spikes can pull it up and then
it'll settle back down slowly.

Bob Pease told me he'd fix it, about 15 years ago. I'm still waiting.

(BTW, all that care that must be taken regarding this device's
sensitivity to capacitive loading and the fact that you need
to included a negative supply or other roundabout methods for
reading temperatures below zero has made me try the LM50 if
I need an analog type sensor.)

Yup. Lately I've been using the SPI parts, which work great once you
write the (expletive deleted) code, or a surface-mount platinum RTD.

John

 

[Joerg]

John Larkin wrote:

[...]

The $190 amp is a tiny IC, not a box.

http://www.hittite.com/product_info/product_specs/amplifiers/hmc465.pdf

Neat. I wonder what their profit margins are on that one. Then again
it's probably a small market.

We use four of them in one of our products.

I'd be tempted to buy a reel of BFG hotrods and try out a discrete
version. Of course only if the production volume is high enough.

 

[John Larkin]

John Larkin wrote:

[...]

The $190 amp is a tiny IC, not a box.

http://www.hittite.com/product_info/product_specs/amplifiers/hmc465.pdf

Neat. I wonder what their profit margins are on that one. Then again
it's probably a small market.

We use four of them in one of our products.

I'd be tempted to buy a reel of BFG hotrods and try out a discrete
version. Of course only if the production volume is high enough.

We attempted a home-made distributed amplifier, using gaasfets, with
mediocre results, and gave up when we found the Hittite gadgets. It's
a distributed amp internally, and they have a lot less parasitics (and
a lot more engineers) than we do.

But they sure get hot!

John

 

[Fred Bloggs]

But you could oversample and average. Temperature changes slowly, and
code is cheap.

John

I don't, and will never, trust averaging an acquisition synchronized to
a digital clock. Use the cheap code to correct for gain and offset of
the cheap gain amp, which can be calibrated with a cheap voltage input.

 

[Anthony Fremont]

John Larkin wrote:

The gain makes the difference between induced noise on the board
amounting to several degrees or a fraction of a degree. The input
range of interest is 50-120 oF or 0.7V so a gain to full range of the
ADC is approximately x14. This means the ADC will see 140mV/oF versus
10mV/oF, making reasonable amounts of onboard noise equivalent to a
fraction of a degree versus several whole degrees.

You seem to be making the assumption that the noise would only be introduced
_after_ the op-amp.

 

[Anthony Fremont]

Oh, yeah, run the opamp from whatever higher supply you have, if you
need to swing higher.

Just messin' with ya.

 

[John Larkin]

I don't, and will never, trust averaging an acquisition synchronized to
a digital clock.

These days, it's hard to make any other kind of acquisition. I almost
always software lowpass filter stuff like this, slow things like
temperature acquired by a fast sampling ADC. There's usually enough
natural noise floating around to dither a 12-bit system up to 14 or
even 16 bits of usable resolution. And if there's a high temp shutdown
or something like that, filtering avoids false alarms from spikes or
whatever.

Use the cheap code to correct for gain and offset of
the cheap gain amp, which can be calibrated with a cheap voltage input.

He can calibrate based on a temperature standard, in the extreme. We
don't know if he wants to store custom cal factors per system.

John

 

[Fred Bloggs]

These days, it's hard to make any other kind of acquisition. I almost
always software lowpass filter stuff like this, slow things like
temperature acquired by a fast sampling ADC. There's usually enough
natural noise floating around to dither a 12-bit system up to 14 or
even 16 bits of usable resolution. And if there's a high temp shutdown
or something like that, filtering avoids false alarms from spikes or
whatever.






He can calibrate based on a temperature standard, in the extreme. We
don't know if he wants to store custom cal factors per system.

John

Since the temperature uncertainty of the '34 is adequate he does not
need a temperature calibration. He can use E96 metal film resistors for
his gain and bias setup of the gain amp and calibrate those errors with
a simple voltage input. The E96 metal film should track in ratio at
5ppm/oC so I don't see reasonable operating environments being a problem.

 

[Fred Bloggs]

You seem to be making the assumption that the noise would only be introduced
_after_ the op-amp.

pleeze...

 

[Joerg]

John, just for everyone's benefit, what have you seen the LM34
do at higher voltages?

(BTW, all that care that must be taken regarding this device's
sensitivity to capacitive loading and the fact that you need
to included a negative supply or other roundabout methods for
reading temperatures below zero has made me try the LM50 if
I need an analog type sensor.)

Or just use a thermistor. They've been good to me, they are cheap, they
don't latch up, don't need a black cat to cross the alley from left to
right for negative temps, and heck, they don't even require a uC with an
ADC on board.

