If we believe Morgan Jones when he writes in "Valve Amplifiers" of the 5-20:
"The input stage is an EF86 pentode, which is responsible for the high gain but poor noise performance...."
"The cathode coupled phase splitter is combined with the driver with the driver circuit using an ECC83...... The output stage has an input capacitance of ~30pF, and combined with the 53k ohm output resistance of the driver stage, this gives a cut-off at ~100kHz, which is quite poor."
"....the driver stage has only 10dB of overload capacity. When the output stage gain begins to fall, for whatever reason - cathode feedback, input capacitance loading the driver, or primary inductance in the output transformer - the global feedback loop will try to correct this by suppling greater drive to the output stage, and the 10dB margin will be quickly eroded, raising distortion."
Jones goes on to discuss the output tube's individual biasing resistors and their bypass capacitors, "The capacitor is a short circuit to AC, and so prevents feedback, but its reactance rises at very low frequencies, so it is no longer a a short circuit, and allows feedback. Because the output stage is load matched, this feedback causes an immediate rise in distortion and reduction of power output due to the mismatch."
There are other points as well, I've just chosen a few.
Isn't it time we moved on from the Mullard 5-20?
Cheers, Steve
"The input stage is an EF86 pentode, which is responsible for the high gain but poor noise performance...."
"The cathode coupled phase splitter is combined with the driver with the driver circuit using an ECC83...... The output stage has an input capacitance of ~30pF, and combined with the 53k ohm output resistance of the driver stage, this gives a cut-off at ~100kHz, which is quite poor."
"....the driver stage has only 10dB of overload capacity. When the output stage gain begins to fall, for whatever reason - cathode feedback, input capacitance loading the driver, or primary inductance in the output transformer - the global feedback loop will try to correct this by suppling greater drive to the output stage, and the 10dB margin will be quickly eroded, raising distortion."
Jones goes on to discuss the output tube's individual biasing resistors and their bypass capacitors, "The capacitor is a short circuit to AC, and so prevents feedback, but its reactance rises at very low frequencies, so it is no longer a a short circuit, and allows feedback. Because the output stage is load matched, this feedback causes an immediate rise in distortion and reduction of power output due to the mismatch."
There are other points as well, I've just chosen a few.
Isn't it time we moved on from the Mullard 5-20?
Cheers, Steve
50-20 demise
Indeed! For audio its relatively poor choice; its beaten hands down by a circuit such as the Williamson class A triode design. By comparison the 5-20 is upper middly and lacks top definition. Having built 5-20's over 55 years ago, they quickly became Williamsons with el34's triode connected and the ef86 and 12ax7 B9A socket holes drilled out to octal and 6sn7's added. An extra small choke was added as per the Wiilliamson circuit. These amps remained unchanged for about 25 years; eventually converted to KT88's with a new output transformer (A&R 4008). They now sit in my workshop, having been replaced by Hiraga class A transistor amps (these are very, very good).
There is no case for valve amps these days!
Indeed! For audio its relatively poor choice; its beaten hands down by a circuit such as the Williamson class A triode design. By comparison the 5-20 is upper middly and lacks top definition. Having built 5-20's over 55 years ago, they quickly became Williamsons with el34's triode connected and the ef86 and 12ax7 B9A socket holes drilled out to octal and 6sn7's added. An extra small choke was added as per the Wiilliamson circuit. These amps remained unchanged for about 25 years; eventually converted to KT88's with a new output transformer (A&R 4008). They now sit in my workshop, having been replaced by Hiraga class A transistor amps (these are very, very good).
There is no case for valve amps these days!
The topology used is good even if the implementation has questionable choices. Eli often points out the HK Citation V as an example of a good implementation of the topology.
I built a 5-20 last year, then quickly modified it. That had more to do with stability problems than the actual sound and performance. It definitely has way too much gain as standard. I ended up replacing the PI with a 6CG7 making it similar to an Eico HF-50, except mine has a CCS in the tail.
I built a 5-20 last year, then quickly modified it. That had more to do with stability problems than the actual sound and performance. It definitely has way too much gain as standard. I ended up replacing the PI with a 6CG7 making it similar to an Eico HF-50, except mine has a CCS in the tail.
If you may think so, then its highly subjective thinking.
I think exactly the opposite way. With me, there is no case for transistor amps anymore. Build a few of them and owned some more. Always sterile sounding, up to the highly famed Mark Levinson & Co. models. No life or any life sucket out of those mega buck models, best thing is to switch them off and enjoy the silence.
