Good Question!
For your particular circut, I suggest you wire it one way, then test the square wave response;
Then wire it the other way and then re-test the square wave response.
Check the ringing.
Check the rise time, and the fall time.
Compare the differences (there will be some).
The reason to do both tests, is that the interstage transformer was Not designed to be used as a choke.
What you get is just a guess until you measure it.
Then please tell us which way had the better square wave response.
A long time ago, I stopped using interstage transformers.
I had used single ended and push pull from 2 manufacturers. I also used one of those manufacturer's 3 wire auto transformer phase splitter.
I considered the tradeoffs between interstage and RC coupling. Now all my amplifiers use RC coupling.
Just my choice, but not every ones choice.
For your particular circut, I suggest you wire it one way, then test the square wave response;
Then wire it the other way and then re-test the square wave response.
Check the ringing.
Check the rise time, and the fall time.
Compare the differences (there will be some).
The reason to do both tests, is that the interstage transformer was Not designed to be used as a choke.
What you get is just a guess until you measure it.
Then please tell us which way had the better square wave response.
A long time ago, I stopped using interstage transformers.
I had used single ended and push pull from 2 manufacturers. I also used one of those manufacturer's 3 wire auto transformer phase splitter.
I considered the tradeoffs between interstage and RC coupling. Now all my amplifiers use RC coupling.
Just my choice, but not every ones choice.
Interleaving reduces leakage inductance, which is irrelevant in a choke. However, it increases capacitance between primary and secondary, which is quite relevant. So, primary sections in series, secondary sections in series, and then primary in series with secondary.
I have used Hammond power supply chokes as signal current sinks and current sources.
5 Henry 200mA, for two 6C45pi triodes, cathode coupled phase splitter (the 5H choke is the "LTP" current sink, no negative supply needed, the choke automatically swings below ground [if necessary]).
5 Henry 200mA, for two 6CK4 triodes, cathode coupled self inverting push pull output stage (the 5 Henry choke is the "LTP" current sink, no negative supply needed, the choke automatically swings below ground [if necessary]).
5 Henry 200mA, for either DHT triode, or triode wired Beam Power tube output stage plate load, in a Parafeed amplifier.
Both the Hammond 20 Henry 100mA choke and 5 Henry 200mA chokes could be used as a constant current sink, or constant current source.
They were designed to work as filter chokes in 50Hz or 60Hz full wave rectified B+ filters (100Hz and 120Hz respectively).
But I consider the 5 Henry 200mA choke to be better than the 20 Henry 100mA choke . . . Less likely to saturate at 100Hz, 50Hz, or 20Hz.
In my circuit applications, the 5 Henry choke does not have significant distributed capacitance, but the 20 Henry choke might have more distributed capacitance; that would affect the high frequency limit of effective CCS.
The other tradeoff is the inductance at low frequencies.
Two 5 Henrys in series would be big, heavy, and expensive, but probably the best combination for some CCS applications.
5 Henry 200mA, for two 6C45pi triodes, cathode coupled phase splitter (the 5H choke is the "LTP" current sink, no negative supply needed, the choke automatically swings below ground [if necessary]).
5 Henry 200mA, for two 6CK4 triodes, cathode coupled self inverting push pull output stage (the 5 Henry choke is the "LTP" current sink, no negative supply needed, the choke automatically swings below ground [if necessary]).
5 Henry 200mA, for either DHT triode, or triode wired Beam Power tube output stage plate load, in a Parafeed amplifier.
Both the Hammond 20 Henry 100mA choke and 5 Henry 200mA chokes could be used as a constant current sink, or constant current source.
They were designed to work as filter chokes in 50Hz or 60Hz full wave rectified B+ filters (100Hz and 120Hz respectively).
But I consider the 5 Henry 200mA choke to be better than the 20 Henry 100mA choke . . . Less likely to saturate at 100Hz, 50Hz, or 20Hz.
In my circuit applications, the 5 Henry choke does not have significant distributed capacitance, but the 20 Henry choke might have more distributed capacitance; that would affect the high frequency limit of effective CCS.
The other tradeoff is the inductance at low frequencies.
Two 5 Henrys in series would be big, heavy, and expensive, but probably the best combination for some CCS applications.
Last edited:
One does not simply connect secondary sections in series in a signal transformer.
You need some empirical knowledge of how distributed capacitance across the P/S layers will manifest itself. Secondary windings connected in series and physically far away exhibit some dip resonances which should be taken care into account, or the project should be designed to take them into account.
Of course, mine was a very general advise. Without knowing how winding sections are arranged, it is hard to predict which connection will result in lowest winding capacitance. 6A3sUMMER gave a very good advice - look at HF response of different connections.
One option is: sell those LL1671s and buy proper chokes or a better IT transformers. Never had good luck with Lundahls with any connection which is not in the datasheet.
People always make the same mistake of trying to adapt an IT transformer to some circuit or tube. It's the other way around if want to get good results.
So the cheaper and, likely, better option is forget the D3a and find a better driver to work with the LL1671 if selling them is not an option. There is plenty of choice at affordable price....
Something like a type 33 in triode connection which is just as linear as the D3a, if not better! At 180V (plate and g2) voltage and 25-30 mA max you will have the 3K plate resistance you need to make the LL1671 work best in ALT S connection. The only downside is the gain but I guess you have much more than 180V so you can do add a preamp stage DC coupled to the 33 and get the gain you want. DC coupling these 2 stages can be tuned to act like one in practice...