What's the best way to measure the envelope temperature of a pivacuum tube? I have tried a infrared thermometer, but the readings fluctuate depending on what part of the envelope I am pointing it at. It's supposed to read only the glass temperature, but i think its picking up point sources of infrared emission inside the tube. Any ideas on a tool and/or a method?
infrared thermometers depend on what is called the emissivity of whatever you are pointing it at. They aren't really "accurate" per se. Different materials emit different amounts of ir vs. temperature.
Get a very small K thermocouple and clamp or tape it to the envelope? I have a couple that came with digital multimeters that have a business end the size of a mustard seed.
Get a very small K thermocouple and clamp or tape it to the envelope? I have a couple that came with digital multimeters that have a business end the size of a mustard seed.
. . . and the mustard seed ones are the ones dipped in epoxy.
This one is more like a poppy seed:
https://goo.gl/HuvORL
This one is more like a poppy seed:
https://goo.gl/HuvORL
Just a thought..
If you had a dark color tape (flat black?) with high-heat adhesive, you could attach a piece to the envelope to use as a target for the IR pyrometer. You'd need a decent instrument, with a good D/S ratio such that only the area of the target was read, but this should eliminate the issues with the transparency / reflectivity of the glass.
Then again, maybe your instrument is +too+ good. If you're picking up point sources, then a poorer D/S ratio might be to advantage - as it would tend to pick up the average of a wider area.
Or increase the S radius by taking the reading with your current instrument from greatert D? No, that would be too easy..
If you had a dark color tape (flat black?) with high-heat adhesive, you could attach a piece to the envelope to use as a target for the IR pyrometer. You'd need a decent instrument, with a good D/S ratio such that only the area of the target was read, but this should eliminate the issues with the transparency / reflectivity of the glass.
Then again, maybe your instrument is +too+ good. If you're picking up point sources, then a poorer D/S ratio might be to advantage - as it would tend to pick up the average of a wider area.
Or increase the S radius by taking the reading with your current instrument from greatert D? No, that would be too easy..
Last edited:
Adding any cover/obstruction to the glass (for IR targeting) will alter the 'natural' temperature of the glass in that region, as the portion of radiation that would normally transmit through the glass would now be stopped. Using as small a patch/dot as possible would be appropriate.
Similarly, any thermocouple should be as small as possible. One method I've seen is to make a thin wire loop around the glass body, and attached the thermocouple tip to that. Another method could be to press the tip against the glass from say a chopstick mounted radially (something non-conductive). The thermocouple lead should initially also stay on the glass, rather than just the tip touching the glass, to minimise thermal conduction along the lead wires.
There are many photos and videos around using thermal cameras, that give an indication.
The glass is just an intermediary between the anode structure and ambient. Some of the energy flow passes straight through the glass and transfers to surrounding objects - so the glass doesn't heat up from that transmitted energy. A small amount of energy gets reflected from the inside of the glass, but as everything inside the valve glass envelope is hot, then the radiation doesn't have a cool spot to flow power to. The glass absorbs some of the energy, especially for the longer wavelengths, but its outside surface temp is very much dependant on ambient air temp in that local region, and how well a chimney effect or breeze can move heat away from the glass.
Ball-park numbers are 500C for a power valve anode near max dissipation rating, and 180C glass surface.
Similarly, any thermocouple should be as small as possible. One method I've seen is to make a thin wire loop around the glass body, and attached the thermocouple tip to that. Another method could be to press the tip against the glass from say a chopstick mounted radially (something non-conductive). The thermocouple lead should initially also stay on the glass, rather than just the tip touching the glass, to minimise thermal conduction along the lead wires.
There are many photos and videos around using thermal cameras, that give an indication.
The glass is just an intermediary between the anode structure and ambient. Some of the energy flow passes straight through the glass and transfers to surrounding objects - so the glass doesn't heat up from that transmitted energy. A small amount of energy gets reflected from the inside of the glass, but as everything inside the valve glass envelope is hot, then the radiation doesn't have a cool spot to flow power to. The glass absorbs some of the energy, especially for the longer wavelengths, but its outside surface temp is very much dependant on ambient air temp in that local region, and how well a chimney effect or breeze can move heat away from the glass.
Ball-park numbers are 500C for a power valve anode near max dissipation rating, and 180C glass surface.
Last edited:
It is not a good idea to put something black to measure temperature, because a black body is the best radiation absorber and the best emitter too, so you would still measure the radiation temperature from the inside.
A small white/silver thermocouple should be better. IMHO.
IR generally is not picked up through glass. You are reading the actual temperature of the glass. However, the spot-to-distance ratio of an IR thermometer will make it difficult to get consistent readings. Also, the temperature at the base of the tube will be much lower than at the top.
Go to OMEGA Engineering : Technical Reference for lots of good info. E
Go to OMEGA Engineering : Technical Reference for lots of good info. E
I asked this question for the purpose of measuring a tube in an amp that I recently completed. This is the link for that build, http://www.diyaudio.com/forums/tubes-valves/305739-5998-421a-build-schematic-feedback-wanted-4.html In the tube data for this tube it states that if it is run above 150 Celsius, it should be provided with forced air cooling. I have seen maximum envelope temperatures on other tubes as well. The entire top cap of this ST tube is covered with silver flashing. This part of the tube measures 138 Celsius with my infrared thermometer, and is therefore below the maximum. Some parts of the envelope, such as when pointing between the two plates, measure 210 Celsius. If you measure the glass in the same part at an angle so you are not pointing at the internal hot spots, the temperature is lower. It seems unlikely based on my experimentation, that the infrared thermometer is reading only the envelope and not the internal parts of the tube.