The very best tube amp models can sound vibrant, full of life and swing, energy that proofs every transistor amp as a faulty model by design and with silky highs and thunderfull lows that no transistor is able to deliver until today. Its just a shame that this powerfull technology maybe hasn't been used to its edge and after all those fails people give their subjective rating on audio forums just showing their inability to use a given technology to its very best.
I don't know whats the best in transistor technology, but I studied electronics and know some of their state of the art circuits and they mostly sound boooooring to me, typical transistor sound. Technically speaking, the transistor was never invented to sound high quality audio like. It took decades before the first transistors were listenable in high end audio. Its exactly the same as with digital technology. It took years and years to crawl from the stone edge to something that is listenable, but with respect to the best analog sources still sucks because it contains the same design flaws like transistorized audio equipment.
Much too overengineered, much too complicated (look at all those dozens of negative feedback and regulator loops in every transistor amp) and invented with a theory behind it, that didn't worked out in real life like it was planned by the designers. Transistors were invented to safe energy and heat, to make things smaller and more comfortable and they have its own right in computers, mobile phones and all those electronic gear where its high priority to be small and low weigth with much computer calculation power. In high end audio they will never cut the edge of whats possible, even when people who have never touched this edge spread the sermon of the technically stone age tube technology.
Btw, to implement some technical equipment is often the same art like its to design a component. Most of the people claiming something sounds worse werent able to enjoy it 100% because of those poor matching with signal sources or other components.
Last edited:
No it isn't. By Mullard 5-20 most people mean the topology, not an exact copy. You're quoting objections to the implementation, specific to the tubes and operating conditions chosen by the Mullard designer. The HK and other implementations are better. With 2021 knowledge and technology you can design much better amps using the basic topology. In fact I find that it can do as well as any topology.
Then again, I'm not looking for the most exotic PPPP SE OTL design using magnetrons
Morgan Jones presents a few "iffy" comments sometimes. For example, in his 3rd Ed. I don't agree with some of his Williamson statements such as class AB1, and his calculations for first stage shelf frequency.
Steve, as you state, you have only quoted a few selective comments from MJ from 2003, but imho have missed the opportunity to clarify when, where and why any stated disadvantage occurs, and rather just gone for a flourish finish about moving on from a 1959 design. It's now 2021, so 18 years after MJ's comments, and 60 years after the original amp.
At least MJ tries to substantiate some of the design aspects of the 5-20, and sometimes goes further in to discussing some improvements, but in such a book there is little opportunity or incentive to pursue the topic of subtle modifications, or to back up any statements with actual measurements (eg. noise floor, or when distortions rise).
Steve, as you state, you have only quoted a few selective comments from MJ from 2003, but imho have missed the opportunity to clarify when, where and why any stated disadvantage occurs, and rather just gone for a flourish finish about moving on from a 1959 design. It's now 2021, so 18 years after MJ's comments, and 60 years after the original amp.
At least MJ tries to substantiate some of the design aspects of the 5-20, and sometimes goes further in to discussing some improvements, but in such a book there is little opportunity or incentive to pursue the topic of subtle modifications, or to back up any statements with actual measurements (eg. noise floor, or when distortions rise).
Last edited:
Granted, there are better ways of implementing a 5-20. Getting rid of the EF86 for a triode - nobody these days needs the input sensitivity that an EF86 yields - and ditching the 12AX7 as a phase splitter is a good place to start.
I avoid 12AX7s as much as possible. One designer some years ago called it a tube for cheap phonographs. In my experience modern 12AX7s expire very quickly even running at a small fraction of their dissipation limit.
But some people are still building slavish copies of the 5-20. Makes no sense to me. And if you are building an amp that uses EL34 or higher dissipation tubes fixed bias is the way to go. IMHO.
S.
I avoid 12AX7s as much as possible. One designer some years ago called it a tube for cheap phonographs. In my experience modern 12AX7s expire very quickly even running at a small fraction of their dissipation limit.
But some people are still building slavish copies of the 5-20. Makes no sense to me. And if you are building an amp that uses EL34 or higher dissipation tubes fixed bias is the way to go. IMHO.
S.
Exactly! Even constrained by cost considerations, Stu Hegeman came up with a winner. Hegeman employed high gm small signal types that allow a "simple" global NFB loop to perform well. Unlike Williamson style, Mullard style is not highly vulnerable to low frequency instability. While there's no changing the reality of the better the "iron" -- the better the amp, Mullard style topology reliably squeezes that which is available out.
