My HiFi stereo 150+150W tube amp weighs a hefty 45kg, (+ 100 Lbs) .... the o/p trannies each weigh 11kg, the rest weight comprised of mains transformer and chassis. I made the chassis from 4mm Aluminium sheet. Round and round circuit earthing busbar is made from solid 6mm copper wire soldered into a grid matrix; .. the point to point construction on both power halves requires good soldering skills and optimal symmetry of the phase splitter and drive circuitry..
With 8 power tubes in P-PP alot of heat is generated (in either fixed or auto bias) and if KT88‘s are used then the 4“ separation rule is used with natural convection, the minimum chassis size will be around 18“x18“ (46cm x 46cm) with alot of ventilation holes drilled around the tube bases and well ventilated underneath base plate and electrolytic caps well away from heat sources.........A fan was not used as this increases sound distortion by adding background noise which marrs sound quality at low listening levels.
The output stage in DUD120 (on another link on same thread) uses separate driver tubes for each output tube. I’m totally against this as this creates delays due to high frequency propagation and differing driver tube characteristics will increase thd at audio high frequencies. Perhaps it is this that resulted in differing zobel component values for each half. Cathode followers wouldn’t be my driver choice either.
As other readers noticed, the zobel networks in the o/p stage with the particular Hammond 1650 transformer aren’t the same values on each p-p output transformer half. This is a constructors nightmare and a chuck in the deep-end. (For MI who cares.) If the output tranny is designed for Hi Fi then it should have symmetrical sections and halves then the zobel values would be equal and square wave response would be identical on an o‘scope..If the hammond o/p transformer is wound to optimum then the differing values may arise from unequal splitting and driving. So who’s right and wrong ? It may be all down to o/p tranny quality.....and price or rough designed circuitry. If electrostatic LS are used , these may represent considerable ringing in the output stage with difficult loads then the zobel networks become even more important for stability.
Some 40 years ago I was a newcomer to the 100W game and to get the feel.....I upgraded an old GEC 88-50.. 3 stage autobias design which used carbon resistors en masse ...It already had a good British HiFi reputation, so I swapped the o/p stage autobias for fixed bias and upped the psu current capability. There’s plenty of possibilities; one can stay with autobias and go P-PP from here on..On my original there’s enough signal headroom to reach 100W staying with a single o/p pair with fixed bias and keeping around 530V B+.. For the driver, it is well known that a 6SN7 sounds dynamically better than an ECC82.
Real parallel push-pull power for the first time ??...put ones money on a design that is uncomplicated, relatively light weight and reliable....it will sound better than one thinks..the bandwidth and dynamics will be noticeable and sound clearer...My parallel P_P 150W per channel at 100 pound-a-weight-piece is a titanic and awesome constructor project. A similiar 100W design would be somewhat reduced weight. One pays for bandwidth and forget up-ending the chassis on a bench (like a duck) for mods..one makes quite sure that the design is electrically stable in any load...i.e experience gets it right the next time. A question many put to me is at loud listening levels does mine sound better than someone elses amp with half the power,complexity and weight ? Probably yes..massive headroom with lower thd...Ahh.....wait for the deep carnival drum to pass then all is revealed in ones guts as it would on the spot. Many amps would hit clipping on such a deep notes even struggle with the lowest organ pedal. In the 1960‘s with period reflex speakers I had to put my Garrard TT on concrete blocks....now it’s the turn of the CD player. Laser loosing tracking..Yup.
Find a P-PP design or any design that is lightweight......i.e has a LF cutoff around 30Hz is quite understandable. You will still hear the deep drum but have to turn the volume down to avoid amp distorting.The size of the output transformer for 20Hz cutoff (as in mine) is 2 times and a bit bigger than for 40Hz MI application. . One pays for iron and bass expansion and beautiful relaxed 2nd bass harmonics but one can’t over-saturate the laws of physics.
Make your mind up which one you want. I went for my optimised design to how I wanted it to sound. I used balanced input transformers for another reason..... I never use unbalanced inputs for high power. Who does ?
