Back to Interview With Gods contents page
Interviews with:
*********************************************************************************************
Interview with Dr Richard Bews - co-founder, owner and designer of LFD electronics. Interview conducted May 29th 2008
Introduction/overview
Earlier today, catalysed by an unexpectedly powerful outcome yesterday comparing the LFD MCT phonostage to a phono stage from a state-of-the-art USA valve maker (through the mighty yet subtle QUAD 2905 speakers, which are superior to any of the speakers I currently stock!) I emailed Dr. Richard Bews, the driving force behind LFD, with a few questions.
To my pleasure and amazement, given that he has a runaway success on his hands due to the incredible review of his Mk3 Integrated amplifier in the Feb. 2008 issue of Stereophile magazine, I received a comprehensive and typically courteous reply – very quickly.
Being a bit of a technical dummy on these sort of things I needed to phone him to gain his authoritative but patiently gently explanations of all the bits I didn’t understand. I came away, after that call, understanding only about 15% more than I did previously.
Thus the Q&As you are about to read is to some extent a compound of Dr Bews’ emailed responses and my own notes from the subsequent conversation earlier today.
As it turns out, this is very much far from being the entire story.
Some of the techniques he uses to get the incredible performance from the MCT are beyond my comprehension. Some are currently beyond belief. The point being that while the MCT is without doubt the most musically credible phono stage I have ever owned, let alone heard, a full interview (complete with my write up of some of the demonstrations he’ll conduct for me in private) will need to follow – in due course.
He’ll demonstrate (and I hope I’ll be able to hear, and he assures me I will – if I don’t go to any more Who gigs for a while) the sonic differences between very high purity silver single strand internal interconnect cables (which he doesn’t believe is good enough) versus his own multi-strand high purity silver interconnect using 5 different and precisely calibrated diameter strands, which he does use in the MCT.
I should explain that the LFD MCT is an anonymous heavy black box with a pair of input RCA sockets and a pair of RCA output sockets and an IEC mains input – and that’s it. No controls, no adjustments and not even an on/off button.
Anyway, here’s what I learned today:
What’s your design philosophy re phono stages in general?
All LFD Products (including phonostages) obey our primary requirements of simplicity of design, obsessive passive component selection and fanatical attention of the layout of PCB and PSU development!
How, if I’ve got this correct, is the MCT’s outstanding sound in part due to your R&D re the battery powered phono stage commissioned by SME?
The Battery Disc preamplifier was the ultimate statement of the LFD philosophy, all other LFD phonostages are compromises. That’s what SME commissioned me to design and built to help their R&D teams to engineer their turntables. This device is NOT available for retail! However, the MCT is trying to get close to the battery disc preamplifier sonically – using all my experience of passive selection (the circuit is, frankly, mundane i.e. just an IC feedback circuit with transformer coupled input to give good noise performance). This circuit should never sound this good, but passive components (important) and multiple diameter silver cables (very important) produce a seriously good sound.
What can you reveal about the input sensitivity and gain?
The gain can be increased by 6dB by soldering link on PCB, but 60dB gain at 1kHz is a very good compromise for a MC phono stage
Why are external loading controls are, in the context of your design, clearly irrelevant? Or put another way, how you achieve this incredible performance without all the other paraphernalia that adorns some other state-of-the-art phono stages?
Circuit complication of expensive phonostage are a DISADVANTAGE. (Dr Bews' capitals, not mine) Take for example input loading. LFD use special resistors like Shinko Tantulum resistors (no longer in production, but I have approx 10,000 in stock, which cost me over £20k) in the MCT which is much more important for sound than changing loading with poor sounding metal film resistors using poor quality switches.
EQ components are very important, and the capacitors used in the MCT are obsolete capacitors. They have a special sound which can never be reproduced by other types of capacitors. It’s lucky I have 1000's of these in stock!. Like cooking, use the best ingredients and cook properly – to produce a very good meal
What currently in your view is the limiting factor on phono stage development – other than R&D time?
Design is probably not the limiting factor – it’s probably the PSU quality, layout and passive component selection. Advances in these areas will push the sound forward until we can hear the circuit – but I think that currently we hear more of the implementation of the circuit design, rather than the sound of the circuit itself
We do a SE version of the MCT where improvements in the PSU (SE is £4k retail) that makes the sound have much more foundation, with blacker silences, i.e. a noticeable and worthwhile improvement.
For people with vinyl-only systems who do not want the additional circuitry of a preamp, what would you advise as a no-compromise external volume control. A passive unit perhaps?
