Hi, in general terms (I don't have a schematic) what variables need to change in an amp originally designed for a 12AU7, utilizing both triodes per channel, to accommodate higher gain preamp tubes?
Would it just be a matter of changing the cathode resistors?
Context:
I'm trying to build an amp to accommodate as many preamp tubes as possible (by setting jumpers). it seems like the only cheap 12AU7 alternative is the russian 6H8C, i've even considered quad 6C4, 7AU7 etc.. but seems like those are quite valuable. I'd like to be able to drop in things like a 6n2 or 6n1 since these things seems to be cheap and plentiful for now.
Edit: I should add - ignoring heater voltage and pin configuration as these are more obvious
Would it just be a matter of changing the cathode resistors?
Context:
I'm trying to build an amp to accommodate as many preamp tubes as possible (by setting jumpers). it seems like the only cheap 12AU7 alternative is the russian 6H8C, i've even considered quad 6C4, 7AU7 etc.. but seems like those are quite valuable. I'd like to be able to drop in things like a 6n2 or 6n1 since these things seems to be cheap and plentiful for now.
Edit: I should add - ignoring heater voltage and pin configuration as these are more obvious
Last edited:
This is like "Jack of all trades, master of none" ...
Of course a compromise circuit could be found but normally you would want the opposite , the best circuit for a specific tube
Of course a compromise circuit could be found but normally you would want the opposite , the best circuit for a specific tube
"Jack of all trades, master of none" is indeed the design goal of this particular amp,
I think what I probably just need to do is use a known signal source like a synth VCO and make sure the P-P coming out of each stage is the same as with the 12AU7 while using R values that won't blow up the tubes (studying the datasheets)
I don't expect to find common component values to work for every "gain class" of tube, however i'd like to be able to make it selectable by jumper to accommodate low, mid, and high gain tubes.
I think what I probably just need to do is use a known signal source like a synth VCO and make sure the P-P coming out of each stage is the same as with the 12AU7 while using R values that won't blow up the tubes (studying the datasheets)
I don't expect to find common component values to work for every "gain class" of tube, however i'd like to be able to make it selectable by jumper to accommodate low, mid, and high gain tubes.
If you are satisfied with the gain of your amplifier, then . . .
Go to a Ham Fest, and purchase either NOS or used: 12AU7, 5814 & 5814A tubes, to last the rest of your life. Or, just purchase new JJ ECC82/12AU7 tubes.
Note, the JJ ECC82/12AU7 uses spiral filaments, that is a plus.
You do not need to change self bias resistors, plate load resistors, B+ volts, fixed bias volts, etc.
Plug and PLAY
The 12AU7, 5814 & 5814A are essentially identical, except:
The 5814 & 5814A filaments require 0.35A at 6.3V, or 0.175A at 12.6V;
ECC82/12AU7 filaments require 0.30A at 6.3V, or 0.15A at 12.6V.
The 5814 & 5814A is a special 12AU7.
If your amplifier has global negative feedback which is fed back to the 12AU7 tube, then . . .
If you increase the gain of that stage, the amplifier may become unstable because you have changed the open loop gain to closed loop gain.
And, a triode with different miller capacitance, u, plate resistance rp, and plate load resistors RL, may change the phase of that stage - that could cause the amplifier to become unstable.
This may become Plug and PRAY.
If you are looking to do tube rolling, do it with near identical tubes.
Otherwise, you are testing the sound of different tubes, while at the same time testing the sound that is affected by the required different surrounding component values for the different tube.
Or, if you keep the surrounding component values, only one type of tube is going to be operated in an optimal setup.
Your Mileage May Vary.
Just my opinions.
Go to a Ham Fest, and purchase either NOS or used: 12AU7, 5814 & 5814A tubes, to last the rest of your life. Or, just purchase new JJ ECC82/12AU7 tubes.
Note, the JJ ECC82/12AU7 uses spiral filaments, that is a plus.
You do not need to change self bias resistors, plate load resistors, B+ volts, fixed bias volts, etc.
Plug and PLAY
The 12AU7, 5814 & 5814A are essentially identical, except:
The 5814 & 5814A filaments require 0.35A at 6.3V, or 0.175A at 12.6V;
ECC82/12AU7 filaments require 0.30A at 6.3V, or 0.15A at 12.6V.
The 5814 & 5814A is a special 12AU7.
If your amplifier has global negative feedback which is fed back to the 12AU7 tube, then . . .
If you increase the gain of that stage, the amplifier may become unstable because you have changed the open loop gain to closed loop gain.
And, a triode with different miller capacitance, u, plate resistance rp, and plate load resistors RL, may change the phase of that stage - that could cause the amplifier to become unstable.
This may become Plug and PRAY.
If you are looking to do tube rolling, do it with near identical tubes.
Otherwise, you are testing the sound of different tubes, while at the same time testing the sound that is affected by the required different surrounding component values for the different tube.
Or, if you keep the surrounding component values, only one type of tube is going to be operated in an optimal setup.
Your Mileage May Vary.
Just my opinions.
I think your goal is wrong.
