43% UL connections

[Hojvaelde]

Hey

Now I have studied some schematics showing SE amps in UL coupling.

I do not understand why the UL connection transformer is almost always 43%.
I have some transformers with 25% UL connections.
What has to change at the schematic to use 25% instead of 43% and what is the theory behind it.

Regards

Benny

[Fuling]

IIRC some early UL output transformers used 7 primary sections and the UL tap was taken between sections 3 and 4.

[Tubes4e4]

Hi Benny

Now I have studied some schematics showing SE amps in UL coupling.
I do not understand why the UL connection transformer is almost always 43%.
I have some transformers with 25% UL connections.

Don´t get fooled by seemingly "magic" numbers. For instance, if an interleaved transformer is wound, those 43% just come out easily by leading out a tap at the end of the 3rd out of 7 primary winding packets. For 25%, the tap is just at the end of the 1st out of 4 primary winding packets - always counting from the B+ connection to the anode connection.

What has to change at the schematic to use 25% instead of 43% and what is the theory behind it.

I´ll start the other way round. For a pentode or BPT in UL connection, upping the winding percentage for the UL tap means (everything else unchanged) more triode like behaviour (100% UL means just a triode-strapped pentode or BPT),
--> less sensitivity,
--> less Po,
--> less THD,
--> less Zo (hence higher DF)
--> higher input capacitance

So, in your case going from 43% to 25% essentially means:

--> power stage is easier to drive due to higher sensitivity and less CMiller,
--> higher Po and THD
--> higher Zout, hence lower DF

Depending on the actual topology, you can derive actual measures to be taken - according to your preferences and goals) to deal with those changes listed above.

If it is a 3-staged design using considerable ammounts of global loop FB, quite likely you will have to fiddle with the time constants in the amplifier path as well as the feedback path to maintain unconditional stability at complex loads like realworld speakers.

Regards,

Tom Schlangen

[SY]

Think of it this way: triode would be 100%, pentode would be 0%. UL is somewhere between. (Oversimplification alert) The closer to triode, the lower the power, distortion, and output impedance. The closer to pentode, the higher the power, but with higher distortion and output impedance. So a 25% tapping will give more power, but a 43% tapping will give lower distortion.

It's a bit more complicated than this, but this is close.

[richwalters]

25% tap runs the screen grid closer to tetrode configuration i.e less AC anode to screen feeedback, so a higher output power is obtainable at higher thd. The 43% tap is closer to running the screen grid to the anode i.e triode performances so less distortion is produced for a lower power. The compromise performance is intermodulation thd being highest in true tetrode configuration 20%, and optimum min at 43%. The output impedance changes with the amount of UL (screen grid/anode)negative feedback.

It is entirely the users choice. I tried 20% taps with ECL82 and found thd too high, so I settled for 40%.
Want reckless power i.e don't care with thd then either choose true pentode or 20% UL. Most tube data sheets do specify impedances for proper output tranny impedance matching. The Radiotron hbk goes into more detail.
Study GEC KT88 data sheets at http:/frank.pocnet.net/

The only change at the schematic is addition of screen grid resistor and perhaps global nfb value.

The article you need to read is the GEC application report No3 GEC amplifier circuits for KT88. dated June 1972. Unfortunately I cannot dive further into this topic as I believe is now copyright protected.

richy

[thevoice]

There has been some confusion over the years between % of turns and % of impedance. They don't match but both ways of talking about UL connections have been used.

I think, from talking to a transformer winder, that 20% turns equals 43% impedance, but I'm a bit hazy on exactly how that went and I'm hungover, so don't quote me.

Have a read of this, for an in depth analysis of why various UL tappings are chosen.

[Tubes4e4]

Hi TheVoice,

talking about impedance doesn´t make much sense in this case, because we are talking about screen feedback (=UL) ratios, which is defined by the voltage swing ratio (=wire turns ratio) between B+ tap, anode tap and screen tap:

UL% = ((number of turns between B+ tap and screen tap) / (number of turns between B+ tap and anode tap)) * 100

or,

UL% = (Vpp,screen / Vpp,anode) * 100

Regards,

Tom

[Eli Duttman]

Benny,

The other posters have summed things up well. A smaller UL % means more pentode like behavior.

The H/K Cit. 2 uses 20% UL taps on the O/P trafos. By way of compensation, the "Duece" employs 3 nested NFB loops. The amount of NFB in the outer, global, loop is surprisingly small. The design is (sic) unconditionally stable. I suggest you follow a similar path, with only a few dB. of NFB in an outer, global, loop and some sort of inner, partial, NFB setup.

[tomchr]

Dav,

I recall seeing a plot of distortion vs UL output tap percentage. I think it was in:

Vacuum Tube Amplifiers; Valley & Wallman; McGraw-Hill (MIT Radiation Labs. Series); 1948

but it could have been in RDH4... I'll see if I can dig it up. Basically, with a screen tap with 40-50 % the number of turns of the anode tap, you get the best power/distortion trade-off.

~Tom

[EC8010]

Here's a plot for KT66 taken from page 52 of "High Fidelity Sound Reproduction," E Molloy (ed.), Newnes, 1958. Beware that although it shows minimum distortion at about 20%, that just means it gave a minimum reading on an (unspecified) total harmonic distortion meter and says nothing about the relative proportions of 2nd versus 3rd harmonic. And don't forget that very few loudspeakers present a resistive load, so many of these "optimum" configurations (which usually rely on cancellation) are pretty tenuous anyway. Realistically, the generalities offered by SY earlier are as good as it gets.