Tone Controls

[Synchrodyne]

This thread derives from comments made about the Quad tilt tone control in “the Audiophoolery Thread”, see post #828 et seq, https://www.vintage-radio.net/forum/...&postcount=828.

I think it could also be seen as a continuation of the earlier thread “Tone controls – old and new”, at: https://www.vintage-radio.net/forum/...d.php?t=133623.

As far as I know, for its tilt control circuit, Quad used the Ambler tone balance control with the Baxandall modification.

The Ambler circuit was published in Wireless World (WW) 1970 March:

Ambler Tone-Balance Control WW 197003.pdf



A point to note is that Ambler intended that it be used additional to, and not in place of a conventional Baxandall tone control.

Baxandall made brief mention of the Ambler tone balance control in Amos, Radio, TV & Audio Technical Reference Book (RTV&ARB) of 1977:



Then he mentioned the modified (more elaborate) version in Talbot-Smith, Audio Engineer’s Reference Book, 2nd edition, 1999, noting that he had developed it in 1979:



Here is the Quad 44 tone control circuit:



I’d say that it was derived from the Baxandall modified Ambler circuit. Given that Baxandall was known to work closely with Quad, that is not so surprising.

The “Tilt” name appears to have originated with Quad, who also evidently decided that the combination of this control with its own bass step and lift control obviated the need for a conventional Baxandall tone control, contrary to Ambler’s original position.

My Quad 44 (dating from 1986) has three steps either way on the tilt control, so at some stage, perhaps following the introduction of the Quad 34, the change was made from the original two steps either way. It certainly is a very useful control, and I like the fact that on the Quad 66, there is a visual indication as to what it is doing.

Generally though, the Ambler tone control does not seem to have been widely used. It arrived just before the time that the anti-tone control movement got started, so that probably did not help its chances. I suspect too that some amplifier makers may have seen its inclusion in addition to regular treble and bass controls as too costly, and its inclusion along with the deletion of the regular tone controls as being too risky in that part of the market that wanted or at least was not averse to tone controls and generally expected the regular form, or perhaps more complexity, but along the vector between Baxandall and graphic equalizers.


Cheers,

[Synchrodyne]

In this recent thread: https://www.vintage-radio.net/forum/...d.php?t=155380 there was some discussion about the passive treble and bass tone controls used in the Barthe record player at interest.



Although the intent was the same, the network used was somewhat different to that of the Voigt tone control of 1940, which is generally reckoned to be the first that provided continuously variable control each of treble and bass without the need to switch between lift and cut. (Although it might not actually have been the first.)



Further searching has produced the James tone control, published in WW 1949 February as a reasonably probable antecedent for the Barthe control.

James Tone Control WW 194902.pdf



The Barthe control had added another resistor between the bass potentiometer and the output, perhaps to further reduce the interaction between the treble and bass controls.

It turns out that James was not the first with this form of tone control, about which more later.

In his WW 1952 October article, which described the second iteration of his tone control, Baxandall referred to both the 1940 Voigt and 1949 James articles. The R-C network used by James has some similarity in general form to that originally used by Baxandall, shown on the right-hand side here:



However, the essential defining feature of the Baxandall control was not so much the network details as the fact that the network was inserted in both the input and feedback arms of a virtual earth amplifier. A notional feedback version of the James circuit could have been be obtained by inserting the network in the feedback path only of a either an inverting or a non-inverting amplifier.


Cheers,

[Synchrodyne]

This letter to WW 1949 April made a prior claim, dating from 1939, for the James tone control circuit, with Volkoff as developer. Thus it predated the Voigt tone control. Probably James was unaware of the previous American work.



In Radio & Television Engineer’s Reference Book, Third Edition, 1960, (RTVERB III) the section on sound reproduction and distribution noted that two kinds of treble and bass tone controls were in common use, one being passive, and the other being the Baxandall.

from R&TVERB III pp.37-20,21,22.pdf

The passive tone control circuit was essentially the James/Volkoff circuit, with the additional standoff resistor. Perhaps the RTVERB editorial staff avoided naming the passive circuit to avoid getting into any issues about who developed it. On the other hand, the Baxandall case was clear-cut, with, as far as I know, no other claimants.




Cheers,

[Craig Sawyers]

The early Quad 44 used a much more complex tilt topology that involved two switch banks per channel to give +/-2dB. But from the first post here at some point Quad went across to exactly the same circuit that was adopted in the Quad 34.

