|
Post by audiobill on Jun 13, 2019 18:27:17 GMT -5
You’re fine, just a longer delay.
|
|
|
Post by ttocs on Jun 16, 2019 20:50:16 GMT -5
As Josh famously said in the movie "Big", "I don't get it".
When I plug a solid state amp into a Kill-A-Watt meter and set it to Amps or Watts, the transients will make the reading change A LOT. But when I do the same with the tube amp it stays at 168 watts, end of story. Can one of you engineer types please explain why there's no change in current draw?
With my Krell the readings would bounce from just under 3 Amps to about 8 Amps, as measured with a Fluke meter which can only record as quick a 1ms. These tube amps are using less than 1.75A each, so basically a little more than the minimum amount for the Krell when using two channels.
|
|
KeithL
Administrator
Posts: 9,934
|
Post by KeithL on Jun 16, 2019 23:52:38 GMT -5
A typical Class A/B solid state amp is about 70% efficient after it's warmed up (in about five minutes). However, it will draw a large surge, for a few seconds, to fill the storage capacitors. After that, other than a small amount of idle current, it is about 70% efficient. In normal operation it draws about 1-1/2 times as much power as it's putting out - plus a small steadly idle power. (Even when you're playing music at a reasonably loud level the average power you're actually putting out is usually only a few watts. A TRUE Class A amplifier, in contrast, is about 25% efficient AT FULL POWER.... However, a Class A amplifier will draw about 4x its rated output ALL THE TIME - REGARDLESS OF ITS OUTPUT LEVEL. The story is quite different for tube amplifiers. For starters, the filament in each tube has to be heated to operating temperature. So, for example, if your amplifier uses KT-120's, EACH ONE uses 6.3v at about 1.75 amps just to heat the cathode (about 10 watts each). And, if it's got tube rectifiers, a 5U4 uses 5v at 3 amps, again just to heat the filament (another 15 watts). The smaller "peanut tubes" draw a few watts each for their heaters. That's power that's used all the time "to light the tubes". Once you get past that, you need to allow some idle current to flow through the tubes, especially the big ones... The "typical" amount of idle current will depend on the design of the amplifier and what the bias is set to. (You can cut the power usage, and actually make the amplifier run cooler, by turning down the bias... However, most tube amps have one particular bias setting where they have the lowest distortion and sound their best... After all that, the overall efficiency is also a bit lower.... But, again, the average audio power you typically use hardly makes any difference.... So MOST of the power used by your tube amp most of the time is due to heater and bias current. The reason that tube amps don't draw a big surge on startup is that tubes only conduct current when they're warm. Therefore, the big power tubes start to conduct gradually, since they take a few minutes to warm up. And, if you have a tube rectifier, it also warms up very gradually, which causes the supply voltage to rise somewhat gradually. If you have a solid state rectifier, it will draw an initial surge, but not a huge one, then the amp will draw gradually more as the big tubes warm up. Most modern solid state amps do have some sort of component that limits the startup surge (called "an inrush limiter"). (I don't have the numbers, but I would expect the inrush on our new XPA Gen3 models to be lower than on our older models, thanks to the SMPS.) As Josh famously said in the movie "Big", "I don't get it". When I plug a solid state amp into a Kill-A-Watt meter and set it to Amps or Watts, the transients will make the reading change A LOT. But when I do the same with the tube amp it stays at 168 watts, end of story. Can one of you engineer types please explain why there's no change in current draw? With my Krell the readings would bounce from just under 3 Amps to about 8 Amps, as measured with a Fluke meter which can only record as quick a 1ms. These tube amps are using less than 1.75A each, so basically a little more than the minimum amount for the Krell when using two channels.
