Hi everyone,
The last few months have brought me some strange experiences in my sourcing of 45 tubes for my DIY amp. I had one seller who refused to provide precise measurements of his tubes, even though he listed them as testing "good". Which to me means nothing. Another one listed measurements and also gave me more details by email, but he backed out once I had purchased the tubes, perhaps because I mentioned that I was going to test them on my Pentavac tube tester.
However, the case that intrigues me the most is an experienced seller with whom I have been doing business for over 5 years. He has been selling tubes for nearly 30 years. He claims that the 45 is an extremely difficult tube to test and that he tests each one on four different calibrated tube tester.
When I ask him for more explanation on these testing difficulties, he vaguely mentions a temperature to obtain at the level of heating the filaments, as well as a "hard" material of it, without really giving more details.
Is the 45 really so difficult to test? I would like to have opinions on the question.
For my part, my Pentavac is equipped with an external power supply in 2.5 or 5V to test among others 45, 2A3, 300B. I have tested several 45 with it and I am always happy with its use. I have several that have tested with excellent emission between 95% and 105% of new for example, but also several that I had to reject. In two recent cases, I have returned them to the seller because they did not match the NOS measurements announced.
Finally, recent experiences have shown me a strong intolerance of some sellers who do not accept that we buy tubes from them if we test them afterwards. De facto, they reject the calibration of our tube tester and do not want to know anything about trading with customers who would be likely to test their tubes later with their own tube tester.
Personally, in this kind of business, it seems to me that we must at best have a common reference. The seller must be able to ensure the quality of the tubes he sells and the serious buyer, to be able to check everything after purchase.
Sébastien
The last few months have brought me some strange experiences in my sourcing of 45 tubes for my DIY amp. I had one seller who refused to provide precise measurements of his tubes, even though he listed them as testing "good". Which to me means nothing. Another one listed measurements and also gave me more details by email, but he backed out once I had purchased the tubes, perhaps because I mentioned that I was going to test them on my Pentavac tube tester.
However, the case that intrigues me the most is an experienced seller with whom I have been doing business for over 5 years. He has been selling tubes for nearly 30 years. He claims that the 45 is an extremely difficult tube to test and that he tests each one on four different calibrated tube tester.
When I ask him for more explanation on these testing difficulties, he vaguely mentions a temperature to obtain at the level of heating the filaments, as well as a "hard" material of it, without really giving more details.
Is the 45 really so difficult to test? I would like to have opinions on the question.
For my part, my Pentavac is equipped with an external power supply in 2.5 or 5V to test among others 45, 2A3, 300B. I have tested several 45 with it and I am always happy with its use. I have several that have tested with excellent emission between 95% and 105% of new for example, but also several that I had to reject. In two recent cases, I have returned them to the seller because they did not match the NOS measurements announced.
Finally, recent experiences have shown me a strong intolerance of some sellers who do not accept that we buy tubes from them if we test them afterwards. De facto, they reject the calibration of our tube tester and do not want to know anything about trading with customers who would be likely to test their tubes later with their own tube tester.
Personally, in this kind of business, it seems to me that we must at best have a common reference. The seller must be able to ensure the quality of the tubes he sells and the serious buyer, to be able to check everything after purchase.
Sébastien
Well, last time I checked, I've got around 90- 45 Triodes from various manufacturers. Manufacturers include: RCA, Sylvania, Philco, Tung-Sol, Ken-Rad, GE, Zenith and probably some others (I'm not home now, so can't go digging in the boxes).
I don't find the 45 a difficult tube to test, however, I don't use any sort of classic tube tester or curve tester. I use an actual SET amplifier on the test bench. I can easily measure plate current, grid bias voltage, grid current draw (if any), gain, distortion, power output, signal-to-noise and frequency response. I find actual operating conditions a better way to test and match them into pairs, etc. I also do additional matching of the input/driver tubes and rectifier tubes to create "sets of matched tubes" for all of my amplifiers.
My main concerns with 45 DHTs are physical construction and signal-to-noise that can be achieved under actual conditions. Old Globe tubes are pretty, but usually quite poor in S/N and generally quite sensitive to any mechanical excitation. I'm not a fan of RCA 45s either. All of my best tubes are: Sylvania, Ken-Rad, Philco and Zenith. Granted, Zenith did no tubes but sourced from Sylvania mostly. The internal construction confirms this.
