• 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.

Interesting Williamson article..

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
As soon as you read about the HF imbalance and distortion from the split-load, you know you're reading a very flawed analysis.

He does manage to mention the one actual mistake in the circuit, an insufficient LF margin because of the two RC couplings with similar time constants. A simple change of one of them by a decade fixes that problem very nicely.
 
Hi,

This article has been discussed here several times before and as SY write there are several factual errors the text. For a better understanding of the Williamson amplifier and also to understand what the errors are in the White article I recommend to read the original build description from Wireless World that can be bought as a reprint.

When it comes to correct the insufficient LF margin I believe the best method is to use a phase corrective network between the splitter and the driver as can be seen here http://www.tubetvr.com/Williamson_comp2.pdf compare with the original amplifier here http://www.tubetvr.com/Williamson.pdf

Regards Hans
 
He's got some legitimate complaints, but he's also over reaching quite a bit.

"As soon as you read about the HF imbalance and distortion from the split-load, you know you're reading a very flawed analysis."

With cathodynes, the unequal plate and cathode impedances become a problem that creates distortion when the load impedances become unbalanced, as they will during an overdrive situation where one grid cuts off as the other begins to draw current. Otherwise, no, it's not a prob. The reason I don't care for it is that a cathodyne always seems to be the "weak link" so far as headroom is concerned (unless you supply it with some insane DC Vpp's). Otherwise, from the balance and distortion perspective, there are lots worse phase splitter topologies out there.

" He also says that 6SN7 have inherently low bandwidth... "

Just plain nutz there. :crazy: He said that this was due to the higher than normal r(p) of the 6SN7. Isn't so: the r(p) of this type is comparable to that of similar triodes: 12AU7 and 6C4. At ~7.0K, it's not all that high. It does, however, seem to have a highish reverse transfer capacitance, but at 4.0pF, it's not really a problem unless your Cmiller is way up there and you're driving from something like a 12AX7A with a very low plate current. Still, that could be solved by using an active load on the 12AX7A to get the plate current up if you have a very limited Vpp.

The 6SN7 was intended to be (amoung other things) an oscillator/buffer tube and/or an astable multivibrator. It could run all the way down to the 10M band with no probs.

"He does manage to mention the one actual mistake in the circuit, an insufficient LF margin because of the two RC couplings with similar time constants. A simple change of one of them by a decade fixes that problem very nicely."

I avoided that problem with one of my designs by using DC coupling as much as possible, and making sure that the one place where a coupling capacitor was unavoidable to make it a big one. A couple of 0.22uF/450Vdc Auricaps aren't all that expensive. Good open loop low frequency response and gNFB that consequently stayed effective made for excellent bass performance. :cool:
 
Administrator
Joined 2004
Paid Member
Just about everything that needs to be said has been said, that said I'll just go ahead and add my bit.. :devilr:

His assertion that two chassis amplifiers (power supply on one, amp on the other) aren't done any more is rubbish. I built quite a small number of them commercially, and I know of others who have (in much larger quantities) as well. Jadis would be just one. I wouldn't recommend it in applications with extremely high B+ like a 211 or 845 amp, but even there it is possible to do it quite safely with a little thought. It's a very good way to end up with an extremely low noise amplifier.

The 6SN7 has relatively low rp compared to many commonly used driver tubes, it may not be the best choice, but many engineers & audiophiles think it is a reasonable choice. I have used transformer coupled 6CK4, 6SN7 in SRPP and choked assisted mu follower, 12AU7/6SN7/12BH7/6CG7 in long tailed driver configuration, etc. All worked well.

His comments about the power supply were a red herring. Williamson used components and practices common in the UK in the late 1940's. The rectifiers common at the time liked small to no input cap..

I am not a big fan of the topology, but with careful attention to supply decoupling and by staggering the two sets of low frequency poles in the loop, stability can be achieved.

This design achieved <0.1% thd @ 12W, a relatively high damping factor, and wide bandwidth of perhaps 100kHz, a major milestone in 1947 when IIRC he published the design. His intention was to demonstrate how judiciously designed output transformers in conjunction with negative feedback could greatly improve the measured performance of audio amplifiers compared to standard practice at the time. The Partridge transformer at the time probably had the least HF phase shift and resonances of any commonly available OPT.

His assertion that all hifi was DIY even in the late 1940's is false, companies like Brook were building amplifiers commercially for home use in those days. Earlier on companies like MCMurdo Silver and EH Scott catored to the budding, well healed audiophile type. I believe the situation in the UK was not too different with companies like Quad getting started right after the war. I'm pretty sure LEAK was around as well.

