• 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.

Conflicting 807 plate impedance

I am having a great deal of trouble finding out the plate impedance of the 807 tube.

I am in the process of building a Williamson amp, and I somewhat remember reading that the plate to plate impedance of this amp was 10kOhm.

Now this is in a push-pull configuration, but nowhere can I find this information to collaborate this. I have read 6kOhm, and various other values.

Is the plate to plate impedance, the sum of the two tubes being used?

And why can I not find this information clearly?
 
Look up the info on the valve.
It's very straightforward.
A-A load very strongly depends on Anode voltage used and class A, AB1 or AB2.

Think of it as a 6L6, (6L6GC even) look that up, and work from there or 5B254M...

You can use the KT66 as a basis or the STC 807.
Triode connection AB1 cathode bias say 4K A-A at 440V, 2,5K at 270V,
(G2 absolute max on 807 is 400V,- less than KT66).

STC 254M /807 says 400V 3K A-A.

AB1 beam Tetrode cathode bias say 7K at 450V, or 9K at 360V fixed bias is always lower.
Fixed bias like 6L6GA, 360V on anode is 6k6.

Point about the 807 and 5933, it's designed to be driven in AB2 so up to +20V on the control grid at which point it can deliver anything from 75-120W per pair dependent on anode voltage, and A-A loads anything from 3k-6k.
 
Last edited:
The original Williamson amp used KT66 wired as triodes (no top cap), at which point it works like a DHT PX25.

The typical application wired how Williamson suggests, is triode strapped at 400V (MAX), 3K A-A, and will give you a max of 15W with lots of 3rd harmonic, (which he fixed with masses of NFB, and a strong attempt to stop the darn things oscillating to destruction).

TBQH I don't see the point of using 807s like that,- maybe the 5B255M but not a TC type valve.
Hence why DTNW opted for the British made KT66, derived from the lousy RF, badly misbehaving KT8 which had a TC.
(After that point there was no longer any market for the orphan KT8/KT8C).

The USA opted for 807s because there were mega zillions of them going for next to nothing post 1945.

NB:-
(Basically AB1 conditions are double the AB2 A-A load, so to give you any idea my Bogen AB2 amps run A-A loads of 3K2 at 600V or 5K2 at 835V. simply put that means in AB1 you should aim at 6k4-7K for 500-600V, and 9-10K at 800V).

Get the 43 page STC PDF, otherwise you are wasting your time and ours.
 
Last edited:
Ahum.
None of those threads mentioned actually is any good at answering the main question.
One wonders if Williamson actually ever read some of the valve manufacturer's graphs and documentation at all?

He then claimed belatedly you could get lot more power by running PPP, by which time of course the Americans with Westinghouse and Sylvania, were into ultra linear and making nice stuff like the 7591 and eventually the 8417, which all run rings around the old low gain high distortion 807 and KT66.
Even Mullard/Leak/Quad/Radford worked on the EL34 and EL84 specifically to put the Williamson design into the dustbin of history by using higher gain pentodes.

I've worked a lot with 807s.
The whole idea of driving them with anything other than a cathode follower is pretty silly, for the precise reason they work best full DC coupled,- which of course avoids all the instability prone phase shifts inherent in DTN's AC coupled designs, as well as making them far less prone to turning them into a short wave FM transmitter by default.
 
Last edited:
If you can find an 8417 that is not gassed!!!!!

I have at least 20, 8417.
Not one of them has gas problems, so I don't know what you are talking about.
Lots of them do however have weak emissions.

In the sample of 20 or more 807 I bought there were worse problems than in any of my 8417s.
The production spread was enormous, 1 had a heater/cathode short, lots of them red plated, and at least 4 of them had loose TC which came off in my fingers.

It's good to know what you are talking about with valves.

My 2 modified Bogen monobloc 8417 based amps are used daily, have much lower distortion than anything I have seen from Williamson and only use CFB, NO global NFB at all despite having dedicated OPT windings for it.
The valves are perfect and 60 years old with who knows how many hours behind them!

