Hi,
Now you would think why no one else is doing this, right?
If only it were this easy...
Parallelling two sections of a twin triode will only work satisfactorily if:
Both sections are absolutely identical on all possible parameters.
Both sections will age in the exact same way and by doing so will remain identical to one another for life...
Anyone wanting to count the odds?
Again, this is something you can try at home if:
You have plenty of the same type of tube or are willing to pay someone to select the tubes for you.
In general, using a lower noise tube type, even if it comes with a lower gain, will be the better solution IMO.
Cheers,
Now you would think why no one else is doing this, right?
If only it were this easy...
Parallelling two sections of a twin triode will only work satisfactorily if:
Both sections are absolutely identical on all possible parameters.
Both sections will age in the exact same way and by doing so will remain identical to one another for life...
Anyone wanting to count the odds?
Again, this is something you can try at home if:
You have plenty of the same type of tube or are willing to pay someone to select the tubes for you.
In general, using a lower noise tube type, even if it comes with a lower gain, will be the better solution IMO.
Cheers,
cool... all is understood. i think... just a couple of things.
just for the benifit of my learning, why must both sides be exactly the same for this to work??? it would at the very least be lowering the output impedance from these tubes, which is a good thing, and will also offer at least some of that 3dB extra over the noise floor... does the imballance effect the quality of the sound???
the way i have done this before is by using all shared components (ie, shared plate load resistor, cathode resistor, bypass cap, etc). Although i did this in a guitar amp, which is of course slightly different form hifi amps... some of the things i do in my guitar amps would make hifi buffs choke... like using unmatched tubes in the output stage! weird stuff if hifi's your thing... i know even a lot of guitar amp builders cough at this one... but i do have my reasons, and i justified everything and it was what i wanted... and i did it, and it sounded how i wanted... so that's all that matters to me! but like i said, hifi and guitar are different, so there must be something i don't see in this... (that is the paralleled preamp stage, not unmatched output tubes)
cheers.
just for the benifit of my learning, why must both sides be exactly the same for this to work??? it would at the very least be lowering the output impedance from these tubes, which is a good thing, and will also offer at least some of that 3dB extra over the noise floor... does the imballance effect the quality of the sound???
the way i have done this before is by using all shared components (ie, shared plate load resistor, cathode resistor, bypass cap, etc). Although i did this in a guitar amp, which is of course slightly different form hifi amps... some of the things i do in my guitar amps would make hifi buffs choke... like using unmatched tubes in the output stage! weird stuff if hifi's your thing... i know even a lot of guitar amp builders cough at this one... but i do have my reasons, and i justified everything and it was what i wanted... and i did it, and it sounded how i wanted... so that's all that matters to me! but like i said, hifi and guitar are different, so there must be something i don't see in this... (that is the paralleled preamp stage, not unmatched output tubes)
cheers.
Hi,
It all depends on how well you want things to work but if both sections aren't identical one is going to take a bigger share of the load than the other creating what is known as "current hogging".
One thing than leads to another and as the load isn't balanced so too the signal will be treated in an unexpected way leading to more distortion than was expected.
I you want to experiment with this, run a 6AS7G with both sections in // in Class A close to maximum dissipation. If both sections aren't matched you'll soon see one section glowing cherry red eventually leading to self-destruction.
Yes, it does. Blurred, unfocused stereo imaging is one...
Slightly? Lot's of slightly if you'd ask me...
Cheers,
It all depends on how well you want things to work but if both sections aren't identical one is going to take a bigger share of the load than the other creating what is known as "current hogging".
One thing than leads to another and as the load isn't balanced so too the signal will be treated in an unexpected way leading to more distortion than was expected.
I you want to experiment with this, run a 6AS7G with both sections in // in Class A close to maximum dissipation. If both sections aren't matched you'll soon see one section glowing cherry red eventually leading to self-destruction.
Yes, it does. Blurred, unfocused stereo imaging is one...
Slightly? Lot's of slightly if you'd ask me...
Cheers,
if grid resistance was matched, wouldn't this lead to a ballance of the signal splitting???? or can the signal 'see' more than just this, and will also split acording to other parameter, like Gm?
i thought both the sections would, once the input has been split, work independantly, untill the waveforms from both tubes are summed together... and this is external to any of the tubes... so where is the distortion coming from??? surely if both tubes are amplifying the same signal, only at different levels, the waves should still be the same shape.... and this is all that's important. If they are the same shape, only different amplitudes, they will still sum together to make a waveform of the same shape... and as long as we make sure no imballance is causing one section to reach clipping, would this be fine???
yes, true... very very different...
cheers
Hello Benny,
Thanks for the fun discussion.
Just one suggestion:
For safety, I suggest that use one cable, with a female connection on the end that connects to your amp, to
route your HIGH VOLTAGES to your amp. Use one of
the cable leads to control a relay, so that if the cable
is disconnected its power is shut off.