|
|
|
Post by ttocs on Jun 17, 2019 8:41:27 GMT -5
The story is quite different for tube amplifiers. Once you get past that, you need to allow some idle current to flow through the tubes, especially the big ones... The "typical" amount of idle current will depend on the design of the amplifier and what the bias is set to. (You can cut the power usage, and actually make the amplifier run cooler, by turning down the bias... However, most tube amps have one particular bias setting where they have the lowest distortion and sound their best... After all that, the overall efficiency is also a bit lower.... But, again, the average audio power you typically use hardly makes any difference.... So MOST of the power used by your tube amp most of the time is due to heater and bias current. So let me see if I get how this relates to noticing no movement on a meter at high volume even on transients. Would the volume need to be much greater (more than I can tolerate), or, over a threshold level to be able to see fluctuation on a meter set for Current? Or, would the bias need to be much lower to then be able to see the amp starving (exaggeration) and needing more when a transient comes along? Bottom line, at 117V at the wall socket, and the amp drawing 1.75A ALL THE TIME AND THE METER NEVER MOVES, the amp is satisfied with the amount of current to be able to produce sound and even loud transients at the levels at which I've been listening, is that about it Bunky? (Only us old folks know Bunky. www.youtube.com/watch?v=ifNyqcpM4k8 )
|
|
KeithL
Administrator
Posts: 9,934
|
Post by KeithL on Jun 17, 2019 11:46:38 GMT -5
You've got the basic idea. The amount of power you're asking the amp to put out simply doesn't require the amp to draw enough additional power to be noticeable.
Pretty much all of what you're seeing is the idle operating current of the amp... which is significant for tube gear.
In order to see any sort of significant difference, you would probably have to be drawing far more power, and playing things a LOT louder.
In most cases, a small fraction of a watt is enough for background levels, and a few watts sounds reasonable, but you may still need a lot of power to get "really loud". The amount of power we use also depends heavily on the efficiency of your speakers, the size of your room, and how loud you listen to things. The amount of power you use is also dependent on the frequencies involved. One watt of midrange is actually quite loud, but loud bass notes require a lot more power. Our hearing is also logarithmic... which means that, in order to make something sound "8x as loud", you need about 1000x as much power... (Amazingly, the sequence 1x - 2x - 4x - 8x in loudness is roughly equivalent to 1x - 10x - 100x - 1000x in power.)
There are a couple of possible things going on - although they all work out about the same. 1) Your meter only displays to two digits and has averaging built in. An extra draw of a few watts, for a few seconds, to power a loud transient, may not be enough to show up on it.
In order to show up the amp would need to draw enough extra power - on average - to change that reading.
(A really accurate meter, with more digits, or a faster response, or one with an actual pointer, might show a tiny twitch when those transients play.)
2) Almost all modern amplifiers operate in Class A/B. This means that, for the first watt or two, they are running in Class A. As long as you actually remain using less than that amount of power they really won't draw more current at all. (Below that point, the power used to make output is simply subtracted from power that is normally dissipated as heat.)
3) If you were to set the bias way down the extra draw would be more noticeable - but not for the reasons you think. First off, by lowering the average power consumption, the small difference you're causing would become a larger percentage of the total, so more likely to be visible on a meter. Second, by lowering the bias, you reduce the area over which the amplifier operates in Class A. (Therefore, if you're not seeing any difference because the amp is remaining in Class A, and there really is no difference, you will force it to go outside of Class A sooner...) You really shouldn't do that though.... Most amplifiers have a specific bias setting where they have the lowest distortion and sound their best. Setting the bias too far below this will save power, and make the amp run cooler, but will probably adversely affect the way it sounds.
(But, in most cases, it won't actually damage anything.)
I have absolutely seen situations where someone thought the LED meters on an amp were broken... Only to discover that they simply weren't playing their music loudly enough to light the first LED...
Most audio meters are calibrated in a logarithmic scale: 0.01 - 0.1 - 1 - 10 - 100 so you have something to watch. (The dB scale used on most audio gear is a variation of this..... for example, compared to 0 dB, -20 dB is 1/10 the voltage, -40 dB is 1/100, -60 dB is 1/1000, etc.)
(In contrast, meters like the Watt Wizard are linear: 1 - 2 - 3 - 4 - 5 because that's the way the power company charges you for power consumption (more or less) ).
If you want to measure what your amplifier is putting OUT... You want an audio output meter - calibrated in some sort of logarithmic or dB range.
The story is quite different for tube amplifiers. Once you get past that, you need to allow some idle current to flow through the tubes, especially the big ones... The "typical" amount of idle current will depend on the design of the amplifier and what the bias is set to. (You can cut the power usage, and actually make the amplifier run cooler, by turning down the bias... However, most tube amps have one particular bias setting where they have the lowest distortion and sound their best... After all that, the overall efficiency is also a bit lower.... But, again, the average audio power you typically use hardly makes any difference.... So MOST of the power used by your tube amp most of the time is due to heater and bias current. So let me see if I get how this relates to noticing no movement on a meter at high volume even on transients. Would the volume need to be much greater (more than I can tolerate), or, over a threshold level to be able to see fluctuation on a meter set for Current? Or, would the bias need to be much lower to then be able to see the amp starving (exaggeration) and needing more when a transient comes along? Bottom line, at 117V at the wall socket, and the amp drawing 1.75A ALL THE TIME AND THE METER NEVER MOVES, the amp is satisfied with the amount of current to be able to produce sound and even loud transients at the levels at which I've been listening, is that about it Bunky? (Only us old folks know Bunky. www.youtube.com/watch?v=ifNyqcpM4k8 )
|
|
|
Post by ttocs on Jun 17, 2019 12:42:58 GMT -5
You've got the basic idea. The amount of power you're asking the amp to put out simply doesn't require the amp to draw enough additional power to be noticeable.
Pretty much all of what you're seeing is the idle operating current of the amp... which is significant for tube gear.
In order to see any sort of significant difference, you would probably have to be drawing far more power, and playing things a LOT louder.
In most cases, a small fraction of a watt is enough for background levels, and a few watts sounds reasonable, but you may still need a lot of power to get "really loud". The amount of power we use also depends heavily on the efficiency of your speakers, the size of your room, and how loud you listen to things. The amount of power you use is also dependent on the frequencies involved. One watt of midrange is actually quite loud, but loud bass notes require a lot more power. Our hearing is also logarithmic... which means that, in order to make something sound "8x as loud", you need about 1000x as much power... (Amazingly, the sequence 1x - 2x - 4x - 8x in loudness is roughly equivalent to 1x - 10x - 100x - 1000x in power.)
There are a couple of possible things going on - although they all work out about the same. 1) Your meter only displays to two digits and has averaging built in. An extra draw of a few watts, for a few seconds, to power a loud transient, may not be enough to show up on it.
In order to show up the amp would need to draw enough extra power - on average - to change that reading.
(A really accurate meter, with more digits, or a faster response, or one with an actual pointer, might show a tiny twitch when those transients play.)
2) Almost all modern amplifiers operate in Class A/B. This means that, for the first watt or two, they are running in Class A. As long as you actually remain using less than that amount of power they really won't draw more current at all. (Below that point, the power used to make output is simply subtracted from power that is normally dissipated as heat.)
3) If you were to set the bias way down the extra draw would be more noticeable - but not for the reasons you think. First off, by lowering the average power consumption, the small difference you're causing would become a larger percentage of the total, so more likely to be visible on a meter. Second, by lowering the bias, you reduce the area over which the amplifier operates in Class A. (Therefore, if you're not seeing any difference because the amp is remaining in Class A, and there really is no difference, you will force it to go outside of Class A sooner...) You really shouldn't do that though.... Most amplifiers have a specific bias setting where they have the lowest distortion and sound their best. Setting the bias too far below this will save power, and make the amp run cooler, but will probably adversely affect the way it sounds.
(But, in most cases, it won't actually damage anything.)
I have absolutely seen situations where someone thought the LED meters on an amp were broken... Only to discover that they simply weren't playing their music loudly enough to light the first LED...
Most audio meters are calibrated in a logarithmic scale: 0.01 - 0.1 - 1 - 10 - 100 so you have something to watch. (The dB scale used on most audio gear is a variation of this..... for example, compared to 0 dB, -20 dB is 1/10 the voltage, -40 dB is 1/100, -60 dB is 1/1000, etc.)
(In contrast, meters like the Watt Wizard are linear: 1 - 2 - 3 - 4 - 5 because that's the way the power company charges you for power consumption (more or less) ).
If you want to measure what your amplifier is putting OUT... You want an audio output meter - calibrated in some sort of logarithmic or dB range.
This is great stuff Keith! Thanks! This answers why the power transformer gets so hot no matter what the volume is set at. And BTW, the limit for me when testing how loud something can play at and still sound good is around 105dB for a very short period. Sounds great, but still doesn't move the meter. Bear in mind that my current speakers only need the electrostatic panel powered by an external amp, each woofer - 2 in each speaker - has its own amplifier. So the external amps only need to supply power for 300Hz and above. The Kill-A-Watt goes 2 places past the decimal point, and when I plugged in my Krell it would fluctuate between 2.xx to about 5.xx Amps. To get closer to an accurate measurement I rigged up a Fluke meter which is able to record down to 1ms and it came up with transients going close to 8 Amps. That's when I was playing with power cords. Thanks again!
|
|
|
Post by ttocs on Feb 4, 2020 20:34:05 GMT -5
I just popped in some Tung Sol KT120 tubes into my M-125's and notice some interesting things.
