Temperature measuring multimeter - recommend

[paolo]

So, my current multimeter is a no-frills thing, aimed at automotive folk, which does it's stuff ok for my needs. In terms of accuracy, when my bench PSU says it's delivering (x), the cheapo multimeter agrees - good enough for my undemanding work.

Now, doing some prototyping, I want to measure component temperature. For this design, it's more about trend lines than ultimately accuracy, so something that can do logging would be nice (but not essential, I can always "sit and observe").

So, recommendations? Something with probes that monitor an SMD component. Accuracy not massively important, but say within a few degrees.

[iloveelectronics]

Martin reviewed this very low cost multimeter a couple of months ago. It does data logging, and it even has .1 degree resolution for temperature measurement. Downside though is that the temperature probe that comes with it might not be the most suitable for SMD components (I suppose you can buy a suitable one cheaply though), and it only measures in Celcius and not Fahrenheit (not a problem if you're not American).

[T4P]

MS8226 : Appears to have 0.1C resolution
And ...
http://pcsensor.com/

They have lots of PC sensors that aren't cheap but are decent

[Jeff1946]

I would suggest get some thermocouple wire, (type K) fine gauge for your SMD. Check on ebay, lots for sale.  You can just connect it directly to your meter and read the voltage.  Place it in water ice mixture to get one data point and boiling water for another.  Compare voltages to table and use as a correction factor.  To make the actual thermocouple twist ends together and strike with a hammer against a metal surface to get good contact.  Thermocouples do go bad, corrosion etc, but is easy to tell as room temperature reading will be off.

You can do the same thing with a  meter with thermocouple readout, in this case you don't need to calibrate if you are only interested in a few degrees accuracy.  For better accuracy you can put two thermocouples back to back.  Put one in an ice bath and one on your unknown.  This way the same type of wire will be connected to your meter and the voltage produced by the thermocouple you are forming between your wires and meter jacks will cancel.

Another way is to use a diode at small current and measure the forward voltage drop.  This decreases with temperature and you could do a simple calibration, which could be good enough for your needs. An SMD one would be good.  Solder fine wire to it and put it back to back against your SMD, thermal grease would help.

[IanB]

I would suggest get some thermocouple wire, (type K) fine gauge for your SMD. Check on ebay, lots for sale.  You can just connect it directly to your meter and read the voltage.  Place it in water ice mixture to get one data point and boiling water for another.  Compare voltages to table and use as a correction factor.  To make the actual thermocouple twist ends together and strike with a hammer against a metal surface to get good contact.  Thermocouples do go bad, corrosion etc, but is easy to tell as room temperature reading will be off.

Remember that thermocouples don't measure temperature, they measure temperature differences. Every thermocouple has two junctions--one at the measurement point, and one at a known reference temperature. The actual measured temperature is found by adding the difference between the two thermocouple junctions to the known reference temperature.

So you can't just connect thermocouple wires to your meter and measure the voltage. That won't work. You need a proper thermocouple interface circuit.

[alm]

The cold junction is effectively at the meter's banana jacks (where the chromel/alumel wires are terminated), so you're measuring the difference between the banana jacks and the hot junction. Many DMMs with temperature option do the same. The net Seebeck voltage over the inputs of the meter is zero assuming no temperature gradients. It consists of nickel-copper-solder-pcb-solder-copper-nickel junctions: the opposing voltages cancel each other out, so it effectively forms a second junction. Similar to this Wikipedia diagram. Just like the ball of solder at the hot junction.

In real life temperature gradients and semiconductors might screw this up, and you will need to apply the conversion curve by hand, but I have gotten some ballpark measurements using this method in the past.

[IanB]

Every multimeter with a thermocouple function contains an internal temperature sensor for measuring the ambient temperature at the banana jacks so it can use this as the offset to compute the actual measured temperature. If the meter doesn't have the ambient temperature sensor built in you have to use an external thermometer and apply the offset by hand (or use an external thermocouple adapter). But if you are only interested in temperature changes and if the ambient temperature remains constant there is no need.

[alm]

I believe the Fluke 189, which is advertised as being able to use thermocouples, has the cold junction temperature as a manual setting in the setup menu. I wouldn't be surprised if many other meters do the same with a fixed cold junction temperature. This is similar to what you would get if you just measure voltage and do the conversion by hand.

[IanB]

The 189 contains an internal reference junction temperature sensor IC that it uses to compute the thermocouple temperature. You can see what it measures as the ambient temperature by shorting the input terminals while in temperature mode. The adjustment available in the setup menu is an offset adjustment that can be used to fine tune the thermocouple reading for additional accuracy.

[SherlockOhms]

Sometimes a separate temperature meter is handy. There was a review here of a 2 channel one -
https://www.eevblog.com/forum/reviews/odd-one-'tondaj'-dm6802-ii-standalone-temperature-meter/msg163872/#msg163872

Or search for TM-902C on ebay (single channel).  I have one, the resolution is only to nearest degree, but the one I have is accurate, at least at freezing and boiling points, it is. 

I was unhappy with the accuracy of the temperature function on my DMM.

[paolo]

Martin reviewed this very low cost multimeter a couple of months ago. It does data logging, and it even has .1 degree resolution for temperature measurement. Downside though is that the temperature probe that comes with it might not be the most suitable for SMD components (I suppose you can buy a suitable one cheaply though), and it only measures in Celcius and not Fahrenheit (not a problem if you're not American).

That looks perfect for my needs - unfortunately it looks like that Amazon based supplier won't ship to the UK.

Anyone know anywhere else that sells it?

[Wytnucls]

http://www.ebay.com/itm/TekPower-TP4000ZC-Auto-Range-AC-DC-Digital-PC-Interface-Multimeter-/251085806183?pt=LH_DefaultDomain_0&hash=item3a75e15a67

[iloveelectronics]

Martin reviewed this very low cost multimeter a couple of months ago. It does data logging, and it even has .1 degree resolution for temperature measurement. Downside though is that the temperature probe that comes with it might not be the most suitable for SMD components (I suppose you can buy a suitable one cheaply though), and it only measures in Celcius and not Fahrenheit (not a problem if you're not American).

That looks perfect for my needs - unfortunately it looks like that Amazon based supplier won't ship to the UK.

Anyone know anywhere else that sells it?

I'm actually going to stock these for my eBay store. But production is going to take some time so it will be at least a month before I get them. Price wise I won't be able to complete with Amazon for the US market but will be very competitive for the rest of the world.

[paolo]

http://www.ebay.com/itm/TekPower-TP4000ZC-Auto-Range-AC-DC-Digital-PC-Interface-Multimeter-/251085806183?pt=LH_DefaultDomain_0&hash=item3a75e15a67

Excellent, thanks - I've just ordered.

Even though shipping is costing more than the meter, it's still looking like a bargain. My jaw dropped when I looked at the prices of Flukes... one that doesn't have data logging! I'm sure they are really well made, but at nearly ten times the price...