Plea for high-quality
valve amplifier

Having recently started a book on amplifier design, I have reviewed a large number of circuits ranging from the original Williamson to your single-ended parallel 6L6 hybrid of August 2005. Although I glanced at your design initially, I didn’t take it in until now.

I accept that, from the objectivist point of view, valve amplifiers are not capable of the accuracy and economy of the best solid-state designs. People who use valve amplifiers do so regardless of cost or inconvenience: Audio Research and Conrad-Johnson (among others) sell the biggest and most satisfying tube amplifiers at premium prices although they also produce solid-state models which are, on paper, a more practical proposition.

The Williamson is probably the most famous circuit of all time. It was not simple or cheap but a vast number were built in various forms. At the same time, Baxandall published a cheaper circuit which was ignored as far as I can determine.

A decade later Baxandall published an even cheaper design which cost slightly less than the (by then established) Mullard 5-10 but produced little more than half the power. This also sank without trace. The clear inference to be drawn is that for at least 60 years, people have been custom-building amplifiers to get better sound quality than that obtainable off the shelf.

Since it is now possible to buy highly linear amplification for less than a dollar a watt (in the form of a 5.1 home-theatre receiver), one must ask why anyone would want to build their own amplification. The answer is obviously still "to get better sound quality than that obtainable off the shelf". While conceding that the sweetest recorded sound I can recall was (small-scale, limited frequency range) music from vinyl on a single-ended EL84 (pentode) stage, I must implore you, in order to maintain Australia’s excellent audio tradition (Benson, Thiele, Small, Langford-Smith, Cherry, et al) to publish at least one quality valve amplifier as a reference.

I would suggest an upgraded Mull-ard 5-20 configuration to get the advantages of zero unbalanced DC in the output transformer primary and cancellation of even-order non-linearity (and ripple) in the output stage. Since class-A operation is desired, cathode-bias can be used and if more than 30 watts per channel is required, paralleled output devices can be fitted, facilitating even better output DC balance with a negligible drop in open-loop gain.

The EF86 input stage (and its partition noise) could well be replaced by a cheaper and easier to obtain 12AX7 configured as a DC-feedback pair. The open loop gain and linearity of the whole amplifier would easily allow for a closed-loop sensitivity of 500mV with 30dB negative feedback, resulting in THD of well under 0.1% at 30W (or 60W with four output devices).

The amplifier needs to be affordable, but not cheap in any sense, and certainly not below contemporary standards for acceptable audio quality. Anything less would be a bad joke which could rebound on Australia’s international audio reputation.

Maybe I’m overlooking some important point but I can’t see how a parallel single-ended output stage can possibly compare with a push-pull stage in any way other than in the elimination of the phase-splitter (this being the excuse given in a circuit for an "economy" radiogram published in NZ about 50 years ago!).

Roger Lowry,
via email.

Error in home
theatre article

I am a subscriber to the magazine and I noticed the following apparent errors in the "Home Theatre Video Projector Survey" article in the August 2006 issue. On page 14, in the paragraph titled "Scanning method", the text states:

(1) "On your television set, half the lines (called a field) are displayed 50 times a second (Hertz) and then the other half are woven in at 50 times per second. . ."

(2) "In PAL we have two interlaced 50Hz displays or 25 complete pictures in each second".

This did not seem right and after much thought the first statement is incorrect. With 625-line TV signals, the first field of a frame occurs at 25 per second in a time period of 1/50th of a second (the odd lines). The second field takes the next 1/50th of a second (the even lines), to give a frame rate of 25 frames per second – as the second quote states.

Thus to be correct, the first statement should read "On your television set, half the lines (called a field) are displayed in a 50th of a second and then the other half are woven in during the next 50th of a second . . ."

David Williams,
Hornsby, NSW.

Tips on restoring 78 RPM discs

I was very interested in the article in the September 2006 issue of SILICON CHIP, detailing ways of re-mastering LPs and removing the various forms of surface noise inherent in them, while transferring them to CD.