Very interesting to note that Baxandall had his hand in the Quad 34/late 44 circuit. And as you say Baxandall worked on a number of things with Quad. He has a superb chapter in Loudspeaker and Headphone Handbook (ed Borwick) in which he goes through the Quad ESL designs, and derives the expected frequency response of the ESL57 and overlays the actual response. From the level of detail I suspect he collaborated on the design of that.

Craig

[Synchrodyne]

Thanks, Craig. I had missed the early Quad 44 tilt control circuit. I have now found it, attached:



I think that the concept aligns with the original Ambler circuit, but that the execution was somewhat different. With the original Quad 44 circuit, the input to the virtual earth amplifier stage was taken from the centre of the feedback resistor ladder, rather than a variable position on its centre section. On the other hand, the two R-C shaping networks were connected to different positions along the resistor ladder according to the control setting. Also, these networks were not in parallel with part of the resistor ladders, as in the Ambler circuit, but were connected between the resistor ladder and ground. Presumably Quad did it this way to obtain the relatively mild curves (as compared with the Ambler original) it desired.

The Quad 44 was released in 1979, so it is possible that its circuit was defined and committed to production before Baxandall developed his modification to the Ambler circuit, also in 1979. Then perhaps Quad chose the Baxandall-Ambler circuit for the Quad 34, following that with a like update of the 44.

I understand that Baxandall was involved to some extent with the Quad 33/303 back in the later 1960s. The Quad 303 was one of the first designs in which the inherent asymmetry of the Lin quasi-complementary output circuit (different asymptotes and different “turn-in” curves for the two-halves ) was tackled head-on rather than avoided or (somewhat) ameliorated. One wonders whether Baxandall had a hand in the resultant output triples, or whether they inspired him to develop his own, simpler solution in the form of the well-know “Baxandall diode”, which appeared in 1969 (at about the same time as the Shaw power diode) and then became the norm for QC output circuits. In his chapter in RTV&ATRB, 1977, Baxandall worked through the problem and the Quad 303 solution (in his usual very lucid way) , but mentioned his diode solution only in passing. The Quad 33 was also used as a worked example. This was the only Quad control unit to have used a fairly conventional Baxandall tone control. It was executed with some gain (not unknown in the valve era, but not part of the original Baxandall circuit), this being done by bootstrapping the collector load of the tone control transistor, a technique for which Quilter later gained the credit. The 33 also used essentially the same approach to ceramic cartridge matching and equalization as later advocated by Burrows.


Cheers,

[Synchrodyne]

From the Quad 44 we can go back to the Acoustical QA12/P of 1947. This had a tone control that predated the 1949 James passive tone control but was generally similar to it.



It included a standoff resistor between the bass control potentiometer slider and the output. Also, capacitor C10 was placed following the treble control potentiometer slider rather than between the lower end of the potentiometer and earth. The net effect was probably the same, in that the capacitor was there to provide a high impedance to bass frequencies, so that they were essentially unaffected by treble control operation. A similar, but not identical variation in capacitor position was also seen between the two variants of the Baxandall control.

Evidently there were several separate but similar developments going on in the field of in the passive tone control circuits back in the 1940s.

And whilst one may find mention of the 1939 Volkoff tone control, there do not seem to be any actual details available. I suspect that where it has been mentioned, the source has been that WW 1949 April letter.


Cheers,

[Craig Sawyers]

Peter Baxandall was indeed very involved in the 303. There is a Linear Audio book called "Baxandall and Self on Audio Power". That is a collection of papers by both designers. And then a long set of notes by Baxandall, hand typed with sketches in his own hand, just before his fairly early death aged 74. Well worth a read, certainly at Euro 9.50 https://linearaudio.net/books/2219 .

Pages 104-5 describe his involvement in the 303. "I was instrumental in introducing them (triples) to Quad" "The use of a regulated power supply was also my contribution" "(this) was done purely because in 1967 we were all scared of busting power transistors!" "At the time noone there, including Peter, knew anything much about transistors"

Based on that, I suspect he had a hand in the 33 too.

The system I have in our kitchen, fed through ceiling speakers with input from an FM4 and an optical/analogue converter from the kitchen TV, is a 33/303. Something I lusted after so badly it hurt in my teens, but now you can buy easily on eBay for reasonable money.

