Hi All,
After building my SE EL84 amp I've decided to build a SE EL34 amp. I'm looking for a little bit more power. I'm considering using a 5687 (I have lots of them) or a 6SN7 in the gain stage of the amp. I'm looking for opinions and suggestions on which tube to use and how to bias it. I'm going to power the plate of the EL34 with 300 volts. If I use a 6SN7 I need the plate resistor to be about 20k. I'm looking for about 10mA current from the 6SN7. I like the sound of the 5687 but I've heard that 6SN7s sound really nice also and an all octal amp appeals to my sense of history. Any suggestions are appreciated. Thanks.
G
After building my SE EL84 amp I've decided to build a SE EL34 amp. I'm looking for a little bit more power. I'm considering using a 5687 (I have lots of them) or a 6SN7 in the gain stage of the amp. I'm looking for opinions and suggestions on which tube to use and how to bias it. I'm going to power the plate of the EL34 with 300 volts. If I use a 6SN7 I need the plate resistor to be about 20k. I'm looking for about 10mA current from the 6SN7. I like the sound of the 5687 but I've heard that 6SN7s sound really nice also and an all octal amp appeals to my sense of history. Any suggestions are appreciated. Thanks.
G
anodresistor value
Hallo I´m building preamps for guitar. I use ECC 81 and sometimes 12AX7. Mullard and Telefunken are good marks. In Marscall and Fender they use 100K for 12AX7. But I get a much cleaner sound with 270K or sometimes larger value. The same with ECC81. About 90K gives the cleanest sound. A very good thing is to use an adjustable katodresistor, minipot. That will make it possible to set a "best workingpoint". Good luck!
Stewen in Gothenburg, Sweden
Hallo I´m building preamps for guitar. I use ECC 81 and sometimes 12AX7. Mullard and Telefunken are good marks. In Marscall and Fender they use 100K for 12AX7. But I get a much cleaner sound with 270K or sometimes larger value. The same with ECC81. About 90K gives the cleanest sound. A very good thing is to use an adjustable katodresistor, minipot. That will make it possible to set a "best workingpoint". Good luck!
Stewen in Gothenburg, Sweden
> I'm looking for about 10mA current from the 6SN7.
That's an awful lot of current for a 6SN7. It will do it, but it won't do much more, which will limit your negative swing.
Not sure why you are starting with a specification of DC conditions (300V, 20K, 10mA). Usually you start with the audio requirements, which might be "70V peak-peak into 100KΩ, under 50K source impedance for HF response". Sure, it is nice if the driver B+ is the same or a bit less than the output stage B+, but why the other restrictions? If there is some reason you need specific DC conditions, maybe you could post a schematic so we understand?
A good classic bias for 6SN7 is 300V B+, 47K plate resistor, and about 2mA plate current. You can steal values directly from the Resistance Coupled Amplifier chart in the back of the tube manual.
If you "must" suck 10mA, Joel's suggestion to try a plate-choke makes a lot of sense, and gives horrendous large available grid drive (possibly 400V p-p from a 300V supply!). And distortion is lower than resistance coupled, if the choke is perfect. It never is, so there is always some "iron sound", which may or may not be to your taste. If you like heavy overall feedback, it is hard to manage that with two transformers/chokes inside the feedback loop, because the LF roll-offs and HF-peaks give complicated phase shifts.
That's an awful lot of current for a 6SN7. It will do it, but it won't do much more, which will limit your negative swing.
Not sure why you are starting with a specification of DC conditions (300V, 20K, 10mA). Usually you start with the audio requirements, which might be "70V peak-peak into 100KΩ, under 50K source impedance for HF response". Sure, it is nice if the driver B+ is the same or a bit less than the output stage B+, but why the other restrictions? If there is some reason you need specific DC conditions, maybe you could post a schematic so we understand?
A good classic bias for 6SN7 is 300V B+, 47K plate resistor, and about 2mA plate current. You can steal values directly from the Resistance Coupled Amplifier chart in the back of the tube manual.
If you "must" suck 10mA, Joel's suggestion to try a plate-choke makes a lot of sense, and gives horrendous large available grid drive (possibly 400V p-p from a 300V supply!). And distortion is lower than resistance coupled, if the choke is perfect. It never is, so there is always some "iron sound", which may or may not be to your taste. If you like heavy overall feedback, it is hard to manage that with two transformers/chokes inside the feedback loop, because the LF roll-offs and HF-peaks give complicated phase shifts.
Actually 10 mA for a 6Sn7 is only about 50% of what the tube is capable of. The 20K plate resistor comes from the old rule of thumb for a triode: "three times the plate resistance". Going from the "datasheet" I can put 300 VDC to the EL34 and use about 23K ohms to drop the voltage to 90VDC for the 6SN7 with a current through the 6SN7 of about 9mA. I'm mainly asking what the tube sounds like compared to a 5687. Thanks.
G
G
> 10 mA for a 6Sn7 is only about 50% of what the tube is capable of. The 20K plate resistor...
If you say 300V supply, 20K resistor, and 10mA bias, the 6SN7 is working at 100V 10mA. My RCA RC-22 graph for 6SN7 shows that point about 1/16 inch away from the zero grid bias line. In fact it cites a test condition of 10mA, zero grid volts, and says the plate drop is 90V. Sitting at 100V 10mA, the negative plate swing is only 10V, while the positive plate swing is well over 150V. An odd bias point.
