hello!
Tubes are quite new to me... but i always wanted to build a tube amp. then a month ago i got a great deal for some parts, some big OPT a PT, 8x6550we, 2xecc82 and some pcb's.
so.. now i have a working tube amp.. but i kinda have the feeling that one ecc82 as driver per side isnt quite up for the job.
i dont have the exact schematic on a computer yet, but it looks alot like the first one on this page, but with different component values ofc.
velleman K8010 kt88 monoblock tube amp,el34,ecc82,ecc88,kt88,300b,845 direct heated triode,mkp, ...
so my question is:
is the ecc82 suitable to drive 4 6550? and if not, then i need some ideers to what to do instead? and i would love some schematics with the ideers.
if needed i will digitalise my schematic and post it tonight.
Tubes are quite new to me... but i always wanted to build a tube amp. then a month ago i got a great deal for some parts, some big OPT a PT, 8x6550we, 2xecc82 and some pcb's.
so.. now i have a working tube amp.. but i kinda have the feeling that one ecc82 as driver per side isnt quite up for the job.
i dont have the exact schematic on a computer yet, but it looks alot like the first one on this page, but with different component values ofc.
velleman K8010 kt88 monoblock tube amp,el34,ecc82,ecc88,kt88,300b,845 direct heated triode,mkp, ...
so my question is:
is the ecc82 suitable to drive 4 6550? and if not, then i need some ideers to what to do instead? and i would love some schematics with the ideers.
if needed i will digitalise my schematic and post it tonight.
How is the output stage configured (i.e., triode, UL, pentode)? Output stage operating conditions (i.e., B+, p-p load, idle current)? That will determinie requirements for the driver.
it is UL,
b+ = 380V
p-p load, is this the impedance of the opt? if so it is 1.75Kohm
bias around 87mA at the moment
i use a resistor between the cathode and ground for my bias, so does that mean that i should subtract the voltage across that resistor to get my effective b+?
b+ = 380V
p-p load, is this the impedance of the opt? if so it is 1.75Kohm
bias around 87mA at the moment
i use a resistor between the cathode and ground for my bias, so does that mean that i should subtract the voltage across that resistor to get my effective b+?
Last edited:
Yes, you do, but this is good enough for estimation. Your bias at those conditions is likely to be about -30V, which means that the drive signal needs to be 60V peak to peak max. Input capacitance works out to something less than 100pF. This is really not a difficult load at all- a good, moderately high transconductance triode configured as a split load inverter is more than adequate.
Your other alternative is a long tail pair input coupled to cathode followers, but that's more complex than needed. If you can provide higher B+, the same level of complication will allow you to configure the output tubes as triodes in AB2 and get 150 watts easily.
Your other alternative is a long tail pair input coupled to cathode followers, but that's more complex than needed. If you can provide higher B+, the same level of complication will allow you to configure the output tubes as triodes in AB2 and get 150 watts easily.
so are you saying that my ecc82 is just fine for the job? or that i should look for a different tube? to use in my split-load topology? (with the rest of component modified ofcourse)
and if i wanted a more powerfull amp coupled in triode mode, i could make a "long tail pair input coupled to cathode followers" driver stage, and up my B+ closer to what the tubes allow?
just want to be shure i understand you 🙂
and thanks for the help.
and if i wanted a more powerfull amp coupled in triode mode, i could make a "long tail pair input coupled to cathode followers" driver stage, and up my B+ closer to what the tubes allow?
just want to be shure i understand you 🙂
and thanks for the help.
In terms of drive, the ECC82 as a split load inverter will work adequately. It is not, however, the most linear choice of tube! At the same operating point and with minimal socket changes, you could substitute a 6CG7/6FQ7. An ECC88 will have even more drive capability, but the operating points and component values will have to be moved around quite a bit.
The take-home point here is that your output stage is not a particularly onerous load, except psychologically- it just LOOKS like all those tubes would be tough to drive.😀
The take-home point here is that your output stage is not a particularly onerous load, except psychologically- it just LOOKS like all those tubes would be tough to drive.😀
Becareful interpreting the datasheet for 6550 series; there many about. Your'e fine with values for auto/cathode bias. An ECC82 cannot drive without distortion driving parallel p-p fixed bias grid leaks when datasheet values give each tube around 50K ohms. There is some leaway on this but this is nasty load for the driver stage.
richy
richy
Becareful interpreting the datasheet for 6550 series; there many about. Your'e fine with values for auto/cathode bias. An ECC82 cannot drive without distortion driving parallel p-p fixed bias grid leaks when datasheet values give each tube around 50K ohms. There is some leaway on this but this is nasty load for the driver stage.
richy
At 50k, the load ends up at ~17k, which is not too terrible for a split load inverter, especially an ECC88.