Another snippet. As a research student in the late 70's I still could not remotely afford Quad gear, but thought I might build the Quad 405 dumper, and the service manual, that they would send on request, had the schematic and circuit layout. I wrote to Quad to ask where I might buy the inductor in the bridge. I got a reply from Ross Walker to the effect that their job was to sell audio product "and not to assist those who would plagiarise our designs". I pinned that up in the toilet until it fell to pieces through age.

I actually have a 405/2, and so far it has blown up twice. First because the dreaded 3.3k resistor with too low a power rating burnt out, which puts one power transistor to rail, crowbar fires, burns out power transistors and drivers, the crowbar itself, then the fuses blow. Then it blew again a year or so later just to be capricious.

Craig

[Craig Sawyers]

Those on the audiophool thread will know that I'm working on a variant of the Quad/Baxandall tilt control.

Anyway, a question. Switched 1dB increments as per the original, or continuous using a dual linear pot (Bourns 91 series conductive plastic)? I can do either.

Craig

Continuous, you can adjust it without clicks (both mechanical and electrical) interfering with you perceived audio impression.

[Radio Wrangler]

I have a 5 position switch. I've never felt a need for intermediary positions. it just switches 4 passive networks and a fixed attenuator.

David

[Synchrodyne]

My inclination in general is towards continuously variable controls. However, my experience with the Quad 44 and 66 tilt controls is such that in this case, I’d say that the switched form with 1 dB increments is fully adequate, so this gets my vote.


Cheers,

[Craig Sawyers]

I *think* that the 34/late44 switched control was actually a custom variable control with detentes. I used to have a 34, but sold it some years ago, or I could have measured it.

Craig

[Synchrodyne]

The schematic tends to support that notion, as the controls have RV numbers.

Cheers,

[Craig Sawyers]

A fairly comprehensive set of measurements, including tone response curves, for the 34 are here https://kenrockwell.com/audio/quad/3...m#measurements

Craig

[Boulevardier]

<< Pages 104-5 describe his involvement in the 303. "I was instrumental in introducing them (triples) to Quad" "The use of a regulated power supply was also my contribution" "(this) was done purely because in 1967 we were all scared of busting power transistors!" "At the time noone there, including Peter, knew anything much about transistors" >>

Interesting Craig. I’m sure I remember reading in one of Douglas Self’s writings (DS has always railed against any need for regulated PSs in properly designed PAs) that he encountered PW at a meeting somewhere, and demanded an explanation for the inclusion of a regulated PS in the 303. He said that PW replied that it was purely down to Quad’s lack of experience in solid-state PAs, and their extreme caution in designing their first one. DS said that PW had admitted that it had been unnecessary overkill…

Mike

[Radio Wrangler]

Overkill or not depends on the variability of the mains supply and how close to the voltage limits the PA transistors are run... and that includes the second breakdown mode which wasn't fully understood until after the genesis of the 303.

We've long been lucky in the UK with a pretty well regulated mains supply free from nasties. With early 2N3055 running that nominal voltage in an amplifier to be exported around the world, having a regulator to absorb the mains variations turned out to be a good idea.

The 303 was probably the first fully accepted transistor amplifier. There were plenty earlier amps, but they were rather poor compared to the 303 and failures were expected with them.

David

[Craig Sawyers]

I have a mint 303, I'm the second owner, and it is still going strong.

One channel occasionally has a bad moment, but its intermittent nature suggests it is one of the open frame pots (for centering the midpoint voltage and quiescent current) is flaky. Must replace them as a precaution.

Craig

[Synchrodyne]

One may wonder if the Baxandall’s association with Quad went back to the time when the Quad II was in development.

The original Q.U.A.D. of 1951 (retroactively the Quad I) had a passive tone control system similar to that used in the QA12/P of 1947:



The QCII, released I think in 1953 September, had an active tone control system. I have endeavoured to extract the essence of this circuit (from the complex-looking actual circuit) in this hand-drawn diagram:



The tone controls were incorporated in the control unit output stage, which was built around a triode pair (ECC83) and so was non-inverting. The feedback was of the series type, fed into the cathode of the first triode. This stage provided quite a reasonable amount of voltage gain, taking an input of 100 mV to a final output of 1.4 V, thus 23 dB. As there was some loss in the output network, the actual output gain would have been higher than 23 dB. Here are the published curves:



Quad described the QCII tone control thus:

“The continuously variable bass and treble controls are so arranged that the central position of their travel will always give a level response with virtually no error. The bass rise operates by varying an impedance in the feedback circuit of the two final stages and bass fall is obtained by varying an impedance in the top arm of the feed to the filters. Those two impedances are of like value so that a linear bass control is used. This is advantageous because a linear control does not appreciably change its scaling and any change in overall value mill be symmetrical. Further. the two fixed capacitors are small and of like value, thus easy to match and extremely unlikely to drift.

