I saw the video last night about the Skunkie 300B and I would definitely go for it. The cascode input offers tuning/tweaking possiblity that both the AN and Sun Audio style driver don't have. In my opinion, the Sun Audio style circuit is generally a poor choice for the 300B unless you add a much higher voltage power supply for the driver alone.
The last video on that amp shows that increasing the bias for the driver tube the amp has low distortion., something like 6.5-7W at 1% THD. And it sounds best!
Moreover, earlier square wave test shows the slew rate is really good so there is no need to increase the current. It's a cascode and linearity can be improved without doubling the anode current, which you cannot do with a simple SE stage.
Moreover, earlier square wave test shows the slew rate is really good so there is no need to increase the current. It's a cascode and linearity can be improved without doubling the anode current, which you cannot do with a simple SE stage.
Last edited:
srpp is a pp stage period , hope you dont want argue with Broskie....
this is diyaudio we dont need audionote
Sorry for the offtopic from OT issues... 
Rather use some lower output generator resstance. Putting in the srpp "upper" tube some high transconductance tube. And set the input section "kower" tube in parallel operation that will significantly boost driving capabilities...
Equivalent output resistance of SRPP is about 1/S...
cheers.
Rather use some lower output generator resstance. Putting in the srpp "upper" tube some high transconductance tube. And set the input section "kower" tube in parallel operation that will significantly boost driving capabilities...
Equivalent output resistance of SRPP is about 1/S...
cheers.
Almost the same thig in the simulations occurred with more complex reactive model of the reactive load.
And should be count plus core and magnezitation issues... That for sure will not improve situation...
That is why so rear we have real measurements vith large signal + current trough the primary. More complicated is to set circuit for that than to use real tube
I don’t want to discuss with anyone also with you
I make normally a lot of test in the lab with many stuff and the Kit one is one of the best as 300B se
And the fact that it is AN is irrilevant due the fact that the diagram is well known so it is possible to build with good OT without difficulty
Bye
Walter
Complete nonsens to say that the backwards test is much better. The transformer is reversible as you mentioned before so what you say is not true. Secondly a test at lower power is far more interesting because at normal listening levels power is more likely lower then 1 Watt, specially with horn speakers.
For SE you need to have a bias current which you not have in your test.
Old traditional tests with tubes as a source is far better then your “backwards” methode. More realistic and reliable and more acepted in the real world.
3.5nf of input capacitance from transformer windings? The primary inductance will decrease as the tube bias increase, that is one reason the single ended produce that fat 2n harmonic bass = good on old paper drivers, not so good on modern cones.
Start with a perfectly linear output tube;
Next, use the tube to drive the primary of a perfect transformer;
Connect the secondary of the transformer to a closed box speaker, for example, with a resonance of perhaps 50 Hz, that also has a mid driver, and a tweeter; and passive crossovers at 2 frequencies.
Now, look at a possible load line that the output tube "sees".
At 20 Hz, it sees a low impedance resistive load
At 40 Hz or 45 Hz, it sees an inductive load (similar to that 16 Henry output transformer you were complaining about).
At 50 Hz, you see a very large impedance resistive load.
At 70 Hz, you see a capacitive impedance load (similar to the capacitance of the transformer you were complaining about.
It is not quite as simple as that. But you get the idea, those loudspeakers are "full of ellipses" at multiple frequencies.
Now, in spite of those "problems", put on a good recording, and "Enjoy the Music".
Next, use the tube to drive the primary of a perfect transformer;
Connect the secondary of the transformer to a closed box speaker, for example, with a resonance of perhaps 50 Hz, that also has a mid driver, and a tweeter; and passive crossovers at 2 frequencies.
Now, look at a possible load line that the output tube "sees".
At 20 Hz, it sees a low impedance resistive load
At 40 Hz or 45 Hz, it sees an inductive load (similar to that 16 Henry output transformer you were complaining about).
At 50 Hz, you see a very large impedance resistive load.
At 70 Hz, you see a capacitive impedance load (similar to the capacitance of the transformer you were complaining about.
It is not quite as simple as that. But you get the idea, those loudspeakers are "full of ellipses" at multiple frequencies.
Now, in spite of those "problems", put on a good recording, and "Enjoy the Music".
A quick question to the experts here. I obtained a pair of LL1623 SE OPTs secondhand, but they are gapped for 120mA, which I realise it not optimum for a single 300B. Would the sound be compromised if I use them for the single 300B? Or should I really plan to use them for a parallel pair or pass them on to someone who will?
They will fit your purpose with a big reserve vs DC saturation if primary Z is right for single 300B. May be your transformers have Z designed for 2 parallel 300B, you need to check.
Look up the LLL1623 specifications:
Inductance of the model that is gapped for 120mA (from the data sheet, I think it is 23 Henrys; but depends on how you wire the primary).
Make an estimate:
A 300B @ 60mA, plate impedance is 700 Ohms.
Suppose your output transformer primary is 23 Henry.