I began tinkering with this sort of thing after reading a SILICON CHIP article about seven or eight years ago, describing a program called DC-Art (or Diamond Cut 32) which was designed for audio restoration. My interest was in restoring a very large library of 78 and 80 RPM recordings which I had accumulated over the years and I bought a copy of DC-Art, then a later version, DC Millennium (1999), then DC5 (2002) and finally, DC Live 6 in 2004. Each program is a big advance on the one before. As an example, DC6 has a 20-band graphic equaliser with some invaluable pre-sets, compared to the 10-band equaliser in Millennium and DC5.

I note that John Clarke describes using Nero 6 and its Wave Editor for removing noise from LPs and it certainly does that very successfully with 45s and LPs. I have tried using Wave Editor in Nero 7 Premium to clean up 78s but I find that DC6 is more flexible for these older discs – not only giving greater control but also providing a valuable group of presets which make the job easier.

In addition, DC6 has a major advantage for someone working with 78 and/or 80 RPM discs, in that it has a speed conversion effect which allows a 78 to be played at 45, or even 33, RPM into the computer hard drive; and then to be converted to 78 speed – invaluable when it is so difficult to buy a 78 turntable these days.

Even better, DC6 has a fine-tuned pitch-control, which recognises that many 78s and 80s were recorded at speeds varying by 2 RPM or so, either way. I use an ancient but still functional Dual 1019 4-speed turntable, so I don’t need the speed conversion but the pitch control in DC6 has improved a number of discs which are "off-speed" even further than the mechanical speed variation on the turntable can correct.

Some of the old discs have been loved to death but it’s amazing how they clean up once you get the hang of using the program.

John notes in his article that the RIAA curve built into most preamps is not suitable for 78s and this becomes very clear when you try using a standard amplifier preamp to boost the signal between pickup and sound card. For a start, the bass is overwhelming and emphasises the rumble built into many old recordings. While DC6 has a frequency-variable rumble filter which can deal with this, it’s better not to have it there in the first place because of the way it loads up the signal.

So I have tried what I am told is a controversial technique, recommended by tracertek.com, who are involved with the DC products. They market a small preamp which amplifies the signal from the pickup without equalisation and I have found, particularly on 80 RPM and very old 78s, it works very well. Their suggestion is to record to the hard drive, unequalised, and then use the playback correction curves provided in DC6 to introduce whatever equalisation works best with that particular record.

I am not a technical person and friends who are far more expert than I have questioned the use of this preamp, saying it will introduce distortion, because its output is too high and will overload the sound card. I have not experienced distortion or apparent overload when using the preamp but I must admit the unequalised sound really is weird to listen to. However it comes good after finding the right equalisation and in the end produces pretty good results to the CD.

I have several 78 pickup cartridges, which give some variation in stylus size, to match the variations in groove diameter found in old records. But the cartridge that works best for me is an Ortofon OM78. It tracks beautifully at 2g or less and seems untroubled by warped records although DC6 even has a way of correcting these.

John Tingle,
Port Macquarie, NSW.

Comment: using a high-gain unequalised preamp running from a low battery supply risks signal overload from the cartridge on loud music passages, although it may not overload the sound card input.

Historical TV sets

I have a Rank Arena C-2251 (NEC Chassis) that is still working although the blue gun is a bit weak. I would like to be rid of it but it seems a shame to destroy a small piece of Australian TV history (especially as it still works).

If anybody would like it they would be very welcome to it. I am in Adelaide. Please contact me at researchnurse@iprimus.com.au

LPG production in Australia
is well in excess of demand

I read with interest your Publisher’s Letter on natural gas in the September 2006 issue and hope to clarify some of the information you cited about LPG.

The Federal Government subsidy for LPG conversion is expected (based on industry estimates) to result in around 80,000 installations by the end of 2006 and up to 120,000 installations per annum during 2007 and 2008, then moderating in subsequent years. Over the life of the grant scheme to 2014, it is estimated that around 750,000 vehicles will be fitted with LPG, leading to over 1.25 million vehicles being LPG-fuelled.