You wanna both loud and clear then it’s going to sound a big hole in your pocket. Long live moderately expanded HiFi.
richj
With 8 power tubes in P-PP alot of heat is generated (in either fixed or auto bias) and if KT88‘s are used then the 4“ separation rule is used with natural convection, the minimum chassis size will be around 18“x18“ (46cm x 46cm) with alot of ventilation holes drilled around the tube bases and well ventilated underneath base plate and electrolytic caps well away from heat sources.........A fan was not used as this increases sound distortion by adding background noise which marrs sound quality at low listening levels.
The output stage in DUD120 (on another link on same thread) uses separate driver tubes for each output tube. I’m totally against this as this creates delays due to high frequency propagation and differing driver tube characteristics will increase thd at audio high frequencies. Perhaps it is this that resulted in differing zobel component values for each half. Cathode followers wouldn’t be my driver choice either.
As other readers noticed, the zobel networks in the o/p stage with the particular Hammond 1650 transformer aren’t the same values on each p-p output transformer half. This is a constructors nightmare and a chuck in the deep-end. (For MI who cares.) If the output tranny is designed for Hi Fi then it should have symmetrical sections and halves then the zobel values would be equal and square wave response would be identical on an o‘scope..If the hammond o/p transformer is wound to optimum then the differing values may arise from unequal splitting and driving. So who’s right and wrong ? It may be all down to o/p tranny quality.....and price or rough designed circuitry. If electrostatic LS are used , these may represent considerable ringing in the output stage with difficult loads then the zobel networks become even more important for stability.
Some 40 years ago I was a newcomer to the 100W game and to get the feel.....I upgraded an old GEC 88-50.. 3 stage autobias design which used carbon resistors en masse ...It already had a good British HiFi reputation, so I swapped the o/p stage autobias for fixed bias and upped the psu current capability. There’s plenty of possibilities; one can stay with autobias and go P-PP from here on..On my original there’s enough signal headroom to reach 100W staying with a single o/p pair with fixed bias and keeping around 530V B+.. For the driver, it is well known that a 6SN7 sounds dynamically better than an ECC82.
Real parallel push-pull power for the first time ??...put ones money on a design that is uncomplicated, relatively light weight and reliable....it will sound better than one thinks..the bandwidth and dynamics will be noticeable and sound clearer...My parallel P_P 150W per channel at 100 pound-a-weight-piece is a titanic and awesome constructor project. A similiar 100W design would be somewhat reduced weight. One pays for bandwidth and forget up-ending the chassis on a bench (like a duck) for mods..one makes quite sure that the design is electrically stable in any load...i.e experience gets it right the next time. A question many put to me is at loud listening levels does mine sound better than someone elses amp with half the power,complexity and weight ? Probably yes..massive headroom with lower thd...Ahh.....wait for the deep carnival drum to pass then all is revealed in ones guts as it would on the spot. Many amps would hit clipping on such a deep notes even struggle with the lowest organ pedal. In the 1960‘s with period reflex speakers I had to put my Garrard TT on concrete blocks....now it’s the turn of the CD player. Laser loosing tracking..Yup.
Find a P-PP design or any design that is lightweight......i.e has a LF cutoff around 30Hz is quite understandable. You will still hear the deep drum but have to turn the volume down to avoid amp distorting.The size of the output transformer for 20Hz cutoff (as in mine) is 2 times and a bit bigger than for 40Hz MI application. . One pays for iron and bass expansion and beautiful relaxed 2nd bass harmonics but one can’t over-saturate the laws of physics.
Make your mind up which one you want. I went for my optimised design to how I wanted it to sound. I used balanced input transformers for another reason..... I never use unbalanced inputs for high power. Who does ?
You wanna both loud and clear then it’s going to sound a big hole in your pocket. Long live moderately expanded HiFi.
richj
I have read this whole thread 10 times and did alot more research this week and i find everything getting more and more complicated. Im glad OLT has been weeded out thus simplifying my brain slightly.