...
i dont have the exact schematic on a computer yet, but it looks alot like the first one on this page, but with different component values ofc.
velleman K8010 kt88 monoblock tube amp,el34,ecc82,ecc88,kt88,300b,845 direct heated triode,mkp, .....
This first one appears to be a simple and forgiving design. I bet it sounds good too. You can push the tubes more but then you add complexity.
The other schematics below show how to add a long tailed pair phase splitter but that adds another tube and many more parts. then you might think about making the bias of each tube independently adjustable and switches to allow use on ultralinier or triode modes but the amp becomes complex. Keeping the parts count minimal and running the tube conservatively is best for the sound quality.
Simple things you might do are, check if the phase splitter is balanced, are AC volts the same on both outputs, adjust resisters until it is. I'd do this style of fine tuning rather than major changes.
hmm would like some confirmation to what i am understanding here (again...)
if i change to fixed bias, then i would need a more powerfull driver stage, like a ecc88, or long tailed pair phase splitter. (i will proberly change to a ecc88 anyway.. since it would force me to learn some more about the circuit around tubes.)
#8
ye i know about the datasheet thing... and the most anying thing is that the only one i found regarding the sovtek 6550we was in russian... not that helpfull to me 😛 so i try to keep the bias current, plate voltage and so.. under the lowest i found in the datasheets.
#10
actually it does sound quite nice already.. a much more pleasent sound that my all digital diy amp (based on some TI evm board)
the things i want to do, is converting it to individual fixed bias, and make that switch between ul and triode. that have been in my mind since i got the parts. but it will not happen until i get i proper case for the amp. (though I find it hard, to find a proper case or cases...)
right know it is build on my table... and not really a good solution.. regarding safety and so on (dont worry i do know what i am doing.. i am a electronic technician.)
if i change to fixed bias, then i would need a more powerfull driver stage, like a ecc88, or long tailed pair phase splitter. (i will proberly change to a ecc88 anyway.. since it would force me to learn some more about the circuit around tubes.)
#8
ye i know about the datasheet thing... and the most anying thing is that the only one i found regarding the sovtek 6550we was in russian... not that helpfull to me 😛 so i try to keep the bias current, plate voltage and so.. under the lowest i found in the datasheets.
#10
actually it does sound quite nice already.. a much more pleasent sound that my all digital diy amp (based on some TI evm board)
the things i want to do, is converting it to individual fixed bias, and make that switch between ul and triode. that have been in my mind since i got the parts. but it will not happen until i get i proper case for the amp. (though I find it hard, to find a proper case or cases...)
right know it is build on my table... and not really a good solution.. regarding safety and so on (dont worry i do know what i am doing.. i am a electronic technician.)
Triode is a whole different kettle of fish. To drive that high capacitance, you'll almost certainly need a cathode (or source) follower.
ohh.. but.. why does the capacity change? i change the connections to my opt. but that doesnt change anything at my control grids? and isnt the circuits around them that defines how hard a load it is for my driver stage? or aperently it isnt... but that was what i thought 🙂
if this is one of the "go read up on some basic tube knowlegde and you know, questions" please do tell me.. dont want to pester anyone with all my questions.
if this is one of the "go read up on some basic tube knowlegde and you know, questions" please do tell me.. dont want to pester anyone with all my questions.
ohh.. but.. why does the capacity change? i change the connections to my opt. but that doesnt change anything at my control grids? and isnt the circuits around them that defines how hard a load it is for my driver stage? or aperently it isnt... but that was what i thought 🙂
if this is one of the "go read up on some basic tube knowlegde and you know, questions" please do tell me.. dont want to pester anyone with all my questions.
The capacitance changes because of Miller effect. Detailed discussions of all these matters can be found in "Valve Amplifiers," 3rd edition, by Morgan Jones. It's a superb book that will answer questions you didn't even think to ask. Strongest possible recommendation.
You can see this dirver for PP 6550, or KT88, El34
PP2010 - Hi-End Push Pull Amplifier
From that article:
"The split load without a regulation give two precise outputs with the same level but the output impedance are completely different so it should not be used to drive directly difficult load like the final stage."
This is thoroughly and completely wrong. The final design is pretty good (though the capacitive coupling of the cathode followers to the output stage is questionable), but there's a lot of misleading information in this article.
ahh, ofcourse.. was one of those questions 🙂
but just ordered that book. and some 6N1P, since they were dirt cheap, and from what i have read, they should do it pretty good where i need them.
and thanks quanghao, not shure i am going to do something like that anytime soon though. first i am going to upgrade my input stage, and then make fixed bias.
but just ordered that book. and some 6N1P, since they were dirt cheap, and from what i have read, they should do it pretty good where i need them.
and thanks quanghao, not shure i am going to do something like that anytime soon though. first i am going to upgrade my input stage, and then make fixed bias.
