I apologize for yet another question regarding output impedance.
Are there any disadvantages to using a greater than specified output impedance other than loss of power?
I have a pair of PP opts with primary impedance 9.8K at 8 ohms. I am going to use them in a PP EL34 triode strapped amp. The circuit also could accept KT88s with only a bias adjustment. 9.8K is high for both of these tubes, but the output, although low is still more than enough power for my needs, and the distortion at maximum output simulates really low (probably meaninglessly low...).
I could go to the trouble of getting other opts, and may eventually anyway, just to upgrade, but mostly i want to know if there are any other concerns with operating at this impedance or am I fine as long as i don't need the additional power.
Are there any disadvantages to using a greater than specified output impedance other than loss of power?
I have a pair of PP opts with primary impedance 9.8K at 8 ohms. I am going to use them in a PP EL34 triode strapped amp. The circuit also could accept KT88s with only a bias adjustment. 9.8K is high for both of these tubes, but the output, although low is still more than enough power for my needs, and the distortion at maximum output simulates really low (probably meaninglessly low...).
I could go to the trouble of getting other opts, and may eventually anyway, just to upgrade, but mostly i want to know if there are any other concerns with operating at this impedance or am I fine as long as i don't need the additional power.
Don't worry about it. If it sounds okay, enjoy it. Can't hurt anything. Just for fun, put an audio generator on and a 'scope and see what you do get.
For triode mode high impedance is better. It gives less distortion, better damping factor, but a little loss of power.
While your speaker is nominally 8 ohm, but could be 6 or lower at some points.
At 6 ohm load, the reflected impedance will be around 7K which might be close to theoretical optimal.
SveinB.
While your speaker is nominally 8 ohm, but could be 6 or lower at some points.
At 6 ohm load, the reflected impedance will be around 7K which might be close to theoretical optimal.
SveinB.
Hello,
My understanding is that there is a point of diminishing returns to increasing the transformer impedance. The load line gets flatter. There is reduced current swing to overcome the increasing interwinding capacitance in the transformer. As a result the bandwidth falls off and there is increasing high frequency distortion.
You are most likely ok. Do it.
DT
All just for Fun!
My understanding is that there is a point of diminishing returns to increasing the transformer impedance. The load line gets flatter. There is reduced current swing to overcome the increasing interwinding capacitance in the transformer. As a result the bandwidth falls off and there is increasing high frequency distortion.
You are most likely ok. Do it.
DT
All just for Fun!
Last edited:
In my rotation, I have a pair of Heath W-4AMs that I have converted over to triode EL34 operation decades ago. Among the many changes made, they still run cathode bias, the same basic circuit topology, and the original OPT, which on this model is 10K P-P -- almost identical to yours.
With 450 vdc B+, this configuration produces 12 watts RMS from 20 Hz to 20 kHz at vanishingly low distortion across this bandwidth at this power level. With 20 db of stable NFB, THD is typically <.5%, and operation is entirely class A over the full output range. I use them with Klipsch Cornwall IIs, and they are one of my favorite amps by far.
Dave
With 450 vdc B+, this configuration produces 12 watts RMS from 20 Hz to 20 kHz at vanishingly low distortion across this bandwidth at this power level. With 20 db of stable NFB, THD is typically <.5%, and operation is entirely class A over the full output range. I use them with Klipsch Cornwall IIs, and they are one of my favorite amps by far.
Dave
If you want triode operation you will want lower impedances. Although you are sacrificing power with this OPT, potentially that is being traded off for lower distortion. Something interesting might happen here though, the 8 ohm load of the speaker is transformed to 9.8K. If you load the 8 ohm tap with 4 ohms, you will *increase* the power of the amplifier without troubles with distortion!
With a small amount of feedback (which I am not a fan of) you might be able to create a tube amp that behaves like a true voltage source. I suspect it will come close without feedback.
With a small amount of feedback (which I am not a fan of) you might be able to create a tube amp that behaves like a true voltage source. I suspect it will come close without feedback.
One says higher for triode, the other says lower, but you seem to say the same thing after that.
I thought using a lower impedance speaker on a larger tap would simply lower the reflected impedance, which would increase power, but also increase distortion.
Not sure distortion matters that much, my design modeled in PP CAD has really low distortion at normal playing powers, still low at max power, but I will never play it there, it would hurt, not to mention freak out the baby....
Not sure power matters so much either. I don't listen to really loud music and plan on using 94-97db sensitivity speakers. Even with my 89db Adagios, 20W has been way more than enough, I probably need 10ish, max to adequately hit transients, maybe less, at my modest listening levels.
So I guess I can go with less distortion even there are not very meaningful differences.
I am not sure what you mean by adding a little negative feedback to approximate a voltage source. Can you educate me more? Thanks?
I am also interested in thoughts on feedback. Definitely have two camps of good vs not good and I really don't know.
My original intention was to build an amp without (intentional GNF), just for the purity of it, but once I got it together I will need some just to reduce the input stage gain to appropriate levels (if the tubecad models are close), or I could use a different input tube with lower gain and have less headroom, but no GNF.
My current plan is to try swapping some tubes with adjustable and turn-off-able GNF.
The input stage is a direct input LTP with CCS tail.
I need a minimum gain (not mu) of 14 to drive my output tubes to clipping (need 40v pk, input 2.8v pk). By my tubecad models I think I can do this with a tube of minimum mu 20 in my topology. i also understand that It is better to have a low gm tube, rather than high.
Current tube choice is 12AT7. mu 60, gm 4. Tubecad gain 36ish so will probably do better with some GNF to allow me to turn my volume knob past 1...
Other options would be 6cg7 or 6sn7, mu 20, gm2.6 (although 6sn7 needs different bases, these are tubes I have...). These would get nominal gains of 14.5 - 15 per tubecad. I hope I get more since tubecad models a tube CCS and I will use sand. These would have little to no room for GNF.
Any other tube suggestions? I can buy other tubes, but would rather not add a driver tube. I suppose I could use a pair of single triode tubes, I am not so far in the process I can't throw in another couple of sockets.
I realize I have just changed the entire gist of my thread.
Thanks for any thoughts and maybe I will start another thread soon regarding direct build questions of which i should have many...😱Have you verified this information? It is quite at odds with published and actual measured performance for these tubes in triode mode. Mullard published data showing the lowest distortion operating point for these tubes in triode mode was achieved with a 10K plate to plate load. If the load is cut in half to 5K, power output in increased by only 5 watts, but distortion is increased over 6 times. In addition, loading a transformer as you suggest will restrict it's power bandwidth, and increase it's insertion losses as well.
Dave
Triodes are generally lower impedance. But the idea is speculation: by starting out with an amplifier wherein the tubes have a load that is too high (and so don't make full power) you then decrease the load on the secondary to get more power => voltage source.
- Status
- Not open for further replies.