Been thinking lately about the tube amps I have built and have come to the conclusion that the only way to do a tube amp right will require a solid state voltage amp/driver stage driving the tube output....(garbadge in garbadge out) as i cannot stand a noisy amp (hum etc...)
Any suggestions and or schematics on how to do this?
Thanks
Lawrence
Any suggestions and or schematics on how to do this?
Thanks
Lawrence
A solid state voltage amp does not bode well with your statement above. You'll also find very little enthusiasm for such a project among tube diyers.
John, not to be rude but.....whats the point of this fourm????
I think there might be merrit to what im trying to do as the voltage gain/driver stage is everhting in an amp IMMHO. I can see increased dynamics less noise, wider bandwith etc.....
Lawrence
LVH-audio
I think there might be merrit to what im trying to do as the voltage gain/driver stage is everhting in an amp IMMHO. I can see increased dynamics less noise, wider bandwith etc.....
Lawrence
LVH-audio
Increased dynamics?! What are you talking about? A simple common catode tube stage can swing a hundred volt easily. Use a choke/transformer load and this doubles. Low distortion, great bandwith and all this without feedback. Solid state can do that?
Noise? Why is noise a consideration in a high level stage? People build microvolt headamps with tubes...
If you use a SS front end which will probably determine the prevailing sonic character, what is the point of putting up with tube output and transformers?
analog_sa
Please don't be so negative. You might not like the idea, and that's fine, but that is only your opinion.
If you don't like it, just don't post in this thread. We like to encourage and guide new projects, who knows what might come of them?
Please don't be so negative. You might not like the idea, and that's fine, but that is only your opinion.
If you don't like it, just don't post in this thread. We like to encourage and guide new projects, who knows what might come of them?
IMHO, ultimately transistor voltage gain stages have lower noise than tube, but design, component selection and construction influence this more than the device technology, and for a power amp both should offer adequate noise performance.
For any device technology, resistive loaded single ended gain stages do not reject power supply noise so a high degree of power filtering is needed for low hum, current source loaded gain stages can have high power supply noise rejection but the current source itself can add excess noise, Gary Pimm is an expert on current source loads. Differential gain stages have good power supply noise rejection if a current source or high value resistance is used between the common cathodes and the negative power supply rail.
Filament supplies for input tubes should be well filtered and preferably regulated DC for low noise and hum, a floating filament supply referenced to about 50VDC above the cathode helps in hum rejection. For lowest input stage hum a separate filament supply transformer may be needed to avoid circulating capacitive currents between transformer windings especially with capacitive input power supply filters.
Ground management is of utmost importance for low hum, you cannot rely on tying all ground points to a steel chassis,Kuei Yang Wang's post is a brief introduction to the correct approach.
There are some examples of solid state drivers for tube outputs, the Sutherland driver board for Dynaco ST-70 comes to mind, I recall a description of its design using JFETs cascoded with MOSFETs, however I don't have a schematic and have not seen one on the web. SY has posted a hybrid driver design here, search for SYclotron. JFETs cascoded with tubes are commonly used for low level preamps and seem well accepted by tubeophiles, Allen Wright for example.
Since you already have some tube amp projects perhaps you can start by examining your ground management and power supply designs for possible improvement. If you want to start over, I suggest you try whatever approach suits your fancy, you will learn more if you try whatever you have time and patience for.
For any device technology, resistive loaded single ended gain stages do not reject power supply noise so a high degree of power filtering is needed for low hum, current source loaded gain stages can have high power supply noise rejection but the current source itself can add excess noise, Gary Pimm is an expert on current source loads. Differential gain stages have good power supply noise rejection if a current source or high value resistance is used between the common cathodes and the negative power supply rail.
Filament supplies for input tubes should be well filtered and preferably regulated DC for low noise and hum, a floating filament supply referenced to about 50VDC above the cathode helps in hum rejection. For lowest input stage hum a separate filament supply transformer may be needed to avoid circulating capacitive currents between transformer windings especially with capacitive input power supply filters.
Ground management is of utmost importance for low hum, you cannot rely on tying all ground points to a steel chassis,Kuei Yang Wang's post is a brief introduction to the correct approach.
There are some examples of solid state drivers for tube outputs, the Sutherland driver board for Dynaco ST-70 comes to mind, I recall a description of its design using JFETs cascoded with MOSFETs, however I don't have a schematic and have not seen one on the web. SY has posted a hybrid driver design here, search for SYclotron. JFETs cascoded with tubes are commonly used for low level preamps and seem well accepted by tubeophiles, Allen Wright for example.