Switching the LTP from a 6CG7 to an ECC99 further emphasizes gm, without substantially altering overall gain structure. A CCS as the tail load of a LTP is an idea that's fast becoming ubiquitous.
If you want a voltage amplifier that employs an in production type, the provided 6922 cascode is "just what the doctor ordered".In a new/from the ground up, build, a single 6922 could service both channels. Use hybrid cascodes, with the triodes at the bottom of the "totem poles" and high voltage MOSFETs as the top, constant current, elements.
Attachments
trobins,
I had considered a 5-20 twenty years ago but dismissed it even then as somewhat of an antique even then. Jones just sort of confirmed my view. Whether or not you buy into his reasoning is open for debate.
Yes, I had been seeing recent forum posts and I just wanted to stimulate a discussion and perhaps learn something in the process.
S.
I had considered a 5-20 twenty years ago but dismissed it even then as somewhat of an antique even then. Jones just sort of confirmed my view. Whether or not you buy into his reasoning is open for debate.
Yes, I had been seeing recent forum posts and I just wanted to stimulate a discussion and perhaps learn something in the process.
S.
For every complex question there is a simple answer- and its mostly wrong.
When looking at the 5-20, I see a classic tube amp, which can be, of course, bettered by this or that particular different circuit detail.
Does this make a classic tube amp a bad one? I don't think so.
But when looking at the build and parts quality today, something has been changed.
So making it sounding more perfect, like a computer PC transistor circuit won't be a problem, but to re-build the original will be nearly impossible today.
So how can one judge an amp that mostly never heard in its original creation?
And when every detail has been changed, how can this sound as good as the original?
People who never heard the original in a perfect implementation try to better something they even have no knowledge from but just the circuit which means nearly nothing.
Gray, dear friend, is all theory but green of lifes golden tree.
Faust, Mephistopheles, Studierzimmer
When looking at the 5-20, I see a classic tube amp, which can be, of course, bettered by this or that particular different circuit detail.
Does this make a classic tube amp a bad one? I don't think so.
But when looking at the build and parts quality today, something has been changed.
So making it sounding more perfect, like a computer PC transistor circuit won't be a problem, but to re-build the original will be nearly impossible today.
So how can one judge an amp that mostly never heard in its original creation?
And when every detail has been changed, how can this sound as good as the original?
People who never heard the original in a perfect implementation try to better something they even have no knowledge from but just the circuit which means nearly nothing.
Gray, dear friend, is all theory but green of lifes golden tree.
Faust, Mephistopheles, Studierzimmer
That's good. But if you want to discuss the merits of Manitoba, you don't start by saying "I heard it's a dump, why would anyone ever want to live there?"
I had an original pair of Mullard 5-10's I received from my grandfather (he built them in the early 1960's following the Mullard chassis and layout instructions). They had way too much gain and were quite noisy, so the Partridge transformers were used as donors for something else. But if I kept them untouched I'm sure they would be worth more if I was to sell them, even though what I built has much better performance and is more useable with modern sources.
Remember, the goal for Mullard fpr publishing the 5-20 and 5-10 designs was to advertise their then new tubes ECC83, EF86, EL34, EL84, GZ34 and EZ81. Hence these amplifiers had to be easily reproducible, and their performance surely could well compete with other designs of those days, given that apt iron was used. I guess they were, and still are, built by the millions.
Best regards!
Best regards!
This means, to give the machine a perfect environment to show its virtues.
With a tube power amp it means, to match input and output impedances and to load it with a speaker that shows the matched load impedance over all of the frequency and show an efficiency in dB which is appropiate to the matched amp.
And, second, but not least important, to audition an amp and a speaker in an environment, that is acoustically optimized for flat sound response with nearly no echoes responding from walls.
In most cases, those circumstances are not being given in a listening shootout and afterwards, one of the components are given a bad rating.
A very unfair method to rate equipment and a false one, too.
You have auditioned an amp thats 50 years old with worn out electrolyts and wonder why it performs badly and has hum?
Do you ever have one idea that this might not be the appropriate method to test something and rate it? Or do you simply have no clue about whatsoever in old tube amps could make them sound worse?
Your rating is simply false and I bet even the speakers and the source didn't match the setup that you used for a qualified rating of components.