Otherwise you risk having a disconnected cable with
high voltage, electrocuting pets, babies, and maybe
even you!
Just a suggestion--not responsible, caveat emptor, YMMV, etc.
Good luck and have fun!
Best,
George Ferguson
Thanks for the fun discussion.
Just one suggestion:
For safety, I suggest that use one cable, with a female connection on the end that connects to your amp, to
route your HIGH VOLTAGES to your amp. Use one of
the cable leads to control a relay, so that if the cable
is disconnected its power is shut off.
Otherwise you risk having a disconnected cable with
high voltage, electrocuting pets, babies, and maybe
even you!
Just a suggestion--not responsible, caveat emptor, YMMV, etc.
Good luck and have fun!
Best,
George Ferguson
Vortex's nimble fingers tapped:
Hey, this site is neat! It nicely explains some
relatively exoteric electronic concepts, with
illustrations and stuff. Check out the "Analog Audio" section:
http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/audio/Analog.html
and the main page has some other
interesting stuff.
Best,
George Ferguson
Hey, this site is neat! It nicely explains some
relatively exoteric electronic concepts, with
illustrations and stuff. Check out the "Analog Audio" section:
http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/audio/Analog.html
and the main page has some other
interesting stuff.
Best,
George Ferguson
Hey guys!
That's a nice topic, nice discussion... WOW!
Thank you all... keep it running please..
I'm going to focus on ordering quality parts (caps and resistors) in Europe somehow from somewhere.
What do you think, which brand of non-ultra-expensive capacitor and/or resistor should I take for crisp clear open sound ?
I've chosen until now Shinkoh Tantalum resistors and ELNA MKP caps for non-electrolytics / Rubycon Black Gate WKZ for that really few electrolytic caps.
They aren't so much expensive, not cheap, okay, but hey, I'm building for 20 years The final amp will be my father's "commemorative edition" for me
You could help me with some brands/types of caps and resistors for clear sound, I don't favor the lush sweet sound of tubes (meant here in good terms now) too much - just a bit.
And don't forget to keep the discussion abot nfb alive, I'm interested. You wrote some really nice things.
George, thanks for the advice I almost forgot that 320V DC ... 
That's a nice topic, nice discussion... WOW!
Thank you all... keep it running please..
I'm going to focus on ordering quality parts (caps and resistors) in Europe somehow from somewhere.
What do you think, which brand of non-ultra-expensive capacitor and/or resistor should I take for crisp clear open sound ?
I've chosen until now Shinkoh Tantalum resistors and ELNA MKP caps for non-electrolytics / Rubycon Black Gate WKZ for that really few electrolytic caps.
They aren't so much expensive, not cheap, okay, but hey, I'm building for 20 years
The final amp will be my father's "commemorative edition" for me You could help me with some brands/types of caps and resistors for clear sound, I don't favor the lush sweet sound of tubes (meant here in good terms now) too much - just a bit.
And don't forget to keep the discussion abot nfb alive, I'm interested. You wrote some really nice things.
George, thanks for the advice
I almost forgot that 320V DC ...
Hi,
Excellent as anode resistors, chose wattage and voltage accordingly....
Where do you buy them from?
For gridstoppers I use ordinary carbon comp, 1/4W.
Cathode resistors, I'd recommend Riken but this is more audible in a preamp.
Most other lower wattages can be ordinary metalfilms from say, Philips.
For coupling caps I really like the ICEL MKPs which are axial and therefore ideally suited for P2P wiring.
Arcotronics is said to be fine sounding as well.
A little more high-end are the Solen FC, Audyn Caps and Mundorf, all of which should be available from European sources.
I don't think there's any electrolytic inside of them so they should last a very long time as they shouldn't dry up.
Be warned that they need a really long break-in period.
For signal path wiring you want to try out pure silverwire...
Cheers,
Excellent as anode resistors, chose wattage and voltage accordingly....
Where do you buy them from?
For gridstoppers I use ordinary carbon comp, 1/4W.
Cathode resistors, I'd recommend Riken but this is more audible in a preamp.
Most other lower wattages can be ordinary metalfilms from say, Philips.
For coupling caps I really like the ICEL MKPs which are axial and therefore ideally suited for P2P wiring.
Arcotronics is said to be fine sounding as well.
A little more high-end are the Solen FC, Audyn Caps and Mundorf, all of which should be available from European sources.
I don't think there's any electrolytic inside of them so they should last a very long time as they shouldn't dry up.
Be warned that they need a really long break-in period.
For signal path wiring you want to try out pure silverwire...
Cheers,
http://ww3.szentes.hu/partnerek/gadget/tube_h.htm
From here (Shinkoh Tantalum).
