Working on another 7C5 amplifier again, and this time am looking to go UL PP output for about 12-14W.
Reading up on UL outputs, I come across several articles from the originators of the Ultralinear moniker, Hafler and Acro Sound, stating that 6V6 tubes should have 24% UL taps. This is correlated by other sources claiming best performance between 21% and 26%, so I start looking for 25% UL tap transformers with a 7 to 9K primary, and there are not many. That's OK, I've done this before successfully, with 50% UL taps, so it should be simple enough to wind a 8KCT primary with 25% taps, right? Well, I'm told that there is more to it than that, but can't find much detail on it.
Here is what I propose, guided by Yves M.'s excellent OPT_DA program. I have not used this program to do a UL transformer, so I will lay out my inputs/outputs exactly as I've put the data into the program so anyone else can have a look.
Specifications Block:
50hz, 15W power.
Rp (Ohms) = 80,000 for 6V6, SE box NOT checked.
Primary Z 8000, Ip0 = 0.032mA, for 346Vrms
2 sections in series by 1 in parallel
Secondary Z 6, 9.5Vrms
1 section in series by 2 in parallel
Wires per section 2 in parallel
Iron Block:
EI66B core stack, 0mm gap/interleaved core.
AFe 7.7cm^2
mFe 0.74kg
MPL 13.2cm
B DC 0T
B AC 1.32T
mu 9515
Fe Losses 0.8W
Copper Block:
4.42 Turns/volt
2A/mm^2
Primary L = 129.9H
0.22mm primary wire diameter, 1530 turns total (32 gauge enameled)
0.71mm secondary wire diameter, 41 turns total (22 gauge enameled)
This gives a recommended winding format of:
Primary: 382 turns/4 layers, Secondary 41 turns/2 layers, Primary 382 turns/4 layers, CT, Primary 382 turns/4 layers, Secondary 41 turns/2 layers, Primary 382 turns/4 layers.
If I remember properly, The center section of the primary winding should be split at the CT, and the inner most section paired with the outer half of the CT winding, and the outermost primary section paired with the inner half of the CT winding. This is to level out the DC resistances due to the change in turn length around the core.
This also means that I should have my 25% UL taps in the CT block, located at 191 turns either side of the CT. That's the location that is 25% of the turns of the winding.
I've also seen UL taps called out by %of the impedance, but the articles I've found call for 25% of the turns.
Is there anything glaringly wrong here?
Reading up on UL outputs, I come across several articles from the originators of the Ultralinear moniker, Hafler and Acro Sound, stating that 6V6 tubes should have 24% UL taps. This is correlated by other sources claiming best performance between 21% and 26%, so I start looking for 25% UL tap transformers with a 7 to 9K primary, and there are not many. That's OK, I've done this before successfully, with 50% UL taps, so it should be simple enough to wind a 8KCT primary with 25% taps, right? Well, I'm told that there is more to it than that, but can't find much detail on it.
Here is what I propose, guided by Yves M.'s excellent OPT_DA program. I have not used this program to do a UL transformer, so I will lay out my inputs/outputs exactly as I've put the data into the program so anyone else can have a look.
Specifications Block:
50hz, 15W power.
Rp (Ohms) = 80,000 for 6V6, SE box NOT checked.
Primary Z 8000, Ip0 = 0.032mA, for 346Vrms
2 sections in series by 1 in parallel
Secondary Z 6, 9.5Vrms
1 section in series by 2 in parallel
Wires per section 2 in parallel
Iron Block:
EI66B core stack, 0mm gap/interleaved core.
AFe 7.7cm^2
mFe 0.74kg
MPL 13.2cm
B DC 0T
B AC 1.32T
mu 9515
Fe Losses 0.8W
Copper Block:
4.42 Turns/volt
2A/mm^2
Primary L = 129.9H
0.22mm primary wire diameter, 1530 turns total (32 gauge enameled)
0.71mm secondary wire diameter, 41 turns total (22 gauge enameled)
This gives a recommended winding format of:
Primary: 382 turns/4 layers, Secondary 41 turns/2 layers, Primary 382 turns/4 layers, CT, Primary 382 turns/4 layers, Secondary 41 turns/2 layers, Primary 382 turns/4 layers.
If I remember properly, The center section of the primary winding should be split at the CT, and the inner most section paired with the outer half of the CT winding, and the outermost primary section paired with the inner half of the CT winding. This is to level out the DC resistances due to the change in turn length around the core.
This also means that I should have my 25% UL taps in the CT block, located at 191 turns either side of the CT. That's the location that is 25% of the turns of the winding.
I've also seen UL taps called out by %of the impedance, but the articles I've found call for 25% of the turns.
Is there anything glaringly wrong here?
I suspect you are trying to describe cross-connection of windings. To break it down for you, if we label the primary layers as the following:
Primary D
----------
Secondary 2
----------
Primary C
CT
Primary B
----------
Secondary 1
----------
Primary A
You will achieve equal resistances by combining primaries B and D and primaries A and C.
If you are looking for perfection into leakage inductance and overall capacitance however, you need to consider symmetrical winding. Either split chamber single bobbin or double bobbin transformer.
Connecting the windings to level the primary resistance is the intended message, yes.
Any thoughts on the UL taps? That's what I am told is "not just a tap" and requires some kind of special magic.
Any thoughts on the UL taps? That's what I am told is "not just a tap" and requires some kind of special magic.
Fail to find any magic in scratching enamel and soldering a wire there.
Some people want Mojo in every aspect of their life 😉
Or CLAIM it for marketing.
A tap is a tap is a tap.
Quite uncritical, since screens take substantially less current than plates.
As of the tap %, you just calculate it, keep an eye on the turns counter, stop and add tap at the proper moment, you decide it, no magic.
Some people want Mojo in every aspect of their life 😉
Or CLAIM it for marketing.
A tap is a tap is a tap.
Quite uncritical, since screens take substantially less current than plates.
As of the tap %, you just calculate it, keep an eye on the turns counter, stop and add tap at the proper moment, you decide it, no magic.
Have you seen the UL curves? It might help you to determine the optimal loadline
https://www.diyaudio.com/community/threads/6v6-line-preamp.102352/post-7045150
In this post I describe how you can record the output of the design and playback. Obviously you can hear the difference between Pentode (no UL) and triode (100% UL). and anything in between. This does not mean final sound, but is a good start, as it depends a lot on OT in addition, you can compare with and without OT. It's fessible to build the UL using active feedback first rather the actual OT UL tap.
https://www.diyaudio.com/community/threads/6v6-line-preamp.102352/post-7045150
In this post I describe how you can record the output of the design and playback. Obviously you can hear the difference between Pentode (no UL) and triode (100% UL). and anything in between. This does not mean final sound, but is a good start, as it depends a lot on OT in addition, you can compare with and without OT. It's fessible to build the UL using active feedback first rather the actual OT UL tap.
Attachments
Last edited:
UL is usually done with taps. It's up to you to solder it within the layers, or take out enamel wires out of the coil and solder them externally. I do the later.
Some insist UL should be in separate layers, but after some study, I find it unnecessary. The g2 impedance is high enough at normal tube operation, so leakage inductance from primary to UL sections is not an issue. On top of that, sectioning UL layers will add additional UL to secondary or primary capacitance and you must do a very careful analysis of the capacitance vs leakage distribution before continuing further.