Question about temp probes/PT100

[cowasaki]

I've only ever had hand held multimeters which had bespoke temperature probes but now I have 2 (soon to be 3) bench multimeters which have proper temperature sensing ie Philips/Fluke PM2525 x 2 and from next week a Keithley 2100. 

Now I've seen PT100 probes starting at under £4, such as https://www.ebay.co.uk/itm/PT100-Platinum-Resister-Temp-Temperature-Sensor-Probe-Waterproof-Anti-rust-UK/122894459724?epid=1163351040&hash=item1c9d14a34c:g:jzwAAOSwDNdVu52J, but are they any good?  How can I check if the probes are even remotely accurate?  I have a laser temperature tool but again I have no way of knowing if it is actually accurate. 

Are all PT100 probes compatible with each other and the circuit diagrams for these DMMs?

Should I buy a genuine Philips or Keithley temperature probe?

[mikeselectricstuff]

Even a bad PT100 should be way more accurate than an IR sensor. Assuming it is actually Pt100 and not some Chinese fake. Platinum is expensive...
 Pt100 is a standard curve, so any Pt100 should work the same, cost differences being down to specified accuracy, packaging etc. 

No reason to by a branded probe, at least not one branded by a meter manufacturer as these will 100% be branded from an established Pt100 probe manufacturer.

[edpalmer42]

No, the cheap PT100 probes on ebay aren't compatible with the Philips/Fluke DMMs.

I have a PM2534 and was in the same situation you are.  I decided to buy the cheap WZP PT100 probes and see what happens.  My measurements suggest that the PM2534 needs a probe with the standard alpha value of 0.00385 and these probes definitely don't meet that criteria.  At a room temperature of 25.5C, the probe measured 107.29 ohms.  The PM2534 reported a temperature of 18.7C which would be correct for a probe with the standard alpha.

Very few of the probes on ebay state what their alpha is or have the 4 wire configuration that our meters require.  One that does is https://www.ebay.com/itm/232066612597 , but they're sold out.

Ed

[cowasaki]

No, the cheap PT100 probes on ebay aren't compatible with the Philips/Fluke DMMs.

I have a PM2534 and was in the same situation you are.  I decided to buy the cheap WZP PT100 probes and see what happens.  My measurements suggest that the PM2534 needs a probe with the standard alpha value of 0.00385 and these probes definitely don't meet that criteria.  At a room temperature of 25.5C, the probe measured 107.29 ohms.  The PM2534 reported a temperature of 18.7C which would be correct for a probe with the standard alpha.

Very few of the probes on ebay state what their alpha is or have the 4 wire configuration that our meters require.  One that does is https://www.ebay.com/itm/232066612597 , but they're sold out.

Ed

Thank you for this information.  To be honest I wanted the temperature probes to work with my PM2525s more than the Keithley because that frees up the better DMM for measurements.....  I will also check the Keithley manual for this same information.

[Vtile]

The trap for young players is that the PT100 comes under different standards ... unfortunately. I recal that atleast IEC and Japanese PT100 standards do have different values. The TC isn't 100% linear, but most things it can be considered to be.

The wire compensation is a must (3-wires minimum), with 2-wire connection and long wires you measure the wire temperature and not the PT100 temperature.

[cowasaki]

The trap for young players is that the PT100 comes under different standards ... unfortunately. I recal that atleast IEC and Japanese PT100 standards do have different values. The TC isn't 100% linear, but most things it can be considered to be.

The wire compensation is a must (3-wires minimum).

Thanks, just checking RS Components at the moment.  That link is for a company who don't post to the UK anyway but it at least gives me an idea.  I will update if I find one.

