I am considering building a SE tube amp with some parts I have laying around. I am not sure what output tube I am going to try but lets assume an EL84 for my question. I am looking at Edcor transformers because the consensus seems to be they are the best bang for the buck and to really get to the next level above them you need to spend about 5X their price. So my question is, is there any down side to using an oversized output transformer over a regular sized one? There are obvious drawbacks to using a transformer that is to small but is there such a thing as to big?
For example are there any downsides to using a CXSE25-8-6.5K instead of a GXSE10-4-5K or GXSE5-8-5.5K. The upside to the larger transformer, that I am aware of, would be greater frequency extension and the ability to be reused on different amps in the future. Short of the added cost are there other issues to consider?
An even more dramatic example would be if I was considering a PP amp. Are there drawbacks to using a CXPP100-MS-5K over a GXPP15-4-5K besides price and the limits on the frequency range?
For example are there any downsides to using a CXSE25-8-6.5K instead of a GXSE10-4-5K or GXSE5-8-5.5K. The upside to the larger transformer, that I am aware of, would be greater frequency extension and the ability to be reused on different amps in the future. Short of the added cost are there other issues to consider?
An even more dramatic example would be if I was considering a PP amp. Are there drawbacks to using a CXPP100-MS-5K over a GXPP15-4-5K besides price and the limits on the frequency range?
too big transformer
In my very limited experience, I have noticed in my particular application that I lost a lot of top end with the extra beef. I was really surprised. I wrote a little about it here:
http://andrewtomasulo.com/electronics/6v6.html
I also have a thread somewhere on the tubelab section. Thread title something like "push pull with differential input" or something like that.
In my very limited experience, I have noticed in my particular application that I lost a lot of top end with the extra beef. I was really surprised. I wrote a little about it here:
http://andrewtomasulo.com/electronics/6v6.html
I also have a thread somewhere on the tubelab section. Thread title something like "push pull with differential input" or something like that.
Looks like browskie has been thinking about the same issue.
Flea-Power OTL Amplifier
That makes sense in theory but I hope some other people weigh in with their own experiences.
Flea-Power OTL Amplifier
That makes sense in theory but I hope some other people weigh in with their own experiences.
I didn't measure anything, but the difference was unmistakable. I think it also depends a lot on your speakers. In my case I have 95db speakers, so I'm not really using a lot of power.
Think about it though, smaller transformer equals more turns, tighter windings, etc. At the same time, transformer will saturate more easily on big bass notes. But if your speakers rock with just a few Watts, you shouldn't saturate. I'm sure the more experienced builders can weigh in some more.
Think about it though, smaller transformer equals more turns, tighter windings, etc. At the same time, transformer will saturate more easily on big bass notes. But if your speakers rock with just a few Watts, you shouldn't saturate. I'm sure the more experienced builders can weigh in some more.
The best compromise for me is the minimum size that will handle the full power at low frequency without significant distortion. Low frequency can be 30Hz for SE up to 20W (or 40 Hz for a cheaper project or big power) and lower frequency for PP. Once you achieve this increasing the size will make it worse and worse at the other end because stray capacitances increase as they are function of winding surfaces, regardless of turns as for best efficiency you want to use all the space.
The problem lies in the basics.
You need a lot of inductance to transfer the bass and that means a lot o wire.
A lot of wire gives capacitance and that makes a filter cutting out the treble.
In comes the fine art of layering the wire to reduce capacitance to give an acceptable compromise between inductance and capacitance.
Optimum irons is a biamped setup with dedicated OPT´S for treble and bass.
You need a lot of inductance to transfer the bass and that means a lot o wire.
A lot of wire gives capacitance and that makes a filter cutting out the treble.
In comes the fine art of layering the wire to reduce capacitance to give an acceptable compromise between inductance and capacitance.
Optimum irons is a biamped setup with dedicated OPT´S for treble and bass.
At some point the power will be the first limiting factor however you can make up to 40W at 30Hz with 10K primary load with excellent frequency response (up to 50-60 KHz). It will be expensive though and will be big (around 8Kg I estimate)! The primary impedance also is an important factor.