They are far more stable than anything from DTNW, and of course use a better lower impedance "shortened Williamson" without the extra pre OPV driver pair which causes all the known problems and instability.

In the other 2 BIG BEASTS, - Bogen H050, I really struggled to get THD under control.
They use AB2 driven 807s at 500V.
I totally changed the lousy fixed bias system and made a totally different NFB loop which altered them out of all recognition. (they originally used the 140V winding for NFB for just part of the input loop, I altered this to earth the 140V CT, and derive 2 counter phase NFB loops with far less NFB than the original amongst other stuff).
That all proves a lot of the distortion in valve amps comes from unstable screen grid supplies, unstable bias, and sagging main PSU.

FYI:-
DTNs criticism of the paraphrase invertor is mostly nonsense.
It permits fully balanced line inputs if you do it right with a switch in/out option.
With his design you cannot do that.

With a switched in/out paraphrase, I could compare the phase inverting perfomance of the sound card OP amps v that of the internal valve version. On a distortion meter once the valve version correctly adjusted and balanced, there was NO measurable or audible DIFFERENCE.)

The output waveform of the H050 807 amps at full power leaves most DTN stuff in the dust, and this being at FULL power, not at a measly 15W where DTN's THD goes out of control, but at 50 where it was totally clean in AB2, so no nasty things happen approaching 0V on the grids.

I want to see any 70yr old amp make those kind of figures, never mind a Williamson.
So far I haven't even seen a modern "hi end" get even close... 🙄, PLUS they blow up within a minute if run as I do, full power square wave tests for 15-20min.

(the "hi end crew" are all "fake it" measure it at 1W stuff).

As I say, I can't imagine why anyone would want to connect a 807 as a triode, never mind claiming some sort of big advantage for deliberately running a KT66 this way.
Maybe it's because the KT66 is well known for producing vast amounts of distortion?

Btw, I tested my 807 based amps before and after swopping to KT8C (a KT66 in reality) and couldn't detect any difference in sound whatsoever. (much to my annoyance after all that work locating perfect s/h original B5 bases and doing all the modification work!)
 

Attachments

  • HO50_1khz_full_power.jpg
    HO50_1khz_full_power.jpg
    138.9 KB · Views: 238
  • HO50_1khz_sq_w_60W.jpg
    HO50_1khz_sq_w_60W.jpg
    139 KB · Views: 250
Last edited:
That is the 43 page doc I was referring to.
STC (Kent) were a great company producing the very last loctal 807 variants.
They sold mostly to military, and were never popular for audio.

Their docs also apply to their miniature "807s" which are available by the 1000 for peanuts.

I wouldn't even bother with anything else as you can fit a full quad of 5b2xxm in the space that 2 807s originally occupy, and it's far more reliable with much lower interelectrode caps.

That's why I bought 70 NOS Mcmurdo Loctal sockets......forget UX5, go miniature!
 
I am fairly sure that is the old Brimar design.
It's a bit too complex and is sensitive to oscillation and OPT quality.

I made one of those in school as a bass guitar amp.
It made about 80W and sounded pretty good.

ie.
You have to be extremely careful with the quality of the Anode 47ohm resistors.
If one of them dies, (cos of oscillation usually) so will the 807 that goes with it...the screen grid going white hot and collapsing in a pile of molten metal.
 
So Sakers, did you see the 1700 ohm plate resistance level for 807 at 200V Vak at 40mA in triode mode in the STC report? With rp increasing to over 30kohm in tetrode mode. Is that what you were after in initial post, or was it just a particular design PP level used, such as when Williamson chose 10k PP output transformer primary impedance as reflected from the secondary side loading for his amp?

A designer doesn't have to choose a particular datasheet's standard parameter benchmark setup levels, and that can very much depend on how well they want to remain in class A and in a low distortion operating range.

Note that Williamson wasn't entertaining class AB or any such variants, and left it for others (like Radiotronics) to try other similar valves like the 807. So if you are planning on using 807's, then I recommend also reading the Radiotronics reports, as they were the instigator to the 807 being rated for the 400V screen level, and gave a reasonable set of operating parameters and measurements.

