6SN7GTB DRIVER STAGE?
- Thread starter Stavee
- Start date
Register to hide this ad
That driver stage has enough output to drive the KT88s right out of their sockets — and rather easily so! A classic Williamson design……..
Dave
Dave
Stavee
Active Member
Thank you!
onemug
Super Member
Curious... how would a 6sn7 do driving a 300b? Is it harder or easier to drive than a KT88?
Asking because I have a 300b SET amp that was designed to have 2 6sn7s per channel and it was modified to where only 1 is driving the 300b now. A quick answer is fine.
300B requires significantly more voltage and current to drive than KT88, but the 6SN7 is a very common choice for that.
I'd like to suggest a modification to the bias arrangement. See attached. This would let it fail safe in the event of a pot failure.
As for the 300b, a quick and dirty way of estimating required drive voltage is to look at the grid bias. If it needs -45 volts, that tube needs pretty darn close to 45 volts peak to drive to full power. A push-pull amp would need double that grid to grid.
As for the 300b, a quick and dirty way of estimating required drive voltage is to look at the grid bias. If it needs -45 volts, that tube needs pretty darn close to 45 volts peak to drive to full power. A push-pull amp would need double that grid to grid.
onemug
Super Member
short answer is yes, long answer is that its not very much gain so expect to need a pretty considerable amount of input to reach full output power. The resistance coupled amp charts for a 6sn7 give conditions to get nearly 100 volts peak output, and the 300B datasheet says about 74 volts peak input in class A to get 7 watts of output.
Keep in mind that a 6SN7 is two triodes in one bottle, so one tube could be a three-stage design, and the two-tube setup could have sections paralleled and also be a three-stage design. The answer really becomes "it depends", depending on the circuit and the goals. Generally speaking, common advice seems to be that to sound best, 300Bs and other power triodes need to be adequately driven so not somewhere to flirt with the minimum.
true, I read that initially as "one" not "one per channel". Per-channel you get 2 triodes and thats plenty.
So, the wiper remains unconnected?
no, the drawing indicates it wired as a variable resistor, not a potentiometer. The wiper is connected to one end of the resistance track. Different schematic symbol for a VR vs a pot. If the wiper lifted, you'd get minimum voltage drop which = max negative grid supply. Tube stops conducting vs starts to cook itself to death.
this also should have a fixed resistor between the end of the variable resistor and ground to limit the adjustment range. As-drawn it could be adjusted to give 0 grid volts which you'd never want or need. Normally you'd want to design it knowing your expected bias voltage and have that be the output voltage with the pot set at approximately mid-travel. Set up the resistors on either end of the pot so the adjustment range is within some sane amount, maybe +/- 10% of the target grid voltage. The more range the pot has, the more touchy it gets so there is always a compromise of adjustment range vs ease of adjustment.
can also get this done with a resistor from the wiper to the bias supply end of the pot to provide a path to negative supply. Both methods work fine, but having almost lost tubes because of pots that decided to behave poorly I'm not in love with having to rely on them not getting flaky.
this also should have a fixed resistor between the end of the variable resistor and ground to limit the adjustment range. As-drawn it could be adjusted to give 0 grid volts which you'd never want or need. Normally you'd want to design it knowing your expected bias voltage and have that be the output voltage with the pot set at approximately mid-travel. Set up the resistors on either end of the pot so the adjustment range is within some sane amount, maybe +/- 10% of the target grid voltage. The more range the pot has, the more touchy it gets so there is always a compromise of adjustment range vs ease of adjustment.
can also get this done with a resistor from the wiper to the bias supply end of the pot to provide a path to negative supply. Both methods work fine, but having almost lost tubes because of pots that decided to behave poorly I'm not in love with having to rely on them not getting flaky.
maxhifi
Addicted Member
Yes, a fail-safe design is always better than a design where a single failure initiates a destructive chain reaction!
Stavee
Active Member
Thank you so much for the visual input. Makes sense and I will absolutely implement that measure. Do you or anyone else have a suggestion on where to order good components that won't break the bank? This will be my first build, so I'm stumbling through the trees right now.....
Things like a good volume pot, tag strips, octal sockets, etc.......
The schematic at the top of this thread uses a 4K ohm 65W UL PP OPT. I've been looking at Edcore for iron, but the only 4K Primary is 60W and I'm not sure at this stage what kind of wattage to expect it to handle. I want to stay well away from any core saturation.
Last edited:
maxhifi
Addicted Member
So, if I understand correctly, you're going to build this PP KT88 Williamson design? A few suggestions:
1) A Williamson is my favorite type of PP amplifier because of the naturalness of the sound. You'll pry my DIY Williamsons out of my cold, dead hands. ;-) But it took a fair bit of learning to get there.
