I tested 34 tubes. Of that 34, 7 tested emission low by amounts ranging from 30% low (which would have no impact in a typical circuit) to 70% low. A few others had emission lower than normal by amounts less than 30%.
Of the 7 tubes that tested bad on emission, all had an increased heater current ratio. Scatter plotting indicates, for example, tubes with emission 50% low will have the heater current ratio about 8% higher than for good tubes.
One tube, British-made Brimar (STC/ITT) has a heater current ratio about 12% high. Further electrical testing and checking the cathode temperature by comparing the red glow in a dark room with an known good AWV tube (AWV is known to have used the correct RCA-spec heater wire alloy) indicates its' heater was made with a different heater alloy.
Another tube tested ok on emission test but had a higher than normal heater current ratio. Further testing proved it to be gassy. Gas is known to confound a standard emission test. So it is reasonable to belive the tube really is low emission, correctly picked by the heater current ratio test. But, as it is gassy, it would be rejected for use anyway.
I entirely agree that more tubes should be tested before putting faith in this, but it IS so far pretty consistent.
The real weaknesses I see is that I tested only one tube type, and b) it may be that that the 7 bad tubes had spray-coated cathodes and the rest slip-coated cathodes. And it may be that very old spray-coated cathodes tend to loose emission more. Spray-caoted cathodes are known to be a little more fragile. LuckyDog wase on the right track, perhaps, in his 10:39 post above.
Of the 7 tubes that tested bad on emission, all had an increased heater current ratio. Scatter plotting indicates, for example, tubes with emission 50% low will have the heater current ratio about 8% higher than for good tubes.
One tube, British-made Brimar (STC/ITT) has a heater current ratio about 12% high. Further electrical testing and checking the cathode temperature by comparing the red glow in a dark room with an known good AWV tube (AWV is known to have used the correct RCA-spec heater wire alloy) indicates its' heater was made with a different heater alloy.
Another tube tested ok on emission test but had a higher than normal heater current ratio. Further testing proved it to be gassy. Gas is known to confound a standard emission test. So it is reasonable to belive the tube really is low emission, correctly picked by the heater current ratio test. But, as it is gassy, it would be rejected for use anyway.
I entirely agree that more tubes should be tested before putting faith in this, but it IS so far pretty consistent.
The real weaknesses I see is that I tested only one tube type, and b) it may be that that the 7 bad tubes had spray-coated cathodes and the rest slip-coated cathodes. And it may be that very old spray-coated cathodes tend to loose emission more. Spray-caoted cathodes are known to be a little more fragile. LuckyDog wase on the right track, perhaps, in his 10:39 post above.
Thanks Kiet, could you post the 'blind' scatter plot, and/or run a swift correlation analysis between ratio and emission without discarding any data points please? I suspect that because variation in 'ratio' you're finding is apparently small, any association with 'emission' may well be random.........though I can see why you believe it might not be, you are also selecting thresholds for 'low emission' and 'ratio' when forming the association. The truth will out in such things just from correlation anaylsis of raw data IME.
I've done some further work on this. More tubes have been tested (directly heated amplifying types). The correlation between filament current ratio and emission degradation still holds, though exceptions do occur.
I've also done a bit more research on the theory side. While DF96's statement that a surfcae whose emissivity changes with temperature necessarily is a coloured (thermodynamic meaning of coloured) surface is wrong, it turns out that with regard to vacuum tube emisssion coatings, that is not the full picture.
It isn't just that the emission coating surface is involved. The oxide coatings used are partially transparent, so that the underlying cathode (or filament) metal plays a part in determinining thermal emissivity. The transparency of typical oxide coatings varies in proportion with wavelength. So, as temperature is raised, the effective emissivity is less determined by the underlying cathode/filament metal and more determined by the emissivity of the oxide coating.
In effect, an oxide coated metal cathode/filament IS coloured. Not because the oxide coating is coloured (it isn't), but because the effective depth of thermal radiation penetration varies with wavelength. A very thick coating would show an emissivity varying less with temperature.
I've also done a bit more research on the theory side. While DF96's statement that a surfcae whose emissivity changes with temperature necessarily is a coloured (thermodynamic meaning of coloured) surface is wrong, it turns out that with regard to vacuum tube emisssion coatings, that is not the full picture.
It isn't just that the emission coating surface is involved. The oxide coatings used are partially transparent, so that the underlying cathode (or filament) metal plays a part in determinining thermal emissivity. The transparency of typical oxide coatings varies in proportion with wavelength. So, as temperature is raised, the effective emissivity is less determined by the underlying cathode/filament metal and more determined by the emissivity of the oxide coating.
In effect, an oxide coated metal cathode/filament IS coloured. Not because the oxide coating is coloured (it isn't), but because the effective depth of thermal radiation penetration varies with wavelength. A very thick coating would show an emissivity varying less with temperature.
It’s interesting that the apparently popular Orange VT1000 Valve Tester https://www.orangeamps.com/products/accessories/amplifier-management/vt1000/ rates tubes operating with 2 Volts on normally 6.3V heaters (see attached).