Also, the factory-fitted LPG vehicles that are eligible for a smaller grant could add a further 250,000 LPG vehicles by 2014. That would mean a significant proportion of motorists – over 1.5 million vehicles (10% of the car park) would have access to LPG, resulting in a saving of some three billion litres of petrol and further reducing Australia’s reliance on imported crude oil and refined fuel.

LPG in Australia is produced from extraction from "wet" natural gas or from the refining of crude oil. Over 80% of the Australian LPG production comes from the "wet" gas extraction process and only around 20% is sourced from the refineries. Major naturally occurring LPG production sources include North West Shelf, Bass Strait and Cooper Basin.

Australia currently exports around 3 billion litres of LPG and this is forecast to grow to over 4 billion by the end of 2008 on the back of further developments in natural gas production. ABARE has forecast a significant surplus of LPG to 2020 and the likelihood that supply will exceed demand for around 50 years.

While CNG is also a resource that could and should be used to further reduce our vehicle fuel reliance, it does have some issues in smaller vehicles. CNG passenger vehicles are used in the US and Europe and have been trialled here in Australia but their wider use has been hampered by the lack of infrastructure. Only about 25 public refuelling facilities exist in Australia compared to the 3240 LPG outlets amongst the 6500 service stations.

Range is another issue and most CNG passenger vehicles are only capable of 200-250km before refuelling is required. Home refuelling devices are being used in the US but can take up to 16 hours to refuel a completely empty tank. This would be OK for a vehicle that’s used to travel from the suburbs to the city and back again each day but would not be suitable for inter-city use.

Installing a large volume CNG refuelling facility (required for public use) is also costly because of the high compression and storage requirements, meaning a cost of around $500,000 plus per site. Recovering these costs on each cubic metre of CNG sold can add a significant amount and reduces the overall CNG benefit. The leasing and operating cost of a home refuelling unit could also result in a similar benefit reduction.

Efficiency when running a vehicle on CNG is also reduced compared to petrol and requires additional timing modules to be installed or other modifications to bring performance up to a similar level. Ford trialled a CNG Falcon wagon some years back but this required engine modifications, including increased compression ratios and ignition timing requirements. This car also needed two under-floor tanks and one large tank mounted behind the seat to give a range close to petrol. This option did not receive much consumer support and so was scrapped.

CNG has similar qualities to diesel and with reasonably minor modifications, can deliver a similar outcome, which is why most CNG trials have been with larger diesel vehicles and in particular government-owned urban buses. These have been combined with depot-based refuelling infrastructure to deliver reasonable outcomes.

A number of trials are now concentrating efforts on using LNG stored in cryogenic cylinders in heavy vehicles, as these have extended ranges. If these are successful and if vehicle manufacturers support such developments, then a number of freight carry vehicles could be switched to natural gas. As these vehicles can consume up to 55l/100km, this would be a significant additional saving on diesel use but infrastructure still needs to be developed. The Federal Government had a scheme in place previously covering half the refuelling site cost but only a few took up the offer.

At this stage LPG is still in my opinion the only "readily available" fuel alternative but I would like to think that Australian ingenuity will overcome the barriers to the use of CNG, LNG and other alternatives to petrol and diesel use in the near future.

More information on LPG is available at our web sites: www.lpgaustralia.com.au and www.lpgautogas.com.au

Phil Westlake,
Industry Development Manager,
LPG Australia.
www.lpgaustralia.com.au

Nuclear power
can be fragile

I just read your "nuclear" editorial in the July 2006 issue. For anyone who has implicit faith in nuclear reactors, we had an experience in Cape Town to keep one on the edge of one’s seat. One of two nuclear generators in the city was (apparently) sabotaged and was out of commission. As a result, the city drew too much power and the whole nuclear power station went down.

Now, either the nuclear power station needs to keep going or it needs to draw power from elsewhere, to avert nuclear melt-down. So they diverted power from the north of the country. That overloaded the grid, so that went down, too.

The third and final line of defence was three large diesel generators, which they powered up. However, as I understand it, diesel fuel production was dependent on the nuclear power station!

My insurer, with amazing rapidity, posted a nuclear exclusion clause to all its clients. They got the power station fixed eventually with the help of the French (it’s called Koeberg – it might well be on the internet).

Thomas Scarborough,
Cape Town, South Africa.

Comment: it sounds as though the country’s electric grid is not very robust if it cannot withstand the loss of one power station.

‘Moon Shots were faked’ idea came from USA

First let me say that your Publisher’s Letters and the Maibag section are, like the rest of SILICON CHIP, always stimulating.

But now let me say that this month’s (October) ‘PL’ deserves a small "oops". I don’t know if your bias is of the ‘fixed’ or ‘self’ varieties but it was very evident. The US has many good things about it but it also has many bad! One of these is its apparent need to infect other cultures with its mores. Where do you think the ‘Moon shots were faked’ idea came from?

A quick look at the net will show you that, like the Christian right’s ideas about religion and "Intelligent Design", the evangelists for this crazy idea come from the USA itself. We are now so attached to that culture, that a large number of otherwise intelligent people believe anything that vocal folk from that country tell them.

Mind you, blaming the extreme left (with all its faults), when the real culprits promoting these things in the USA seem to be more aligned to the right, will not help.

Bruce Bowman,
via email.

6V to 12V car radio conversion

I refer to the September 2006 "Ask SILICON CHIP" item regarding changing a 6V car radio to 12V. I remember doing this about 40 years ago. Three things need to be done. The first, as M. S. says, is to get a 12V vibrator. The next is to get out the valve data books and check the heater currents for all the valves and work out the best matching arrangement to connect the heaters in a series/parallel circuit across 12V.

The third is to rewind the vibrator transformer primary. I was fortunate with this as the transformer laminations came apart quite easily and the transformer winding had not been varnished too heavily. Normally, the primary is wound on the outside so the secondary doesn’t need to be disturbed. My memory of vibrator circuits is a bit hazy but I think the primary is centre-tapped. The aim is to change the turns ratio, so after counting the turns as the original primary comes off, half the turns need to go back on and I seem to remember using a slightly heavier gauge wire. I was fortunate when I did this as I was working in a workshop that did motor winding at the time.

Bill Adams,
via email.

Comment: unfortunately, your method would be too involved for most restorers as they would not have access to valve data and would not be able to rewind a vibrator primary. The 6V regulator method has the virtue of simplicity and requires no modifications to the radio.

Better circuit for suppressed zero meter

In "Ask SILICON CHIP" in the September 2006 issue, a correspondent (P. F.) asks for an expanded scale voltmeter. In my estimation, your answer is useless! While the zener diode works by suppressing the zero as stated, it also works as a very good thermometer. A much better system to use, if you want repeatability, was described in "Circuit Notebook" in the June 1995 issue.

The only point not mentioned in that article is that the tabs of the 7805/7905 should be thermally bonded together. The only other components required are two ferrite beads on the supply rails if the voltmeter is being used near any transmitter.

Tests I have done show no change in meter reading if the regulators are placed in melting ice water and then removed and placed in boiling water! I have been using this system since (I think) the mid-70s, not long after 3-terminal regulators become available.

Secondly, on page 95 of the same edition (September 2006), some poor beginner IS going to kill a perfectly good meter by connecting it up as an ammeter as described! For an ammeter to continue to read the current (repeatably) as calibrated, it must have four separate connections to the shunt: two on the outside ends for the high current, and two separate connections for the voltage (Low/Meter current) connections.

Wal Douglas,
Marian, Qld.

Comment: you are right – the June 1995 circuit is a good solution but we had long forgotten it. Here it is again.