[cowasaki]

Nope RS do not show the alpha of the PT100s they sell.......

https://docs-emea.rs-online.com/webdocs/15e5/0900766b815e5304.pdf

[cowasaki]

Well I've found something:

https://www.ebay.com/itm/4-Wires-Class-A-RTD-PT100-Temperature-Sensor-Probe-50-200-C-58-392-F-2-M/152737213382?hash=item238fd8ebc6:g:LCAAAOSwP4FaAt9Y

[cowasaki]

Oh and the Keithley needs the same alpha :-)  Glad I posted or I would have had no idea!

[mikeselectricstuff]

Nope RS do not show the alpha of the PT100s they sell.......

https://docs-emea.rs-online.com/webdocs/15e5/0900766b815e5304.pdf

They do quote IEC751 / BSEN60751

Here is some more info
https://www.omega.co.uk/prodinfo/rtd.html

[cowasaki]

Nope RS do not show the alpha of the PT100s they sell.......

https://docs-emea.rs-online.com/webdocs/15e5/0900766b815e5304.pdf

They do quote IEC751 / BSEN60751

Here is some more info
https://www.omega.co.uk/prodinfo/rtd.html

Thanks. That helps as well I can now check RS’s site and Ebay with that extra info

[edpalmer42]

Well I've found something:

https://www.ebay.com/itm/4-Wires-Class-A-RTD-PT100-Temperature-Sensor-Probe-50-200-C-58-392-F-2-M/152737213382?hash=item238fd8ebc6:g:LCAAAOSwP4FaAt9Y

That looks more interesting.  Too bad that it only goes from -50C to 200C, but the official probe is -60C to 200C so it's a good match.  Now the only question is which of the 3000 variations of DIN connectors is the one you need!  Note that you need the same connector to do 4-wire ohms measurement.  Actually, the temperature measurement is just a 4-wire ohms measurement that is then scaled to match a PT100 sensor.

Ed

[edpalmer42]

I did some more digging and found two more probes that have an alpha of 0.00385.  They don't actually say that, but they give resistance vs. temperature values that are consistent with that value of alpha.  They're only 3 wire instead of 4.  I don't know how big an error that would cause.  The price is quite a bit lower than the other probes you found - less than half.

https://www.ebay.ca/itm/PT100-Temperature-Sensor-Waterproof-Probe-Silicone-Cable-1M-2M-Resistance-RTD/263271057813

https://www.ebay.ca/itm/0-2-Class-A-Platinum-Resistance-Thermometers-Senser-Probe-PT100-3-Wire/181743141868

Ed

[cowasaki]

Thanks. That is also helpful. I will have a look later.

[knotlogic]

cowasaki, have you come across Labfacility before:

https://www.labfacility.com/

They're based in the UK IIRC, and Farnell carries their products.  Or used too at least, when I was sourcing for PT100s.

I think they did some kind of revamp, and changed part numbers on a bunch of items.  Datasheets too, and unfortunately the new ones   don't seem to have as much useful information as before.     So Farnell still lists some items with the old part numbers, but on the plus side also has some of the old datasheets.

[nfmax]

cowasaki, have you come across Labfacility before:

https://www.labfacility.com/

They're based in the UK IIRC, and Farnell carries their products.  Or used too at least, when I was sourcing for PT100s.

I think they did some kind of revamp, and changed part numbers on a bunch of items.  Datasheets too, and unfortunately the new ones   don't seem to have as much useful information as before.     So Farnell still lists some items with the old part numbers, but on the plus side also has some of the old datasheets.

Labfacility are very helpful in providing support - you can email or phone and get through to someone technically knowledgable who can answer questions. They are temperature measurement specialists. Their 'shark tail' range of probes are really cheap and include PRT models: https://www.labfacility.com/temperature-sensors/handheld-temperature-sensors/shark-tail-style/general-purpose-probe-iec.html

Give them a call to check suitability for your meter

[Vtile]

3-wire is no-go for meter if the meter doesn't say otherwise (..or you end up using it as 2-wire probe, with cost of 3-wire). The 3-wire connection actually forms one of the four nodes of the Wheatstone bridge setup. 4-wire probe is in Kelvin setup, just like the 4-wire resistance measurement in the higher end meters.

