There used to be a choke made by Tango, the TC-160-15W (http://www.tubebooks.org/file_downloads/tango_tamura/Tango_TC-160-15W.pdf), which had a winding inductance of 160H at 15mA, with a center-tap. Some folks liked to use it as a push-pull plate choke, as the plate load for a long-tailed pair phase splitter/driver stage. The idea was to be able to DC-couple the LTP plates to the output stage grids. The plate choke doesn't drop much voltage, so the plate voltage to the LTP stage is very close to the cathode voltage of the output stage. A stacked power supply can then be used with a cathode bias resistor to develop the bias for the outputs.
But I never had nearly enough money to buy one of those back when they were available.
Instead of the pricey Tango choke, I used a mystery-model push-pull OPT that was designed for PP EL84's, in the plates of an LTP using 5687 (later changed to a 6N6P). This stage drives a pair of 2A3. The amp works very well, sounds nice. If I use a 12BH7A instead, I notice that the bass response is lost. Obviously the plate resistance of the 12BH7A is too high for the primary inductance of my mystery OPT-as-plate-choke. But the bass extension measures well when used with a 5687, so the primary inductance has to be 30H or so. I'd like to improve on that.
So, I have a question. If I wanted to use an OPT as a poor man's center-tapped plate choke, what specifications would I look for? Highest primary impedance possible? All other things (like bandwidth at a certain signal level) being equal, does a 16k:4 ohm OPT have higher primary inductance than a 10k:4 ohm OPT? I suspect the answer will be complicated, but I'm hoping it's a simple 'higher Z primary = higher L primary for the same secondary (and load) Z at the same signal levels.'
Edcor makes some interesting OPTs. One is 10k ohm primary to 4 ohm secondary (https://www.edcorusa.com/gxpp10-4-10k). They don't give a primary inductance spec, but they do say the transformer passes 40Hz to 18kHz <1dBu at its rated output of 10W. I'd be using it at much lower levels, with about +/-50V signal on the primary, so the bandwidth should be wider, correct? The step-down ratio is probably 50:1, so that means 1V on the secondary, or the equivalent of about 250mW at most.
This guy (https://www.edcorusa.com/cxpp25-6-12k) has a 12k primary, is bigger (25W) and has wider bandwidth (20Hz - 20kHz <1dBu). Would that equal higher primary inductance?
I wonder if Edcor would make a pair of OPTs without the secondary and without the UL screen taps on the primary.
I've read that the bandwidth-widening tricks like interleaving or bifilar winding for OPTs aren't necessary for a center-tapped plate choke for PP driver (see: http://www.diyaudio.com/forums/tubes-valves/104191-pp-plate-choke.html). True?
I hope these aren't dumb questions. Thanks.
- RG
But I never had nearly enough money to buy one of those back when they were available.
Instead of the pricey Tango choke, I used a mystery-model push-pull OPT that was designed for PP EL84's, in the plates of an LTP using 5687 (later changed to a 6N6P). This stage drives a pair of 2A3. The amp works very well, sounds nice. If I use a 12BH7A instead, I notice that the bass response is lost. Obviously the plate resistance of the 12BH7A is too high for the primary inductance of my mystery OPT-as-plate-choke. But the bass extension measures well when used with a 5687, so the primary inductance has to be 30H or so. I'd like to improve on that.
So, I have a question. If I wanted to use an OPT as a poor man's center-tapped plate choke, what specifications would I look for? Highest primary impedance possible? All other things (like bandwidth at a certain signal level) being equal, does a 16k:4 ohm OPT have higher primary inductance than a 10k:4 ohm OPT? I suspect the answer will be complicated, but I'm hoping it's a simple 'higher Z primary = higher L primary for the same secondary (and load) Z at the same signal levels.'
Edcor makes some interesting OPTs. One is 10k ohm primary to 4 ohm secondary (https://www.edcorusa.com/gxpp10-4-10k). They don't give a primary inductance spec, but they do say the transformer passes 40Hz to 18kHz <1dBu at its rated output of 10W. I'd be using it at much lower levels, with about +/-50V signal on the primary, so the bandwidth should be wider, correct? The step-down ratio is probably 50:1, so that means 1V on the secondary, or the equivalent of about 250mW at most.