As for overall performance, the best 45s will generally output 2.25 watts before clipping, have less than 0.5% THD at 1-watt output and have a S/N over 80dB referenced to 1-watt output. I generally run the 45 at 34ma current and ~300 volts on the plate. Note that later 45 specs show plate voltage at 300 Volts and plate dissipation at 10 watts. With adequate ventilation, running them at the above ratings is perfectly fine, albeit likely not with the older globe variants. My current matched tube sets are running close to 90dB S/N referenced to 1-watt and around 0.35% THD at 1-watt at 1KHz.
Of course, your results might vary, but this is how I match 45 and 2A3 DHTs for all of my amplifiers. Hope this helps... have fun!
I don't find the 45 a difficult tube to test, however, I don't use any sort of classic tube tester or curve tester. I use an actual SET amplifier on the test bench. I can easily measure plate current, grid bias voltage, grid current draw (if any), gain, distortion, power output, signal-to-noise and frequency response. I find actual operating conditions a better way to test and match them into pairs, etc. I also do additional matching of the input/driver tubes and rectifier tubes to create "sets of matched tubes" for all of my amplifiers.
My main concerns with 45 DHTs are physical construction and signal-to-noise that can be achieved under actual conditions. Old Globe tubes are pretty, but usually quite poor in S/N and generally quite sensitive to any mechanical excitation. I'm not a fan of RCA 45s either. All of my best tubes are: Sylvania, Ken-Rad, Philco and Zenith. Granted, Zenith did no tubes but sourced from Sylvania mostly. The internal construction confirms this.
As for overall performance, the best 45s will generally output 2.25 watts before clipping, have less than 0.5% THD at 1-watt output and have a S/N over 80dB referenced to 1-watt output. I generally run the 45 at 34ma current and ~300 volts on the plate. Note that later 45 specs show plate voltage at 300 Volts and plate dissipation at 10 watts. With adequate ventilation, running them at the above ratings is perfectly fine, albeit likely not with the older globe variants. My current matched tube sets are running close to 90dB S/N referenced to 1-watt and around 0.35% THD at 1-watt at 1KHz.
Of course, your results might vary, but this is how I match 45 and 2A3 DHTs for all of my amplifiers. Hope this helps... have fun!
New tubes differ by as much as 20% when measured at data sheet conditions. Tube testers differ - the operating point (which is nothing like the operating point in an amplifier) is different for every tube you plug in. I've had a tube that read higher than another on one tester read as the lower of the two on another tester. If you only accept tubes that test 95% to 105% on your tester, nobody will sell to you. I have tubes that measured 40% and 80% on a Hickock that deliver the same undistorted output (after bias adjustment).
My Pentavac tube tester, which is pretty similar to an Amplitrex, allows me to test my 45 at the three operating points described on 45 datasheets. This allows me to properly check the output at different values. In fact, it does not work with a percentage, but rather with the output measurement in mA for a given grid bias and plate voltage.
Of course, I end up using my tubes in my amplifier, but the passage to the tube tester is necessary, because as a friend who knows tubes well and in particular low-power triodes mentioned:
"To really know the quality of a tube, you have to be able to measure the transconductance and the vacuum (often the weak point on the 45). For that, you have to be able to vary the polarization and be able to insert a high resistance on the grid. It is possible to tinker with this on an amp but not really practical. It is also useful to check the output at different HT/polarization values in accordance with the datasheet as well as the regularity of the "cut-off" for different polarization values and only a tube tester can do that."
Sébastien
Of course, I end up using my tubes in my amplifier, but the passage to the tube tester is necessary, because as a friend who knows tubes well and in particular low-power triodes mentioned:
"To really know the quality of a tube, you have to be able to measure the transconductance and the vacuum (often the weak point on the 45). For that, you have to be able to vary the polarization and be able to insert a high resistance on the grid. It is possible to tinker with this on an amp but not really practical. It is also useful to check the output at different HT/polarization values in accordance with the datasheet as well as the regularity of the "cut-off" for different polarization values and only a tube tester can do that."
Sébastien
Tested less than a dozen, but in my limited exprience the globes with their lack of support come with noise and pops, even if they test close to NOS. I'd put more stock in their in circuit performance than their emission. What's left on the market is more likely to be rejects that haven't seem much use.
I have a very old simple tube tester.
I tested used 45 tubes for emission, and for shorts.
I designed and built a very small stereo SE 45 amplifier. It only had a B+ secondary, and a 6.3V filament winding.
I paralleled the 45 filaments, used 2 matched dropping resistors from the ends of the filament secondary, to the filament connections.