I did some more digging and found two more probes that have an alpha of 0.00385.  They don't actually say that, but they give resistance vs. temperature values that are consistent with that value of alpha.  They're only 3 wire instead of 4.  I don't know how big an error that would cause.  The price is quite a bit lower than the other probes you found - less than half.

https://www.ebay.ca/itm/PT100-Temperature-Sensor-Waterproof-Probe-Silicone-Cable-1M-2M-Resistance-RTD/263271057813

https://www.ebay.ca/itm/0-2-Class-A-Platinum-Resistance-Thermometers-Senser-Probe-PT100-3-Wire/181743141868

Ed

[mikeselectricstuff]

3-wire is no-go for meter if the meter doesn't say otherwise (..or you end up using it as 2-wire probe, with cost of 3-wire). The 3-wire connection actually forms one of the four nodes of the Wheatstone bridge setup. 4-wire probe is in Kelvin setup, just like the 4-wire resistance measurement in the higher end meters.

yes - 3-wire needs some specific calculations to effectively measure and cancel the line resistance, working on teh assumption that two of the feed wires have equal resistance, so only useful for systems specfically designed for 3-wire probes.

[edpalmer42]

3-wire is no-go for meter if the meter doesn't say otherwise (..or you end up using it as 2-wire probe, with cost of 3-wire). The 3-wire connection actually forms one of the four nodes of the Wheatstone bridge setup. 4-wire probe is in Kelvin setup, just like the 4-wire resistance measurement in the higher end meters.

yes - 3-wire needs some specific calculations to effectively measure and cancel the line resistance, working on teh assumption that two of the feed wires have equal resistance, so only useful for systems specfically designed for 3-wire probes.

Well, that depends on your accuracy requirements.

On a 3-wire probe you can cancel out the resistance of one lead, but not the other.  This adds a constant resistance to the measurement.  I don't know, but let's assume that the lead is 28 gauge.  That means 1 meter has a resistance of ~0.2 ohms.  At -50C, the probe has a resistance of 80.3 ohms and at 200C it's ~176 ohms, so the extra resistance introduces an error of between 0.25% and 0.11%.  If the lead is 2 meters, double those numbers.  For most of us, we'll be working at the high end of the temperature range where the error is lowest.  But if those numbers are a problem for your application then you can't use 3-wire.

Ed

[mikeselectricstuff]

3-wire is no-go for meter if the meter doesn't say otherwise (..or you end up using it as 2-wire probe, with cost of 3-wire). The 3-wire connection actually forms one of the four nodes of the Wheatstone bridge setup. 4-wire probe is in Kelvin setup, just like the 4-wire resistance measurement in the higher end meters.

yes - 3-wire needs some specific calculations to effectively measure and cancel the line resistance, working on teh assumption that two of the feed wires have equal resistance, so only useful for systems specfically designed for 3-wire probes.

Well, that depends on your accuracy requirements.

On a 3-wire probe you can cancel out the resistance of one lead, but not the other.  This adds a constant resistance to the measurement.  I don't know, but let's assume that the lead is 28 gauge.  That means 1 meter has a resistance of ~0.2 ohms.  At -50C, the probe has a resistance of 80.3 ohms and at 200C it's ~176 ohms, so the extra resistance introduces an error of between 0.25% and 0.11%.  If the lead is 2 meters, double those numbers.  For most of us, we'll be working at the high end of the temperature range where the error is lowest.  But if those numbers are a problem for your application then you can't use 3-wire.

Ed

If you don't have the 3-wire measurement capability you may as well just use it in 2-wire mode and apply a fixed offset.
3-wire systems use the third wire to measure the drop over one of the wires, then subtract this to null out the drop over both.

[edpalmer42]

3-wire is no-go for meter if the meter doesn't say otherwise (..or you end up using it as 2-wire probe, with cost of 3-wire). The 3-wire connection actually forms one of the four nodes of the Wheatstone bridge setup. 4-wire probe is in Kelvin setup, just like the 4-wire resistance measurement in the higher end meters.

yes - 3-wire needs some specific calculations to effectively measure and cancel the line resistance, working on teh assumption that two of the feed wires have equal resistance, so only useful for systems specfically designed for 3-wire probes.

Well, that depends on your accuracy requirements.

On a 3-wire probe you can cancel out the resistance of one lead, but not the other.  This adds a constant resistance to the measurement.  I don't know, but let's assume that the lead is 28 gauge.  That means 1 meter has a resistance of ~0.2 ohms.  At -50C, the probe has a resistance of 80.3 ohms and at 200C it's ~176 ohms, so the extra resistance introduces an error of between 0.25% and 0.11%.  If the lead is 2 meters, double those numbers.  For most of us, we'll be working at the high end of the temperature range where the error is lowest.  But if those numbers are a problem for your application then you can't use 3-wire.

Ed

If you don't have the 3-wire measurement capability you may as well just use it in 2-wire mode and apply a fixed offset.
3-wire systems use the third wire to measure the drop over one of the wires, then subtract this to null out the drop over both.

Yes, the only reason to do it the way I described is that it reduces the error by cancelling out the effect of one of the leads.  Depending on your application, that might be good enough.

Looking around on the net, it appears that the 3-wire probe is normally used in a bridge arrangement which allows it to completely cancel out the effect of lead resistance.  Unfortunately, the Fluke/Philips multimeters are designed for the more expensive 4-wire probe.

Ed

[Vtile]

Looking around on the net, it appears that the 3-wire probe is normally used in a bridge arrangement which allows it to completely cancel out the effect of lead resistance.

Yes!! Wikipedia.org do have wrong wiring diagram, this is the correct one for 3-wire bridge setup (Whetstone). http://www.thermometricscorp.com/3-wire-rtd.html

[knotlogic]

Labfacility are very helpful in providing support - you can email or phone and get through to someone technically knowledgable who can answer questions.

Thanks, that's good to know.  Beats having to guess at what I'm trying to look for!

[edpalmer42]

I'm going to withdraw a comment I made earlier in this thread.  I bought the really cheap 2-wire WZP sensor that claimed to be PT100 and didn't like what I saw when I checked them.  I decided it was long overdue to do some more testing on them.

In ice water, they measured 99R8 or -0.3C on my PM2534.  That's where the '100' in PT100 comes from.  The sensor should be 100R @ 0C.  As the temperature increases, so does the resistance.

I then wrapped the sensor in one of those soft heat-conducting pads you often find in laptops and taped it to the side of a power resistor.  I measured the temperature of the flat metal end of the WZP sensor with my Fluke 80T-150 probe.  I wrapped everything in some old towels for insulation.

As the resistor heated up, I tracked the two temperature measurements.  Between room temperature and 90C, the difference between them was rarely more than 1C and never more than 2C.  Above 90C, there may be more error, but my limited technology and even more limited patience made the readings suspect.

Therefore, within the limits of my technology, calibration and patience I have to say that the WZP sensor seems to be in the ballpark for a standard PT100 sensor.  I would be very interested if someone else with better capabilities could make some measurements.  Are there any temperature-nuts around here?

Ed

[cowasaki]

cowasaki, have you come across Labfacility before:

https://www.labfacility.com/

They're based in the UK IIRC, and Farnell carries their products.  Or used too at least, when I was sourcing for PT100s.

I think they did some kind of revamp, and changed part numbers on a bunch of items.  Datasheets too, and unfortunately the new ones   don't seem to have as much useful information as before.     So Farnell still lists some items with the old part numbers, but on the plus side also has some of the old datasheets.

CPC.Farnell are based about 10 miles away.  I will check CPC's site.  I need some more solder too so might pop in on Friday.