“In the maximum rise and fall positions. the maximum slope is 5db/octave and in intermediate settings the slope is gradually reduced to zero and at the same time the turnover point reduces in frequency. This is preferred to the step obtained by variation of amplitude or the constant slope obtained by variation of capacity alone.

“The treble control operates both rise and fall in the feedback circuit. The slope is always asymptotic to zero so that a boost in musical brilliance in the treble musical register is not accompanied by boosted distortion at very high frequencies, nor is a reduction in the treble musical register accompanied by a complete loss of harmonics. This feature of musical balance control is. of course, only made practicable because of the independent filtering controls provided.”

The bass control looks to me as if it is approximately where one would arrive if the objective was to adapt the Baxandall control for use in the series feedback loop of a non-inverting amplifier. In the Baxandall case, the tone control potentiometers are partly in the input arm and partly in the feedback arm, where they effectively act as frequency-selective active gain controls. In the series feedback case, the input arm and the feedback arms do not connect (in a signal sense), but the output arm and the feedback arm do, so the tone control potentiometers could be place there.

So perhaps the QCII bass control at least was an adaptation of the Baxandall circuit. The latter was published in 1952 October, although Baxandall is on later record as saying he developed it in 1950. So the timing would be right for the QCII tone control to have been a derivative. The QCII/II probably had a longish gestation period. For example, Quad had asked G. Horn to develop an FM tuner to match the QCII/II (as well as an AM tuner), and the prototype of that was first shown in mid-1952. (The final design work for the production version – released in 1955 - was done by J. Collinson of Quad.)

I imagine that Quad’s use of a treble control entirely in the feedback loop was done as the best way of achieving its desired curves – with some eventual flattening. A pointer to that may be found in an unusual place, namely Stephen Spicer’s excellent book on the history of Leak. On page 133 the Baxandall tone control (adopted by Leak in 1954) was discussed, and there was a side panel contributed by Australian audio enthusiast Peter Stinson, from which I quote:

“An important characteristic of the Baxandall circuit is the sliding turnover point, which differentiates it from the fixed hinge point of conventional systems. However, Peter Walker would say, and I would agree with him, that while the Baxandall circuit is almost ideal in the bass, it is not so well suited to treble, particularly treble boost, where it can result in excessive boost at the extreme top when only a little mid-top lift is wanted”

Actually, with the Baxandall tone control, some flattening of the treble curves could be obtained by omitting the treble potentiometer centre-tap of the 1952 circuit. This was a later refinement, probably not available when the QCII was designed. However, the Quad 22 of 1959, whilst having essentially the same bass control as the QCII, had a different form treble control, one that was partly in the output arm and partly in the feedback arm, thus making it analogous to the bass control, and so perhaps the series-feedback analogue of the Baxandall variant without the centre-tap. Here is the hand-drawn simplified circuit:



And the resultant curve-set:



The Quad description was as follows:

“Bass and treble control is effected in two frequency discriminating networks of identical impedance, one in the signal chain and one in a feedback chain. Boosts and cuts are therefore symmetrical. The bass control varies both slope and turnover. The treble operates midway between variable step and variable slope in order to facilitate adjustment of musical brilliance while maintaining natural harmonic balance.”

That description of the treble control is, I think, consistent with the curves provided by the non-centre tap Baxandall.


Cheers,

[Synchrodyne]

The Baxandall tone control curves with and without the treble potentiometer centre-tap are shown in this excerpt from Radio, TV and Audio Technical Reference Book, 1977:



Cheers,

[Craig Sawyers]

OK - I bought a Bourns 91 series dual gang 10k linear conductive plastic pot. I've measured matching between sections at resistance values that correspond to the Quad increments and then bunged the values into the Spice simulation. That shows that the zero error is <0.1dB, and the matching between channels at any point is also <0.1dB.

Which is pretty darned good for an off the shelf pot that cost £7.42 plus VAT from Farnell.

So I'm going for the pot rather than switched resistors.

Craig