 

[LVMarc]

OK, I've got a basic problem. I've got an LM34D (32 - 212 deg F) temp
sensor <http://www.national.com/ds/LM/LM34.pdf> and I need to connect it
to a 0-10V A/D.

To make things more interesting, I am only looking for a relatively
narrow range - say from 50 deg. F to 120 deg. F.

I've got 12 bits of A/D, which, in theory, gives me lots of resolution,
but the LN34 is only good to about a degree. It outputs 10 mV/degree.

The way I read the LM34 spec, the range I am interested in is about
500mV to 1.2V, so I figure I need to amplify that by, say, 8 to get good
use of my A/D accuracy.

I'm mostly a software guy, although once in a while I can solder
something without serious injury.

I've been told I need an op-amp to make the above work, but I have no
idea how to get an op-amp to work in the above scenario.

Could someone please clue me in? Thanks,

--Yan

Yan,

You need to adjust the gain ad offset of the base lm34 output to fit the
ad input range. You can achieve digitization past the 10mV/deg that the
lm34 device outputs!

If you would like some assitnace with the hardware for you and the
organization please contact me directly. I have significant experience
with the device and the process you describe!

Marc

 

[Spehro Pefhany]

I don't, and will never, trust averaging an acquisition synchronized to
a digital clock. Use the cheap code to correct for gain and offset of
the cheap gain amp, which can be calibrated with a cheap voltage input.

Or just use an op-amp with better DC characteristics for about the
same price, and 0.1% resistors. The latter are no longer very
expensive.


Best regards,
Spehro Pefhany

 

[John Larkin]

Just messin' with ya.

Well, I suppose you did have a valid point. Grumble.

Somebody *does* make an opamp with a charge pump inside, BB I think,
but it's only for the front end, so they can go r-r on the input
without all the usual crossover artifacts.

John

 

[John Larkin]

You seem to be making the assumption that the noise would only be introduced
_after_ the op-amp.

Several LSB's of onboard noise? That's perfect for dithering the ADC!

John

 

[Jamie]

Oh, yeah, run the opamp from whatever higher supply you have, if you
need to swing higher.

John

It's funny what people use simple things like the LM34 for.

We have an oven system (large) that required even heating through
out. The selected range goes from 180..220 F, So we redid the duct work
and placed array of duct panels with DC servos motors to position them.
Off the side of each Panel is a LM34 which gets tied into a major
wire way over to a control center. each LM34 and motor is connected to a
board in a card rack. Each Card has calibration points etc..

The LM34 goes into a comparator for a +/- pulse that drives a mini
pic chip which controls the motor position.
We bought the motors that had limits switches already in the motor
mounts. The rest, we designed. had the boards etched for us and we
assembled them etc..
This system regulates the heat radiation on the oven so perfectly now
that each zone only varies no more than +/- 1 degree F from each other.

We didn't experience any of the oscillating or latch up problems that
many have spoke of here.

 

[Tim Shoppa]

I wonder what's the most expensive IC you can buy. I've heard that
some high-end FPGA's, not even rad-hard, run close to $10K.

The first next-gen Intel/AMD/etc. processor off the production line
has a cost measured in the few billion dollars.

If they manage to sell more than a couple, of course, the cost of the
development and the new-tech fab can be spread out over more than one
chip!

I brush up against a few particular sectors of the semiconductor-
processing-equipment companies (many of them in Jim's neighborhood)
that sell equip that goes into fabs, and am continually astonished at
the money scales involved in the big picture. Usually we're dickering
over a few K$ in billing, not a few billion in billing, though!

Tim.

 

[LVMarc]

It's funny what people use simple things like the LM34 for.

We have an oven system (large) that required even heating through
out. The selected range goes from 180..220 F, So we redid the duct work
and placed array of duct panels with DC servos motors to position them.
Off the side of each Panel is a LM34 which gets tied into a major
wire way over to a control center. each LM34 and motor is connected to a
board in a card rack. Each Card has calibration points etc..

The LM34 goes into a comparator for a +/- pulse that drives a mini
pic chip which controls the motor position.
We bought the motors that had limits switches already in the motor
mounts. The rest, we designed. had the boards etched for us and we
assembled them etc..
This system regulates the heat radiation on the oven so perfectly now
that each zone only varies no more than +/- 1 degree F from each other.

We didn't experience any of the oscillating or latch up problems that
many have spoke of here.

well you figrued out exactly where to put the cazapatators, to make em
stable

Marco

 

[Jamie]

well you figrued out exactly where to put the cazapatators, to make em
stable

Marco

well,. maybe it was the 75 feet of 3 wire #26 of Signal Cable for
each LM34 that contributed to it?