I am not certain whether or not commercial amplifiers were exported from the UK to the US using the Williamson design, but variants of it were widely disseminated here in the late 1940's, Sarser and Sprinkle's "Musician's Amplifier" would be a good example - later articles even dealt with converting them to UL/partial triode operation.

You can find Williamsons in the Audio Anthology: When Audio Was Young, Volume 1 which is a reissue of 1947 - 1950 articles originally published in Audio Engineering. (Incidentally the first 4 Audio Anthology series are well worth it, and I would rank them up there with Morgan Jones on the list of must have audio books.)

Several early Heathkit amplifiers were pure Williamson (W2A, W3AM, W4, W5, etc..) Lots of other companies did too.

The Partridge transformer was sold here in the North East, I have seen several pairs, the early ones if good now go for hideous money..

Quicksilver has built many amplifiers in the past couple of decades based on the Williamson design, and may still. Their home page is: http://www.quicksilveraudio.com

My Dad actually knew D.T.N. Williamson personally during a stint with Hazeltine in the UK during the early 1950's. He definitely was a hardware fellow, and liked building and testing stuff. I am also given to understand that he liked music and was interested in high quality sound reproduction. I am not sure where he was employed at the time my dad knew him..

The Williamson is a landmark, and in my mind at least marks the beginning of the true hifi era. That it is not the perfect design, and even offends the sensibilities of some people today isn't really the point..


Hope some of this is illuminating. :D
 
Hi,

A few comments, (there are many more errors to point out but these are some of the important ones)

In the original article in Wireless world of May 1947 there is nothing mentioned about using a 2 chassis design with a separate power supply. In the description of a new version in August 1949 it is mentioned and described 2 recommended layouts for the amplifier, one with a combined chassis and one where 2 chassis are used. In the same article it is also described that a 5V4 is a better choice for use as rectifier.

My conclusion is that White is wrong in saying that Williamson adviced only the use a 2 chassis design and he is also wrong to write that Williamson didn't realised the problem with using a directly heated rectifier.

White is also writing a lot of comments regarding the "american version of the Williamson"? what is this? Williamson never designed an amplifier using 807 tubes or any other than KT66 can not be hold responsible for everyone that used part or the whole of his circuit sometimes without understanding the design in detail. The only true Williamson amplifier is one designed according to the original article, other amplifiers can be influenced by or use part of the circuit but it is wrong to call it Williamson amplifiers.

Regards Hans

BTW, real triode connected KT66 are very linear, fully comparable with 2A3 so it was indeed a very good choice of output tubes.
 
"I am not certain whether or not commercial amplifiers were exported from the UK to the US using the Williamson design, but variants of it were widely disseminated here in the late 1940's, Sarser and Sprinkle's "Musician's Amplifier" would be a good example - later articles even dealt with converting them to UL/partial triode operation. "

Yes there were Williamson amp variations being sold here in the US before 1950 anyways. My Dad worked for Stancor before and after WWII and in the late 40s Stancor sold a 2 chassis Williamson type "kit" using 807s and 6sn7 that was probably very much like the original design. Later they sold an Ultralinear version of it but at first it was 8 watt troide mode. They also published the schematic and parts list in their transformer catalog for people that wanted to collect up all the parts themselves.

Dave
 
White is also writing a lot of comments regarding the "american version of the Williamson"? what is this?

The Radiocraftsman 500 was a true Williamson amp (using KT-66s). It was also one of the few (if not the only one) that used a choke input power supply. It was "designed" by the same guy that did the famous old Marantz gear.

John
 
The Radiocraftsman 500 was a true Williamson amp (using KT-66s). It was also one of the few (if not the only one) that used a choke input power supply.

So it didn't follow the original schematic? Then it is not a true Williamson amplifier. But to be fair what I questioned as true Williamson amplifiers are the one built with other output tubes or even with ultralinear connection of the output transformer.

Here is another article on improving the Williamson which may be of interest.

Yes, it is interesting and cover some of the problems that "bastard" Williamson amplfiers have, the problem described here is that the driver section doesn't give enough voltage to drive 807 tubes to high power. Remember that the Williamson was designed as a 15W amplifier with KT66 tubes, for this configuration the original design is functioning well and there is no need to do the modifications described in the article.

The problem described in the article is is a good example of what I wrote before, that some used the Williamson design ideas without fully understanding the design, it is obvious that you need to redesign the driver stage if you change the output tubes or want to increase the output power, this is of course true for any other original design also.

Regards Hans
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.