> the old rule of thumb for a triode: "three times the plate resistance".
Uh? 3X is a minimum, usually gives significant distortion. And in an R-C Coupled stage you have to account for AC load being smaller than DC load.
My first-crack guess would be "plate resistor 5 times higher than plate resistance and next grid resistor 5X plate resistor".
If the next-stage grid is 100K, then we might use a 20K plate resistor. But then we want a tube resistance of 20K/5= 4K. The large-signal resistance of 6SN7 at 10mA zero-bias 90Vp is 9K. Too high. It will work, but for good output level you bias a triode plate to about 0.6 of B+, or 180V on a 300V supply. Under that condition the 6SN7 will swing over 70V peak. The RC-Amp table suggests 47K plate load (about 5X plate resistance) for 100K grid load and promises 73V peak.
> what the tube sounds like
If biased so it does not have to strain: pretty dull, in the sense that a 6SN7 has little "character". It's just a sturdy little workhorse with some 2nd Harmonic but no kinkiness. Driving a EL34, you should not have to swing a 6SN7 enough even to cancel some of the EL34's 2nd Harmonic.
Working at 300V supply, 20K, 10mA, it will fuzz. Go up the other way (starve it), and it will get sweet. That's part of the sound of the US adaptations of the Williamson amplifier: we used the 6SN7 which is not quite equivalent to the tubes Williamson used, and they ended up under-biased. In push-pull the 2nd Harmonic cancelled but the peaks got soft, moreso than in a happy Williamson. Not bad, in fact I rather enjoyed it.
If you say 300V supply, 20K resistor, and 10mA bias, the 6SN7 is working at 100V 10mA. My RCA RC-22 graph for 6SN7 shows that point about 1/16 inch away from the zero grid bias line. In fact it cites a test condition of 10mA, zero grid volts, and says the plate drop is 90V. Sitting at 100V 10mA, the negative plate swing is only 10V, while the positive plate swing is well over 150V. An odd bias point.
> the old rule of thumb for a triode: "three times the plate resistance".
Uh? 3X is a minimum, usually gives significant distortion. And in an R-C Coupled stage you have to account for AC load being smaller than DC load.
My first-crack guess would be "plate resistor 5 times higher than plate resistance and next grid resistor 5X plate resistor".
If the next-stage grid is 100K, then we might use a 20K plate resistor. But then we want a tube resistance of 20K/5= 4K. The large-signal resistance of 6SN7 at 10mA zero-bias 90Vp is 9K. Too high. It will work, but for good output level you bias a triode plate to about 0.6 of B+, or 180V on a 300V supply. Under that condition the 6SN7 will swing over 70V peak. The RC-Amp table suggests 47K plate load (about 5X plate resistance) for 100K grid load and promises 73V peak.
> what the tube sounds like
If biased so it does not have to strain: pretty dull, in the sense that a 6SN7 has little "character". It's just a sturdy little workhorse with some 2nd Harmonic but no kinkiness. Driving a EL34, you should not have to swing a 6SN7 enough even to cancel some of the EL34's 2nd Harmonic.
Working at 300V supply, 20K, 10mA, it will fuzz. Go up the other way (starve it), and it will get sweet. That's part of the sound of the US adaptations of the Williamson amplifier: we used the 6SN7 which is not quite equivalent to the tubes Williamson used, and they ended up under-biased. In push-pull the 2nd Harmonic cancelled but the peaks got soft, moreso than in a happy Williamson. Not bad, in fact I rather enjoyed it.
A question: Is one 150H choke good enough as plate load for a 6SN7 with both sections in //??
I´m thinking new driverstage for my 6B4G SE.
B+ = 250-300V and Ia= 8mA
(you guessed it, Hammond 156C)
When I think about it it feels like I´ve asked this Q before, but I don´t remember the answer...
I´m thinking new driverstage for my 6B4G SE.
B+ = 250-300V and Ia= 8mA
(you guessed it, Hammond 156C)
When I think about it it feels like I´ve asked this Q before, but I don´t remember the answer...
Well, maybe not too high, but way higher than I need.
I did some tests yesterday on the actual circuit, and at first I was a bit disappointed with the freq. responce.
Things got alot better with two of those chokes in series, luckily I have four more of those chokes coming in on the mail any day now.
I did some tests yesterday on the actual circuit, and at first I was a bit disappointed with the freq. responce.
Things got alot better with two of those chokes in series, luckily I have four more of those chokes coming in on the mail any day now.
In general, you want all the devices doing as little work as possible. 
Operating them in their most linear regions with as little swing as possible lowers distortion.
It's very enlightening to look over the Sylvania RC coupled amplifier charts, since they provide signal levels and distortion levels. The hated 12AX7 will do 0.3% distortion under the right conditions. Even a beloved 6SN7 can't do that.
Gain is good.
Operating them in their most linear regions with as little swing as possible lowers distortion.It's very enlightening to look over the Sylvania RC coupled amplifier charts, since they provide signal levels and distortion levels. The hated 12AX7 will do 0.3% distortion under the right conditions. Even a beloved 6SN7 can't do that.
Gain is good.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.