It's a Hungarian store, but prices are also listed in Dollars and Euro. (Check the bottom).
What about those really expensive Caddock resistors ?
I've some places where the everywhere-used 0,5W-1W Shinkoh wouldn't fit... here I need 5W resistors. Actually a very few need to be 5W rated - that's all.
If you're interested in other resistor sources, try these:
http://www.triodes.nl/
http://www.kyoto-electro.com/
http://www.octave-electronics.com/
http://www.donberg.ie/
http://www.reichelt.de/ -> extremely lots of parts here! nice.
http://www.diyhifisupply.com/
.
.
.
I know the Black Gates aren't real electrolytics, they're just told to be.. I think I can afford them, I have to buy 8 pieces or so all together so no prob.
Pure silver wire is still a problem, where to order from. What cable diameter do you think will suit 320V / 110mA ? (In mm2 please or effective diameter of the metallic wire, so NOT including insulation).
I don't really want to buy teflon insulated ones, just normal pure silver wire. I'll insulate them myself with a nice thick layer of silicone paste usually used in bathrooms (So theoretically there'll be even less air between insulation-cable so silver won't begin to erode so easy).
From here (Shinkoh Tantalum).
It's a Hungarian store, but prices are also listed in Dollars and Euro. (Check the bottom).
What about those really expensive Caddock resistors ?
I've some places where the everywhere-used 0,5W-1W Shinkoh wouldn't fit... here I need 5W resistors. Actually a very few need to be 5W rated - that's all.
If you're interested in other resistor sources, try these:
http://www.triodes.nl/
http://www.kyoto-electro.com/
http://www.octave-electronics.com/
http://www.donberg.ie/
http://www.reichelt.de/ -> extremely lots of parts here! nice.
http://www.diyhifisupply.com/
.
.
.
I know the Black Gates aren't real electrolytics, they're just told to be.. I think I can afford them, I have to buy 8 pieces or so all together so no prob.
Pure silver wire is still a problem, where to order from. What cable diameter do you think will suit 320V / 110mA ? (In mm2 please or effective diameter of the metallic wire, so NOT including insulation).
I don't really want to buy teflon insulated ones, just normal pure silver wire. I'll insulate them myself with a nice thick layer of silicone paste usually used in bathrooms
(So theoretically there'll be even less air between insulation-cable so silver won't begin to erode so easy).
not such a good idea if you are concerned about noise. carbon comp resistors can contribute a fair bit of noise to a circuit, especially if used in any of the earlier stages of the amp. I personally use 1W and 2W carbon film resistors. I think these are a better choice for a few reasons.
1) they are a bit more stable than carbon comps, and will drift less over time. they can also be bought to within 5% acuracy, whereas (as far as i know and can source), carbon comps are only available to 10%.
2) they still use carbon, so they avoid using metal resistors in your signal path. this is of course subjective wether you prefer carbon or metal based resistors. i don't know which one is more acurate, but generally, metal resistors don't have as warm a sound as carbon based ones. also carbon comps are thought to be a bit warmer than carbon films.
3) The carbon films have less noise than carbon comps. while both suffer the same level of thermal noise, and neither is suffering from shot noise (well, they are a little, but in grid stopper position, not really enough to worry about), the carbon comp will have more contact noise than the carbon films. Contact noise is partly dependant on the current flowing through resistor, but we can't change this in our design (we can, but it would mean modifying the design)... what we can change is two other important things... power rating and material. resistors are noisiest to quietest from carbon comp, carbon film, metal oxide, metal film, and then wirewond doesn't suffer from contact noise. carbon film are the best as far as the carbon resistors go in this.... the other point of the power rating is that the higher the power rating, the less contact noise. The reduction of noise acording to extra power rating is not a direct realationship, so 4 times the power won't give four times less noise... it will be closer to 3 times less noise. so this is good! this is not three times less noise in the amp, just three times less resistor noise... but every little bit of noise you can get rid of is an advantage, so do everything you can! also, paralleling resistors reduces contact noise.
of course, for even less noise, use metal film... wirewonds are hard to get in high resistance values (at least where i source components from, no one stocks anything higher that 5k)
The reason i use 1W and 2W is out of economy... for the majority of the amp, i will use 1W, but in high resistance posistions, and early stages i will pay the extra for 2W... simple as that.
as far as caps. my best recomendation is use polypropelene caps... people say to use these for reasons of better sound, but as everything, this is subjective... a better reason to use them over say, polyester is they are more tolerant to varying temperatures, and won't drift much with temp... and drif is a bad thing, as especialy true with caps, it will change the sound of your amp. you might thinl that temp variation might not be to great in your amp, but firstly, while the inards of your amp might not get really hot, they will get a lot warmer the longer they're used... and it doesn't take much to change polyester... trust me, i know this from another one of my hobbies, which is bodyboarding. When you buy a board, you buy a material suitable for temp of water your gunna surf in, or else your board will be to stiff or too flexy to manoevuer properly with. Polypropele is a widely used material becuase of its stability across a wider range of temperatures... and trust me, it doesn't take much a temp change to change stiffnes of it... and this all relates to properties of caps... so yeah.... kinda rambled there, but anyway. moving on.
yeah, i wasn't talking about phase splitters... sorry if i implied it somehow... but from my interperatations of what you said earlier, it seemed like you were saying that the two sides must be matched, or else one will hog more of the current...
when we parallel two devices, the source current is split between them... as we all know.. and of course this is what's happening here... our source's current is split between the two sides of the tube when we parallel them. as we've biased both of these sides the same, as long as the current is split fairly evenly, and the tube's biased corectly, one side shouldn't dissiapte much more power than the other... unless of course we have fairly unequal current split... what i'm asking is, seeing as the two sides are in //, shouldn't the current be split acording to the grid resistance??? or is the current able to 'see' more than this, and will split unevenly.
Once both tubes have amplified signal, it comes out on the anode, and as they are also parallel, the two waveforms from the tubes add together. as long as neither side of the tube has distorted because the current was split too unevenly, and the two signals from the anodes being added are still in phase, there shouldn't be any reduction of sound quality as oposed to just the one section on it's own.
well at least that's how i see it.
cheers
I believe you are thinking in transistor terms where there is a base current flowing, tubes don't generally draw grid current and therefore the grid voltage is equal when 2 tubes are connected in parallell. Current hogging occur as the transconductance in the tubes are not equal so the one with highest transconductance gets the larger part of the current.
It is no real problem to connect tubes in parallell but you should have resistors in series with each cathode or anode before connecting in parallell, this is something that is mandatory when connecting power tubes in parallell, (as is usually done in OTL amps) but sometimes it is forgotten or the resistors are too small. Also tubes in parallell must be derated as some current hogging can not be avoided if the series resistors should have reasonable values, derating of parallell connected tubes is something that many manufacturers of OTL amps especially have forgotten or chose to ignore.
Regards Hans
Hi hans,
thanks for that. i understand now.
so baisically, as long as all the right precautions are taken, apart from the added cost, paralelling of tubes is actually a good thing like i originally bought up because of lower noise floor and lower output imp.... so is there still some degeneration of the quality of the sound somehow, or is it all good as long as it is implemented properly?
i've used this in a guitar amp before, but i haven't used it in a hifi amp before... so i don't know yet from practical experience.
just one other thing,
what's OTL stand for??? i'm not familiar with this abreviation.
thanks
thanks for that. i understand now.
so baisically, as long as all the right precautions are taken, apart from the added cost, paralelling of tubes is actually a good thing like i originally bought up because of lower noise floor and lower output imp.... so is there still some degeneration of the quality of the sound somehow, or is it all good as long as it is implemented properly?
i've used this in a guitar amp before, but i haven't used it in a hifi amp before... so i don't know yet from practical experience.
just one other thing,
what's OTL stand for??? i'm not familiar with this abreviation.
thanks
hi vortex,
just had a look over @ your website... even though not really a website, i still had a look... and i found your schematic for this amp. i must say, i quite like the look of this design.
i also see what you mean when you said you're father laughed at you for just doubling all caps... not that i'm laughing, but there are a few caps you've doubled in there that don't need to be!
anyway, while i'm thinking of it, i wish you all the best, and hope it turns out well for you. Just remember to make sure you really know what you're doing if this is your first amp, or else you might be in for a nasty surprise...
later
just had a look over @ your website... even though not really a website, i still had a look... and i found your schematic for this amp. i must say, i quite like the look of this design.
i also see what you mean when you said you're father laughed at you for just doubling all caps... not that i'm laughing, but there are a few caps you've doubled in there that don't need to be!
anyway, while i'm thinking of it, i wish you all the best, and hope it turns out well for you. Just remember to make sure you really know what you're doing if this is your first amp, or else you might be in for a nasty surprise...
later
well, never thought of that one... thanks for the tip!
do you know how carbon film compares in this respect to carbon comp? obviously it's not as good if carbon composite's are the best, but is it close?
maybe we could have the best of both worlds??? by paralleling a couple of resistors, say two of double the value we need, we drop contact noise, while still retaining the low self inductance of carbon comp? Even if this is more expensive, it could be a good idea early on... especially on that big 1M resistor right on the I/p!
cheers,
Yes and no, parallelling is better as you say because of lower nosie and lower output impedance but if it s done the right way with all precautions then the efficiency is reduced and there is still the risk of runaway and current hugging.
At least in power amplifiers I prefer to use a pair of larger tubes than using several smaller tubes in parallell.
Regards Hans
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.