The self bias resistor was 820 Ohms, went to two 25 Ohm resistors, to the filament ends (45 filaments in parallel), there was a 1000uF bypass cap across the 820 Ohm resistor to ground.
The channel separation was -40dB from 20Hz to 20kHz, as a result of the un-bypassed 25 Ohm resistors to the bypassed 820 Ohm resistor.
Nobody ever complained about the -40dB channel separation;
That is better than any phono cartridge that I ever heard, all the way from 20 to 20k.
Many SE output transformers do not have exactly equal DCRs. So put a series resistor from the lowest DCR, you can match them that way.
The 45 plate impedance is in series with the primary DCR, make the DCRs equal.
Get set to measure the voltage drop from each plate to B+
I = Vdcr / DCR
Push pull output transformer's push and pull winding's DCRs are not always equal. With low rp triodes, you can use that trick of a series resistor to the lowest DCR winding.
With the high rp's that you get with pentodes and beam power output modes, you do Not need to match the DCRs.
I put in a pair of 45 tubes, and replaced them one at a time, rotating out the first one after comparing it to the other 3, rotating the other tubes again until you find a pair that has similar plate currents. You have to record the currents of each unique pairing of quite a few 45 tubes.
Hard to write up the exact procedure, but you get the idea.
Stereo SE is so much easier, plate currents with air gapped transformers that do Not have to have exact plate currents in the L and R channels.
Push pull currents have to be very close, or your non-air-gaped push pull output transformer swill run away with your spouse who is now with a better amplifier designer.
Everybody at work loved the "sound" of my system . . . CD player, stereo SE 45 amplifier, and Usher S 520 two way loudspeakers.
Conclusion:
The amplifier you build and put your 45 tubes into is the very best tube tester.
I hope some of you are learning about things you never thought about before.
Have Fun!
I tested used 45 tubes for emission, and for shorts.
I designed and built a very small stereo SE 45 amplifier. It only had a B+ secondary, and a 6.3V filament winding.
I paralleled the 45 filaments, used 2 matched dropping resistors from the ends of the filament secondary, to the filament connections.
The self bias resistor was 820 Ohms, went to two 25 Ohm resistors, to the filament ends (45 filaments in parallel), there was a 1000uF bypass cap across the 820 Ohm resistor to ground.
The channel separation was -40dB from 20Hz to 20kHz, as a result of the un-bypassed 25 Ohm resistors to the bypassed 820 Ohm resistor.
Nobody ever complained about the -40dB channel separation;
That is better than any phono cartridge that I ever heard, all the way from 20 to 20k.
Many SE output transformers do not have exactly equal DCRs. So put a series resistor from the lowest DCR, you can match them that way.
The 45 plate impedance is in series with the primary DCR, make the DCRs equal.
Get set to measure the voltage drop from each plate to B+
I = Vdcr / DCR
Push pull output transformer's push and pull winding's DCRs are not always equal. With low rp triodes, you can use that trick of a series resistor to the lowest DCR winding.
With the high rp's that you get with pentodes and beam power output modes, you do Not need to match the DCRs.
I put in a pair of 45 tubes, and replaced them one at a time, rotating out the first one after comparing it to the other 3, rotating the other tubes again until you find a pair that has similar plate currents. You have to record the currents of each unique pairing of quite a few 45 tubes.
Hard to write up the exact procedure, but you get the idea.
Stereo SE is so much easier, plate currents with air gapped transformers that do Not have to have exact plate currents in the L and R channels.
Push pull currents have to be very close, or your non-air-gaped push pull output transformer swill run away with your spouse who is now with a better amplifier designer.
Everybody at work loved the "sound" of my system . . . CD player, stereo SE 45 amplifier, and Usher S 520 two way loudspeakers.
Conclusion:
The amplifier you build and put your 45 tubes into is the very best tube tester.
I hope some of you are learning about things you never thought about before.
Have Fun!
Last edited:
I once had two different tube testers that were both popular and well respected at the time. After buying 100 new boxed 6U8 tubes for an amp I built several of long before Tubelab existed I learned that both tube testers were good at weeding out tubes that were bad, like non functional kind of bad. The numbers read from the meter on the front had poor correlation with how that tube worked in a typical audio amp circuit. The circuit used the pentode for voltage gain and the triode for the phase inverter. The triode section was capacitor coupled to the pentode's plate so DC offsets were not a factor. There were a few tubes with low gain compared to the average of the rest, and a few with high THD. The two stage circuit offered THD readings in the 0.5 to 1% range on most tubes, but a few were way off, like 4 to 10%. All of them read good on both testers. This, and a few other similar experiences led to the sale of both testers. The best tester is the actual circuit in which the tubes will be used. When I have a lot of tubes to test, I rig up such a circuit.