In triode connection, the screen grid is no longer shielding the control grid from the plate. The result is that the grid-to-anode capacitance (Cga) is multiplied by the effective amplification factor -- known as the Miller effect.ohh.. but.. why does the capacity change?
Your drivers will have to overcome not only the grid-leak resistors of the OP tubes but also their input capacitances, both in parallel. As SY says, this can be achieved most effectively with FET source follower drivers directly coupled to the OP tubes or, if you are unhappy with using SS, cathode followers (a valid use for those ECC82s).
In either case, you would have the added benefit of avoiding unpleasant blocking disrortion on transients and it would work perfectly well for either triode mode or UL. As SY also mentions, it would enable you to use Class AB2 if necessary. You would need fixed bias but it would be a great solution!
Last edited:
About parallel push pull o/p stages with fixed bias.
If one is going to the expense using extra power tubes, and using an LTP for a driver, one would expect each half driver to suck at least 10mA to deal with Miller effect and be capable of sinking a fair grid leak resistor value (6550 series) with low distortion. The prize and the snag, roughly 50+50V rms at low thd drive into an effective 40K load plus Miller effect dump for 100W o/p is quite a challenge. Remember the output stage should be the limiting performance factor not the driver nor the front end and so many designs fail in this aspect.
Parallel push pull has many performance advantages, the effective lower Z output tranny perfomance is vasty improved, then USE it and not slew limit the output stage by using a lousy driver. A high current Williamson diff driver is the ultimate perfomance goal, and it pays to get this part right. A good quality E&I o/p tranny for p-p-p should see a min -3dB around 65Khz. Use it to the fullest as one has paid for it !
I'd admit that optimising the design can be a challenge; the b/w trade off vs. stability and squarewave shape are all factors to take into account. The power supply is another area for study. Make sure it can supply the current.
Keep at it !
Richy
If one is going to the expense using extra power tubes, and using an LTP for a driver, one would expect each half driver to suck at least 10mA to deal with Miller effect and be capable of sinking a fair grid leak resistor value (6550 series) with low distortion. The prize and the snag, roughly 50+50V rms at low thd drive into an effective 40K load plus Miller effect dump for 100W o/p is quite a challenge. Remember the output stage should be the limiting performance factor not the driver nor the front end and so many designs fail in this aspect.
Parallel push pull has many performance advantages, the effective lower Z output tranny perfomance is vasty improved, then USE it and not slew limit the output stage by using a lousy driver. A high current Williamson diff driver is the ultimate perfomance goal, and it pays to get this part right. A good quality E&I o/p tranny for p-p-p should see a min -3dB around 65Khz. Use it to the fullest as one has paid for it !
I'd admit that optimising the design can be a challenge; the b/w trade off vs. stability and squarewave shape are all factors to take into account. The power supply is another area for study. Make sure it can supply the current.
Keep at it !
Richy
ok, so basicly the first thing i want to do is to improve my input stage. before i try and do anything else..
so if i understand it correctly... (lets pretend my schematic is the same as the first one in the link in #0) then i would be able to add 2 cathode followers between the splitter and the power stage, and if i did that, it would be ready for fixed bias for my power tubes, or triode connection (or both...)?
right now i am thinking of using a 6N1P instead of my ecc82 as first stage+phase splitter, and then add the ecc82(or would something else do a better job here?) as cathode follow, to use as a driver.
i still need to upgrade my powersupply aswell though... i have pushed it abit to far already... have to use a smps for my heaters... else the PT just gets to hot.... so i am looking for 2x PT so i can make some proper monobloks.
hmm this small project is growing more and faster then i had expected... hope my wallet can sustain the need for new stuff for it 🙂
on a side note... i do feel abit ashamed to use this thing... it uses almost 400W all the time.... and right next to it, i have my digital amp 2x300W... that have a idle consumption around 10-20W...
so if i understand it correctly... (lets pretend my schematic is the same as the first one in the link in #0) then i would be able to add 2 cathode followers between the splitter and the power stage, and if i did that, it would be ready for fixed bias for my power tubes, or triode connection (or both...)?
right now i am thinking of using a 6N1P instead of my ecc82 as first stage+phase splitter, and then add the ecc82(or would something else do a better job here?) as cathode follow, to use as a driver.
i still need to upgrade my powersupply aswell though... i have pushed it abit to far already... have to use a smps for my heaters... else the PT just gets to hot.... so i am looking for 2x PT so i can make some proper monobloks.
hmm this small project is growing more and faster then i had expected... hope my wallet can sustain the need for new stuff for it 🙂
on a side note... i do feel abit ashamed to use this thing... it uses almost 400W all the time.... and right next to it, i have my digital amp 2x300W... that have a idle consumption around 10-20W...
- Status
- Not open for further replies.
- Home
- Amplifiers
- Tubes / Valves
- driver stage for ppp 6550's