Firstly, at the same bias as was used for the KT88's the output is very slightly less, maybe about 1/2dB using tones.
Secondly, and frankly very happily, everything is running cooler!! The transformer with the KT88 tubes would be over 120 degrees after only a half hour and would increase to almost 150 after a few hours. I was so concerned about longevity that I bought some really cool looking retro minifans that run pretty quiet, just aimed at the trans. This kept things under control, about 110 degrees after 4-5 hours. edit: Now the transformers are at about 112 degrees after a couple hours of running!!
The KT120 tubes also run cooler, at about 275 for the rear pair and around 200 for the front pair, measured on the silvered top. Internally the highest temp I can measure is only around 300 degrees. Heck, the KT88's ran at 340 measured on the top and more when pointing the temp gun inside! This is all without fans.
As Tom Cruise said in the movie "A Few Good Men": "Can you explain that?" (When asking Col. Jessep why Santiago wasn't packed).
I expected things to run hotter, not cooler. The KT120 tubes and the driver tubes all exhibit less glow, not as bright as with the KT88's. It's kinda like this circuit was designed for these KT120 tubes.
Sound. It's easier. More comfortable. More effortless. Synergy. "I love it when a plan comes together!" Me very happy. It just keeps getting better!
|
|
novisnick
EmoPhile
CEO Secret Monoblock Society
Posts: 27,223
|
Post by novisnick on Feb 4, 2020 23:18:08 GMT -5
I just popped in some Tung Sol KT120 tubes into my M-125's and notice some interesting things. Firstly, at the same bias as was used for the KT88's the output is very slightly less, maybe about 1/2dB using tones. Secondly, and frankly very happily, everything is running cooler!! The transformer with the KT88 tubes would be over 120 degrees after only a half hour and would increase to almost 150 after a few hours. I was so concerned about longevity that I bought some really cool looking retro minifans that run pretty quiet, just aimed at the trans. This kept things under control, about 110 degrees after 4-5 hours. edit: Now the transformers are at about 112 degrees after a couple hours of running!! The KT120 tubes also run cooler, at about 275 for the rear pair and around 200 for the front pair, measured on the silvered top. Internally the highest temp I can measure is only around 300 degrees. Heck, the KT88's ran at 340 measured on the top and more when pointing the temp gun inside! This is all without fans. As Tom Cruise said in the movie "A Few Good Men": " Can you explain that?" (When asking Col. Jessep why Santiago wasn't packed). I expected things to run hotter, not cooler. The KT120 tubes and the driver tubes all exhibit less glow, not as bright as with the KT88's. It's kinda like this circuit was designed for these KT120 tubes. Sound. It's easier. More comfortable. More effortless. Synergy. "I love it when a plan comes together!" Me very happy. It just keeps getting better! I’ve been running Tung Sol KT120 for a very long time and can concur with your reports. Hot, but not too hot. Sound is just devine. Time to warm up my cave a little. Love your plan! 👍
|
|
|
Post by ttocs on Feb 5, 2020 21:35:53 GMT -5
I found a reply from Bob Latino to a question about KT120 tubes in the M-125 amps: "My answer to your question about longevity is most probably > YES - the Tung-Sol KT120 or KT150 should last longer than any KT88 in either the VTA ST-120 or VTA M-125. Yes - These tubes would be loafing at KT88 voltages and bias points.
Bob"
The KT120 tubes "loafing" would seem to answer why things are running cooler.
|
|
|
Post by ttocs on Feb 9, 2020 23:34:35 GMT -5
Some interesting developments since going to KT120 tubes. But first a bit of background. My main L&R speakers have dual powered woofers, so my tube amps only power the electrostatic panels from 300Hz and up. The day before installing the new tubes I had done a ARC calibration which only calibrates the woofers and is entirely local to the amps built-in to each speaker, nothing else is affected by this particular calibration. The KT120 tubes I get are burned-in at normal operating voltage for 24 hours prior to shipping, so I wasn't expecting much, or any change (maybe) during the initial few days of usage.
