Thermal Image of a Quad 2

[_Clint_]

I have couple of Quad 2's I started restoring a couple of years ago (long story) I only got as far as stripping them down, anyway to help reassembly I brought another one, I could not resist taking a look through my thermal camera which I thought I would share:

[Nickthedentist]

Fascinating. Thanks for sharing, Clint.

[Herald1360]

An underside view including the questionable cathode resistor might be interesting.......

[GrimJosef]

It might also be interesting to leave it running for three hours and then photograph it again. The transformers and choke do warm significantly. It's remarkable that they've lasted as long as many of them have given that they were often (usually ?) squirreled away in the bottom of some more or less well ventilated cabinet.

Cheers,

GJ

[_Clint_]

Will Get the shots tomorrow

[GrimJosef]

I guess the valve glass is opaque to much of the infrared, so the bright colours reflect the temperature of the glass envelopes, not the internal structures. It does show how cool the KT66s' getters (right at the bottom of the envelope) are. They work much better as gas scavengers when they're hot and I've tried turning amps upside down, so the external air convection heats the bottom of the glass, to help with this.

Cheers,

GJ

[julie_m]

Originally Posted by GrimJosef

I guess the valve glass is opaque to much of the infrared, so the bright colours reflect the temperature of the glass envelopes, not the internal structures.

I thought ordinary quartz glass blocked UV, but not so much IR? The radiant tube heaters you see in railway waiting rooms and pub smoking areas sometimes have a glass screen. And since radiation is the only way any energy can get through the vacuum to the outside world, it would make the most sense for valves to be as IR-transparent as possible.

[Bazz4CQJ]

I never looked at any valves with a thermal camera but have used one to identify faults on large PCB's. Especially in solid-state equipment, components that are running hot can be quickly and easily identified even when the applied voltage is much lower than normal operating voltage. Applying 40 volts from a Variac to mains driven equipment, I've seen resistors "glowing" brightly against an otherwise blank background within a few seconds of switch-on. In reality, they are just a few degrees above ambient.

B

[trobbins]

There was a long thread discussion where the topic of valve glass and anode temperatures and the transparency of the glass to that emitted radiation and the glass temperature dependence on different cooling mechanisms was discussed.

http://www.diyaudio.com/forums/tubes...nks-tubes.html

The physics gets a bit involved. The valve glass is not 'quartz' type. The plate radiates a 'spread' of radiation frequencies, and the glass acts as a sort of a high-pass filter, and the amount of heat transferred through the glass, or absorbed by the glass, depends on the frequency of the radiation, so it is all a bit of a mathematical integration nightmare. And that is just from the plate to the glass, let alone having other heat sources radiating on to the glass from the outside, and the affect of air cooling.

That said, a thermal camera would be great to play with and try and deduce what temp things like getters reach.

[GrimJosef]

Originally Posted by ajs_derby

I thought ordinary quartz glass blocked UV, but not so much IR?

Ordinary glass is not strictly quartz (actually pure quartz is a relatively good UV transmitter) but it does absorb across a range of wavelengths longer than about 2um or so.

The radiant tube heaters you see in railway waiting rooms and pub smoking areas sometimes have a glass screen. And since radiation is the only way any energy can get through the vacuum to the outside world, it would make the most sense for valves to be as IR-transparent as possible.

Radiation is not the only way to the outside world. Radiation transports heat to the inside of the glass where it's absorbed, heating the glass up (as you'll know if you've touched it !). The heat can then travel through the glass by conduction and be lost to the outside world by a combination of all three major transport processes - conduction to the air, radiation into the space surrounding the valve and convection of the air past the glass surface. There may well be a good case for making the envelope as IR-transparent as possible. But in practice if we want to use glass then it's simply impossible to make it completely transparent. I used to rub shoulders at work with IR spectroscopists and they go to enormous lengths to find window materials which transmit better than glass http://infrared.als.lbl.gov/content/...ndow-materials. Large salt crystals were among their favourites. Woe betide anyone who ever let these go cool enough for condensation to form on them though ...

Cheers,

GJ

[Tractorfan]

Hi,
I would have expected the small valves to the left of the photo to have shown up more than they do. Unless the pic was taken very soon after switch-on.
Cheers, Pete.

[_Clint_]

I can not seem to work out how to put images inline with the text so I will create a few posts with the relevant photos attached.

