HP3450A Reference Peltier Chambers
I've been looking at the various voltage reference schemes used in HP's 345x
line of DMMs. I was intrigued by the use of a Peltier chamber to control
the temperature of the reference element in the HP 3450A DMM.
I looked in the service manual for that unit, but I couldn't find a good
picture of the reference assembly. Would anyone having a 3450A be
adventurous enough to pull it apart enough to take a few good pictures of
the innards? I found a Sales Amplifier paper from HP that roughly describes
the reference assembly, but no good pictures.
I'm wondering if the diode under test and a thermistor should be mounted
directly to the surface of the Peltier device or to a copper or aluminum
interface between the Peltier and the thermistor & diode. Any thoughts?
How did HP do it in the 3450A chamber?
I have a number of 1N825 and 1N827 diodes that I'd like to toy with. In
particular, I'd like to build such a Peltier controlled chamber to ease the
task of finding the zero tempco spot on the zeners.
Dave M
In message F57B1617EE984B10AAD5379A2402DFE6@D77M7BF1, "Dave M" writes:
I was intrigued by the use of a Peltier chamber to control
the temperature [...]
It's a very obvious idea, but it runs into one of the widely unknown
footnotes about peltiers: Don't feed them AC.
The thermal/mechanical stress when you change direction of the
current significantly shortens their life.
I should add that I have not found any studies which say to what
degree this depends on the magnitude of the current, so an oven
balancing around a couple of mA may not be a problem, but switching
polarity on several amps will be.
One complication is the difference in directional efficiency: The
warm side receives about 4 times as much energy as is removed from
the cold side.
Unless you electrically compensate for this, your PID will be
really confused.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
Yes, I'm aware of the problems with AC drive to a Peltier. However, the
3450A drives the chamber to 43 degC with DC, so that isn't a problem. In
fact, it's a pretty simple drive circuit, runing stable state at 1.5 - 2
volts.
My interest is in how best to fix the components to the Peltier. Seeing a
picture of the innards of the 3450A chamber would help.
I have a few 12V Peltier junctions of various powers ratings, and have no
other thoughts as to what to use them for. Making a small chamber to
ascertain the zero tempco point of my reference zeners seems to be an easy
project. This isn't for a constantly-on instrument; it's for short-term
data gathering.
Thanks,
Dave M
Poul-Henning Kamp wrote:
In message F57B1617EE984B10AAD5379A2402DFE6@D77M7BF1, "Dave M"
writes:
I was intrigued by the use of a Peltier chamber to control
the temperature [...]
It's a very obvious idea, but it runs into one of the widely unknown
footnotes about peltiers: Don't feed them AC.
The thermal/mechanical stress when you change direction of the
current significantly shortens their life.
I should add that I have not found any studies which say to what
degree this depends on the magnitude of the current, so an oven
balancing around a couple of mA may not be a problem, but switching
polarity on several amps will be.
One complication is the difference in directional efficiency: The
warm side receives about 4 times as much energy as is removed from
the cold side.
Unless you electrically compensate for this, your PID will be
really confused.
In message CE0EFF06B1234752A6A2039737AF2587@D77M7BF1, "Dave M" writes:
My interest is in how best to fix the components to the Peltier.
Copper - it has high thermal conductivity and high thermal capacity.
Junked electrical bus-bars are a good place to start.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.