If you are straying well off Williamson's path, as 6Vheater banters on about, then you are not building a Williamson. And if you are building a Williamson then you really do need to start with an output transformer that meets or exceeds the best known commercial examples of what has been shown to work well, or you may encounter issues.

It is interesting to appreciate why Williamson chose his topology, and that can be a help when reading broadbrush claims and raves from others about how the Williamson design can be improved. Even some history awareness about DTN working in MOV's applications lab at the time he finished off the prototyping and initial demonstrations of the amp can usually shut down drivel comments suggesting he may never have looked at a manufacturers docs at all. DTN was a very smart cookie - he prepared the Williamson amp design by the time he was just 22, and one doesn't go on to achieve a Royal Society fellowship by sitting on your hands.
 
Last edited:
Note that Williamson .....left it for others (like Radiotronics) to try other similar valves like the 807.

.... if you are building a Williamson then you really do need to start with an output transformer that meets or exceeds the best known commercial examples of what has been shown to work well, or you may encounter issues.

It is interesting to appreciate why Williamson chose his topology, .....Even some history awareness about DTN working in MOV's applications lab at the time he finished off the prototyping and initial demonstrations of the amp can usually shut down drivel comments

DTN was a very smart cookie - he prepared the Williamson amp design by the time he was just 22, and one doesn't go on to achieve a Royal Society fellowship by sitting on your hands.

I love the myths and legends from down under.
PERHAPS, you could do with a trip back home to get your facts right?

1/ Williamson did nothing original at all.
He copied and pasted someone else's design, claiming it was his own work.
In the typical 2 faced class ridden Britain of the 30s this was par for the course.

Partridge even say so:-
The
Partridge Amplifier Circuits’ manuals from 1938. Describes a 40dB feedback 15W amp but no details of output transformer used.
His main marketing innovation was reduction of distortion, by introducing masses of sound deadening NFB, which was also not new, as well as opting for a triode input stage and a "concertina" type phase splitter.

It had to use the extra driver stage because nothing from MOV/GEC bottom drawer had anything even close to enough gain (unlike Mullard).
The extra driver stage eventually went obsolete in the 1960s when much higher gain valves appeared, like the 8417 & 7591.
In fact the Bogen MO200 is in fact a shortened "williamson" without the instability hassles.
It does all the "willy" stuff in a very simple 7247 (12DW7), and needs no phase messing extra driver or huge NFB values.
(Unsurprisingly Quicksilver cloned it down to the last nut and bolt, but introduced the dreaded compression inducing. valve melting Ultralinear stuff tapped into the mix.)

The DTN idea was actually quite stupid because it was well known that speakers & sound sources at that time were absoutely unable to give the resolution required for performance compatible with an amp making only 0.2% THD, but there you go, legends live on.....and wireless world was a great magazine.

It was the first application of PR & marketing that I can think of post war.
It's not what actually happened that makes history, it's what people believe happened.

Luckily for DTN, there was a London based transformer manufacturer called Partridge, who made the best transformers in the world for many decades, and the BBC literally around the corner with a huge latent demand for high quality broadcast audio.
Wireless world advertised all this stuff.

The BBC knew very well, the future was moving from 4V heaters (the vast majority of radios at the time), to 6V miniaturised high efficiency designs which Mullard had already started churning out by the million.

The BBC went their own way, even designed their own speakers and opted for amplifiers which ultimately didn't use the Williamson designs.
I wonder why?

It must be also noted it was the Germans during the war that invented the tape recorder, so that the German Rundfunk could broadcast Hitler's speeches as though they were live.
Those in the west post war, that saw these tape recorders for the first time were astonished.
Once copied all over the world it was the tape recorder propelled high quality audio along, not the amplifiers, Along came 50s-60s rock and DTNs amp was totally unable to compete with the needs for 50-100-200 and 500W amps.

2/ Why did he chose the KT66 which was a Hammersmith Marconi-Osram design?
Because he knew full well GEC had a totally failed 807 on their hands, which the military found to be utterly useless. (It wouldn't run at over 15mhz!).

