You need to compare it to a live source to be sure, real instruments rarely sound flat..
.. but we do not know if the flattening happened in the amp or elsewhere in the chain - so we still can't say for sure.
The answer may just be to buy a more suited tube
The sound of real instuments is totally dependant on the acoustic envoironment. For example,put a piano whose soundboard has flattened out -simply because of Anno Domini- in a church and it is surprising how much bloom is restored to the perceived sound whereas it may sound thin and tired in your living room. Valves are not musical instruments. It would be surprising if EH made poor quality 6SN7's with much higher distortion, higher than usual Miller capacitance or poor microphonic characteristics. Surely sonic characteristics will only be greatly different with different brands of good quality valves if the circuit design is really critically dependant on exact biasing,Miller capacitance, ra, etc. with little tolerance for small variations in these?
There are ''sweet spots'' i.e. where the anode loop cancels with the grid loop non linearity in a certain way. Each make has its own little curve quirks. So different voicing can be tuned out of the same amplifying element on bias level. There is a Pass article expanding on that for all the main active component families. So its the engineering way and the costly roll way. Guess what is going to happen for 99.9% of users not with engineering experience and no FFT. Rock & Roll.
The "roll way" may be a good way to get a pleasant sound, and sometimes it may even give similar results to the "engineering way", but the problem comes when people believe that it is in fact a superior engineering way. This happens when people can't distinguish between preferences and improvements, and won't even admit that there might be a distinction to be drawn.
When somebody says Valve A is better than Valve B, what he may really mean is that in one particular circuit at one particular bias point with particular items upstream and downstream that valve gives a sound which he prefers. This may be treating an amplifier as an instrument rather than a reproducer. Someone else will have different bias points, and different preferences, so will prefer different valves. If you swap valves and can't hear any difference then it may simply be because the amplifer is well-designed (although this is not necessarily the case).
When somebody says Valve A is better than Valve B, what he may really mean is that in one particular circuit at one particular bias point with particular items upstream and downstream that valve gives a sound which he prefers. This may be treating an amplifier as an instrument rather than a reproducer. Someone else will have different bias points, and different preferences, so will prefer different valves. If you swap valves and can't hear any difference then it may simply be because the amplifer is well-designed (although this is not necessarily the case).
Its only after tuning out on bench level and ear each scenario we can know other tone traits between brands and vintage. Certainly they are smaller than they appear in gross roll situations. But they do exist. Cathode chemistry, noise, overload, and absolute THD aren't the same between brands and samples. But its generally a small part of the main picture. Takes a system with thorough resolution and solid criterion of tone from the user's side. I.e. it always takes knowledge, tools, work, and being serious so to state something serious. That's a fact in life for all things. Else it all falls to the misunderstanding and claim pit. No accident no tube roll review satisfactorily agrees. But its a hobby right? Just take everything stated off hand with a pinch of salt.
I had this friend who used to be regarded a young prodigy pianist but he would not look for a career in that later on. So we were out once and we passed a girl playing the violin in open air classical tunes basking. He told me, can you listen to the thin tone, that's a chamber violin, can't cut it in the open. He went on explaining to me that each acoustic instrument of the classic orchestra has been evolved in tandem with the usual space it was performing in. That thing he said lasted for centuries and it was a trial and error between requests of maestro's, musicians, and instrument makers. You probably think a Stradivari is so famous and mythically expensive because it must be the best hands down, he told me.
Well, no, he continued. Its the best loud violin for soloists in halls. If a Stradivari will play amongst Amatis in a chamber orchestra it will shadow 3 of them for volume, but it will sound positively gross also. The Amati is the delicate one. Each one to its own function he concluded. Don't know if this story ties here, but stories are nice.
In other words, tube roll is not what it's typically praised to be by online reviews?
How much can the variation in bias for a tube span from one vendor to another, and one time period to another?
What is the likelyhood that a lot of "Tube Rolling" results in tubes biased far enough off the sweet spot that it is noticable? If the tubes were individually biased for their own best point there wouldd there be near the difference if any at all?
Unfortunatly this would take a lot of effort to determine wheyther true or not.
Just a comment, when people say the tube sounds "warm" it it generally means the sound has add 2nd and maybe 3rd order harmonic distortion. When they say "flat" and detailed they mostly main very low THD. Many people prefer "warm"
About the quote above....
What I tell people who want to do home recording is that what matters the most to the the end result is what is closest to the source of the sound. The #1 thing is the player and his instrument, then comes the space, this is the room acoustics and then the mic and where it is placed and then we get into mic preamps, A/D converters and software. For recording I'm sure that you are best to invest your money and your time in that which is closest to the source. If I'm recording my self I think adding practice time helps the most
For playback I think the opposite is true. The direction runs in reverse. That which is closest to the listener's ear matters the most. The room being the most important followed by the speakers and then the amp and finally the quality of the source. I think most agree, that speakers matter the most.
So if you look at the total system from musician to listener the amp is in the mid section of the chain when it is the ends of the chain that make the most difference.
Thanks for explaining
I messured 2,6V at that point so it's correct. And there is no cap to next stage
I think i leave it as it is
Regards //Daniel
The 125V that is supposed to be on the grids of the succeeding ECC99 driver stage is kind of important. It looks to me if the information you have provided is correct that a plate load resistance of 43K might be a better compromise for this particular circuit.
I have found that there are significant differences in the way that 6SN7 sound not necessarily based on less than ideal operating points for one brand vs another.. Some just do sound better IMLE than others regardless of operating point. My personal favorite is the Sylvania, more or less any version made prior to the late 1980s - with earlier ones being slightly preferred.
