van Dorn's Octode design - recommendations for a 10W/ch version

[BoilermakerFan]

I'm captivated by the Octode design and van Dorn's tube choices for the standard version are some of my favorites as well. I happen to have E-P 10W OPTs that are 8200R:8R so they are perfect for an Octode.

I wish to keep the 6922 as the input tube since I have some nice 6N23Pi and I love the sound of the 6V6 family of tubes so I'm considering the 6V6 in triode mode for that portion of the amp, but I'm not sure what to use for the pentode tube to keep the power down around 7W-9W for the pentode since the 6V6 will deliver about 2W in triode. Would the 6W6 be a good choice or would the EL84/6BQ5 be a better choice?

Thanks!

[kevinkr]

How about a schematic and/or a link to more information on the Octode. This is the first time I have even heard of it, and have little time at the moment for searching..

[smoking-amp]

Octode Topology

Seems to be a standard buffer circuit, but using the LTP triode's plate for the feedback instead of it's grid, so the reverse Mu acts as the feedback attenuator. Net effect, a buffer with gain of Mu.

[BoilermakerFan]

The article is in the Nov. 2009 audioXpress. It ties a triode to a pentode for the output tubes. So the power of the two is added together. I believe the triode/pentode portion was the new portion of the circuitry. John's article in Tubecad isn't quite the same as the amp, as he was modifying the circuit a bit and wouldn't fully disclose van Dorn's design.

The article isn't on the aX site, so if you really want to read it Kevin to help me out, I can scan it in and email it to you.

I tried to find contact info for A.J. van Doorn, but only came up with a lady that appears to be in the medical field. So either JB is assuming A.J. is a man, or there are two A.J van Doorns.

[BoilermakerFan]

Since I'm a noobie, especially WRT understanding design variations, what would JB's question below do to the sound and/or performance of the Octode?

"I should stop right now, as I want to encourage you to read Mr. van Doorn’s article first. I will, however, leave with one small thought experiment: What if we attached the second triode’s grid in the Octode amplifier not to ground but to the pentode’s unbypassed cathode resistor?" - JRB

[relder]

Originally Posted by BoilermakerFan

The article is in the Nov. 2009 audioXpress. It ties a triode to a pentode for the output tubes. So the power of the two is added together.

Unfortunately AX goofed and the major schematic is found as errata in the Jan 2010 edition.

Originally Posted by BoilermakerFan

I tried to find contact info for A.J. van Doorn, but only came up with a lady that appears to be in the medical field. So either JB is assuming A.J. is a man, or there are two A.J van Doorns.

Adriaan is the first name, which I found out when I contacted Pieter at Tribute to see how much transformers would cost.

The pentode action seems quite similar in concept to Nelson's Aleph current source since it's trying to present a negative impedance to the power triode, making it's life easier.

I'm definitely going to build one at some point. Need to finish other projects 1st.

I was surprised no one mentioned this circuit before, I find it quite clever.

[BoilermakerFan]

Originally Posted by relder

Unfortunately AX goofed and the major schematic is found as errata in the Jan 2010 edition.

Adriaan is the first name, which I found out when I contacted Pieter at Tribute to see how much transformers would cost.

The pentode action seems quite similar in concept to Nelson's Aleph current source since it's trying to present a negative impedance to the power triode, making it's life easier.

I'm definitely going to build one at some point. Need to finish other projects 1st.

I was surprised no one mentioned this circuit before, I find it quite clever.

Yes, I have the schematic from the Jan '10 issue too.

I do plan to eventually build the amp as designed, but for now, I have the OPTs and 2/3s of the tubes and I could even use 6V6s for the pentodes if I knew what values to tweak to bring the NFB up to reduce their output to 7W-8W. In that case, I'd only need PS iron.

[DF96]

Quite a clever circuit, but I doubt if it will sound like a triode because the cathode-coupled amp is not a triode. The triode curvature of the first triode is largely cancelled by the second triode, which will increase the proportion of odd-order products.

Putting the grid of the second triode to an unbypassed pentode cathode resistor introduces an extra feedback path which will raise the output impedance - the opposite of what the first triode connection will do, and generally the opposite of what people want.

[relder]

Originally Posted by DF96

Quite a clever circuit, but I doubt if it will sound like a triode because the cathode-coupled amp is not a triode. The triode curvature of the first triode is largely cancelled by the second triode, which will increase the proportion of odd-order products.

Putting the grid of the second triode to an unbypassed pentode cathode resistor introduces an extra feedback path which will raise the output impedance - the opposite of what the first triode connection will do, and generally the opposite of what people want.

He claims the opposite:

The main advantage of this system is that without any negative feedback there isn't any crossover distortion but a very high damping of the speaker. Because any current variation of the triode is amplified by the pentode, the plate resistance is very low as well. Say the plate resistance of the triode is 900ohm and the total current variation is three times the current of the triode, then the plate resistance will only be a mere 300ohm.

In the octode mode the two triodes are assisted by a pair of pentodes, not only to increase the power output, but also to lower the output impedance. Say the OPT has a primary impedance of 3200ohm and the primary AC current is 100mA; then the primary AC voltage is 320V, giving a power output of 32W. With 10W output the triode stays in class A mode with a plate current of 32mA AC, so the pentodes must deliver 100-32 = 68mA AC. This means that the plate load seen by the triodes is 10kOhm, and 4700ohm for the pentodes.

[Wavebourn]

Now, let's expect the guy who enlighten us about how to create crossover distortions by the presence of feedback, and how presence of feedback means "without any feedback".

Show must go on!