You cannot reduce overall stray capacitances. You can only re-distribute them and optimize frequency response.
Not necessarily if you don't need more than 40W for SE and 120W for PP.
It's possible to get extension at both ends with a very large core, specialized winding techniques, optimized dielectric materials, etc. Monolith Magnetics is making me a SE OPT with 7Hz to 66kHz -3dB FR, and 32Hz saturation point at 200W output power. It is however, huge, and costly. Here's an in-construction update from them with test results and photo of the core and windings:
http://www.diyaudio.com/forums/tube...-crisis-my-833c-amp-build-18.html#post3490757
To a large degree, what is possible depends on how much you're willing to spend to get there.
http://www.diyaudio.com/forums/tube...-crisis-my-833c-amp-build-18.html#post3490757
To a large degree, what is possible depends on how much you're willing to spend to get there.
If nothing else, since core losses exist and are a funcion of the amount of iron in a transformer, I would expect there to be a point at which those losses would overcome any gain in a larger transformer.
Whether this would result in greater distortion I do not know.
It seems to me that there is an optimum size for the transformer which trades off copper loss for core loss , as well as bandwidth and distortion.
How one would determine that size is beyond me.
Whether this would result in greater distortion I do not know.
It seems to me that there is an optimum size for the transformer which trades off copper loss for core loss , as well as bandwidth and distortion.
How one would determine that size is beyond me.
i based my choices as per RDH4, a formula for core power handling for 60hz is given there, i just scale it down to 20 or 30 hz....that is minimum requirement, you can also choose to go bigger if you can afford it...
The OP ends his first paragraph:
"but is there such a thing as to big?"
I agree that bigger is better for low end as a general rule, but where does
this end? How big is too big relative to a size that gives the optimum for say 20Hz -f3?
Certainly there has to be some point at which a transformer begins to lose detail as the size is increased due to core losses and other factors(money excluded as a factor).
Is this ratio so large as to be moot?
"but is there such a thing as to big?"
I agree that bigger is better for low end as a general rule, but where does
this end? How big is too big relative to a size that gives the optimum for say 20Hz -f3?
Certainly there has to be some point at which a transformer begins to lose detail as the size is increased due to core losses and other factors(money excluded as a factor).
Is this ratio so large as to be moot?
A big Edcor CXSE25-8-5K definately gives up some HF and detail compared to a small XSE15-8-5K when driven with a low powered tube like a 45 or a triode wired 6V6GT. The differences are much smaller when the tube is something bigger like a 300B or triode wired KT88.
I prefer the little Edcor or a mid sized Transcendar, One Electron or Electra Print with a 45 or 6V6GT. I have not tried the mid sized Edcors. They didn't exist when I did this.
I tried all of these transformers with 300B's, KT88's and triode wired 307A's and decided to use the big Edcors, they sounded and measured the best. The big Hammond 1628SE looses detail with every tube tried. They are still looking or a home.
I have measured data somewhere, but I am 1200 miles from home right now. Some of it is on my web site.
I prefer the little Edcor or a mid sized Transcendar, One Electron or Electra Print with a 45 or 6V6GT. I have not tried the mid sized Edcors. They didn't exist when I did this.
I tried all of these transformers with 300B's, KT88's and triode wired 307A's and decided to use the big Edcors, they sounded and measured the best. The big Hammond 1628SE looses detail with every tube tried. They are still looking or a home.
I have measured data somewhere, but I am 1200 miles from home right now. Some of it is on my web site.
What do you mean by the 'amount' of iron? Surely the volume of physical steel has nothing to do with loss. What matters is the dimensions and dielectric between windings (capacitance), and the permeability (inductance).
Well it seems we have some conflicting opinions. I used my example transformers because I wanted to keep variables to a minimum. This way we are just focusing on size and the changes directly related to size and nothing else. I don't wand to focus on different brands, different winding techniques, different insulators, etc.
The amount of iron has everything to do with loss. All books give the loss as function of frequency and the mass of the core. ( and core material )
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.