From my reading of credible old papers, it's a bad idea to run oxide cathodes at saturation, eg by partially heating the cathode and measuring saturation current. This is because only a small part of the cathode might be emissive, so carries all current, so gets hotter so becomes more emissive so more current so gets hotter etc etc until thermal gradient within the cathode might become damaging. So I read. NB this doesn't apply to Keit's method, BTW.
Three questions:-
1. Where did the attachment come from? Who is Gerald Mackelburg?
2. Why test at heater voltage of only 2 V? Why?
3. Noting that in accordance with the well known Richardson-Dushment equation and the Stefan-Boltzman law, a tube cathode designed for an maxium emission of 300 mA at 6.3 V will at 2V have an emission of roughly 30 nA, and the notes on page 2 state that testing was done at a cathode current of around 17 mA, is this attachment complete bull dust?
Why indeed.......? I'm just guessing but the notion might be so as to keep saturation current to a non-destructive level by intentionally having a low emission cathode to keep it cool. But I don't think it works like that, more likely the cathode becomes non-uniformly emissive.
If that method does work in predicting low emission cathodes, perhaps the parts of the cathode which are naturally hotter anyway become low-emission first ? Not advocating, just curious. I still don't like the idea of running oxide cathodes at saturation current for very long.
Last edited:
You've mised my point. Here it is agian in more detail: The optimum operating temperature for an oxide cathode is 1050 K. This is the temperature attained at nominal heater voltage (6.3 V). The temperature attained is the temperature that electrical input power equal to that lost by thermal radiation. (Emission current also supplies a very small amount of heat, and electron emission cooling occurs, analogous to the cooling by evaporation of any substance - but these effects are very small and tend to balance out themselves). Electric power in equals V x I. Radiation is in accordance with the Stefan-Boltzman Law. Heater resistance is almost proportional to temperature.
Hence the cathode temperature with 2 V heating can be readily calculated, and it is about 560 K.
The saturation emission of a cathode at any temperature is given by the Richardson-Dushman equation. Plugging in the standard constants for typical oxide cathodes, for a cathode than can emit 300 mA at 1050 K will only emit of the order 30 nano amps at 560 K.
In other words, a current so low it would be difficult to measure. Yet the document jmack posted says the tests are done with a cathode current around 17 mA.
Either there is a mistake or the doco is bull dust. I was hoping jmack would say where he got it. Or who Gerald Mackleburg is - his name is embedded in the document properties as author.
That depends on why the parts are hotter. There are competing mechanisms. For example, if emission falls because oxide material is lost, thermal emissivity goes down, so it gets hotter - however the effect on temperature is mimimal - it has to be as the oxide surface is held close to teh filament temperature. If emission falls because of iron burns, thermal emissivity increases, lowering temperature.
That's what the tube testers sold in the 1930's through to the 1960's did. The tube industry standard way to test emission is to run the heater at rated voltage (eg 6.3 V) and connect 30 V DC to the anode and all grids. The current that flows is the saturation current.
However in tube testing they only applied anode voltage long enough to get a reading - a second or so. Providing its only for a moment, a good tube will recover while in service.
Drawing off saturation current at rated heater voltage is thus potentially harmfull. But doing it at lower heater voltage is even more harmfull.
It is posible to warm up a battery tube (eg 1T4, 1R5) on a high heater voltage (around 1.8 V for a 1.4 V heater), draw saturation current from the anode - this current will be substantially greater than the 1.4 V saturation current - and turn off the heater. Some tubes will keep going! They can keep going if the emission current is large enough to cause I-squared-R heating of the filament. However:-
a) Doing that is VERY bad for the tube; and
c) A tube that doesn't keep going is not necessarily a bad tube (or at least it wasn't bad before you tried this trick)
And the tester isn't designed for such tubes anyway.
Last edited:
Here’s how Orange responded to my questions about that and other issues in my first email to them (attached):
“Hi Gerald, thank you for contacting us regarding your VT1000 Valve Tester.
Whilst we cannot go into specific details of its operation and design (for obvious intellectual property reasons) I can hopefully address some of your comments.
The maximum voltages and currents used do not approach those in a high power amplifier for a lot of reasons; primarily for user safety and this is also the case with many traditional commercial valve testers. However, the VT1000 overcomes this through a highly complex series of software algorithms and thousands of measurements. For this reason, individual static measurements do not accurately reflect the full picture regarding the software’s operation. On the subject of the emissions testing, providing the valve’s heater filament is glowing your unit is functioning correctly; at 1.7V the filament would not glow.
The whole philosophy behind the VT1000 was to provide the user with a fast, low cost, portable and simple to use tester which can be operated by professionals and laymen alike and this has been borne out by the many good magazine reviews it has received. It is designed as a straight forward solution for guitarists, retailers, technicians and valve amp enthusiasts who are looking for a handy tool to test, grade and match valves easily and accurately, without having to interpret the complex results themselves. Thousands of hours of testing, research and development have gone into making and refining the VT1000 to a point where many hobbyists and professionals are using it to grade and test valves on a daily basis.
I hope this answers your questions.”