Anyway I hope you find this interesting.

Most glass is not transparent to the infrared sensor, even the Germanium lens used on the actual camera is not fully transparent but is about as good as you can get. However the closer to the lens, the object glass is inside the cameras focal length, the less of an affect it has; So if your getting chased by a helicopter with a thermal imaging gimbal then hide in a greenhouse.

The camera I use overlays the outlines seen from its inbuilt standard camera onto the thermal image so its easier to see where you are. You can see this in the first picture attached to this post. In the second I have removed the overlay, and the third image shows a combined standard camera and pure thermal image, that clearly shows the edge of a plate. The forth is a top view of the heater element.

Picture 5 clearly shows a dead tube, who needs a valve tester (I have one so no need to answer that)

You can also clearly see the image is coming from the plates (anodes) rather than the surrounding glass.

[_Clint_]

As people asked about the EF86's I took a couple of images for you.

[_Clint_]

Now one interesting this is that people seem to think the mains transformer gets hot and is therefore under rated, now I am not saying it isn't but from my investigation with the thermal imaging the reason it gets hot is clearly heat soak from the tubes, these shots are 2 hours running.

[_Clint_]

Now here is a biggie, R12 - this shows how seriously out of specification the standard 3.75W resistor is running; now I do not know the original specification but most modern good resistors have a safe MAXIMUM running temp of 200 degrees. As you can see the standard resistor is at 280 degrees with normal running.

I swapped the resistor to a Panasonic 5W and then a 7W.

Here are the results:

Standard 280 degrees
5W Pana 137 degrees
7W 99 degrees

[GrimJosef]

Originally Posted by _Clint_

... Picture 5 clearly shows a dead tube, who needs a valve tester (I have one so no need to answer that) ...

Yes, but which one ? Is the left one dead with too much gas, so it's passing too much current, or is the right one dead with too little emission, so it's not passing enough current ? You're right about the valve tester though. It may not get the valve hot enough for the gas pressure to rise as it does when running at full power in the amp. A DMM will tell us quite quickly. Measuring the voltages across the output transformer windings UV and VW gives the cathode currents in the valves and measuring Vgk for each of them allows those currents to be compared with the datasheet.

Originally Posted by _Clint_

Now here is a biggie, R12 - this shows how seriously out of specification the standard 3.75W resistor is running; now I do not know the original specification but most modern good resistors have a safe MAXIMUM running temp of 200 degrees. As you can see the standard resistor is at 280 degrees with normal running ...

Well ... if we don't know the specification then strictly speaking I don't think we really can tell from the picture how far out of spec the resistor is running. In fact I think it was specified at 3W and it actually runs at a little under 3.8W quiescent if everything else is working properly, so about 27% above nominal maximum. One way of working out if it has been over-stressed would be to check if, after 50 years or so, it has burnt out or the casing has cracked or the resistance has changed significantly. If you get the chance to check I think you might be surprised at the result .

EDIT: It's interesting to see the detailed differences between the old and new R12s. The original one and the 5W Panasonic appear to be roughly the same physical size and, in the steady state, they have to be getting rid of the same electrical power through radiation, conduction and convection, all of which depend primarily on physical size and surface temperature. So we might think that for the same power they should both be at the same surface temperature. It's clear though that the new resistor is nicely designed so it is fairly uniformly hot (or cool, if you like) all over whereas the original one is cooler at the ends and has an unwelcome hotspot in the middle. Technology does indeed advance !

VB

[trobbins]

The valve glass will be more opaque for radiation coming from lower temperature surfaces. The IR camera image of the bottom/base and top of the KT66 would then be mainly representing the radiation coming from the outer surface of those areas of the glass envelope.

In the central section with the anode, radiation from inside the glass envelope from surfaces that are a lot hotter will be more 'visible' in the camera image. The anode is obviously visible, but also the inner surface of the glass envelope. The outer surface of the glass envelope is being cooled by the air.

Luckily you're not using pre-1961 KT66s !

[_Clint_]

Oh nearly forgot one of the Tag board too

[GrimJosef]

The thermal image is quite displaced relative to the non-thermal one on the tagboard. The hot resistors are not the 1Ms in the EF86 screen circuits, towards the 'top' of the board, but the 180k (well, they would have been 180k once) anode loads which are the second ones down.

Cheers,

GJ