From: Jonathan Kirwan
Subject: Re: question on measuring temperarures at high voltages
X-Newsreader: Forte Agent 1.92/32.572
NNTP-Posting-Date: Wed, 23 Oct 2002 19:10:03 GMT
Organization: AT&T Broadband
Date: Wed, 23 Oct 2002 19:10:04 GMT
On 23 Oct 2002 10:10:00 -0700, firstname.lastname@example.org (Kooner) wrote:
>Well, I do have a little training working with high voltage. I'm a
>junior EE studying/measuring someone else’s design of a 10kV
>power supply used to power a traveling wave tube (TWT). Our end
>product is a microwave amplifier.
>The LT1084 is used to power the filament or "heater" in a TWT. I can
>power up the LT1084/filament circuit without turning on the high
>voltage, and safely measure the power dissipation at 4.5W at worst
>case and a case temperature at 115C. The LT1084 is also mounted to a
>small crappy heat sink and there's no room for a better one.
>The problem is I need to measure the temperature of the LT1084 in the
>end product - at high voltage. Thermal dots have a 15C window which
>in not accurate enough. Infrared sensors can be better if they have an
>emissivity setting on them. But I couldn't use it anyway because the
>LT1084 is part of a high voltage module that gets incased in a red
>block of potting material to prevent arcing and corona. The block of
>potting material has other power dissipating components in it also has
>a thermal conductivity of 1 W/(mK) which is a lot better that air. I
>can estimate the LT1084 temperature, but with these other variables it
>really would be guess. My guess is it gets to hot, but I can't prove
>it without a valid measurement.
You might consider contacting Luxtron Corp for their M600 or M60 product. It's currently used for microwave products and high power transformers, for example. A fluorescent material is used at the tip of a fiber optic cable. There is nothing electrical near the target you are measuring. In fact, it works perfectly well inside an operating microwave oven. It also works inside the core of large neighborhood-scale power transformers, including at very high voltages, to measure core oil temperatures.
Target operating temperature range is from -40 C (actually, much lower) to some 400 C (+/- 50 C depending on probes), with a single sigma noise figure of about 0.1 C/rthz for the M600. That's a 4-channel device, sold as 1, 2, or 4 channel, I believe. The M60 is a single channel system with somewhat different noise specs, but roughly the same target temperature range.