I have NOS Sylvania, US made cylindrical plate Tungsols (Not the New Sensor stuff), GE, RCA, Westinghouse and a whole lot others. I genuinely prefer the Sylvania to any of the others. I also like the Brimar version quite a lot, but they are rather costly these days, so consequently I currently have none.
I run 6SN7 at typically 8 - 10mA where they seem to provide the best linearity, and this has proven to be the case with any I have tried. Much above 10mA - 12mA you start to reach the point of diminishing returns and plate dissipation limits also have to be observed.
Last edited:
In a power amp with LTP driving push-pull triode-wired EL34's, I tried the old New Sensor "Sovtek" 6SN7 and found that they did sound a little different than my pile of 6SN7GT, 6SN7GTA and 6SN7GTB tubes. In that amp, the 6SN7 anode load resistors are 36k ohms, with 7.5mA current through each triode side. If I remember correctly, bias is about -5.5V, so that means anode voltage is probably about 185V.
In this particular amp, the standard Sovtek 6SN7's sound a little thin, less "powerful" than the USA 6SN7's. But the difference is very small. The USA-made 6SN7's all sound pretty close to each other. I found the Sylvania 'GTBs to sound very good, and GE 'GTBs also sound good (the GE might be my fave American version, but that's a personal taste thing, I think). My absolute favorite is a pair of rare, Soviet-era, metal-based 6SN7 equivalent (6N8S? I'm not home so can't double-check) that is by far the best of the bunch. Those are really a cut above, but they're expensive these days.
I think that with only 5.2mA going through the 6SN7 in your amp, you could be accentuating the differences between makes/types of 6SN7. Some are probably better suited to low-current, low-voltage operation. Some probably perform worse used under those conditions. If the circuit could be optimized so that there was more current going through the first stage 6SN7, then I have a hunch you'd notice less difference between the various makes/types of 6SN7, and that stage would begin to sound more full and less 'thin.'
Careful attention must be paid to the relationship between the anode voltage at the 6SN7 and the voltage at the joined ECC99 cathodes. The ECC99 cathodes will need to be a few (like 8 to 10?) volts higher than the 6SN7 anode. That creates the bias for the ECC99 triodes, since the two stages are DC-coupled.
--
In this particular amp, the standard Sovtek 6SN7's sound a little thin, less "powerful" than the USA 6SN7's. But the difference is very small. The USA-made 6SN7's all sound pretty close to each other. I found the Sylvania 'GTBs to sound very good, and GE 'GTBs also sound good (the GE might be my fave American version, but that's a personal taste thing, I think). My absolute favorite is a pair of rare, Soviet-era, metal-based 6SN7 equivalent (6N8S? I'm not home so can't double-check) that is by far the best of the bunch. Those are really a cut above, but they're expensive these days.
I think that with only 5.2mA going through the 6SN7 in your amp, you could be accentuating the differences between makes/types of 6SN7. Some are probably better suited to low-current, low-voltage operation. Some probably perform worse used under those conditions. If the circuit could be optimized so that there was more current going through the first stage 6SN7, then I have a hunch you'd notice less difference between the various makes/types of 6SN7, and that stage would begin to sound more full and less 'thin.'
Careful attention must be paid to the relationship between the anode voltage at the 6SN7 and the voltage at the joined ECC99 cathodes. The ECC99 cathodes will need to be a few (like 8 to 10?) volts higher than the 6SN7 anode. That creates the bias for the ECC99 triodes, since the two stages are DC-coupled.
--
Hi!
I've tried to do some maths Have no idea if this is rights
But if it's right around the 6SN7 tube, what about the ECC99? Should something be changed? Like R12?
Thanks //Daniel
I've tried to do some maths
Have no idea if this is rights But if it's right around the 6SN7 tube, what about the ECC99? Should something be changed? Like R12?
An externally hosted image should be here but it was not working when we last tested it.
Thanks //Daniel
13mA for the pair of ECC99 gives 6.5mA per tube.
top anode is at 440-(.0065*30000)=245V
A-K = 245-130 = 115V
Bottom anode is at 440-(0.0065*37170)=198.4
A-K = 198.4-130 = 68.39V
I believe the lower ECC99 will start saturating before you drive the KT88 fully and you will get high distortion.
I would lower the current through the pair by increasing R12, or change the anode resistor values to give less drop, or both. Depends on how the ECC99s are best operated.
I would have to look up the ECC99 and KT88 specs to give more detail.
I looked them up.
The KT88s are going to take around 110V p-p on the grids. so you need 55V peak before saturating or cutoff on the driver.
ECC99 I would not bias any colder.
I'd change the anode resistors. Maybe 20K total on the top and andjust the bottom accordingly, say 24.66K if you scale by your current selection.
You have 130V available in the cathode circuit. How about a current source instead of R12?
top anode is at 440-(.0065*30000)=245V
A-K = 245-130 = 115V
Bottom anode is at 440-(0.0065*37170)=198.4
A-K = 198.4-130 = 68.39V
I believe the lower ECC99 will start saturating before you drive the KT88 fully and you will get high distortion.
I would lower the current through the pair by increasing R12, or change the anode resistor values to give less drop, or both. Depends on how the ECC99s are best operated.
I would have to look up the ECC99 and KT88 specs to give more detail.
I looked them up.
The KT88s are going to take around 110V p-p on the grids. so you need 55V peak before saturating or cutoff on the driver.
ECC99 I would not bias any colder.
I'd change the anode resistors. Maybe 20K total on the top and andjust the bottom accordingly, say 24.66K if you scale by your current selection.
You have 130V available in the cathode circuit. How about a current source instead of R12?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.