|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
71.234.12.142
In Reply to: Re: Your bypass C's posted by Tre' on April 11, 2007 at 14:25:06:
i wonder why people have never looked at the converse of this which is what happens at high frequency? (particularly wrt iron)
what follows is probably OT so apologies in advace.
to be flat from 20-20K as you state (on a first order) a tranformer must deliver 2-200K bandwidth. People always seem to chase the low end striving for F3 of say 5hz but readily accept a -f3 of 40Khz which should still leave you -.1dB at 4khz.
If we consider transformers in general "fixed bandwidth devices" in order to gain an octave on the bottom you need to lose an octave on the top. Obviously different people have different perspectives on this. Looking at the plots for much of the japanese iron (tango tamura hashimoto) a -2dB point of 20hz is often listed and this comes with an extended frequency response on the top end. Listed -2dB points of 20 and 80K are not uncommon, and this must be a conscious choice since those numbers could easily be converted to 10 and 40Khz.
I'm not exactly sure what my point is other than i have always noticed this and found it to be interesting.
Follow Ups:
Dave, I don't know. I've thought about that also. Maybe it's a matter of "some things we can control and some we can't" as in: A good output transformer that will do 2-200k is hard (expensive) to find?I'm just throwing this out there, I don't know the answers.
Even if we are using a transformer that is rolling the top (or bottom), do we want the circuit to roll the top (or bottom) more?
Maybe, to keep the transformer out of "trouble", we do want to roll the extremes?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
2-200K is really easy as long s you keep the DC out and work at 600 ohms :-)Like anything else in audio i suspect we all get caught looking at single parameters in a one-dimensional manner. Sometimes i find i need to step back and conider the whole picture as a reality check.
DS wrote::::2-200K is really easy as long s you keep the DC out and work at 600 ohms :-):::
there are examples, again, of pretty good FR responses in tranneys with nominal primary impedances well above 600 ohms.
for example, I can think of a samll signal PP trans that Peerless made for the Fort Wayne Steel Co. that had a 30K PP primary and was rated at plus or minus 1 db from 5 hertz to 75 kilohertz.
And there are many examples of large signal OT that have bandwidths from 10 hertz to 75KHZ or 100KHZ (and some even higher) that also have primary impedances in the 3K to 10K range and have normal secondary speaker load impedances.
keeping (or minimizing the magnitude of) unbal DC out of the cores makes life much, much easier. Unbal dc plate currents consume the core's flux density and makes the iron work harder... when the iron has to work harder (operate at larger and larger flux densities) so doesn't the core induced harmonic distortions go up. It is just one good reason to strictly limit or to totally eliminate DC currents from transformer windings.
Also... dc unbal current will tend to rob the core of it's perm... and thus make the L go down... thus contribute to moreso poorer quality LF response.
I do agree with you on the point that FR is but one of the many factors that must or should be considered... but definitely not the sole factor in acheiving a good design. And the transformer with the widest BW may not always be the best sounding tranney--- frequency response is after all--- basically a quantitative measure not inherently a qualitative measure.
MSL
that large signal power response may be more important than small signal FR.MSL
Well put Mikey !! Thanks.
Hi Tre:just a few comments\observations
:::A good output transformer that will do 2-200k is hard (expensive) to find?:::
I don't know... 10hz to 100khz =/- 1db tranneys are out there. And I've seen PP output transformers with even wider bandwidth.
It takes a special design... a good design... to acheive this extension of bandwidth on top and bottom but it can be done.
Heck, there is a drawing in the Peerless archives... where on the bench a small cap was put in para across one half of the pri winding to take out a small blip in the FR (out at 250 khz)....
there is a Freed PP output trans.... that sails way out over 200khz and goes down below 10 hertz--- it was a very evloved coil design with all of the capacitive and leakage gradients carefully controlled and thought out. Had to be a "stellar" OT cause it was used in an amp with 80db of feedback!!!!
that said--- i've said many times--- to me freq response is about the thirteenth most important thing in a transformer design--- and getting good bw top and bottom is not gained by aiming at that parameter directly but usually is the result of doing the first twelve most important things well.
and with all of this said--- it is perhaps moreso important to get the bottom end "tuned up" properly--- i.e., to focus on good bottom end response--- simply because that is where the core has to work the hardest--- i.e., where the flux density will be the greatest. So a wimpy bottom end not only has poor FR--- but will also tend to exhibit a much greater degree of core generated harmonic distortion as well. so if you want good distortion free response at 40, 30, or 20 hertz you better design the iron to behave well an octave or two below that....
:::Even if we are using a transformer that is rolling the top (or bottom), do we want the circuit to roll the top (or bottom) more?::::see my notes above re: distortion and bottom end response.
where and how the top end rolls off is largely determined by coil geometry characteristics... above 1, 2, 3, 5 or 10khz the core can conceptually be removed from trans in explaining the behaviour of the device.
And at high frequencies the flux density of the core is very, very low and tends not to be a large contributor to the generation of core-induced harmonic distortion.
:::Maybe, to keep the transformer out of "trouble", we do want to roll the extremes?::::
nah... more important to tailor your circuit to the capabilities of the transformer. If your transformer has good phase response (and good FR and low THD) from X to Y... then it won't be the limiting factor. And you should be able to deploy a wider BW circuit without running into gremlins.
again... all of this said about bandwidth and freq response--- I'm responding to the technical issues--- not endorsing or claiming that FR in-and-of-itself is the barometer (nor sole determinant) of sound quality in an amplifier.
msl
"here is a Freed PP output trans.... that sails way out over 200khz and goes down below 10 hertz--- it was a very evloved coil design with all of the capacitive and leakage gradients carefully controlled and thought out. Had to be a "stellar" OT cause it was used in an amp with 80db of feedback!!!!"Those transformers from the Harmon Kardon Citation 2 are very good and in fact what I'm using but without the feedback.
Thanks for the great explanation.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hi Tre:The Freed's I mentioned were built for a different company. Interestingly the pri impedance of the one's I mentioned was about half of the Citation 2's... the trans I mentioned was designed for PPP KT88's.
The Freed designs were very nice... their coil design (in the QGA series) were quite different than other mainline, highline transformer winders from the golden era.
There are lots of great vintage designs--- some of the Triad's looked suspiciously like Acrosounds. The Dyna's were optimized very, very well for UL operation. Both the Dyna and Acrosound transformer patents are interesting and instructive reads. Having copies of the actual factory blueprints has been a real treasure.
msl
I should learn not to make assumptions. Sorry.Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
This post is made possible by the generous support of people like you and our sponsors: