Elvee,
That is very interesting.
I hope someone builds a complete 2 transformer circuit, with all the extra components . . .
I would like to see an oscilloscope screen shot of something like a 5k Ohm primary z to 8 Ohm secondary z:
from a square wave signal source impedance of 1k Ohm driving that dual transformer circuit, and with an 8 Ohm resistor load.
Having a dual transformer circuit like that, which has a frequency response that starts far below 20Hz, and extends to far above 20kHz . . .
does not necessarily guarantee a good square wave response (it may, or may not be very pretty).
That is very interesting.
I hope someone builds a complete 2 transformer circuit, with all the extra components . . .
I would like to see an oscilloscope screen shot of something like a 5k Ohm primary z to 8 Ohm secondary z:
from a square wave signal source impedance of 1k Ohm driving that dual transformer circuit, and with an 8 Ohm resistor load.
Having a dual transformer circuit like that, which has a frequency response that starts far below 20Hz, and extends to far above 20kHz . . .
does not necessarily guarantee a good square wave response (it may, or may not be very pretty).
IMHO if the amplifier isn't capable then we have failed before we even started. We aim for perfection, not for failure, and get as close as we can. A 25 Hz tone through my speakers with an amplifier that can handle it (I used to have a 300 watt x 2 solid state amp) will leave all of the pictures on the walls sideways. Play a pipe organ recording and get the same result.
Never Get Old,
I like that, EL84 pentode mode in push pull; and EL34 pentode mode in single ended.
Unless you have a loudspeaker that has a flat impedance across the audio range, and an intrinsically well dampened woofer (woofer and cabinet / open baffle / etc. that is designed to be self dampened) . . .
You will need to use a signal sample for negative feedback, which comes from either the output transformer primary, or the output transformer secondary.
Use primary or secondary based global negative feedback with pentode mode output tubes (beam power tubes too).
Without global negative feedback:
Ultra Linear feedback, especially at medium to large %, will have some damping for the woofer and amplifier control versus the loudspeaker impedance variations versus frequency.
Without global negative feedback:
Triode Wired pentodes have the highest damping factor, for woofer damping and for better results versus the loudspeaker impedance variations versus frequency (Triode Wired beam power tubes too).
I like that, EL84 pentode mode in push pull; and EL34 pentode mode in single ended.
Unless you have a loudspeaker that has a flat impedance across the audio range, and an intrinsically well dampened woofer (woofer and cabinet / open baffle / etc. that is designed to be self dampened) . . .
You will need to use a signal sample for negative feedback, which comes from either the output transformer primary, or the output transformer secondary.
Use primary or secondary based global negative feedback with pentode mode output tubes (beam power tubes too).
Without global negative feedback:
Ultra Linear feedback, especially at medium to large %, will have some damping for the woofer and amplifier control versus the loudspeaker impedance variations versus frequency.
Without global negative feedback:
Triode Wired pentodes have the highest damping factor, for woofer damping and for better results versus the loudspeaker impedance variations versus frequency (Triode Wired beam power tubes too).
Just a basic question, which may be difficult to answer ("it depends...").....
I use SE amplifiers which are around 5W output. I've noticed with a bunch of around 10 OPTs I've had over the years that the small ones have better high treble. Noticeable on instruments like cymbals and brushes in a jazz drum kit. The only exception is my pair of amorphous core NP Acoustics OPTs which are large but have excellent treble, I imagine because of the amorphous core.
Is this generally true? I know very little about winding transformers but I'm sure the boffins here do.
I use SE amplifiers which are around 5W output. I've noticed with a bunch of around 10 OPTs I've had over the years that the small ones have better high treble. Noticeable on instruments like cymbals and brushes in a jazz drum kit. The only exception is my pair of amorphous core NP Acoustics OPTs which are large but have excellent treble, I imagine because of the amorphous core.
Is this generally true? I know very little about winding transformers but I'm sure the boffins here do.
If I find the right components, I may try to apply the scheme to what has become my "universal SE test-bed":
https://www.diyaudio.com/community/...rging-a-e-pcl82-amplifier.351667/post-6134055
I have read this, so it's not just you. I read somewhere this week that it's one reason not to use an output transformer that is much larger than is necessary to achieve the low frequency design goal. I'll leave it up to the experts to explain what may be going on.