After all the advise i think i should build a 70w/channel amp. My options would be expanded exponentially. But that 400watt looks
um ... I have two worries about it. I can't find the power supply tubes and some tubes arnt even named on that scematic. Second worry is transformers, think they will all have to be custom wound? If the tubes and transformers are hard to come by i will shift my view to a smaller amp. Anyone have any idea how that big one will sound compaired to a 70-100watt?
After all the advise i think i should build a 70w/channel amp. My options would be expanded exponentially. But that 400watt looks
um ... I have two worries about it. I can't find the power supply tubes and some tubes arnt even named on that scematic. Second worry is transformers, think they will all have to be custom wound? If the tubes and transformers are hard to come by i will shift my view to a smaller amp. Anyone have any idea how that big one will sound compaired to a 70-100watt?
Those GEC amp projects
Hi. I'm one of those who bought parts a few years go for an amp project that never got built. It 's actually the GEC circuit on the other page - http://www.bonavolta.ch/hobby/en/audio/kt88_5.htm
Are those amps any good?
Can anyone help me with a parts layout?
Hi. I'm one of those who bought parts a few years go for an amp project that never got built. It 's actually the GEC circuit on the other page - http://www.bonavolta.ch/hobby/en/audio/kt88_5.htm
Are those amps any good?
Can anyone help me with a parts layout?
Re: Those GEC amp projects
Yup. the 50W works. There is alot going for a simple 3 stage design.
1st; Build modification is to delete 1M res from 2nd (follower) ECC83/12AX7 to ground. This will improve CM balance at high frequencies. The ground is accomplished by 4M7's in the interstage HP network via 470K.
2nd; the 12AU7 isn't my favourite sounding tube. A 6SN7 is far better.
3rd; The 25K balance pot in driver stage anodes can be omitted if 2% resistors are used.
4th; For cathode decoupl use switchmode electrolytic caps instead of cheapo types. They are sonically better.
Despite it's simplicity a well designed layout will return 0.2% thd at 50W. That's better than a complicated inter feedback Citation 2.
O/P tranny, readily available.
Drawbacks for beginner;
Ace of hearts, trump slam into 500V B+. Double/series psu caps.
Expect s/n ratio around -65dB below full o/p. (perceptible hiss 30cms from tweeter)
Output stage runs hot; nearly class A at 90mA per tube quies= 40W expect cherry red plates.
O/p stage can be modified to run in fixed bias to give 100W, change o/p tranny/req - bias circuits/PSU upgrade. Front end can stay as it stands
richj
fmw63 said:
Yup. the 50W works. There is alot going for a simple 3 stage design.
1st; Build modification is to delete 1M res from 2nd (follower) ECC83/12AX7 to ground. This will improve CM balance at high frequencies. The ground is accomplished by 4M7's in the interstage HP network via 470K.
2nd; the 12AU7 isn't my favourite sounding tube. A 6SN7 is far better.
3rd; The 25K balance pot in driver stage anodes can be omitted if 2% resistors are used.
4th; For cathode decoupl use switchmode electrolytic caps instead of cheapo types. They are sonically better.
Despite it's simplicity a well designed layout will return 0.2% thd at 50W. That's better than a complicated inter feedback Citation 2.
O/P tranny, readily available.
Drawbacks for beginner;
Ace of hearts, trump slam into 500V B+. Double/series psu caps.Expect s/n ratio around -65dB below full o/p. (perceptible hiss 30cms from tweeter)
Output stage runs hot; nearly class A at 90mA per tube quies= 40W expect cherry red plates.
O/p stage can be modified to run in fixed bias to give 100W, change o/p tranny/req - bias circuits/PSU upgrade. Front end can stay as it stands
richj
The real story...
I was planning on running higher voltages & bought KT90's (maybe they were KT100's), and was thinking of using the power supply from the other schematic for the fixed bias 100w amp. I even got some 700v transformers, but I've since change my mind (maybe...) and will go with stock 50w version, unless there is a way to make the higher powered version work
I was planning on running higher voltages & bought KT90's (maybe they were KT100's), and was thinking of using the power supply from the other schematic for the fixed bias 100w amp. I even got some 700v transformers, but I've since change my mind (maybe...) and will go with stock 50w version, unless there is a way to make the higher powered version work
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.