The 45 is no different from many tubes in this respect. There are a wide range of possible uses for a 45, and many tubes will not work in all of them. Build a SE amp that runs the 45 in class A fixed bias SE with a 5000 ohm OPT and 280 or so volts of B+ and bias it up to somewhere between 27 and 32 mA and most old tubes will give a useful performance. I run my TSE-II with 320 volts of B+ and about 27 mA of current. Again most tubes will work OK, but some may not make 2 watts of undistorted power output, while a few will make over 3 watts.
Wire up a pair in fixed bias push pull run in class AB2 with a 3300 ohm OPT running from 300 volts and see how many tubes cam make over 15 watts. I have about 12 old, well used tubes, but I can't get two useful pair from my collection in this circuit, I am not abusing the tubes, since this setup comes directly from the 1941 vintage Australian data sheet for the "type 45 Radiotron." Many of the well used tubes have ample emission for class A SE, but not enough to meet the 65 to 70 mA peak plate current demands of the push pull class AB2 circuit.
The 45 is no different from many tubes in this respect. There are a wide range of possible uses for a 45, and many tubes will not work in all of them. Build a SE amp that runs the 45 in class A fixed bias SE with a 5000 ohm OPT and 280 or so volts of B+ and bias it up to somewhere between 27 and 32 mA and most old tubes will give a useful performance. I run my TSE-II with 320 volts of B+ and about 27 mA of current. Again most tubes will work OK, but some may not make 2 watts of undistorted power output, while a few will make over 3 watts.
Wire up a pair in fixed bias push pull run in class AB2 with a 3300 ohm OPT running from 300 volts and see how many tubes cam make over 15 watts. I have about 12 old, well used tubes, but I can't get two useful pair from my collection in this circuit, I am not abusing the tubes, since this setup comes directly from the 1941 vintage Australian data sheet for the "type 45 Radiotron." Many of the well used tubes have ample emission for class A SE, but not enough to meet the 65 to 70 mA peak plate current demands of the push pull class AB2 circuit.
Tube testers are useful but they have their limits. The AVO Mk4 is very good for finding shorts and leakage, and will tell me if something is useful.
However for power tubes, especially low Rp triodes, lashing up a bench jig with powers supplies and a signal generator to inject a known AC voltage over the bias works. With a reasonably calibrated 10 ohm resistor in the anode circuit and an AC+DC dual display DMM (I use a Fluke 45), can read anode current and interpolate Gm in one measurement. Otherwise, a known amplifier is the best test as remarked by others.
However for power tubes, especially low Rp triodes, lashing up a bench jig with powers supplies and a signal generator to inject a known AC voltage over the bias works. With a reasonably calibrated 10 ohm resistor in the anode circuit and an AC+DC dual display DMM (I use a Fluke 45), can read anode current and interpolate Gm in one measurement. Otherwise, a known amplifier is the best test as remarked by others.
Use a VOM to monitor the filament voltage of the 45 tube while testing. If your tester does not supply 2.5V when set to 2.5V filament or if the filament voltage drops more than .1V while testing, your tube tester is having a hard time testing this tube. Most vintage tube testers do a poor job of supplying 2,5V to the filament.
The more a 45 tube is used the more critical filament voltage is in testing. A new tube may pass at a filament voltage of 2V.
The more a 45 tube is used the more critical filament voltage is in testing. A new tube may pass at a filament voltage of 2V.
I don't believe the 45 is a difficult tube to test, maybe the seller doesn't have a tester that can give results that relate back to the makers data, simply emission tube testers have meter scales marked simply good/bad and some 'high end' tube testers are able to check anode current and transconductance but the results are abitary numbers that don't really relate back to the tube makers data. I'm not implying that the seller is dishonest - maybe they simply can't give you the info you want.
I've 2xAvo's and also a home built 'RAT' tube tester as designed by Steve Bench, all of these will give data that relates back to the makers data sheet ie anode current and Gm at a specific bias voltage - the RAT tube tester is my go-to as it just so quick and easy to use.
I've 2xAvo's and also a home built 'RAT' tube tester as designed by Steve Bench, all of these will give data that relates back to the makers data sheet ie anode current and Gm at a specific bias voltage - the RAT tube tester is my go-to as it just so quick and easy to use.