Since you already have some tube amp projects perhaps you can start by examining your ground management and power supply designs for possible improvement. If you want to start over, I suggest you try whatever approach suits your fancy, you will learn more if you try whatever you have time and patience for.
Hi Acoustat,
Don't let anyone put you off.
I built a fantastic 4x KT88 AB1 40% Ultralinear this way some 30 years ago. It had part self + part fixed bias.
It is still in my roofspace ... ! somewhere !
Each 2x KT88 output half was driven by 1/2 a Mullard ECC82 (12AU7) cathode follower, which was directly coupled to and bootstrapped its first 1/2 operating as conventional common cathode amplifier. The bootstrap completed the 500uF+1/4A choke+500uF smoothed power rail rejection at the driving amplifier. Yes I was operating beyond manufacturers ratings, but genuine Mullard's were okay.
I measured to the ECC82 cathode outputs without the output valves, and without any NFB found less than 0.1% distortion for 100Vrms drive, which was far more than the KT88s would ever need. The amplfier was humless (no 6.3Vac at low level circuitry), and almost hissless without signal input.
Each ECC82 input grid was fed by the collectors of a simple single differential stage, and there was enough spare gain for NFB. The transistors were powered by the output stage fixed grid bias supply.
Sure we still use differential input stages in high power SS amplifier designs, and it is not this stage that imparts characteristics. Besides, the input transistor stage was so much faster than the following valve stages or any alternative input tube type that they did not - could not - impart any sonic character.
This was my reference amplifier for many years, and I felt particulary decieved by Quad when their supposedly 'ideal' (at the time) 405 amplifier was so inferior beside it. The Quad was loud enough alright, and okay for pop parties, but for serious listening - yeough.
I did try driving output valve grids with BF259s but they were not as clean with the grid loading.
Cheers .......... Graham.
Don't let anyone put you off.
I built a fantastic 4x KT88 AB1 40% Ultralinear this way some 30 years ago. It had part self + part fixed bias.
It is still in my roofspace ... ! somewhere !
Each 2x KT88 output half was driven by 1/2 a Mullard ECC82 (12AU7) cathode follower, which was directly coupled to and bootstrapped its first 1/2 operating as conventional common cathode amplifier. The bootstrap completed the 500uF+1/4A choke+500uF smoothed power rail rejection at the driving amplifier. Yes I was operating beyond manufacturers ratings, but genuine Mullard's were okay.
I measured to the ECC82 cathode outputs without the output valves, and without any NFB found less than 0.1% distortion for 100Vrms drive, which was far more than the KT88s would ever need. The amplfier was humless (no 6.3Vac at low level circuitry), and almost hissless without signal input.
Each ECC82 input grid was fed by the collectors of a simple single differential stage, and there was enough spare gain for NFB. The transistors were powered by the output stage fixed grid bias supply.
Sure we still use differential input stages in high power SS amplifier designs, and it is not this stage that imparts characteristics. Besides, the input transistor stage was so much faster than the following valve stages or any alternative input tube type that they did not - could not - impart any sonic character.
This was my reference amplifier for many years, and I felt particulary decieved by Quad when their supposedly 'ideal' (at the time) 405 amplifier was so inferior beside it. The Quad was loud enough alright, and okay for pop parties, but for serious listening - yeough.
I did try driving output valve grids with BF259s but they were not as clean with the grid loading.
Cheers .......... Graham.
SS vs tube
Sometimes I was talking with chiefdesigner from KR Enterprise and he said to me, that in their amps is all ( except output tube ) SS, 'cos it give better sonic results. 😉 " Let's think make it better " ( Philips 😎 ) and don't stay in " old fashion " 😎 .
Sometimes I was talking with chiefdesigner from KR Enterprise and he said to me, that in their amps is all ( except output tube ) SS, 'cos it give better sonic results. 😉 " Let's think make it better " ( Philips 😎 ) and don't stay in " old fashion " 😎 .
Notes and Questions
Acoustat,
There was a good SS front end, Ultralinear connected EL34 Output design in Electronics World about 12 months ago. Used Lundahl Output Transformers.
Graham,
Can you describe the bootstrap connection from the direct coupled cathode follower back to the common cathode previous stage in your KT88 Amp - there are many ways to skin the cat..
Did you split the load resistor of the common cathode stage and capacitively couple the cathode follower output to the junction of the 2 resistors?
OR
Did you sit the cathode follower on top of the common cathode stage ie a Mu Stage connection?
OR
Did you AC couple the cathode follower output back to the cathode of the Common Cathode Stage?
OR
Something else entirely?
