I have seen Pete's low mu triode preamp in which he uses CCS load to linearize the pass tube triodes that are the main active devices. This got me to thinking, what if we use a triode that is already quite linear and use regular resistive loading. You will be dropping significant power in the load resistors but the active load dumps a lot of power too.
So I looked at the data sheet for the power triode in the 6EM7. According to the rp plots and plate curves we ought to get pretty low Zout (well under 1K) and linearity running a b+ around 300V or a bit more with 140V on the plate and around 40mA of current. We would need to dissipate about 5 or six watts in the plate load (around 5K ohms IIRC) which wouldn't be hard to do with a hand full of paralleled MF resistors. I was thinking fixed bias to make the maximum use of the available B+ and avoid having to dissipate large wattage in the cathode resistor too. This also would also allow easy experimentation with different bias levels.
The extra small signal triodes can be used for buffers for a subwoofer output or as part of a built in crossover or phono preamp.
Thoughts?
So I looked at the data sheet for the power triode in the 6EM7. According to the rp plots and plate curves we ought to get pretty low Zout (well under 1K) and linearity running a b+ around 300V or a bit more with 140V on the plate and around 40mA of current. We would need to dissipate about 5 or six watts in the plate load (around 5K ohms IIRC) which wouldn't be hard to do with a hand full of paralleled MF resistors. I was thinking fixed bias to make the maximum use of the available B+ and avoid having to dissipate large wattage in the cathode resistor too. This also would also allow easy experimentation with different bias levels.
The extra small signal triodes can be used for buffers for a subwoofer output or as part of a built in crossover or phono preamp.
Thoughts?
A couple of things I disagree with. First, the question of linearization. No matter what triode you choose, the linearity will be worse with a resistor load than with a CCS. The transfer function is mu times load resistance divided by the sum of load and plate resistances. Plate resistance varies strongly with current, load resistance and mu are relatively constant. The variation in the TF is minimized with higher load resistance.
Second, the CCS load allows you to reduce the dissipated power by working with a reduced B+ for a given output swing.
Second, the CCS load allows you to reduce the dissipated power by working with a reduced B+ for a given output swing.
There is 50+ pages discussing a resistively loaded 12B4 line stage here. The 12B4 is a vertical sweep triode similar to the big half of a 6EM7. I played with it for a while and then lost interest. Most of my 6EM7 experiments have used the big half with a CCS load and a lot of B+ to make lots of drive voltage to feed a cathode follower or screen driven output stages. My issue with the 6EM7 / 6EA7 is not the big half. It is the small half. They are very inconsistent and the gain can vary over a 2 to 1 ratio from tube to tube in a given circuit. I have tried a lot of them >50.
Another trick that can be used is to put a CCS and a resistor in parallel. On some triodes this can be more linear than either by itself. On many pentodes it can increase the gain over a resistor without increasing the distortion too much.
http://www.diyaudio.com/forums/tube...r-than-grounded-grid.html?highlight=12b4+line
Another trick that can be used is to put a CCS and a resistor in parallel. On some triodes this can be more linear than either by itself. On many pentodes it can increase the gain over a resistor without increasing the distortion too much.
http://www.diyaudio.com/forums/tube...r-than-grounded-grid.html?highlight=12b4+line
Forgot about that one. I will have to page through that one again. Looks like he had good sucess with it. 12B4 has about half the max plate dissipation of the 6EM7 and a bit higher rp than the 6EM7 but similar enough. In a quick perusal of the first couple of posts it appears that he was able to run at about half the current that I was calculating for the 6EM7. The graphs for the 6EM7s show that the plate resistance is substantially increased at lower current. I haven't seen rp graphs for the 12B4 so I don't know how close to optimum the 12B4 design is.
Gain variation on the small signal triode would not be too important for use as a subwoofer output buffer but might make it a bad choice for use in a phono pre.
I didn't mean to imply that a CCS would not improve linearity on virtually any triode but it was my understanding that pass tubes that Pete was using had significantly greater distortion than most power triodes and thus more of a need for the CCS.
I hadn't considered the ability to lower B+ (at least on SS CCS) and thus reduce the power dissipation. In any case 12W for the pair doesn't really bother me too much.
Gain variation on the small signal triode would not be too important for use as a subwoofer output buffer but might make it a bad choice for use in a phono pre.
I didn't mean to imply that a CCS would not improve linearity on virtually any triode but it was my understanding that pass tubes that Pete was using had significantly greater distortion than most power triodes and thus more of a need for the CCS.
I hadn't considered the ability to lower B+ (at least on SS CCS) and thus reduce the power dissipation. In any case 12W for the pair doesn't really bother me too much.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.