My infrared thermometer (DT-380) is at the simple end of the range. It appears that its sensor has a nominal wavelength response of 8 to 14um. Glass is effectively opaque in that range, but I would anticipate it responds to some extent to a wider range and so may 'see' a higher temp surface behind the glass.
But I would also anticipate that such a focusing sensor pointed at a glass surface would be sensitive to angle of incidence, and reflections, and distance to sensor.
Without a supporting measurement using eg. a thermocouple, your reported temperatures and comments are suspect.
If you apply a small dot on the glass, and then heat up the valve in a controlled temperature 'space', and use a thermocouple for reference, and take multiple readings as temperature is increased, with a consistent positioning and pointing, then that should provide some confidence.
But I would also anticipate that such a focusing sensor pointed at a glass surface would be sensitive to angle of incidence, and reflections, and distance to sensor.
Without a supporting measurement using eg. a thermocouple, your reported temperatures and comments are suspect.
If you apply a small dot on the glass, and then heat up the valve in a controlled temperature 'space', and use a thermocouple for reference, and take multiple readings as temperature is increased, with a consistent positioning and pointing, then that should provide some confidence.
Last edited:
My measurements with my cheapy infrared thermometer are totally suspect. So much so that they appear to be useless. I don't have a thermocouple handy, but I did try pressing the tip of a digital probe thermometer against the glass at various points. This resulted in a painfully slow rise time that ended up showing a fairly consistent temperature over the entire surface of the valve. The bottom was cooler than the top by about 20 degrees C. All the temperatures were under 145 C using this method. The infrared gun showed 125 to 220, which would seem to indicate an issue with the device and/or my use of it
As others have said, this is a difficult thing to do. Placing anything on the glass will affect the local temperature, and will be slow to respond because glass is a poor conductor of heat. An IR measurement could be confused by shorter wavelength IR coming through the glass from the anode. Where there is internal silvering you could get a low reading because part of what the sensor is 'seeing' is a reflection of the surroundings. Ideally, you want an IR sensor which only responds to the wavelengths where the glass is completely opaque.
Be aware that 150C is not a cliff edge. It doesn't mean that 145C is fine and 155C is disaster. Hotter valves have shorter lives. My guess is that you are sailing close to the edge so some extra cooling would help.
Be aware that 150C is not a cliff edge. It doesn't mean that 145C is fine and 155C is disaster. Hotter valves have shorter lives. My guess is that you are sailing close to the edge so some extra cooling would help.
Emissivity varies. IR thermometers aren't "accurate" without correction for what they are pointed at.
http://www-eng.lbl.gov/~dw/projects/DW4229_LHC_detector_analysis/calculations/emissivity2.pdf
http://www-eng.lbl.gov/~dw/projects/DW4229_LHC_detector_analysis/calculations/emissivity2.pdf
They didn't have Silicon IR sensors in 1959.
They did have TempSticks, and that is probably what you should be using.
Inweld Corporation - Welding Accessories
Thermomelt® | Markal
Tempilstik® | Tempil
https://www.grainger.com/category/ecatalog/N-1z0dneq
https://www.fastenal.com/products/w...ryl3:"603273 Temperature Indicating Sticks"|~
They did have TempSticks, and that is probably what you should be using.
Inweld Corporation - Welding Accessories
Thermomelt® | Markal
Tempilstik® | Tempil
https://www.grainger.com/category/ecatalog/N-1z0dneq
https://www.fastenal.com/products/w...ryl3:"603273 Temperature Indicating Sticks"|~
Only a few output stage valves identified max bulb temperature. KT66 datasheet identified a design max of 250C and absolute max of 280C. A 1972 datasheet for 6550 also identified 250C (measured at that time by an infrared thermometer).
Pearl coolers have a good reference link and background info, although the technical effort stopped a long time ago so does not include modern techniques or materials, and the articles are devoid of measurement technique. There are also some doubtful comments, especially about the amount of heat absorbed by the glass. Of interest is the use of tempilaq paint back in the 50's. The reliability aspects related to redplating and outgassing if a valve can't be adequately cooled is certainly a salient reminder about caring for those expensive output valves.
Pearl coolers have a good reference link and background info, although the technical effort stopped a long time ago so does not include modern techniques or materials, and the articles are devoid of measurement technique. There are also some doubtful comments, especially about the amount of heat absorbed by the glass. Of interest is the use of tempilaq paint back in the 50's. The reliability aspects related to redplating and outgassing if a valve can't be adequately cooled is certainly a salient reminder about caring for those expensive output valves.
> Of interest is the use of tempilaq paint back in the 50's.
TempiLaq is TempStick in a bottle; available from the same company.
You may be right that the lacquer would be more common on tubes. TempStick has remained a front-line tool in welding and other ironmongery, being quick and clean enough for the job, while the lacquer has become more niche.
TempiLaq is TempStick in a bottle; available from the same company.
You may be right that the lacquer would be more common on tubes. TempStick has remained a front-line tool in welding and other ironmongery, being quick and clean enough for the job, while the lacquer has become more niche.
You might want to look into this thread: http://www.diyaudio.com/forums/tubes-valves/174246-reduced-heater-voltage-reduce-noise.html
To eliminate the effect of emissivity of the radiator, consider using IR thermometer which measures at two wavelengths.
To eliminate the effect of emissivity of the radiator, consider using IR thermometer which measures at two wavelengths.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.