Designing an amp for one tube might be optimised. But to accept wildly different tubes has no benefits unless one
is prepared to change most or all of the resistors. Anything else will create a mediocre unit that won't fit any demands.
Well, if my goal is wrong, then so be it. I’m stubborn and need to learn the hard way 🙂 even if it means breaking some things. I will eventually believe you that what i’m doing is pointless, but I need to actually arrive to the point where i’m tell myself “oooooh, this is what all those people were warning me about.” I hope you understand and hope this thread doesn’t become a dogpile against me. I just need to know what sort of things to consider and watch out for.
I don’t have any hamfests near me and so my tube sources are very limited to what I can find online. For some reason all the money in the world can’t get the people on fb marketplace/craigslist etc… to actually respond to tube collection ads.
This amp is more of a test bench for tube experiment/learning, so it’s ok for things to get unstable, I need to see that happen in action.
I have other amps for that’s are “optimized” for specific tubes
For the day when 12AU7s are 300B money, I’d rather know what to do now, than figure out it later.
I don’t have any hamfests near me and so my tube sources are very limited to what I can find online. For some reason all the money in the world can’t get the people on fb marketplace/craigslist etc… to actually respond to tube collection ads.
This amp is more of a test bench for tube experiment/learning, so it’s ok for things to get unstable, I need to see that happen in action.
I have other amps for that’s are “optimized” for specific tubes
For the day when 12AU7s are 300B money, I’d rather know what to do now, than figure out it later.
Even if this could be done, what about the overall stability of the circuit?
You'd have to switch the feedback components also.
Why not have several daughter boards that plug into a motherboard,
with each optimized for a particular tube?
You'd have to switch the feedback components also.
Why not have several daughter boards that plug into a motherboard,
with each optimized for a particular tube?
Actually, yes, that’s the idea, it’s a semi modular amp with the test points easily accessible.
Currently the preamp circuit is on the same board as the auto bias circuitry for the power tubes but I can separate them and make the the preamp section swappable. The preamp tubes are also mounted on a separate plate so i can mount the board with it.
Currently the preamp circuit is on the same board as the auto bias circuitry for the power tubes but I can separate them and make the the preamp section swappable. The preamp tubes are also mounted on a separate plate so i can mount the board with it.
There's a well-known design for this that might help you figure things out:
https://www.cascadetubes.com/the-universal-preamp/
https://www.cascadetubes.com/the-universal-preamp/
The simple answer to your question is that you can do it by changing both the cathode and plate resistors of a simple common cathode gain stage for each different tube. The downside is the gain of the preamp will depend on the tube and could vary from 5 to 50 times. Also, it will only be able to drive short cable lengths and light loads.
Cheers
Ian
Cheers
Ian
Several people have stated in posts that they prefer the E80CC to the 12AU7. This would be one of my tubes of choice, and I find it sounds better than 6N1P or ECC88 for instance. It's more expensive but not stupidly so. I prefer the Tungsram version with nickel pins to the Philips version with gold pins. It will require a higher value anode resistor since the Ri is 10K as opposed to the roughly 7.5K of the 12AU7. It also has a heater current of 600mA instead of 300mA so the heater specs need to be checked. Mu is 27.
For higher gain my choice after trying several alternatives would be the E180CC. Heater current 400mA, Ri=7.5K so that's the same, though the operating point would be different. Both tubes have the same pinout as the 12AU7.
I personally don't like or use active loads - I prefer a simple resistor load. For a B+ of 300v you are looking at around 4 to 6mA through the tube to get a sensible value anode resistor of 27K to 34K. You can use a cathode bypass cap but another trick which makes life easier is a SIC diode in the cathode which gives you 0.8v to add to the bias. I use a single diode to ground with a resistor on top of it of 180 ohms which gives a bias of around 1.7v or 1.8v. So no cathode bypass cap. A single SIC diode seems relatively smooth with an added resistor, though I find SIC diodes are prone to a slight edge without a resistor and sound better with some tubes than others.
Attached a circuit which I use, and I use this same circuit for 6SN7, E80CC and E180CC tubes. So I have the possibility of mu=20 for 6SN7 up to mu=46 for the E180CC. Sound is pretty good to my ears, and it drives my 2a3 outputs in a 2 stage SE.
This is just one way of doing it but it may interest you in your experiments. I also like the concept of a modular build where you can drop in different alternatives, and I found this circuit to work with each of the three alternatives above plus various 6J5 types which can sound very nice, in addition to 7N7 and 7A4. So I have a few different socket types wired together which all use the same circuit.
For higher gain my choice after trying several alternatives would be the E180CC. Heater current 400mA, Ri=7.5K so that's the same, though the operating point would be different. Both tubes have the same pinout as the 12AU7.
I personally don't like or use active loads - I prefer a simple resistor load. For a B+ of 300v you are looking at around 4 to 6mA through the tube to get a sensible value anode resistor of 27K to 34K. You can use a cathode bypass cap but another trick which makes life easier is a SIC diode in the cathode which gives you 0.8v to add to the bias. I use a single diode to ground with a resistor on top of it of 180 ohms which gives a bias of around 1.7v or 1.8v. So no cathode bypass cap. A single SIC diode seems relatively smooth with an added resistor, though I find SIC diodes are prone to a slight edge without a resistor and sound better with some tubes than others.