2) Williamson amplifiers are challenging to build. Because of the number of frequency poles in this amp, they can be difficult to stabilize.
3) Because of the amount of feedback employed in a typical Williamson, the output transformer is critical. It must have enough bandwidth to tolerate the amount of feedback employed. You can't drop in any old output transformer and expect the amp to perform properly. It might even oscillate and potentially damage your speakers.
This being your first build, I'd ask the following questions, because a PP feedback amp, and especially a Williamson, can be a challenge:
Do you have any test equipment, like an oscilloscope and a signal generator? Nothing fancy, just the basics. An entry level USB Picoscope is not expensive and is a *very* handy tool.
Also, do you really need 60wpc? What speakers are you driving? I ask because a 25wpc cathode-bias Williamson is not only a nicer-sounding amp, IMO, it is also easier to build from known schematics. But a 60wpc Williamson like the early Dynaco designs is also quite nice.
Before you spend a dime on parts and ***especially*** before you buy output transformers, it would be a very good idea to find a schematic that specifies readily-available components and has been carefully designed around those components.
There are folks here and at diyaudio who have designed good feedback amplifier around commonly-available transformers from Hammond, Edcor and other suppliers.
If you have some test equipment and want to build from "scratch," that's great and I know a lot of folks here would be willing to help and would be interested in your results.
My guess is that a Hammond 1650N would be your best bet for this circuit, and you could probably find a tested schematic using it. The Hammonds are variable, but the 1650 series seems to be pretty well-regarded.
I got really interested in Williamsons a few years ago but had built NOTHING more than a simple SE 300B amp. I have some really terrific sounding amps now and I owe that to the incredibly generous folks in this group.
1) A Williamson is my favorite type of PP amplifier because of the naturalness of the sound. You'll pry my DIY Williamsons out of my cold, dead hands. ;-) But it took a fair bit of learning to get there.
2) Williamson amplifiers are challenging to build. Because of the number of frequency poles in this amp, they can be difficult to stabilize.
3) Because of the amount of feedback employed in a typical Williamson, the output transformer is critical. It must have enough bandwidth to tolerate the amount of feedback employed. You can't drop in any old output transformer and expect the amp to perform properly. It might even oscillate and potentially damage your speakers.
This being your first build, I'd ask the following questions, because a PP feedback amp, and especially a Williamson, can be a challenge:
Do you have any test equipment, like an oscilloscope and a signal generator? Nothing fancy, just the basics. An entry level USB Picoscope is not expensive and is a *very* handy tool.
Also, do you really need 60wpc? What speakers are you driving? I ask because a 25wpc cathode-bias Williamson is not only a nicer-sounding amp, IMO, it is also easier to build from known schematics. But a 60wpc Williamson like the early Dynaco designs is also quite nice.
Before you spend a dime on parts and ***especially*** before you buy output transformers, it would be a very good idea to find a schematic that specifies readily-available components and has been carefully designed around those components.
There are folks here and at diyaudio who have designed good feedback amplifier around commonly-available transformers from Hammond, Edcor and other suppliers.
If you have some test equipment and want to build from "scratch," that's great and I know a lot of folks here would be willing to help and would be interested in your results.
My guess is that a Hammond 1650N would be your best bet for this circuit, and you could probably find a tested schematic using it. The Hammonds are variable, but the 1650 series seems to be pretty well-regarded.
I got really interested in Williamsons a few years ago but had built NOTHING more than a simple SE 300B amp. I have some really terrific sounding amps now and I owe that to the incredibly generous folks in this group.
Love that tube
Stavee
Active Member
Yes, I have a tektronix 2213 oscilloscope. Been looking for a signal generator at some ham swaps; no luck yet.
I don't always crank it up, but there are some times when I like to "feel" the music. I all ready bought a quad of Psvane KT88 Mk2 Collection and have a quad of Tung-Sol 6SN7GTB's on the way. Right now I have an older pair of Klipsch book shelf size speakers that really sound good. They can handle some power. But when I finish this amp, I have been really interested in building one of CSS Audio's kit speakers. Heard a LOT of good reviews about them. I think StephK built a pair on her channel. (learned a lot from her videos)
I have Looked at the Hammond 1650N before but leaned toward Edcore for a little savings. But I might just have to go with what works best. It may be a long time before I build another and I just don't want to be left wishing I had done something a little more. (wattage) But then again I don't have to max them out and maybe they last a long time. Just because you have a rocket doesn't mean you have to go to the moon!
I also read a thread where GordonW was talking about combining global and local feedback into a 6SN7 driver stage and got super low distortion and better bass response. I think that thread was started by StephK
Thank you
P.S. I saw some TherionicLab OPTs where you can tie together some of the secondary terminals and utilize all the turns on the output for which ever speaker load you decide to run.............Thoughts?
Last edited: