Apologies but these are almost certainly stupid questions displaying my complete ignorance.
I have a couple of SE output transformers with 5K and 7K inputs and 4, 8, and 16ohm outputs.
Can I use the 5K tap as an ultra-linear feed to the screen of the output pentode to give a roughly 30%/70% input. I assume this would mean connecting B+ to the 7K tap, the screen to the 5K tap and the anode to the B+ connection.
If I connect 8 ohm speakers to the 16ohm output does this then mean the tube sees a 3K5 impedance?
The OTs are 12VA with a maximum bias current of 60mA
Thanks (ducking )
I have a couple of SE output transformers with 5K and 7K inputs and 4, 8, and 16ohm outputs.
Can I use the 5K tap as an ultra-linear feed to the screen of the output pentode to give a roughly 30%/70% input. I assume this would mean connecting B+ to the 7K tap, the screen to the 5K tap and the anode to the B+ connection.
If I connect 8 ohm speakers to the 16ohm output does this then mean the tube sees a 3K5 impedance?
The OTs are 12VA with a maximum bias current of 60mA
Thanks (ducking )
Sure, give it a shot.
If the transformer is perfect, yes it does. If the transformer is a budget offering from China, you may get some weird high frequency response.
The UL ratio is the ratio of the transformer turn ratio, not primary impedance. If the transformer is 7K at 16 ohms load, the turn ration is square root (7K/16) or about 21 to 1. The 5K tap is about 17.7 to 1. UL is roughly 16%/84%.
The OPT has primary inductance which limits the low frequency and a mutual inductance which limits the high frequency. The frequency response of a OPT is given for that précis indicated load impedance. If you load more than indicated than the high frequency will be lower in same ratio but the low frequency will be lower also by the same ratio.
If you load the transformer output tap with 1/2 the rated tap impedance (8 Ohms on the 16 Ohm tap), you will have 2X the insertion loss at mid frequencies.
At low frequencies 1.; and at high frequencies 2., calculating the insertion loss is more complex:
It depends on the inductance 1.(low frequencies); the leakage inductance, and the distributed capacitance 2.(high frequencies), respectively.
The results with a non-inductive load resistor is one thing.
The results with most loudspeakers are another thing.
Over the audio frequency range, loudspeakers have varying capacitive reactance; varying inductive reactance; and varying resistance versus frequency (a woofer and a tweeter often have different resistance).
At low frequencies 1.; and at high frequencies 2., calculating the insertion loss is more complex:
It depends on the inductance 1.(low frequencies); the leakage inductance, and the distributed capacitance 2.(high frequencies), respectively.
The results with a non-inductive load resistor is one thing.
The results with most loudspeakers are another thing.
Over the audio frequency range, loudspeakers have varying capacitive reactance; varying inductive reactance; and varying resistance versus frequency (a woofer and a tweeter often have different resistance).
Last edited:
I took a push pull transformer with no air gap.
I removed the E's and I's.
I re-stacked all the E's on one side, and all the I's on the other side, and put a piece of kapton tape over the end of the E's.
That created an "air gap".
Then I used the transformer and a Beam Power tube in Ultra Linear mode.
The center tap became a 50% UL tap.
Then I used a 100 Ohm resistor between the screen and a switch. The switch either connected to the UL tap, or to the Beam Power tube plate. (UL mode, or Triode mode respectively).
Caution: Only change the switch setting with the amp cold (B+ at zero volts).
I removed the E's and I's.
I re-stacked all the E's on one side, and all the I's on the other side, and put a piece of kapton tape over the end of the E's.
That created an "air gap".
Then I used the transformer and a Beam Power tube in Ultra Linear mode.
The center tap became a 50% UL tap.
Then I used a 100 Ohm resistor between the screen and a switch. The switch either connected to the UL tap, or to the Beam Power tube plate. (UL mode, or Triode mode respectively).
Caution: Only change the switch setting with the amp cold (B+ at zero volts).
Last edited:
It's not abusing the OPT and I have done exactly this, likely with the exact OPT the OP is inquiring about. My equipment is mediocre and my techniques likely little better, but I took crude SMPTE IMD measurements of the various configurations with 46, 47 and 307A tubes, and crudely graphed the results, and also with the Edcor XSE-15 5K for comparison, so probably slightly better information than speculation. I have not tried it with tubes that were actually made for UltraLinear operation. The results might be different for those tubes.
The Edcor was markedly better in all configurations as best I recall. That said, I am still running these OPT's as a 7K load for a ( wait for it ) 2A3. The 5K load they present is terrible. Best I recall.
The data and graphs are in a thread somewhere down in Tubelab's sub forum.
w5jag,
A possible reason the 2A3 works better on the 7k than on the 5k is because:
With the 2A3 connected to the 5k tap, there are 18% more turns from the 5k tap to the 7k tap that are not connected to anything . . . and they are "flapping in the breeze".
An unterminated winding is probably going to make for a very poor square wave response.
Also, depending on the transformer winding techniques, there may be less leakage reactance from the 7k tap to the secondary windings; versus the leakage reactance from the 5k to the secondary windings.
A possible reason the 2A3 works better on the 7k than on the 5k is because:
With the 2A3 connected to the 5k tap, there are 18% more turns from the 5k tap to the 7k tap that are not connected to anything . . . and they are "flapping in the breeze".
An unterminated winding is probably going to make for a very poor square wave response.
Also, depending on the transformer winding techniques, there may be less leakage reactance from the 7k tap to the secondary windings; versus the leakage reactance from the 5k to the secondary windings.
I'd not given the free floating winding much thought, but the Hammond 125 type universal OPT's do have one primary connection, and get the different loads off the secondary taps.
So, short the 5K and 7K together to use it at 5K?