I remember when bi-amping exploded in popularity and some people were mixing solid state for the bottom end and tube for the top end. I thought about it at the time. I don't remember why I never tried it. All I was missing was a crossover for the two amplifiers. An all-tube stereo "bi-amplifier" would be interesting with two channels for lows and two channels for highs using optimal output transformers for each.
Last edited:
Right after the war (1946) Paul Klipsch built an amplifier intended for use with a special Klipschhorn ("model B"). The amplifier, push-pull 2A3 types, ended at the output valve plates, and the plates and B+ appeared on a 3-terminal output strip. Yikes! Inside the speaker was a two pole 500Hz crossover with characteristic impedance of 5000 Ohms followed by two separate OPTs, each dedicated to its frequency range. This was before DTN Williamson, and wideband transformers were really hard to make back then.
When we learned of this, my building partner and I built a reproduction (with all the high voltage stuff inside the amplifier chassis) with help from Dave Slagle (donated time and effort for custom inductors and OPTs) for the Klipsch Museum to run the recently restored very first Klipschhorn prototype. Jim Hunter of the Museum:
https://www.itishifi.com/hifi/1-klipschorn-rides-again
All good fortune,
Chris
When we learned of this, my building partner and I built a reproduction (with all the high voltage stuff inside the amplifier chassis) with help from Dave Slagle (donated time and effort for custom inductors and OPTs) for the Klipsch Museum to run the recently restored very first Klipschhorn prototype. Jim Hunter of the Museum:
https://www.itishifi.com/hifi/1-klipschorn-rides-again
All good fortune,
Chris
Always wanted a pair of Cornwall II speakers. Had an opportunity to buy a mint condition pair in 1998 but had no room for them given that my current speakers are even larger and I never have another room. I never cared for the Klipschorn though. I had the Heresy II for a while. Nice, but no low bass so I got rid of them. They were powered by my Dynaco Stereo 70 series ii (Panor Corp circa 1993). I need to put new PS caps in that amp. I took size measurements for them yesterday and put it back on the shelf.
Serial Xover feeding dedicated opts primaries of a high power, high impedance SE should offer some advantages...
If anyone wants to try, i could offer some big pot cores including bobbins ( same cores have been used by philips for a 400w 20kHz ultrasonic cleaner driven from pp TB-something transmitting tubes). Please notice, those 400W was rated output, the resonant power the core had to handle was much higher in order to sustain oscillation.
If anyone wants to try, i could offer some big pot cores including bobbins ( same cores have been used by philips for a 400w 20kHz ultrasonic cleaner driven from pp TB-something transmitting tubes). Please notice, those 400W was rated output, the resonant power the core had to handle was much higher in order to sustain oscillation.
Last edited:
Back then, they used low Dk paper and iron with higher si content, today we mostly use higer Dk mylar and iron with lower si content wich translates into comparable higher capacitance and higher eddy current losses. Also, formvar wire isolation seems to be favourable in respekt to todays wire isolation dielectric losses. Sure, voltage isolation properties improved enourmously but all in all, not much easier now than it was back then to make a good opt. And one thing is sure, a lot of experiance got lost during the decades and have mostly only be regained today by studying books of the past...
Last edited:
Yes, I answered that question almost one year ago today. That solid state amplifier is what caused me to stop listening to music for almost 10 years. I have owned the finest solid state over the decades including Threshold. Yes, solid state is powerful, but never again will I buy any solid state amp or preamp. All tube all the time for the remainder of my life.
Gee, some decades ago i bougth a claimed class A fet-amp module that sounded to me just as amazing as my el84 pp 50/50 split loaded triode amp but in contrast to that lill tubeamp offered all the power one could ever wish for. Back then i thougth, i'll never need any tube amp again. A few years later, after that thing decided to stop working and turned out unrepairable, i changed back to tubes, and today, its still tubes only that never disappointed me and keep up my fascination with DIY.
Offcourse, the fact that this otherwise really good sounding amp module, not only was unrepairable, but also took out my speakers when it poofed, was big part of my love returning to tubes...
Offcourse, the fact that this otherwise really good sounding amp module, not only was unrepairable, but also took out my speakers when it poofed, was big part of my love returning to tubes...
