Best portable multichannel thermocouple thermometer

Den tirsdag den 31. december 2013 02.16.32 UTC+1 skrev Phil Hobbs:

fluke 52 or fluke 54 ?

dual channel and I think ~$300/~$400

-Lasse

Some two channel J K T H E loggers...

Maybe you need a silicon bandgap temp sensor?

Cheers

That's your best deal there I think, it gives you SD card for recording and it has a long data-logging run time.

That comes with 3 channels with K probes, etc..

Jamie

So how did you temperature test it in the first place?

I've never cranked a test oven that high. Probably the max I ever did was 170 deg C testing thermal shutdown.

The Pico Thermocouple data logger is good and has moree than two channels. Comes with software and connects to the PC via USB. About half your budget level I should think. Last we used this was with a small experiement that was monitoring thermal transfer between two metal components in a high power scanning electron microscope. As the experiement started melting one of the metal components and the readings were still coming I think your range is well within its capabilities.

Omega does a lot of temperature stuff and seems to be reasonable about having things in stock. I have used their handheld meters and thermocouples before in a portable application (college engineering lab stuff) and they worked OK, but the requirements there (+/- 1 C or so, single channel) weren't as tight as yours. I also used one of their bench-type meters - really an industrial panel meter in a box - with thermocouples at work for product testing (+/- 1 C or so, multiple channels) and it worked OK.

Looking at their online catalog... HH501DK. 4 channels, Type K,

comes with four simple Type K thermocouples and a battery, for $125.

it via USB. No mention of isolation. It comes with Windows software, one simple Type K thermocouple, and batteries, for $300.

really be -50.1 to +999.9 C.) It logs data to an SD card and also has a USB connection, but the Windows software is apparently separate at $100. No mention of isolation. It comes with a 2 GB SD card, no batteries (6 AA) or thermocouples, for $300.

There are some other models that may fit the bill, including HH1384 ($375), HH309A ($315), HH374 ($300), and HH378 ($340).

This may be in the "grandmothers and eggs" category, but one thing I have found, with any portable thermocouple application, is that the screws inside the thermocouple connectors at the meter end tend to loosen up as things are banged around. A lot of meters can tell if the thermocouple is completely open circuit, but I have found that they can't detect slightly-loose screws that give unstable readings. One way I have found this problem is to put the thermocouple junction in a place where the temperature should be pretty stable, and then tap and tug on the connector with my fingers.

Standard disclaimers apply: I don't get money or other consideration from any companies mentioned.

Matt Roberds

If you were willing to use RTDs, something like an Adam 6015 (Advantech) would handle 7 channels, have 10/100 Ethernet interface, and fits your price budget. No display, but I think they have free software for viewing the readings on a netbook or whatever.

Accuracy should be considerably better than T/C over that range, particularly if the ambient temperature is not closely regulated.

Best regards, Spehro Pefhany

On a sunny day (Mon, 30 Dec 2013 20:16:32 -0500) it happened Phil Hobbs wrote in :

You are a buyer, but I designed and build this, 4 channels, wrote the soft too, add a Rasberry Pi, or like I use it with an eeePC:

Well, mine was cheaper, you could make a real peeseebee.

ebay in any other case?

Any free Modbus/TCP software.

Or maybe only web browser will be necessary.

None, (sorry) But I'm interested in what others offer. One issue I found with a cheap omega (panel mount) was that the cold junction temp sensor seemed to be on the opposite end of the pcb from the TC input. I guess not that much of an issue if you mostly care about relative temps and not absolute. (I can send a bucketful* of pnp's in to-220 packs that are calibrated to ~0.3C from 77K to 400K.. with most of the uncertainty at the low temp end.)

• well ok not a bucket full... I have to do a single point room temp measurement on each one...but 2-3 is easy.

Thanks, George. It's a halogen bulb spectrometer, so being able to measure the bulb envelope temperature would be useful, which is one reason why I want to use thermocouples. (Air temperature is easier with thermocouples as well--you can make the leads really skinny and put a bit of copper tape on it to make the air temperature dominate the lead conduction.)

Cheers

Phil Hobbs

(Inveterate hater of thermocouples in normal situations)

As I said, "bulb envelope temperature". One of those retro tungsten things that emits from the far IR to the near UV, not a circuit board.

Cheers

Phil Hobbs

Normally I'm a big fan of RTDs for temperature measurement, and a big hater of thermocouples. In this instance, though, I need to measure things like the air inside the box, gradients in the centreless ground steel rod chassis, and the bulb envelope. Thermocouples are better for those jobs.

Cheers

Phil Hobbs

With a thermocouple mounted on the envelope and at that close aproximity, wouldn't radiation effects get into play?

I have to get on a plane to Orange County this Sunday, so I need it fairly fast. Also I have lots of other science projects, most of which are on somebody else's nickel. ;)

Cheers

Phil Hobbs

Sure. What I'm trying to track down is a DC drift problem in a chopped AC measurement. One of those effects you don't usually think about too much is the effect of ambient temperature on bulb output.

If you take a tungsten bulb at constant input power, and raise its ambient temperature by 10 degrees, the filament will get hotter by probably 2 or 3 degrees, so it gets a bit brighter. The thermocouple will see both effects, but it'll at least allow me to estimate whether they're plausible causes.

It's a fibre-bundle spectrometer, so there are other possible issues, e.g. mechanical drift plus a moire pattern between the fibre ends and the image of the helical filament. The particular bulb has a box-wound filament, and is canted a few degrees so that the wire in the back of the helix fills in the spaces in the front, but there could still be some residual effect.

The first thing you look for in debugging thermal problems in an optomechanical system is bending due to temperature gradients or dissimilar CTEs.

My original prototype was built open-style, using Microbench optical rails and an RC-airplane servo motor for turning the grating, with a CMOS mux as the phase detector and a LabJack U6Pro for data acquisition. The new swoopy one has a very nice 3D-printed case, a proper encoder on the shaft, and an all-digital back end.

But it drifts, and mine didn't (despite being made of toy parts and dead-bug circuitry). We're going to spend next week going through it in the goriest possible detail, because I have no interest whatsoever in doing it over again.

Cheers

Phil Hobbs

A tungsten bulb inside a (plastic?) case is going to make all the dimensions change, during warmup and as external air wafts around.

What sort of dimensional changes are tolerable?

It has two fans, and the optical bits are located by centreless-ground steel rods. The three fibre bundles take a lot of the slop.

Fibre-bundle spectrometers are sort of fun--you rearrange the fibres to fill the desired area. Fanned out in one row they make a spectrometer slit; in two or three rows they match the bulb filament; bunched in a disc, they image well onto a round photodiode.

You do have to manage the area vs NA tradeoff carefully, or you lose a lot of light.

Cheers

Phil Hobbs