This guy (https://www.edcorusa.com/cxpp25-6-12k) has a 12k primary, is bigger (25W) and has wider bandwidth (20Hz - 20kHz <1dBu). Would that equal higher primary inductance?
I wonder if Edcor would make a pair of OPTs without the secondary and without the UL screen taps on the primary.
I've read that the bandwidth-widening tricks like interleaving or bifilar winding for OPTs aren't necessary for a center-tapped plate choke for PP driver (see: http://www.diyaudio.com/forums/tubes-valves/104191-pp-plate-choke.html). True?
I hope these aren't dumb questions. Thanks.
- RG
Attachments
Edcor was very helpful to me. You can e-mail or call them.
The impedance is just a rating based on the expected load resistance.
My guess is generally the highest impedance with the lowest power rating would be the highest inductance.
Just a guess but pick one to handle more than the needed power at the lowest frequency.
Low frequency response of power bandwidth goes as so, according to Vanderveen:
(2) (sqrt 2)
2.5W 20Hz
5W 28.28Hz
10W 40Hz
20W 56.56Hz
40W 80Hz
The impedance is just a rating based on the expected load resistance.
My guess is generally the highest impedance with the lowest power rating would be the highest inductance.
Just a guess but pick one to handle more than the needed power at the lowest frequency.
Low frequency response of power bandwidth goes as so, according to Vanderveen:
(2) (sqrt 2)
2.5W 20Hz
5W 28.28Hz
10W 40Hz
20W 56.56Hz
40W 80Hz
There's also the Lundahl LL1667:
http://www.lundahl.se/wp-content/uploads/2013/05/1667_68.pdf
Not cheap, but probably less than the Tango used to be...
Pete
http://www.lundahl.se/wp-content/uploads/2013/05/1667_68.pdf
Not cheap, but probably less than the Tango used to be...
Pete
Thanks, that does look good. $230 for a pair is pretty steep, but not that bad...
It says that for the LL1667 the max standing current for full bandwidth is 15mA. That's comparable to the Tango TC-160-15W. Does that mean a max of 15mA out-of-balance current is allowed, or that the plate current for the two triodes used would have to be no more than 7.5mA each?
--
That would be max unbalanced current in P-P.
I'm using one for SE. The datasheet doesn't show it, but there's also a 5mA version (I'm using that with ~3-4mA SE). Gapped for 5mA, it has some absurdly high inductance (700H!) and very wide bandwidth.
Pete
There's also a "PP" version of the LL1667, which has a higher inductance yet, but will require better balance (about 2.5mA) even than the 5mA gapped version. I normally stock the PP, 5mA, and 15mA versions, but don't have the 5mA at the moment.
Thanks, I didn't know about that.
I wonder... It seems to me that using a tube with higher rp and mu, like 12AT7 for example, you'll have less plate current, maybe about 2 or 3mA per triode. At that kind of plate current drawn through the CT choke, how much imbalance can there possibly be?
The choke would need higher inductance to get the same f3 at low frequency, but the upside would be the higher gain possible from the higher mu/higher rp tube.
So, how much standing current from imbalance can something like a 12AT7 possibly cause in the coils of the CT plate choke?
--
Last edited:
If each triode is only drawing 3mA, then the max offset is one tube conducting 6mA and the other 0mA. That would only happen if there was a wiring mistake or one triode was dead, so it's not really a concern. I would imagine that dual triodes with a mismatch of 2mA in one and 4mA in the other under identical bias conditions might be a worse case.
I have a question for PP:
Why go for the LL1667/1668...when you can get 210H with the 1692 or 290H with the 1660 ?
I mean for mainly the same price you get for free an IT ??? Or did I not understand something right ?
Why go for the LL1667/1668...when you can get 210H with the 1692 or 290H with the 1660 ?
I mean for mainly the same price you get for free an IT ??? Or did I not understand something right ?
The LL1667/LL1668 is lower cost than the LL1660 or the LL1692, but as you point out, not by much. Otherwise, the primary windings can be used as a plate choke.
- Status
- Not open for further replies.