• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Need help starting

grungeman91

Member
2010-08-18 11:11 pm
I'm trying to design a 25 W tube amplifier with push pull output for my senior project in electrical engineering. I'm using 6l6 for the output tubes and I know I want my output power to be 25 watts, but I don't know how to start designing the amp from there. Any help would be greatly appreciated
 

bob91343

Member
2010-03-11 10:43 pm
Start with the tube manual, which will give you the voltages and currents needed to get the power you want. You will see which class of operation is most suitable.

Next step, design a power supply to provide those voltages. Then see what amplitude signal you need to drive it, and select a tube to do that (assuming you are building a complete amplifier).

You need to select a phase inversion scheme. Then you need to drive the phase inverter with a lower level stage. Finally, the feedback network.

I didn't mention the output transformer but that's part of the design of the push pull output stage.

Much of the project may be dictated by available parts, although nothing would stop you from winding your own transformers.
 
Work Backwards....

Start with the operating conditions of the output tubes, determine B+ voltage and required input swing, then do the input stage/splitter, and finally, the power supply.

Check out Duncan Amps TDSL pages for the 6L6. There is a handy table for typical SE and PP operating conditions.

TDSL Tube search

Your operating choices are Triode, Ultralinear (with appropriate output transformers), and pentode.

While you're there, you may also want to download and use PSUDII (power supply designer II) free software for the design of the PS.

PSUD2

Turner audio also has a great section on designing amps, among the zillions of resources out there.

education+diy

Also, once you get the soldering iron warmed up, there is always this forum.................................
 
Last edited:
If you already have a grasp for how vacuum tubes work (for example in a simple single ended amplifier), you could start reading up about the classic PP amplifier topologies, e.g. the Mullard 5-20 and the Williamson, their advantages and shortcomings.

Then you could start improving these designs- a lot of people now use mosfets, silicon rectification, high voltage regluators, constant-current sources and so forth to optimize the performance of the classic circuits and even provide protection for the expensive NOS valves. You might like to take a look at tubelab.com for an example.
 

taj

diyAudio Member
2005-02-23 8:49 pm
There is an excellent cut-to-the-chase, step-by-step guide called Designing Vacuum Tube Amplifiers, by Randy Couch. The best one I've read yet (with due respect to Morgan Jones.) It's a free e-book found here:

https://acrobat.com/#d=UjgTcfpvfqvboIiQLKn5rQ

The book is targeted at designing tube amplifiers for Jazz guitar, which are as clean as hi-fi amps (purposely low distortion), but usually bandwidth limited (no effort spent getting the response much beyond 50Hz - 6kHz). Clean and dark sounding. But the book certainly discusses methods of extending that bandwidth. It's is my favourite book so far.

Morgan Jones's Valve Amplifiers goes into much more detail and analysis, covering more situations, thus is slower reading if you're in a hurry, but still a must-have.

Pete Millet's website (Technical books online) has a gazillion antique books available to download. Norman Crowhurst is my favourite author among those.

..Todd
 
Last edited:
I'm trying to design a 25 W tube amplifier with push pull output for my senior project in electrical engineering. I'm using 6l6 for the output tubes and I know I want my output power to be 25 watts, but I don't know how to start designing the amp from there. Any help would be greatly appreciated


Dude,

Start with the 6L6GC data sheet. Notice the condition set for PP AB1 using 360 V. on the plate, 270 V. on g2, and a 6.6 KOhm (end to end) O/P trafo. That yields 26.5 W. ;)

The open loop 2% THD appears discouraging, but it's not really an issue. Use ultralinear mode "finals", instead of pure pentode, and that open loop number will come down. Add some GNFB and you'll get down to an inaudible 0.3% THD easily enough. Be certain to place current limiting resistors in the lines leading from the screen grids to the O/P trafo's UL taps.

Edcor's model CXPP60-8-6.6K O/P trafo will do very nicely. Yes, you do need the magnetic headroom that comes with the power handling shown. The low freq. error correction signal GNFB generates requires that magnetic headroom be present.

Soviet surplus 6Π3C-E (6p3s-e) tubes have decent sonics and an attractive price.

You also asked about gain structure. Sufficient open loop gain is needed to raise the -22.5 V. to zero and drive the GNFB loop. "Classic" Mullard style circuitry will do the job for you.
 
How do I design for class AB1 operation? Do I need any extra transformers?

What are the schools teaching new EEs these days? No, you don't need any extra trafos. Class "A" operation means the devices have a 100% duty cycle. Class "B" means the devices have an exactly 50% duty cycle. Naturally, Class "AB" means that the duty cycle is > 50% and < 100%.

The 1 or 2 seen in tube operating classes refers to presence of absence of positive control grid current, Class "x2" means positive grid current is present all or part of the time, while Class "x1" means no positive grid current is present. It's more or less a given that SS operates in Class "x1", which is why you may not have seen the 1 or 2 before. If it says Class "A" ... the 1 is implicit.

The data sheet has your "recipe". All you have to do is follow the directions. :D A fair amount of drive voltage is needed before the tubes alternately enter a cut off state.
 
The open loop 2% THD appears discouraging

No it isn't. For audio finals, it's quite impressive. What counts is the harmonic content, and, unforch, this type likes to make lots of h3 and higher order harmonics. It definitely needs the help of local NFB. Open loop, they sound nasty with some program material.

Use ultralinear mode "finals", instead of pure pentode, and that open loop number will come down. Add some GNFB and you'll get down to an inaudible 0.3% THD easily enough. Be certain to place current limiting resistors in the lines leading from the screen grids to the O/P trafo's UL taps.

I favour parallel local NFB (what a lot of folks around here have taken to call "Schade feedback"). You don't need the special UL OPTs, and you can regulate the screen voltages.

This type also tends to like to Barkhausen oscillate, so you will need screen stoppers as well as plate stoppers. 1K5 on the screens prevents that.
 
I previously recommended Mullard style circuitry. I've uploaded Mullard's seminal 5-20 schematic. Please observe the key features, which are a common cathode voltage amplifier DC coupled to a differential, AKA LTP, phase splitter and a GNFB loop.

Use a different small signal complement, as the EF86/12AX7 combination shown yields too much gain for mating to today's digital signal sources and both types are low transconductance (gm). High gm is protection against slew limiting, in a circuit with loop NFB.

Another improvement is to replace R8 (the LTP tail resistor) with a 10M45S constant current sink (CCS). A CCS in the LTP's tail, instead of a resistor, forces symmetry between the 2 sides.
 

Attachments

  • Mullard 5-20 Schematic.jpg
    Mullard 5-20 Schematic.jpg
    380.3 KB · Views: 83
Next question. What tube and configuration is good for an input stage? Would a simple common cathode triode work?

Highly unlikely, especially if you include NFB.

As far as the front end goes, there are lots of possibilities.

Ex 1
Ex 2

Other possibilities include the "Williamson" topology (uses cathodyne splitter behind a second stage of differential gain). There are various other kinds of splitters, but none as good as either the cathodyne or LTP.