Cheers,
Ian
Acoustat,
There was a good SS front end, Ultralinear connected EL34 Output design in Electronics World about 12 months ago. Used Lundahl Output Transformers.
Graham,
Can you describe the bootstrap connection from the direct coupled cathode follower back to the common cathode previous stage in your KT88 Amp - there are many ways to skin the cat..
Did you split the load resistor of the common cathode stage and capacitively couple the cathode follower output to the junction of the 2 resistors?
OR
Did you sit the cathode follower on top of the common cathode stage ie a Mu Stage connection?
OR
Did you AC couple the cathode follower output back to the cathode of the Common Cathode Stage?
OR
Something else entirely?
Cheers,
Ian
Hi Ian,
Okay, this is 30yr memory because I have not been able to find my notebook or the amplifier since the birth of my son, which meant that the roofspace became storage with a top layer of kiddies/teenage toys/books etc.
I used a solid state rectified HT rail with capacitve charging delay to the driver tube, which would still have strained the second halves of each ECC82 at start-up, but I did not observe any problems.
Possibly 350V, to give 150V p-p output voltage per ECC82.
I split the anode resistor (for the first half common cathode) and cannot remember the exact voltage drops. I would have had equal resistors to start with, but cannot remember whether I left them like that.
The first half had simple cathode self bias, but I don't remember using a decoupling capacitor.
I directly coupled the second triode grid to the first triode anode and had an appropriate cathode resistor, the second cathode voltage must have been between 120 and 150 volts. The three resistors were large 5% carbons.
The second anode went to stage HT while the second cathode was bootstrapped to the split anode resistor junction via a high voltage electrolytic.
There was one ECC82 per push-pull half, each driving 2x KT88 grids via separate 400V polystyrenes, possibly 1uF or 2,2uF (four in total) for I still see these in my parts boxes, and the KT88 grid resistors were either 22k or 47k.
This arrangement was surprisingly stable and linear, and there were not any reliability problems. I know I was on the limit for grid voltage during switch-on, and heater-cathode during operation, but it ran trouble-free.
For sure I'd have to increase the supply voltage very slowly if I ever do turn this amplifier on again.
Cheers ......... Graham.
Okay, this is 30yr memory because I have not been able to find my notebook or the amplifier since the birth of my son, which meant that the roofspace became storage with a top layer of kiddies/teenage toys/books etc.
I used a solid state rectified HT rail with capacitve charging delay to the driver tube, which would still have strained the second halves of each ECC82 at start-up, but I did not observe any problems.
Possibly 350V, to give 150V p-p output voltage per ECC82.
I split the anode resistor (for the first half common cathode) and cannot remember the exact voltage drops. I would have had equal resistors to start with, but cannot remember whether I left them like that.
The first half had simple cathode self bias, but I don't remember using a decoupling capacitor.
I directly coupled the second triode grid to the first triode anode and had an appropriate cathode resistor, the second cathode voltage must have been between 120 and 150 volts. The three resistors were large 5% carbons.
The second anode went to stage HT while the second cathode was bootstrapped to the split anode resistor junction via a high voltage electrolytic.
There was one ECC82 per push-pull half, each driving 2x KT88 grids via separate 400V polystyrenes, possibly 1uF or 2,2uF (four in total) for I still see these in my parts boxes, and the KT88 grid resistors were either 22k or 47k.
This arrangement was surprisingly stable and linear, and there were not any reliability problems. I know I was on the limit for grid voltage during switch-on, and heater-cathode during operation, but it ran trouble-free.
For sure I'd have to increase the supply voltage very slowly if I ever do turn this amplifier on again.
Cheers ......... Graham.
Graham,
Thanks for the description of your bootstrap arrangement. I'll give it a try.
Acoustat,
I went thru' my stack of Eelctronics World last night. The article you would be looking for was called "Improved Hybrid Amplifier". It had 15dB of AC feedback plus an idle current balance DC feedback path so that a Amplimo (Van der Venne) Toroidal Output Transformer could be used. Specs were excellent. I meant to bring in in today so I could give you the exact volume and issue number and forgot BUT it was 2nd half of 2003.
Cheers,
Ian
Thanks for the description of your bootstrap arrangement. I'll give it a try.
Acoustat,
I went thru' my stack of Eelctronics World last night. The article you would be looking for was called "Improved Hybrid Amplifier". It had 15dB of AC feedback plus an idle current balance DC feedback path so that a Amplimo (Van der Venne) Toroidal Output Transformer could be used. Specs were excellent. I meant to bring in in today so I could give you the exact volume and issue number and forgot BUT it was 2nd half of 2003.
Cheers,
Ian
- Status
- Not open for further replies.