Something changed.
During the first 30 hours I felt the stat panels were being overpowered by the bass coming from the woofers compared to what I heard from the KT88 tubes before making the swap. So I altered two settings on each speaker. One is a switch which subtracts or adds 2dB of Mid Bass, I changed this to -2dB. The other is a dial which reduces or adds 10dB of Bass below 75Hz, and I set this to -1dB. Ok, now things sound normal.
After about 8 hours using these settings it seemed like the bottom end was missing, so I changed both settings back to Zero. Now it sounds normal again, or at least closer to normal - but different, not worse, I'm perplexed. So at this point the tubes have about 62 or so hours on them (vendor burn-in plus my hours), plus last night and today I put another 10 hours on them. The sound is now incredible! The top end is back, and maybe even a bit brighter than what the KT88's sounded like, not sure yet but very happy.
Can tubes become brighter sounding as they break-in? Can tubes exhibit more power as they break-in? Because it's one or the other, or a little of both.
|
|
|
Post by audiobill on Feb 10, 2020 6:03:08 GMT -5
New tubes can show extended frequency response compared to used ones. How old were the KT88s?
|
|
|
Post by ttocs on Feb 10, 2020 9:38:40 GMT -5
New tubes can show extended frequency response compared to used ones. How old were the KT88s? When I installed the KT120 tubes they immediately sounded less bright than the KT88's, not a huge difference, but noticeable. After about 30 hours I made the smallest adjustment possible on my speakers for less bass, but after only 8 hours of this adjustment I set the controls back to zero because of a lack of bass and found that the upper freqs were able to compete with the lows. I'm not saying that the issue was suddenly 100% fixed, but it was different (better) from when the tubes were first installed, and seemed to improve with each power-on and more hours. The Mid Bass switch controls freqs between 140-300Hz, and the Bass Level dial controls freqs below 75Hz. In all of this the speakers haven't been moved, nor the speaker cables been touched. I use a Variac to control the incoming voltage to the tube amps so the voltage is always in the same range, and the digital voltage display is always in view. Over time the new tubes seem to have gotten brighter sounding, so now they are probably brighter than the KT88 tubes, and frankly they seem more articulate and a bit more dynamic (maybe articulate and dynamic go together?). The differences are small, but noticeable listening to the same source material. The KT88's have around 1500 hours. Wanting to try the KT120 is why I replaced them, not because they were exhibiting any poor performance. What kind of changes are normally expected to happen with new tubes during initial break-in?
|
|
KeithL
Administrator
Posts: 9,934
|
Post by KeithL on Feb 10, 2020 9:44:57 GMT -5
As tubes age their electrical characteristics change... which can cause a variety of different minor changes in performance. Exactly what those will be depends on the actual tubes involved and the circuitry around them. What you've experienced is pretty typical.
New tubes can show extended frequency response compared to used ones. How old were the KT88s? When I installed the KT120 tubes they immediately sounded less bright than the KT88's, not a huge difference, but noticeable. After about 30 hours I made the smallest adjustment possible on my speakers for less bass, but after only 8 hours of this adjustment I set the controls back to zero because of a lack of bass and found that the upper freqs were able to compete with the lows. I'm not saying that the issue was suddenly 100% fixed, but it was different (better) from when the tubes were first installed, and seemed to improve with each power-on and more hours. The Mid Bass switch controls freqs between 140-300Hz, and the Bass Level dial controls freqs below 75Hz. In all of this the speakers haven't been moved, nor the speaker cables been touched. I use a Variac to control the incoming voltage to the tube amps so the voltage is always in the same range, and the digital voltage display is always in view. Over time the new tubes seem to have gotten brighter sounding, so now they are probably brighter than the KT88 tubes, and frankly they seem more articulate and a bit more dynamic (maybe articulate and dynamic go together?). The differences are small, but noticeable listening to the same source material. The KT88's have around 1500 hours. Wanting to try the KT120 is why I replaced them, not because they were exhibiting any poor performance. What kind of changes are normally expected to happen with new tubes during initial break-in?
|
|
KeithL
Administrator
Posts: 9,934
|
Post by KeithL on Feb 10, 2020 11:22:18 GMT -5
You can expect tubes to continue to change, although progressively less and less, for as long as the first few hundred hours of use. (The composition of the "vacuum" in the tube, and even the coating on the cathodes, can continue to change for that long.)
Some interesting developments since going to KT120 tubes. But first a bit of background. My main L&R speakers have dual powered woofers, so my tube amps only power the electrostatic panels from 300Hz and up. The day before installing the new tubes I had done a ARC calibration which only calibrates the woofers and is entirely local to the amps built-in to each speaker, nothing else is affected by this particular calibration. The KT120 tubes I get are burned-in at normal operating voltage for 24 hours prior to shipping, so I wasn't expecting much, or any change (maybe) during the initial few days of usage. Something changed. During the first 30 hours I felt the stat panels were being overpowered by the bass coming from the woofers compared to what I heard from the KT88 tubes before making the swap. So I altered two settings on each speaker. One is a switch which subtracts or adds 2dB of Mid Bass, I changed this to -2dB. The other is a dial which reduces or adds 10dB of Bass below 75Hz, and I set this to -1dB. Ok, now things sound normal. After about 8 hours using these settings it seemed like the bottom end was missing, so I changed both settings back to Zero. Now it sounds normal again, or at least closer to normal - but different, not worse, I'm perplexed. So at this point the tubes have about 62 or so hours on them (vendor burn-in plus my hours), plus last night and today I put another 10 hours on them. The sound is now incredible! The top end is back, and maybe even a bit brighter than what the KT88's sounded like, not sure yet but very happy. Can tubes become brighter sounding as they break-in? Can tubes exhibit more power as they break-in?Because it's one or the other, or a little of both.
|
|
|
Post by ttocs on Feb 10, 2020 12:48:33 GMT -5
You can expect tubes to continue to change, although progressively less and less, for as long as the first few hundred hours of use. (The composition of the "vacuum" in the tube, and even the coating on the cathodes, can continue to change for that long.)
Thanks Keith. I found during searching that Jolida says their amps need 250-300 hours of use so that the tube can reach a stable thermal operating temp, which agrees with your statement. I've read before that tubes can exhibit a lack of bass initially so I was surprised that there'd be a reduced amount of treble, and in my case the bass below 300Hz is handled by the built-in amps in each speaker so reduced bass is something I wouldn't have noticed. I've also read that tubes can be a bit grainy when new, and I've had a little of that with these tubes but it's in a narrow range of freqs and has reduced. Frankly, late last night the sound was very impressive! Earlier today I read a little something that mentioned about the coating on the cathodes which can alter the sound while the burning off phase is happening. (Seinfeld: The Burning) So now I'm wondering if what I experienced was a lack of range but with an emphasis on lower mid/upper bass? Don't know, just curious.
|
|
|
Post by ttocs on Feb 13, 2020 21:02:43 GMT -5
Ok, so I've been trying to figure out why things are so much better in these M-125 amps with the KT120 tubes vs KT88 mated to my electrostatic speakers which can dip to 0.7 Ohms impedance.
I haven't checked the internal voltages yet to get a glimpse of what might be different but want to do that soon. What I do know is that with KT88 tubes the power transformer would get really hot, over 150 degrees was the highest temp I saw before getting some really quiet fans to keep the temp below 120 degrees. Now with KT120 tubes and no fans the temp hasn't gone over mid 130 range temps.
The output transformer has never gotten hot.
The sound is incredible! Slightly brighter, in all good ways. More dynamic, and with stat panels that's quite a feat.
I'd like to understand why this tube is a better match, both for the fact that the PT isn't as hot as before and the improved sound. Is part of the equation the fact that the KT120 has lower Load Resistance p-p ( 3000 Ohms if the info I've found is accurate vs 4500 Ohms for KT88) making it a better fit for my low impedance speakers? The output transformer Primary Impedance is 2200 Ohms and my speakers are on the 4 Ohm tap.
|
|
KeithL
Administrator
Posts: 9,934
|
Post by KeithL on Feb 14, 2020 10:53:57 GMT -5
In general the KT-120 is usually considered to be "a better tube"...
However, there are a lot of things that determine the sound, and some of them are quite complex. For example, in very general terms, the output transformer is an impedance matching device, which matches the impedance of the speakers to one which the tubes are comfortable driving. At the very highest theoretical level, and in a perfect world, this is just a matter of having the proper turns ratio. However, in practice, there are a whole list of other variables, which also matter, and will affect the sound. So, in your case, those transformers are either matched to 3000 Ohms primary, or they're matched to 4500 Ohms primary, or somewhere in-between, and that will affect how they sound with different tubes.
Being "off match" in certain ways will affect the sound in different ways... and how exactly that works will... you guessed it... depend on other things. By choosing different transformer taps you are "fiddling with the match" (only one is actually correct for a specific load impedance... and there is no 0.7 Ohm tap),
However, the single biggest factors in terms of how hot the tubes will run are the voltage they're running at, and the idle current they're set to draw.... The supply voltage is determined by the power supply.... But the idle current is controlled by the bias voltage... which is a DC voltage that is applied to the grids along with the audio signal. The relationship between bias voltage and idle current depends on the tube... both the tube number and the individual tube... which is why bias gets adjusted. In all normal circuits higher bias voltage (that's a higher negative voltage) means lower idle current and vice versa (the bias voltage acts to "squeeze off the current").
If your amplifier uses some sort of circuit that automatically determines the tube bias... then the results may be very different with a different type of tube. And, of course, different amounts of idle current are recommended for different tubes... (So, if you are able to adjust the bias, you may see different recommendations for different tubes.)
And which one sounds the best, or a certain way, depends on the design of the amplifier itself, and to a degree on personal preference. In general, with all standard amplifier designs, increasing the idle current... - makes the amp run closer to Class A - makes the tubes burn more power and run hotter - shortens the life of the tubes - makes the transformers also run hotter - puts more wear and tear on the power supply And, in general, reducing the idle current... - makes the amp run "further into Class A/B" - makes the tubes run cooler and last longer - makes the transformers run cooler - reduces stress on the power supply With most amplifiers there is a recommended "happy medium" where the distortion is at its lowest and the amp doesn't run especially hot. Going too far above this may make the amp sound a tiny bit better - but puts a lot of wear on all of the components. Going too far below this can increase distortion, and reduce maximum power output, but will usually make the tubes run cooler and last longer.
Of course.... - different tube numbers also consume different amounts of power to heat the cathodes - different tube numbers are rated to be safe to operate at different voltages - different tube numbers are designed to run at different temperatures (it's safe to run a 6550 at a temperature where you can light a match off the glass; don't do that with a 12AX7) - there is even some variation between manufacturers Also note that maximum limits are really something to be avoided. So, for example, many people replace a tube that's "lower down in the family tree" with a new KT-120. Even though you may need to adjust the bias, this is usually fine, because the KT-120 can tolerate a higher plate voltage, and handle more power. However, going the other way could be quite dangerous, and could even damage or destroy the amplifier. Whether that's the case or not depends on the specific details of the amplifier itself. If your amplifier runs at a plate voltage of 550V then it's perfectly safe to use a tube rated for 600V or one rated for 800V. However, if your amplifier runs at a plate voltage of 700V, then one will be just fine, but the other will NOT, and the results could be very bad. Different tubes also require different heater currents... and the same situation applies. Replacing the recommended tube with one that draws less current is usually OK, but whether it's OK to use one that requires more current depends on what your amp can safely supply.
NOTE THAT, BECAUSE DIFFERENT TUBE TYPES CAN DRAW DIFFERENT IDLE CURRENT WITH THE SAME BIAS SETTING, IN SOME CASES YOU MUST ADJUST THE BIAS WHEN CHANGING TUBES. (Failing to do so CAN lead to premature failure of the tubes and CAN damage or even destroy the amplifier.)
If all of this sounds complicated and confusing then your best option is to consult someone familiar with YOUR PARTICULAR MODEL OF AMPLIFIER. Since people just love to "roll tubes" there's usually plenty of specific information available. If it's a current model then there will be plenty of information available from the manufacturer and from other owners.
Ok, so I've been trying to figure out why things are so much better in these M-125 amps with the KT120 tubes vs KT88 mated to my electrostatic speakers which can dip to 0.7 Ohms impedance. I haven't checked the internal voltages yet to get a glimpse of what might be different but want to do that soon. What I do know is that with KT88 tubes the power transformer would get really hot, over 150 degrees was the highest temp I saw before getting some really quiet fans to keep the temp below 120 degrees. Now with KT120 tubes and no fans the temp hasn't gone over mid 130 range temps. The output transformer has never gotten hot. The sound is incredible! Slightly brighter, in all good ways. More dynamic, and with stat panels that's quite a feat. I'd like to understand why this tube is a better match, both for the fact that the PT isn't as hot as before and the improved sound. Is part of the equation the fact that the KT120 has lower Load Resistance p-p ( 3000 Ohms if the info I've found is accurate vs 4500 Ohms for KT88) making it a better fit for my low impedance speakers? The output transformer Primary Impedance is 2200 Ohms and my speakers are on the 4 Ohm tap.
|
|
|
Post by ttocs on Feb 14, 2020 13:45:06 GMT -5
If all of this sounds complicated and confusing then your best option is to consult someone familiar with YOUR PARTICULAR MODEL OF AMPLIFIER. Thank you very much for the incredibly detailed info! It mostly confirms what I've been thinking about and that maybe I'm on the right track about the "why" things are better with the KT120. I run these amps through a Variac and strive to keep the voltage between 115-117. When the amps get 120V the internal voltages go way high!! Some are 30 volts too high according to the schematic, but at 117 they are within the correct range. Very soon I want to recheck the voltages to find out what the min/max can be for the supply voltage and for bias as well, and to check the output power differences between the 88 and 120 tubes. Maybe when I ship my XMC-1 back for the Trade-In.
|
|
KeithL
Administrator
Posts: 9,934
|
Post by KeithL on Feb 14, 2020 14:34:40 GMT -5
Glad to be of assistance...
Even though I'm not a big fan of listening to tube equipment... I used to repair it and even occasionally design it... further back than I care to think about... In general, tube designs are rather simpler to understand than discrete solid state designs, which makes them sort of fun to look at.
Most equipment should be designed to operate over a relatively wide range of voltages... because the line voltage itself is subject to significant variation. Power in the US is generally specified at 120 VAC +/-5% (114V - 126V) so equipment designed for use in the USA should be safe to operate over that range. And, to be honest, in most cases there is a lot of leeway around some of the voltages shown on equipment schematics.
The catch is that, in some situations, it's supposed to be obvious whether that's the case or not, and critical voltages are supposed to be indicated, but it doesn't always work out that way.
(And, with most modern designs, the critical voltages are regulated, so they stay spot-on regardless of supply voltage variations...)
However, you can't always count on that, and there are exceptions, so it never hurts to check...
If all of this sounds complicated and confusing then your best option is to consult someone familiar with YOUR PARTICULAR MODEL OF AMPLIFIER. Thank you very much for the incredibly detailed info! It mostly confirms what I've been thinking about and that maybe I'm on the right track about the "why" things are better with the KT120. I run these amps through a Variac and strive to keep the voltage between 115-117. When the amps get 120V the internal voltages go way high!! Some are 30 volts too high according to the schematic, but at 117 they are within the correct range. Very soon I want to recheck the voltages to find out what the min/max can be for the supply voltage and for bias as well, and to check the output power differences between the 88 and 120 tubes. Maybe when I ship my XMC-1 back for the Trade-In.
|
|
KeithL
Administrator
Posts: 9,934
|
Post by KeithL on Feb 14, 2020 15:02:45 GMT -5
Incidentally....
Anyone interested in some of the finer points of tube amplifier design should absolutely check out this site: tubecad.com/(they also have kits and PCBs...)
I would, however, add one note here....
These guys like to discuss interesting, unique, and sophisticated designs for tube-based circuits. The reality is that the vast majority of vintage tube amplifiers, and most modern ones as well, are based on a few very basic designs. Probably 2/3 of the decent tube amps ever made are basic push-pull cathode bias designs... most of those are of the Williamson Ultralinear type....
You can see a huge number of example schematics if you Google: "push pull tube amplifier schematic" ...
The main reference book for working with tubes for quite literally decades was "The RCA Receiving Tube Guide". It was published every year and included data sheets for all common tubes, examples of schematics, and lots of useful extra information. (Of course it's way out of date now... but it's interesting to see how much HASN'T changed.)
You can find lots of copies, in various forms, available for free download...
Check out the one on the right side of this page:
(you'll find it looks quite familiar in the essentials)
|
|