BTW, 47 and 307A do not respond well to distributed load ( ultra linear ).
So, short the 5K and 7K together to use it at 5K?
BTW, 47 and 307A do not respond well to distributed load ( ultra linear ).
Last edited:
1. No, do not ever short two different impedance taps together.
2. I did not realize one of the transformers in question was a 125 series.
Since there are only two ways to connect the primary (only 2 wires, you can only change the phase driving it), then all of those many primary turns are always in the circuit.
There are lots less turns in the 125 series secondary, so I would not worry about any serious "flapping" from all the taps that are disconnected. There is of course the issue of leakage reactance depending on how many of the secondary turns are used, and the winding techniques.
The 125 series transformer is OK, but leaving several secondary taps disconnected is far from being the major issue, versus its overall performance.
I have a pair of 125DSE, but have not used them in an amplifier yet, I did some tests on them some time ago.
If I get the time, would like to build a lightweight small stereo amp, or lightweight pair of mono-block amps with them.
2. I did not realize one of the transformers in question was a 125 series.
Since there are only two ways to connect the primary (only 2 wires, you can only change the phase driving it), then all of those many primary turns are always in the circuit.
There are lots less turns in the 125 series secondary, so I would not worry about any serious "flapping" from all the taps that are disconnected. There is of course the issue of leakage reactance depending on how many of the secondary turns are used, and the winding techniques.
The 125 series transformer is OK, but leaving several secondary taps disconnected is far from being the major issue, versus its overall performance.
I have a pair of 125DSE, but have not used them in an amplifier yet, I did some tests on them some time ago.
If I get the time, would like to build a lightweight small stereo amp, or lightweight pair of mono-block amps with them.
It's not, I was just pointing out that Hammond did their multi load OPT's in a different way.
I have some Hammond 125ESE and they are good OPT's, imo. I've used the extra secendary taps for cathode feedbck.
Yes, I should have purchased the 125ESE instead of the 125DSE.
A little more laminations, a little more max current, but a little more expensive and heavier.
A little more laminations, a little more max current, but a little more expensive and heavier.
Here is a poor picture of my dual primary OPT's; apparently a bit different than the OP's, as mine lack a 16 ohm secondary.
Looks like I also ( tried to ) measure the primary inductance, but I can't read my crappy handwriting in the picture ( 3.0 H ? 10 H ? ) and I am at our lake house today, so I can't just go look at them.
Looks like I also ( tried to ) measure the primary inductance, but I can't read my crappy handwriting in the picture ( 3.0 H ? 10 H ? ) and I am at our lake house today, so I can't just go look at them.
Attachments
Last edited:
"breadboard"-ing works.
And there is no magnetic coupling between all the transformers and choke(s).
What are the tube types, if you remember?
And there is no magnetic coupling between all the transformers and choke(s).
What are the tube types, if you remember?
5V4, 307A/VT-225, probably a 12BH7 - hard to tell from the angle.
I usually lay down lay down power transformers, but in this case, I didn't ... This hefty old timer has dual 2.5 VCT secondaries, but 307A is 5.5 volts, so some improvisation was required. So far as I can tell, I've never had magnetic coupling problems at AF, regardless of transformer orientation.
edit: the breadboard actually simulates the interior dimensions of a Hammond 12 x 10 inch chassis, in case I wanted to drop it into an existing chassis I had. It's still running on the same breadbord with the same OPT's, although the output tubes are now some CBS 2A3's.
I usually lay down lay down power transformers, but in this case, I didn't ... This hefty old timer has dual 2.5 VCT secondaries, but 307A is 5.5 volts, so some improvisation was required. So far as I can tell, I've never had magnetic coupling problems at AF, regardless of transformer orientation.
edit: the breadboard actually simulates the interior dimensions of a Hammond 12 x 10 inch chassis, in case I wanted to drop it into an existing chassis I had. It's still running on the same breadbord with the same OPT's, although the output tubes are now some CBS 2A3's.
Last edited:
Since we have determined that no OPT's are being "abused" it might fair to say the an SSE board IS.
I don't know much about the #47, but I can confirm that 307A's do not like UL. In class A SE, or in class AB2 P-P, the output power is about the same for triode or UL, but the distortion and DF are better in triode. I have several different brands of 307A's and all behave the same way.
As for intentionally miswiring an OPT to get a different impedance reflected to the tube, every OPT responds differently, and often the same OPT will behave differently when the driving tube is changed, biased differently, or configured differently. The shortcommings in an OPT are less obvious when the plate impedance of the driving tube are lowered. This is why the results are usually better in triode mode, and with the lowest Rp.
I have a bunch of 6600 ohm push pull OPT's that were designed for guitar amps. They work far better as 3300 ohm OPT's when driven by a pair of big TV sweep tubes with lots of local feedback than in the usual 6600 ohm connection with just about any tubes.
I don't know much about the #47, but I can confirm that 307A's do not like UL. In class A SE, or in class AB2 P-P, the output power is about the same for triode or UL, but the distortion and DF are better in triode. I have several different brands of 307A's and all behave the same way.
As for intentionally miswiring an OPT to get a different impedance reflected to the tube, every OPT responds differently, and often the same OPT will behave differently when the driving tube is changed, biased differently, or configured differently. The shortcommings in an OPT are less obvious when the plate impedance of the driving tube are lowered. This is why the results are usually better in triode mode, and with the lowest Rp.
I have a bunch of 6600 ohm push pull OPT's that were designed for guitar amps. They work far better as 3300 ohm OPT's when driven by a pair of big TV sweep tubes with lots of local feedback than in the usual 6600 ohm connection with just about any tubes.
That SSE board is still working like a champ, though - it's in a pretty harsh environment out in my radio room. It's the one that has been dishwashered at least three times to clean experiment residue off of it.