Last edited:
OPT limitations:
1. Low frequency roll-off
a) Depends on the primary inductance vs driving impedance. The lower in driving impedance you get, the more you can get away with lesser primary inductance
b) At high power, the core can begin saturating and primary inductance gets reduced.
2. High frequency limitations.
a)Leakage inductance vs load. The more the transformer is loaded, the sooner the roll-off
b) Shunt capacitance vs driving impedance. The higher the driving impedance is, the sooner the roll-off.
1. Low frequency roll-off
a) Depends on the primary inductance vs driving impedance. The lower in driving impedance you get, the more you can get away with lesser primary inductance
b) At high power, the core can begin saturating and primary inductance gets reduced.
2. High frequency limitations.
a)Leakage inductance vs load. The more the transformer is loaded, the sooner the roll-off
b) Shunt capacitance vs driving impedance. The higher the driving impedance is, the sooner the roll-off.
< 2. High frequency limitations.
a)Leakage inductance vs load. The more the transformer is loaded, the sooner the roll-off
b) Shunt capacitance vs driving impedance. The higher the driving impedance is, the sooner the roll-off.>>
You know about winding transformers. Is it easier to get the best high frequency response from smaller 5W SE transformers?
a)Leakage inductance vs load. The more the transformer is loaded, the sooner the roll-off
b) Shunt capacitance vs driving impedance. The higher the driving impedance is, the sooner the roll-off.>>
You know about winding transformers. Is it easier to get the best high frequency response from smaller 5W SE transformers?
Going with a larger core and reducing the number of turns is the way to go get better high frequency response.
-Power capability increases linearly by the core area (Afe). And core area is square, compared to Mean Turn Length (MLT), which is an enemy for frequency response and losses. For example, a core 10x10mm has a zero MLT of 40mm and an Afe of 100mm2. A core of 20x20mm has a zero MLT of 80mm, but an Afe of 400mm2. Double amount of MLT, but a 4xAfe, which dictates you can reduce the winding turns 4 times to get the same power output.
-Leakage inductance decreases by a square factor of winding turns
-That gives the advantage of less interleaving, hence less shunt capacitance can be achieved.
PP transformers are the easiest, because:
-No DC magnetizing current, lesser turns needed for the same core Afe
-Interleaving capacitance reduction bonus in PP transformers.
-Symmetrical capacitance distribution, little to no dipping resonance in the frequency response.
In terms of DC magnetizing current, SE transformers are even worse/harder, because customers often require an overexaggerated Idc bias point value for the sake of using the tube into a more linear loadline. Which eats flux density capability even futher.
There are some tricks to get away with a huge amount of turns, I've attempted successful projects with SE OPTs having 9000 turns. However, it requires much more insight when designing the interleaving geometry.
-Power capability increases linearly by the core area (Afe). And core area is square, compared to Mean Turn Length (MLT), which is an enemy for frequency response and losses. For example, a core 10x10mm has a zero MLT of 40mm and an Afe of 100mm2. A core of 20x20mm has a zero MLT of 80mm, but an Afe of 400mm2. Double amount of MLT, but a 4xAfe, which dictates you can reduce the winding turns 4 times to get the same power output.
-Leakage inductance decreases by a square factor of winding turns
-That gives the advantage of less interleaving, hence less shunt capacitance can be achieved.
PP transformers are the easiest, because:
-No DC magnetizing current, lesser turns needed for the same core Afe
-Interleaving capacitance reduction bonus in PP transformers.
-Symmetrical capacitance distribution, little to no dipping resonance in the frequency response.
In terms of DC magnetizing current, SE transformers are even worse/harder, because customers often require an overexaggerated Idc bias point value for the sake of using the tube into a more linear loadline. Which eats flux density capability even futher.
There are some tricks to get away with a huge amount of turns, I've attempted successful projects with SE OPTs having 9000 turns. However, it requires much more insight when designing the interleaving geometry.
Last edited:
Console manufacturers used this idea in the ‘60’s. Such as Motorola HS-768, where higher frequencies were amplified in stereo by EL84 SE amplifiers (filtered before amplification) and LF amplified by a single PP EL84 amp with much heavier transformer.