After providing my measurement data to Orange and pressing for an answer Orange responded: “Hi Gerald, yes it is functioning as it should.”
To which I responded: “The attached modified Owner’s Manual Page 12 reflects these facts. Note that I only removed those claims that were easy to disprove and blatantly deceptive. The second page of the attachment provides justification for the modifications (in a format that should be easy for “professional users [and other interested parties] to understand” if necessary).”
Orange then responded: “Hi Gerald, whilst we thank you for your comments unfortunately we are not at liberty to discuss the specific workings of the product and the processes involved in conducting the tests listed in the literature as this would compromise both the design and intellectual property associated with the product. Sorry we cannot be of any further assistance.”
“Hi Gerald, thank you for contacting us regarding your VT1000 Valve Tester.
Whilst we cannot go into specific details of its operation and design (for obvious intellectual property reasons) I can hopefully address some of your comments.
The maximum voltages and currents used do not approach those in a high power amplifier for a lot of reasons; primarily for user safety and this is also the case with many traditional commercial valve testers. However, the VT1000 overcomes this through a highly complex series of software algorithms and thousands of measurements. For this reason, individual static measurements do not accurately reflect the full picture regarding the software’s operation. On the subject of the emissions testing, providing the valve’s heater filament is glowing your unit is functioning correctly; at 1.7V the filament would not glow.
The whole philosophy behind the VT1000 was to provide the user with a fast, low cost, portable and simple to use tester which can be operated by professionals and laymen alike and this has been borne out by the many good magazine reviews it has received. It is designed as a straight forward solution for guitarists, retailers, technicians and valve amp enthusiasts who are looking for a handy tool to test, grade and match valves easily and accurately, without having to interpret the complex results themselves. Thousands of hours of testing, research and development have gone into making and refining the VT1000 to a point where many hobbyists and professionals are using it to grade and test valves on a daily basis.
I hope this answers your questions.”
After providing my measurement data to Orange and pressing for an answer Orange responded: “Hi Gerald, yes it is functioning as it should.”
To which I responded: “The attached modified Owner’s Manual Page 12 reflects these facts. Note that I only removed those claims that were easy to disprove and blatantly deceptive. The second page of the attachment provides justification for the modifications (in a format that should be easy for “professional users [and other interested parties] to understand” if necessary).”
Orange then responded: “Hi Gerald, whilst we thank you for your comments unfortunately we are not at liberty to discuss the specific workings of the product and the processes involved in conducting the tests listed in the literature as this would compromise both the design and intellectual property associated with the product. Sorry we cannot be of any further assistance.”
That would be only for continuity testing. The unit would switch to full heater voltage after that for the emissions rating. No other way to light the heater. The manual states it uses design voltage ratings.
Should say? You'll have to explain it in a post because clicking on links to untrusted pdf files is a big no-no. Is it your pdf? The 2v. , if it's really 2v. would be continuity testing before any other high voltage test would be initiated.
20to20
Yes, as Keit pointed out, I forgot to remove the PII from my posted attachments - I did make the measurements and produce those PDFs as well as this one. (The “my measurement data to Orange” referred to in post #70). I guess it’s up to you/others to determine if my posts are to be trusted. Do your work or distribute for Orange?
Sorry, I had to split the attachment into 6 parts to meet the .pdf attachment size constraints! So here they are.
Yes, as Keit pointed out, I forgot to remove the PII from my posted attachments - I did make the measurements and produce those PDFs as well as this one. (The “my measurement data to Orange” referred to in post #70). I guess it’s up to you/others to determine if my posts are to be trusted. Do your work or distribute for Orange?
Sorry, I had to split the attachment into 6 parts to meet the .pdf attachment size constraints! So here they are.
Nope, only know them from my neighbor's guitar amp. But if your tester lights the heaters then there is more than 2v. there. There isn't a 6v. heater in any tube that'll light on 2v. So when the Orange rep said that if the tube lights, then it's working right, he would know you're getting more than 2v. there. If not, it's broke.
Just responding to your post about the tester running the heaters at 2v during the testing. That's all. Do the heaters light to orange/yellow? Let's get that one out of the way and go from there....
I'm an engineer so I prefer objective measurements over subjective observations whenever possible. If you have access to a VT1000 and a voltmeter (the Harbor Freight (HF) freebees are fine) and a tube - simply measure the DC voltage on the heater during a tube test. I just measured 1.89 VDC with both my Fluke and freebee HF digital multi-meters while testing an EL34 (1.5A @ 6.3 VRMS heater) that received a "Matching Value" of 11. Those meters seem to do a surprisingly good job measuring the RMS content of PWM voltages. If you have access to an oscilloscope you should be able measure/calculate the RMS heater voltage directly from the PWM waveform (as I did in the .pdf).
Those HF meters multiply like mice. I have at least 6 with dead batteries... Can't resist those coupons. But I use a Fluke 77, and an Elenco LCM-1950 meter, mostly. I don't use my scope for checking heater voltages. If they are glowing I move to the next step... That's hot enough for the K to do its job. Hot enough for a tester to get a proper emission reading. I'm going to interpret your reply as a, yes, glowing just fine.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.