The original amp was based on the only things that Osram and the GEC conglomerate had, which was the PX series of triodes,- another 4V heater design.
They were already obsolete in the 1930s.

They were desperate at Hammersmith because of Philips/Mullard 6.3V EL37 and EL31.
Philips/Mullard had the rights to the pentode, which was making far better figures, so that they needed someone to help them compete, as their tooling was not up to the precision of the Americans.

They were at that time clearly unable to produce a 807 (which the USA had perfected, and which enabled PA manufacturers immediately post war to start making 807 based amps hitting 120W quite easily).

The KT66 was found to behave very much like the PX25 once triode strapped, and had a modern octal base, which the 807 didn't.
The PX25 is directly heated and feeding the filament to minimise hum is not straight forward. Williamson, who worked for M-OV and then Ferranti, indicated that the modern indirectly heated KT66 when strapped as a triode was almost identical in performance to the PX25.

3/ Blumlein was the "smart cookie". He even had been working and testing the "ultra linear" concept, as well as developments to introduce stereo sound.
Had he not died during his testing of the magnetron, Williamson would certainly have been eclipsed completely.
I have a lot of respect for Blumlein.
Most of his inventions are vital parts of our lives today.

4/ Williamson developed some sort of ridiculous snobbery over the years that followed condemning improvements to "his idea" as being "non-purist".
When Keroes claimed their (Blumlein based and other) patents were original work you can imagine how Williamson talked about those "damn colonials" again, competing against this "purist" English work, and how bad it was.

I reckon snobbery dies hardest especially in England.
All you need is an Eton education and you can even be a Bojo 🙄
 

Attachments

  • willy_part.jpg
    willy_part.jpg
    343.1 KB · Views: 169
Last edited:
Here was Mullard's idea, a high gain pentode:-
mullard_el31.jpg

which much like the EL34 that was developed from it, would blow up and melt. (thanks to secondary emission).
Some sort of myth carries on about Blackburn as a result.
The early EL34 had a metal ring octal base which used to squeeze the glass and make it shatter.
The British do have a unique way of being able to screw stuff up, and it's still in fashion*.

MOV/GEC made this thing which luckily doesn't melt, but is not as linear & on an obsolete usually ceramic base.
The base plays up and usually comes loose.
That ended up being the KT66 which sells for totally silly money.
gb_marconi_kt8_1.jpg

It's easy to see today how the EL34/6L6 wars played out, and why they are still being made 70yrs on.
Williamson's amp still carries on generating myths and legends much like all that other British museum stuff.
Long may it carry on producing knighthoods from her majesty.
(*but not publish their names and addresses)
 
Last edited:
DTN was making and testing a variety of electrical projects from circa 1937 - he was just a 14yo teenager then, but luckily had a family friend who was an elec eng. Elec projects, including audio based projects was just a hobby for him as he progressed through an elec eng degree, and got to work for MOV in 1944. By that stage he was obviously honing in on what he thought was a good amp circuit configuration, and likely now had test gear and free valves and a growing electronics awareness to continue to prepare audio amps after hours at work. DTN didn't put the amp article in to WW - that was his old boss's doing, as he had asked DTN to prepare a paper on the amp given that DTN had been allowed to, and then asked to set up in-house demos. Those in-house demo's would have had the latest and greatest supporting audio gear of the mid-1940's, and had astute professional listeners, including a demo to the MOV board. It was only after DTN left MOV at start of 1946 that MOV sent the article in to WW.

DTN didn't manufacture or market his hobby amp after his article was released in WW. Of course he used MOV valves, as he apparently had free access to them, and likely a custom output transformer made with the help of Vortexion. The design was free for anyone to use and abuse. There was no gain for DTN, just a flush of notoriety and door opening, but by that time he was in to other things and preparing a fresh update set of articles was more a chore it seems. If you want to venture in to marketing, and ripping off designs then put that in to the context of all the commercial enterprises that ran with the basic design, including the US based technical improvements where patenting and showcasing was very much a commercial venture. Many companies gladly put 'Williamson' in to the marketing text, and many espoused improvements but were short on technical details (including Keroes).

