While testing some thermocouples yesterday, both grounded and ungrounded, I noticed something quite peculiar. When the sensitive skin on the underside of my arm touched the metal braid on the grounded thermocouple, I could feel a bit of a tinge. Not a big tinge, but enough to know there was voltage present. I tried 3 other cheap "ebay" power supplies (different brands) with the same results.
I quickly grabbed my Oscilloscope to analyze just what was going on. There is obviously a potential difference between the HM ground and earth (house ground), and the scope showed something I didn't expect. 100Vrms (50 Hz) from ground to earth (no wonder I could feel it). Below is the scope capture.
I had an old linear supply sitting around, so I decided to compare. The linear supply had a substantially less voltage reading (12Vrms) as seen below. I contribute this to the fact that linear supplies are truly "isolated".
From these results, I wanted to test a switching PS that I thought might be more reliable, so I tested my Laptop supply, and the readings were nominal (480mVrms).
So what does this mean? It means that if you use grounded thermocouples with cheap switching power supplies, there is a voltage of as much as 100VAC on the metal braid (with 240V house voltage), and ultimately YOU when you touch it (albeit a very small current on the order of a few mA).
I have found that when using grounded thermocouples, it gets rid of some of the noise introduced from cheap power supplies, but now i'm a bit weary of using them. The problem for other HM users that purchase TCs is that they may or may not be grounded. The only way to tell is to check for continuity between the braid and any of the connector wires. Grounding is just when the measurement junction is in contact with the metal braid via the sleeve on the end.
If I had my druthers, I'd only use linear supplies, but these are getting harder to find as new regulations have been imposed on manufacturers regarding efficiency.
Any ideas? I'll start by trying to source a reliable switching PSU. It would also be great if someone here could replicate my results, maybe with 120 mains.
I quickly grabbed my Oscilloscope to analyze just what was going on. There is obviously a potential difference between the HM ground and earth (house ground), and the scope showed something I didn't expect. 100Vrms (50 Hz) from ground to earth (no wonder I could feel it). Below is the scope capture.
I had an old linear supply sitting around, so I decided to compare. The linear supply had a substantially less voltage reading (12Vrms) as seen below. I contribute this to the fact that linear supplies are truly "isolated".
From these results, I wanted to test a switching PS that I thought might be more reliable, so I tested my Laptop supply, and the readings were nominal (480mVrms).
So what does this mean? It means that if you use grounded thermocouples with cheap switching power supplies, there is a voltage of as much as 100VAC on the metal braid (with 240V house voltage), and ultimately YOU when you touch it (albeit a very small current on the order of a few mA).
I have found that when using grounded thermocouples, it gets rid of some of the noise introduced from cheap power supplies, but now i'm a bit weary of using them. The problem for other HM users that purchase TCs is that they may or may not be grounded. The only way to tell is to check for continuity between the braid and any of the connector wires. Grounding is just when the measurement junction is in contact with the metal braid via the sleeve on the end.
If I had my druthers, I'd only use linear supplies, but these are getting harder to find as new regulations have been imposed on manufacturers regarding efficiency.
Any ideas? I'll start by trying to source a reliable switching PSU. It would also be great if someone here could replicate my results, maybe with 120 mains.
Last edited: