Hi!
I'm thinking about building a tube power amplifier with the following configuration:
a) -10db audio input to a triode connected EF86
b) Interstage transformer working as a phase splitter.
c) Ultra-linear push-pull EL84 quad configured to AB operation.
Is this doable? If yes, are there major inconveniences?
Could the EF86 work in ultra-linear if configured as a pentode and the interstage transformer had the proper taps?
Are there any examples of this configuration?
What would be the specification for the interstage transformer?
Thank you!
I'm thinking about building a tube power amplifier with the following configuration:
a) -10db audio input to a triode connected EF86
b) Interstage transformer working as a phase splitter.
c) Ultra-linear push-pull EL84 quad configured to AB operation.
Is this doable? If yes, are there major inconveniences?
Could the EF86 work in ultra-linear if configured as a pentode and the interstage transformer had the proper taps?
Are there any examples of this configuration?
What would be the specification for the interstage transformer?
Thank you!
Why those particular design choices? It strikes me that you're doing this the hardest, most expensive possible way, and compromising performance. An EL84 quad is trivially easy to drive without using expensive, temperamental, high distortion, bandwidth-limiting interstage iron. Two triode-connected EF86 (if that's the tube you want to use) in a long tail pair will run circles around the configuration you're proposing.
I can see an interstage if the output tubes need a lot of voltage swing, but for 10-11V?
I can see an interstage if the output tubes need a lot of voltage swing, but for 10-11V?
SY said:
Hi!
Thank you for the reply!
I understand that using a transformer may compromise the quality of the audio. I came up with this simply because I can have any transformer I like winded at a reasonable cost. The goal was to do something different and see how it sounds.
Do you know of an article explaining interstage transformer theory? I thought that the IS should not be a step-up transformer.
What about an IS with a ultralinear tap? Would it improve it's audio performance?
Thank you!
Interstage transformers are the most difficult transformers to design and unless your source is a very good winder, it is unlikely that you will get a good result. Coupled to this the fact that the EF86 is a low current driver it is very unlikely to be able to overcome the interwinding capacitances your transformer will present. I would want at least 10mA into any interstage transformer. One possability of making things work satisfactorily is to generate twice as much gain in your driver stage, and then step down the gain to half in the interstage. This will reduce all of the inherent design constraints imposed by the transformer.
Shoog
Shoog
Thank you for the reply!
How would one calculate the core size, primary and secondary impedance of the transformer?
Ideally shouldn't the interstage transformers be 1:1?
Would it be a better Idea to use a 12AU7 or 12AT7 to drive the transformer?
Would the ultralienar tap in the interstage primary make any difference?
I understand that using an active phase splitter would be simpler and in most cases better, but, why do some high end manufacturers use IS and isolation transformers for ac coupling between stages?
Thank you
How would one calculate the core size, primary and secondary impedance of the transformer?
Ideally shouldn't the interstage transformers be 1:1?
Would it be a better Idea to use a 12AU7 or 12AT7 to drive the transformer?
Would the ultralienar tap in the interstage primary make any difference?
I understand that using an active phase splitter would be simpler and in most cases better, but, why do some high end manufacturers use IS and isolation transformers for ac coupling between stages?
Thank you
The winder should calculate the core size once you specify what you want the transformer to do. Impedances are your job- you need to get the curves of the tube of interest, lay out some load lines, and make some impedance decisions. Once you've determined what you need for impedance, you then are faced with bandwidth tradeoffs- a bigger core means better low frequency extension, but worse leakage inductance (and hence worse HF response). Interstages are particularly bad in this respect because they operate at high impedance and must have gapped cores because of the primary DC. Which is more important to you, LF or HF extension? With a transformer, you can't have both.
Of course, with a simple phase splitter, you can have orders of magnitude lower distortion and much better extension on both ends. If you just can't help the burning desire for an unnecessary chunk of iron, then think about using an input transformer instead. It can do the phase splitting, it can have better bandwidth and distortion than an interstage, and if the winder is experienced and competent, it can bring important advantages, notably common mode noise rejection and the ability to run the input balanced or unbalanced.
A 12AT7 is an excellent input/driver tube if loaded correctly (high load impedance) and used for the very moderate swings required for EL84. 12AU7 have excessive high order harmonic distortion when used as voltage amps.
Either because the output tubes require significant drive current, high drive voltages (neither is the case for EL84), fashion, or incompetence. I'm quite serious about the last two.
Of course, with a simple phase splitter, you can have orders of magnitude lower distortion and much better extension on both ends. If you just can't help the burning desire for an unnecessary chunk of iron, then think about using an input transformer instead. It can do the phase splitting, it can have better bandwidth and distortion than an interstage, and if the winder is experienced and competent, it can bring important advantages, notably common mode noise rejection and the ability to run the input balanced or unbalanced.
A 12AT7 is an excellent input/driver tube if loaded correctly (high load impedance) and used for the very moderate swings required for EL84. 12AU7 have excessive high order harmonic distortion when used as voltage amps.
Either because the output tubes require significant drive current, high drive voltages (neither is the case for EL84), fashion, or incompetence. I'm quite serious about the last two.
Hearing one when it works makes it all perfectly clear why you would bother.
I had an amp with interstage splitting and it sounded lovely - but it did have some high frequency role off. I was pushing a whopping 30mA of current through my driver, and was using a parafeed arrangement. I was never happy with that amount of wasted current, though I was mostly happy with the result. Let it be said I was experimenting with very suboptimal transformers and you may achieve better. If you can face the intrusion of cap in the signal path then parafeed helps a lot.
I would second SY's opinion about the use of an input transformer as a better option. I have attached a schematic which shows that it is possible to make a two stage amp without caps in the signal path, to give you pause for alternative approaches.
Shoog
I had an amp with interstage splitting and it sounded lovely - but it did have some high frequency role off. I was pushing a whopping 30mA of current through my driver, and was using a parafeed arrangement. I was never happy with that amount of wasted current, though I was mostly happy with the result. Let it be said I was experimenting with very suboptimal transformers and you may achieve better. If you can face the intrusion of cap in the signal path then parafeed helps a lot.
I would second SY's opinion about the use of an input transformer as a better option. I have attached a schematic which shows that it is possible to make a two stage amp without caps in the signal path, to give you pause for alternative approaches.
Shoog
Attachments
Hi Joao Pedro,
I agree with SY's opinions, especially that the EF86 is unsuitable to drive an IT.
Re: your question about a tapped IT primary allowing UL, the answer is that it's unnecessary because you can achieve UL operation of a small signal pentode perfectly satisfactorily using a simple resistive network -- something you cetainly could not do with a power pentode -- explained at this link.
Forget the 12AU7/ECC82. It's not an audio tube and it behaves poorly in that role. I think it may originally have been misguidedly introduced to audio as a 9-pin equivalent to 6SN7, but it's not. You have to go to 6CG7 or 6FQ7 for that.
If you want to see articles by an acknowledged expert on transformer coupling, you could try Lynn Olson's website. His 'Karna' Amplifier uses both input and interstage transformers. It's all PP, which nicely avoids problems of DC saturation of the transformer core.
I agree with SY's opinions, especially that the EF86 is unsuitable to drive an IT.
Re: your question about a tapped IT primary allowing UL, the answer is that it's unnecessary because you can achieve UL operation of a small signal pentode perfectly satisfactorily using a simple resistive network -- something you cetainly could not do with a power pentode -- explained at this link.
Forget the 12AU7/ECC82. It's not an audio tube and it behaves poorly in that role. I think it may originally have been misguidedly introduced to audio as a 9-pin equivalent to 6SN7, but it's not. You have to go to 6CG7 or 6FQ7 for that.
If you want to see articles by an acknowledged expert on transformer coupling, you could try Lynn Olson's website. His 'Karna' Amplifier uses both input and interstage transformers. It's all PP, which nicely avoids problems of DC saturation of the transformer core.
I would say forget all 12A.... tubes. There are so many better tubes out there.
Go for 6DJ8 if you want more mu and Gm than 6SN7. This will give you some gain-reserve for a small amount of NFB.
To get good results with transformer-loading, Ri has to be low: E182CC, 6H30, 5687, ECC99 are good alternatives.
Go for 6DJ8 if you want more mu and Gm than 6SN7. This will give you some gain-reserve for a small amount of NFB.
To get good results with transformer-loading, Ri has to be low: E182CC, 6H30, 5687, ECC99 are good alternatives.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.