DTN and Walker (Quad) put out an article on their view of how UL fitted in. I don't have a problem with them pointing out that 'ultra' is just a marketing superlative - we could have ended up with super, or mega, or hyper ... And the transformer required an extra level of complexity, which although fine for commercial products, may have been onerous for DIY at that time.

Comments on what you imagine DTN would have thought or did or didn't do, as a means to support your views, are pretty much drivel for a technical forum.
 
DTN was making and testing a variety of electrical projects from circa 1937 - he was just a 14yo teenager then, but luckily had a family friend who was an elec eng.

Comments on what you imagine DTN would have thought or did or didn't do, as a means to support your views, are pretty much drivel for a technical forum.
Well, "drivel" would be how I describe how a 14year old is supposed to have changed audio history.

Fact is, Williamson copied someone else's design then claimed it was his idea.
I like very much how you walk around that fact.

Believe what you like.
From what I know it was Mullard who were making all the running from post war on, and Blumlein did all the inventing up until his death.
Blumlein was not a 14yr old kid, was 38 in 1942 at his death and had masses of patents to his name.
He is widely credited with being the inventor of Stereo.

Mullard's amp designs are by far and away the most common of old British amp designs, scale much better and are usually a lot more stable.
It's also well proven that pentodes have much lower IMD than beam tetrodes.

As I said, it's a total nonsense to be using a top cap anode transmitter valve triode strapped when it's really unsuitable, and designed for cathode follower drivers.
I don't care what radiotron may have claimed.

That's not drivel, but maybe you don't understand the difference between common sense, and legend or rumour?

I rate Alan Blumlein, Rupert Neve, Michael Gerzon all as great British genius.

DTNW as a "bricoleur".
 
Fact is, Williamson copied someone else's design then claimed it was his idea. I like very much how you walk around that fact.
Yup, you got me there - cos so far I have yet to come across 'someone else's design' that is the same as Williamson's. Maybe a link or reference would help.

Sure, there were no patentable aspects in the amp - all the circuitry sections had been seen before 1944 when the Williamson circuit was demoed and the article written.

Maybe you are thinking of the Quality Amplifier articles in WW, such as Dec 1943 - certainly shows similarities - I don't see a copy, but maybe in your eyes?
 
MO-valve had an intern paper with the "Williamson design" as early as 1944.
, the design was developed by Cocking, and also appeared via a guy called Mitchell.
Ie. from a
series of articles began in 1924, lead by W.T. Cocking.
.

The basic building block was freely being used in 1939 in OZ made A503, which is very similar to the high gain version used in my modified Bogen. Good stuff, drive output valves with a nice low impedance.
It's ultra stable.
The AW515 picked up where the A503 left off, this time adding masses of NFB, but using those massively oscillation prone 807s.

As I stated earlier, MOV were forced to incorporate an extra stage of gain before the output valves, for the simple reason they had such low transconductance, and as I also stated forced them to rework their terrible KT8 to turn it into a proper audio valve the KT66.

That has to mean huge output transfomer kludges which Radford clearly demonstrated.
Radford's own design was just an improvement on the Mullard 5/10 with much improved OPTs.

That by comparison is the achilles heel of 6V6 & 6L6 designs, and why they work so badly with screen grid feedback.
(called UL).
They have too low a transconductance & fundamentally the screen grid is low sensitivity, but very fragile,- a problem MOV fixed for good with the KT88 and TT21.

Who after all is going to get any decent power from a valve at only 400V with a 10K A-A load?
It's impossible, even the 6V6 is able to do that.

I stated,- basically attempting to do a 807 based Williamson design with such a high impedance output transformer, (up to 70:1 winding ratios) with its attendant high parasitic losses is going backwards.
Go to a modern valve like the 7591 or similar - it makes life so much easier.
Get the impedances down, treble the OPV gain, increase the current density and forget that dumb 1940s idea.
(usually a TV scan valve will do the job a ton better).
 
Last edited: