transformer impedance and cheese

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
and just to clearify, when i say roll off im talking about -3db...


which cheese can lure the bear out of the woods is the come on bear!
 
5v333 said:
why is that a problem?

perhaps the 2,5hz comes from an earlier stage by the way (before the input transformer is my passive eq with another input transformer as the front end of everything ).
Frequencies below cut-off defined by Ra/Zp should not be permitted. Somehow it should be controlled. The only safe way is by inserting a HP filter earlier in the chain. If such a filter exists in your chain, it's fine. If not, there's a risk that  an uncontrolled signal pushes the transformer into saturation.
 
alright. once again i apologies that i cant present a drawing yet. its coming! think pultec amp without a feedback winding...

so the bandwidth is lower on the input secondary then on the output secondary in open loop. we dont like that!

but, when we close the loop this difference almost evens out (2,5hz vs 3,5hz) isnt that fixing our concerns(almost)?  perhaps i can bias the 6sn7 just a mA more to get the output impedance a little lower.

or do we have to  see all frequency points from an open loop perspective, even if the loop is closed? if so, could the interstage RC component be in charge of this HP control ?



 
5v333 said:
alright. once again i apologies that i cant present a drawing yet. its coming! think pultec amp without a feedback winding...

so the bandwidth is lower on the input secondary then on the output secondary in open loop. we dont like that!
That's the way it should be. The OT is the weakest link. One has to make sure offending signals are prevented to reach it.

but, when we close the loop this difference almost evens out (2,5hz vs 3,5hz) isnt that fixing our concerns(almost)? 
NFB does not improve the saturation characteristics of iron. It tends to "iron" them out by the use of "moderate brute force" and succeeds in terms of linearizing the low-level frequency response, but not much in terms of power bandwidth.

perhaps i can bias the 6sn7 just a mA more to get the output impedance a little lower.
Yes; boosting the output stage power increases the amount of "brute force" available.

or do we have to  see all frequency points from an open loop perspective, even if the loop is closed?
In terms of power bandwidth, yes.

  if so, could the interstage RC component be in charge of this HP control ?
I wrote earlier that's a receipe for creating other problems. The close loop will try to compensate the weaknesses of both the OT and the RC roll-off.
[/quote]
 
you are providing some really interesting stuff about theese kinds of circuits!
i do sence that im repeating my questions :)

so this is starting to look like a small disaster :)

biasing my 6sn7 an mA or so wount get me far...

putting a pair of 13uf polyprops after the input transformer is not very appealing.

i wonder if a reconfiguration of the input transfomer can give a different lo rolloff.

i might have to look for other irons.


oh and the cheese which lures the bear out of the woods: CAM N BEAR!
 
5v333 said:
putting a pair of 13uf polyprops after the input transformer is not very appealing.
Why 13uF? You would have to have grid-leak resistors, then with 470k and 22nF caps, you would get a 6dB/octave roll-off at 15 Hz.

i wonder if a reconfiguration of the input transfomer can give a different lo rolloff.
You could have a capacitor at the input; you would have to adjust the value, but then you could end up with a 12dB/octave slope, which would allow choosing a turnover frequency perfectly matching that of the Ra/Zp.
 
ignore the big cap. its early still...

thanks for the suggestion! i will consider it and some other stuff.

just saw the ll2735 which looks tempting. dont know if it works as a phase splitter.
 
here is how my tube gain stage looks like basicly right now.
 

Attachments

  • NYDEQ TUBE aMP2.pdf
    12.5 KB · Views: 12
i had a look on my scope yesterday, checked the low end respones on the output, with the DC function in place.

im pretty sure i saw a more pronounced bump after 12hz with a peak at about 6hz that was about +5db over the initial +-0db line. after that  there was some shapes of saturation  down to  2hz.

maybe this could be of advantage aswell. for thoose times when we wanna lively up a tune that needs some help in the bottom. and when we dont need that we can just insert a highpass at a previous analogue EQ or plugin.

anybody playd whith this kind of harmonic processing?

after i got rid of some rf resonances caused by the  diodes in the psu, i had a listen, sound is very sweet now, very close to the final!
 
5v333 said:
i had a look on my scope yesterday, checked the low end respones on the output, with the DC function in place.

im pretty sure i saw a more pronounced bump after 12hz with a peak at about 6hz that was about +5db over the initial +-0db line. after that  there was some shapes of saturation  down to  2hz.
This is exactly what a good designer would endeavour to eradicate. That means any schock  to the mic, handling noises and foot noises are exacerbated and will intermodulate with the signal.

maybe this could be of advantage aswell. for thoose times when we wanna lively up a tune that needs some help in the bottom. and when we dont need that we can just insert a highpass at a previous analogue EQ or plugin.
Can you hear 6Hz? (I doubt your speaker have significant response down there)  "Help in the bottom" is something you're gonna find in the 40-80 Hz range, sometimes even higher, not in a range that is inaudible, but creates ugly saturation and intermodulation effects.

anybody playd whith this kind of harmonic processing?
These effects are well-known, and nobody in his right mind would use it for artistic purpose.
 
hi again!

been working some live gigs latly and havnt been connected with the web .


i guess your right, below 10hz is probobly too low of interest!

and the word intermodulation doesnt ring nicley in my ears.

ive askd lundahl about their ll2735, which states a bandwidth of 12hz - 65Khz and seems perfect in this circuit, as long as it works well for phase splitting duty with a source of 600ohm.

to bad my function generator died on me the other day so i could not tweek the last stuff...

thanks for all the knowledge!!!
 
5v333 said:
ive askd lundahl about their ll2735, which states a bandwidth of 12hz - 65Khz and seems perfect in this circuit, as long as it works well for phase splitting duty with a source of 600ohm.
It's a power output xfmr 16k:8ohm, intended for single-ended stages. Indeed, the primary is constituted of two halves, so it may be used in a push-pull configuration, but in fact, the construction is slightly assymetrical.
For these reasons, I don't see it fit for splitting a line level source.
Model LL7905 seems more appropriate, even LL1922.
 
hi there!

i dont think theres any meaning with exchanging my ll7903 with LL7905 or LL1922. the lowend is more or less the same...

i would like to have a input transformer that matches the output transformers lowend and stay away from any more RC poles, especialy since theres no dc to block.

the size would be a bit tricky cause i prob need to move my tube sockets and rewire it.

would you think there could be problems ahead when using an iron with that high capacity with mic/line level signals? such as pickup effects etc?

cheers!
 
5v333 said:
i dont think theres any meaning with exchanging my ll7903 with LL7905 or LL1922. the lowend is more or less the same...
There basically all the same, with different ratios. I thought you may prefer the additional gain provided by the 7905, but if you're satisfied with the 7903, so be it.


i would like to have a input transformer that matches the output transformers lowend and stay away from any more RC poles, especialy since theres no dc to block.
I think you're in a wild goose chase here. the low frequency response of these transformers depends so much on what source they see it's almost irrelevant. I would think with a 100 ohm source (typical of most modern equipment, the LF response goes down to 2Hz.


would you think there could be problems ahead when using an iron with that high capacity with mic/line level signals? such as pickup effects etc?
No, there is no issue, except the waste of copper and iron and money and bulk.
 
ive just realized that i have been seeing stuff that was never there...

the ll2753 was said -3db at 12hz with a 3k source. not 600r as i thought...

i need to get smart... soon!!!
 
Back
Top