Attached a circuit which I use, and I use this same circuit for 6SN7, E80CC and E180CC tubes. So I have the possibility of mu=20 for 6SN7 up to mu=46 for the E180CC. Sound is pretty good to my ears, and it drives my 2a3 outputs in a 2 stage SE.
This is just one way of doing it but it may interest you in your experiments. I also like the concept of a modular build where you can drop in different alternatives, and I found this circuit to work with each of the three alternatives above plus various 6J5 types which can sound very nice, in addition to 7N7 and 7A4. So I have a few different socket types wired together which all use the same circuit.
Attachments
Last edited:
I agree with everything Andy has said about the sound quality of the various tubes he mentioned. I've used the E80CC as driver in a PP 2a3 amp and the E180CC ( or its equivalent the 5965) in a preamp ( I also dislike active loads with transistors as they affect the naturalness of the sound of instruments compared to resistors).
I'd like to add a few more good sounding tubes here. Instead of the E88CC use the 6BQ7 6BK7 and ECC85 which have the same pin out. Only as cascode the E88CC sounds all right but not as a single triode. The ECC 40 (mu=33) is very good too but has an 8 pin socket unfortunately. Then the 6211A (mu=27 like the E80CC) and also the little used ECC 84 (mu=24) only this tube has a different pin out
I once made a simple test apparatus to compare the gains of double triodes which can also be used as a line or driver stage. The reason was that there is no 100% correlation between gain and emission. It can also be used to judge the sound quality of tubes.
It consists of a chassis with 9-pin 7-pin sockets one anode resistor (100k) and one cathode resistor (1k) plus 220mu Ck. The 100k goes to the switch with 12 contacts where all the anodes go to. All cathodes are connected. So are the grids with one Rg= 470k. A 1: 10 resistor divider is placed over the 6.3v (10k and 90k in series) resulting in about 0.6v ac. which goes to the grids. Connect your dvm with an output C of 0.1 muf apply the hv (200v or more) switch from one anode to the other and you know the gain in that particular situation. Of course it's relative but it's very useful especially in stereo amps without nfb. I hope you can do something with this efm 7.
I'd like to add a few more good sounding tubes here. Instead of the E88CC use the 6BQ7 6BK7 and ECC85 which have the same pin out. Only as cascode the E88CC sounds all right but not as a single triode. The ECC 40 (mu=33) is very good too but has an 8 pin socket unfortunately. Then the 6211A (mu=27 like the E80CC) and also the little used ECC 84 (mu=24) only this tube has a different pin out
I once made a simple test apparatus to compare the gains of double triodes which can also be used as a line or driver stage. The reason was that there is no 100% correlation between gain and emission. It can also be used to judge the sound quality of tubes.
It consists of a chassis with 9-pin 7-pin sockets one anode resistor (100k) and one cathode resistor (1k) plus 220mu Ck. The 100k goes to the switch with 12 contacts where all the anodes go to. All cathodes are connected. So are the grids with one Rg= 470k. A 1: 10 resistor divider is placed over the 6.3v (10k and 90k in series) resulting in about 0.6v ac. which goes to the grids. Connect your dvm with an output C of 0.1 muf apply the hv (200v or more) switch from one anode to the other and you know the gain in that particular situation. Of course it's relative but it's very useful especially in stereo amps without nfb. I hope you can do something with this efm 7.
I'm glad our listening experiments seem to have given the same results. I've spent a lot of time trying out different driver tubes and listening carefully to them.
I have some 6BQ7 and 6BK7 I haven't got around to yet. Don't have any 6211A, though. There's a little family around the E180CC - 6414, 12AV7, 5965, 7062, 6829, 5062. They all sound good but to my ears the E180CC is the best of them. That's a very cheap and available tube and deserves to be better known. It seems perverse to use the 12AT7 instead of this little family. Same pinout.
I have some 6BQ7 and 6BK7 I haven't got around to yet. Don't have any 6211A, though. There's a little family around the E180CC - 6414, 12AV7, 5965, 7062, 6829, 5062. They all sound good but to my ears the E180CC is the best of them. That's a very cheap and available tube and deserves to be better known. It seems perverse to use the 12AT7 instead of this little family. Same pinout.
You might find my old Tube Taster interesting which was a line-stage where I could easily switch between 6 different tubes using parafeed topology. You wouldn't need the output transformer.
http://audioratbag.blogspot.com/2012/02/tube-taster-linestage.html
However, I doubt you could emulate the build quality.
ray
http://audioratbag.blogspot.com/2012/02/tube-taster-linestage.html
However, I doubt you could emulate the build quality.
ray
Essentially, 'separate circuits' is the topology except that the B+ has to be the same for all 6 tubes. The plate load can be resistor or CCS, the bias can be provided by resistor, resistor - capacitor, diodes or whatever. The OP is trying to work out what works for him and this topology is just a step along the way.
ray
ray