True! Is the 6C33 any better?
I bought 100 tubes
But I've had some amps with unmatched tubes, and honestly it's hard to tell.But some mismatch is ok. When current is drawn, the weaker sections catch up to the stronger. This is where the relatively large cathode resistors come in, tho it happens even without them.
I've found that sections 1,2,3 in one bottle is easier to match with same section in another. So I use sections 1,2,3 in the upper triode, and sections 4,5,6 in the lower.
SemperFi Thx again.
Are u saying that you just try to match the top sections with each other and the say for the bottom sections on a totem otl? Not sure I am gettin g you here?
Also I have tought about just paralleling the 2 sections of a 6as7/6080 and using that as 1 section. Thoughts?
As always your insight is greatly appreciated
Semper,
You mentioned the EF86 as a reason not to build the amp. In my limited experience, the EF86 is a very, very nice sounding pentode. I wondered if it may be misapplied here, since it only runs at 12AX7 type currents, and for a driver, especially a 6C33C driver where grid current is to be expected, I just don't see the available current to keep the output grid supplied. I was thinking of something many times more robust, along the line of Tube Lab George's PowerDrive circuit....
Your amp design is very interesting. i must go back and read up on current mirrors again. I have to admit, although I use lots of sand in my tube equipment, bipolar junctions make me twitch.
Stuart
You mentioned the EF86 as a reason not to build the amp. In my limited experience, the EF86 is a very, very nice sounding pentode. I wondered if it may be misapplied here, since it only runs at 12AX7 type currents, and for a driver, especially a 6C33C driver where grid current is to be expected, I just don't see the available current to keep the output grid supplied. I was thinking of something many times more robust, along the line of Tube Lab George's PowerDrive circuit....
Your amp design is very interesting. i must go back and read up on current mirrors again. I have to admit, although I use lots of sand in my tube equipment, bipolar junctions make me twitch.
Stuart
For me, I have zero personal experience in building OTL amps, but they are definitely on my bucket list. I just want to do it right and am examining all options I can find. I hope this amp sounds great, be nice if someone had one in the area.
I wonder if Mr. Mellow reads this board? Really could use his input in clarifying his choices.
Stuart
Well, I cannot remember which, but in those tubes it seems one section always measures a little stronger than the other. And its the same in each bottle. I am likely to recall wrongly now, but I think section 123 almost always has a little more Ic than section 456. So I feel sections 123 in one tube should be paralleled with same section in another tube. Not sure it matters, but thats how I do it. It can seem like this gives grossly different bias for the upper and lower, but no since they are in series. The NFB provides reasonable offset.
Using all sections in one bottle (paralleling 123 with 456) may be better for avering out the mismatches. I dunno...
Dynamically I am sure it does not matter. DC match does not guaranty AC match. And like I said, I dont think matching is critical since it appears in average they even out. Not counting extreme mismatch that is.
No problem SemperFi, I was just wanting to get your opinion of this amp. I have 6C33C already on the shelve and always wanted an OTL. I also have a chassis that will work nicely. In your opinion, if I am tri amping, where would this amp shine....tweet, mid, bass?
I've never need Mr. Mellow on any board....and I've looked.
OK, jumped the gun in my last post. As soon as I get these other amps done, I guess I will find out.
Think I gotcha and yes you are right one section is 98% of the time higher then the other and yes it seems to almost always be the same section from tube to tube. SO I see where your going. Your saying keep all the 123 sections with similar matching 123 sections for the top and same for the other section 456 on the bottom. All good on that. I may also try paralleling the two sections of one tube and trying that.
No problem SemperFi, I was just wanting to get your opinion of this amp. I have 6C33C already on the shelve and always wanted an OTL. I also have a chassis that will work nicely. In your opinion, if I am tri amping, where would this amp shine....tweet, mid, bass?
QUOTE]
I'm sure Mellows amp is fine, my way of using mirrors is not necessarily better, just another way of doing it.
I use OTL on the bass, in my case two series connected 8" drivers in a closed box, for a nice OTL friendly 16ohm load, crossed over at about 300Hz to the Fostex fullrange (6") driven by my SE 6AS7 (transformer coupled) amp.
Two main reasons for using the OTL on the bass: 1, Being PP it has higher order harmonics than the SE amp, these higher order harms are filtered a bit by the LP filter in the speaker. 2, No output transformer with saturation problems in the low end.
Not sure this is the best solution, but works for me.
To be honest I havent had time to make a 'proper' version of the OTL for my own HiFi. I mainly build OTLs for guitar amps, and simply have no time left for my own system. My own system tends to be a collection of odd projects, and not always optimized for serious listening
Just to add...some of my observations on what 'traits' are difficult to 'iron out' with a totempole output stage.
The lower tube has gain relative to the top tube. In the case of the 6AS7 with a mu of only 2, this gain isn't a whole lot, but enuff to make it an issue. This gives both 1, poor power supply ripple rejection, and 2, the usual benefits of a PP amp where even harmonics are cancelled does not occur as well.
1: The power supply can either be forced clean with extreme amounts of capacitance, or regulated actively. Both of which is expencive, complex, and little elegant. I find a better way using less capacitance in the upper rail than the lower. The positive rail gets about 60% the capacitance of the negative rail. For instance using 4700uF on the negative rail, I use 2700uF in the positive. (Of course using several smaller caps in parallel). It can also be fixed by injecting some ripple from the neg rail into the PI, but I prefer the former method.
2:Can be remedied with adjusting gain for the lower tube, and having some NFB. In the circuit I posted I'd use 2kohms for R13.
I believe the circlotron being a better design for those two reasons.
The lower tube has gain relative to the top tube. In the case of the 6AS7 with a mu of only 2, this gain isn't a whole lot, but enuff to make it an issue. This gives both 1, poor power supply ripple rejection, and 2, the usual benefits of a PP amp where even harmonics are cancelled does not occur as well.
1: The power supply can either be forced clean with extreme amounts of capacitance, or regulated actively. Both of which is expencive, complex, and little elegant. I find a better way using less capacitance in the upper rail than the lower. The positive rail gets about 60% the capacitance of the negative rail. For instance using 4700uF on the negative rail, I use 2700uF in the positive. (Of course using several smaller caps in parallel). It can also be fixed by injecting some ripple from the neg rail into the PI, but I prefer the former method.
2:Can be remedied with adjusting gain for the lower tube, and having some NFB. In the circuit I posted I'd use 2kohms for R13.
I believe the circlotron being a better design for those two reasons.
Last edited:
Sory Guys if is this statement litlle of Topic:
6AS7G/6080/6h13c Tubes have to be internaly `divided` for OTL Amps beacose of internal Tubes Temperature unload,less for electrical charackteristic matching.
For example:
When 6AS7G /6080 is divided, where `123` triode section is conected for Top part of Totem Pole or Futterman PP OTL output power stage and triode section `456`is conected for Lower part of the same Output power stage,than internall 6080 Tube sections `123`and `456` allways work in oposite(PP) regime, resulting in less internal Peak Temperature Tube load.
Best regards
6AS7G/6080/6h13c Tubes have to be internaly `divided` for OTL Amps beacose of internal Tubes Temperature unload,less for electrical charackteristic matching.
For example:
When 6AS7G /6080 is divided, where `123` triode section is conected for Top part of Totem Pole or Futterman PP OTL output power stage and triode section `456`is conected for Lower part of the same Output power stage,than internall 6080 Tube sections `123`and `456` allways work in oposite(PP) regime, resulting in less internal Peak Temperature Tube load.
Best regards
Thanks for your thoughts and observations Semper. I guess I will just have to build these amp and see how it goes. I can't even imagine that it would suck. I have been assuming that I will use this on my tweets, but one cool thing about tri-amping is all the amp/driver match-ups I can do to get the best sound.
I'm certain it won't "suck". Just be aware that it may have a "sound" that might not mesh with your system. Even the best 12AX7s tend to have a relatively thin and bright sound. The EF86 driver might mean there may not be as powerful a drive (on powerful transients) as another type amp, etc. But I'm sure it won't suck. Build away, then experiment with alterations.
Keep us up to date, and please include pictures!
Thanks, and best of luck,
Stuart
Thanks Stuart. I have heard both: OTLs can sound thin and OTLs are great for bass. I agree, it won't suck, just a matter of what kind of sound it produces. I can live with any type/color/style of sound as long as it is clean and undistorted. My tweets are 94db so it should push them well enough. My mids are 105db so that is not any issue and niether is my woofs, they are 99db.
DJN,
You're in great shape with those efficiencies, just so long as the impedance isn't too low.
I'm not so fortunate: 83dB and 3-4 ohms. Yeow!
Kevin Covi published a great design where he got 100WPC from 4 6C33C per channel, in to 4 ohms, but 6C33C reliability was a problem. He switched to 12 6AS7G per channel and solved the problem. That's still the design i'm considering first.
Stuart
You're in great shape with those efficiencies, just so long as the impedance isn't too low.
I'm not so fortunate: 83dB and 3-4 ohms. Yeow!
Kevin Covi published a great design where he got 100WPC from 4 6C33C per channel, in to 4 ohms, but 6C33C reliability was a problem. He switched to 12 6AS7G per channel and solved the problem. That's still the design i'm considering first.
Stuart