23 Henry is 2890 Ohms at 20Hz.
The 300B plate impedance, rp, is 700 Ohms
2890 / 700 = 4.1 : 1
I think that is good enough.
Others will not think that is good enough.
Try it and listen.
I never had any single ended Lundahl output transformers.
I never had any push pull Lundahl output transformers.
I might try using a push pull Lundahl output transformer.
I used Lundahl interstage transformers, but I no longer use interstage transformers.
Inductance of the model that is gapped for 120mA (from the data sheet, I think it is 23 Henrys; but depends on how you wire the primary).
Make an estimate:
A 300B @ 60mA, plate impedance is 700 Ohms.
Suppose your output transformer primary is 23 Henry.
23 Henry is 2890 Ohms at 20Hz.
The 300B plate impedance, rp, is 700 Ohms
2890 / 700 = 4.1 : 1
I think that is good enough.
Others will not think that is good enough.
Try it and listen.
I never had any single ended Lundahl output transformers.
I never had any push pull Lundahl output transformers.
I might try using a push pull Lundahl output transformer.
I used Lundahl interstage transformers, but I no longer use interstage transformers.
Definitely better for 300B PSE. Use 3K primary with 420V/60mA for each 300B. You should be able to get 17W. This way should result in best distortion figures at low frequency.
SE will be ok but not as good as the PSE at low frequency.
I'll try giving my 5 cents of opinion from a practical point of view in favor to the manufacturer. Once a product, in this case an OPT, lands into the hands of the customer, the later usually uses it the way he/she wants. Potentially different tubes, operating modes, difference in speaker loads, which all change the behavior of the transformer. So in a way, the manufacturer needs to find the most covering universal method. Which for most cases is low-level signal frequency response plots and/or square waves onto a resistive load. In practice, you'll never get the same results with a speaker load.
So I find it a bit meaningless to start arguments about best measurement practices published. However, I find it mandatory for any transformer manufacturer to study well his materials, in this case the magnetic properties of his cores, in order to predict the picture of the whole transformer behavior as detailed as possible.
I still find it worrisome that some transformer designers do not measure inductance using high voltage excitation, quoting a low-voltage signal primary inductance value in the zero cross low-permeability knee region. That becomes a recipe for too much inductance in the optimum working region and a sooner core saturation in SE operation due to practically less DC flux density headroom being available.
P.S. In 2016, on my working bench I studied the behavior of different cores, HiB, and nanocrystalline under CCS DC magnetization halfway through 1/2 of their maximum flux density. Although not taking detailed notes, I realized that under low voltage excitation, there was still a form of hysteresis, giving a lower value of inductance compared to higher voltage excitation, despite the presence of DC bias.
I did not compare the level of hysteresis of a biased and gapped core compared to a non-gapped, non DC biased core, but it was still quite significant to take a look at. Perhaps it would be wise to re-do the measurements and publish them.
Last edited:
Measuring the inductance with small signal, possibly a signal that results in 50 Gauss (i.e. 0.005 T) is a practical way to know the DC inductance (often called minimum inductance). Even in a SE transformer, despite the presence of the air-gap, the permeability will normally increase 20-30% with signal in most cases. Much less than PP type but still following the same trend.
Why is it important knowing the DC inductance? Because DC inductance that sets the power rating. Bac must be less than Bdc (at worst they are equal but distortion will rocket). Then Bdc+Bac also need to say within saturation, of course.
On the other end, max inductance with large signal should happen at max rated power so that the ratio between reactive component at a certain low frequency XL = 2*pi*f*L is much greater than Req (source resistance in parallel with transformer primary impendence) to ensure low distortion. In practice 8-10 times (or more) is only achievable with low plate resistance tubes (triodes or any device with local feedback).
This MUST be true for the optimal DC current by design. That's where the optimal choice for the air-gap enters in. If DC current is lower inductance will be basically the same and the transformer will work at lower induction with lower distortion and the power rating will depend on the smaller power tube.
If one transformer is claimed to be good with, say, 100 mA, but then inductance goes down for such DC current respect to 70mA then it's 70mA....in my book.
Why is it important knowing the DC inductance? Because DC inductance that sets the power rating. Bac must be less than Bdc (at worst they are equal but distortion will rocket). Then Bdc+Bac also need to say within saturation, of course.
On the other end, max inductance with large signal should happen at max rated power so that the ratio between reactive component at a certain low frequency XL = 2*pi*f*L is much greater than Req (source resistance in parallel with transformer primary impendence) to ensure low distortion. In practice 8-10 times (or more) is only achievable with low plate resistance tubes (triodes or any device with local feedback).
This MUST be true for the optimal DC current by design. That's where the optimal choice for the air-gap enters in. If DC current is lower inductance will be basically the same and the transformer will work at lower induction with lower distortion and the power rating will depend on the smaller power tube.
If one transformer is claimed to be good with, say, 100 mA, but then inductance goes down for such DC current respect to 70mA then it's 70mA....in my book.