MCT with any external passive volume control will compromise the sound of the MCT because you will hear the pot (it’s a passive preamp effectively). The tonal balance of passive preamplifiers are typically not sonically correct (lacking in bass slam, but are often open and transparent), and this bass slam deficiency would need to be addressed. This can be done, i.e. switched attenuators and multiple diameter silver wiring. Unfortunately though – to do this properly, this section would cost more than the MCT itself! Most passive preamplifiers are not good enough in some sonic respects and the best of them are as expensive as mains-driven linestages. My advice is to choose the best sounding mains preamplifier you can afford, since you not going to save money in the long run. As the Americans say, there are no free lunches.
Any advance information re an Anniversary phono stage – even though it doesn’t appear in the current product listing?
Anniversary phono stage, if I get around to it – which is doubtful currently – is going to be based on a mains powered version of the battery disc stage. Unfortunately it is going to take a VERY long time to produce anything sonically better than the MCT (or SE version of it) as it is currently, since the these are rather good, i.e. they are the best sounding mains powered phono stages regardless of price – as you and your customers have discovered.
Would an MCT with onboard volume control (not remote though) for dedicated minimalist vinyl enthusiasts ever be considered by you?
A MCT with volume control may be possible with modification to MCT phonostage section, so that the volume control matches with MCT phonostage section to give very good sound that approaches MCT in combination with very good linestage (or very good passive pre). Do not expect this product to be under £5,500 retail.
A bit steep for a volume control isn’t it i.e. another £2.5k?
The cost of the volume control is miniscule. So are the mods to the case work. However for reasons I’ll explain at some later point on your site Howard, the physical position of the circuit board has to be changed. This in turn requires a lot more of the complex 5 different diameter high purity silver strands I mentioned earlier. It is a complex task. And that’s just the tip of this iceberg.
*******************************************************************************************
LFD's founder Dr Richard Bews - interviewed by Stereonow's founder Howard Popeck in 2006
LFD is an enigmatic company. In many ways their public face and their market position mirror that of Michell Engineering in that technical excellence is paramount but the enthusiasm for playing the hi-fi PR game is non existent. The driving forces at LFD have impeccable engineering pedigrees, just like the late John Michell in his organisation. LFD’s managing director and principal designer is Dr Richard Bews. The technical director is Professor Malcom Hawksford.
Their public presence in the audio world is virtually non-existent, yet their cult following is ultra loyal.
Again, like Michell Engineering, they punch far above their weight. They don’t have a production line. Each unit is hand built and while some of their range command high prices, the lower end of the range offers remarkable and, arguably, peerless value for money in terms of sound quality, but unlike Michell Engineering, not aesthetics. Stark is perhaps the most appropriate word to describe LFD’s casework.
LFD argue, why produce great boxes and average chocolates when with a little thought you can produce the reverse? Their products rarely appear on the second hand market primarily because of their peerless attitude when it comes to customer upgrades to currently owned units.
Quietly and without fuss, this unassuming little company are slowly but surely driving forward progress to making sounds approaching live music and speech, via digital and analogue in the home.
Dr Bews is not a man to suffer fools gladly. In hi-fi terms where for so many years a loud voice and monumental ego are the primary tools for publicity, Bews is in our industry a virtual recluse. But as you’ll read in this exclusive interview by Howard Popeck, he is not without humour, and certainly not short of practical, intelligent and focused thought.
Rumour has it that you’re undertaking pioneering work regarding silver interconnects and speakers. Can you enlighten us?
I've been interested in the sound of cables from our inception, since it was obvious that in highly transparent systems, cables are very audible. As a general rule, our electronic products reduce the signal tracks on the PCB to a minimum. Typically our signal tracks are 5 to 10 times shorter than most of our competitors.
This is all well and good, but the signals have to travel through wiring within the equipment. Therefore special wiring is required to reduce the degradation present in the cabling used. That is why cables are very important to us.
From the developments of our electronics came cables that we sell as accessories. We do not use standard copper or silver conductors and often use hand assembly, to produce the sound we require from the cables.
Silver versus copper?
Although pure silver is a better conductor than pure copper, poor quality silver often sounds a lot worse than poor quality copper.
Firstly, choose a very pure silver and make sure that the crystal structure is defined and not distorted by the mechanical processes used to form the strands. We generally prefer continuous cast silver of at least 5 nines pure, however I don’t believe purity of the silver (or copper) beyond 6 nines pure has any benefit.
General experimentation, listening tests, shows that a given diameter of silver or copper has a subjective ‘frequency response’. Unfortunately I've no technical explanation of why this appears to occur, except that the electromagnetic radial field structure, that gives rise to skin effect within the conductor when a current flows, may be responsible.
The subjective ‘frequency response’ of a conductor is tilted towards the low frequencies as the diameter increases. This is why certain diameters are chosen for single strand cables; i.e. for interconnects around 0.6 mm diameter for copper and 0.9 mm diameter for silver. I must emphasis that the dielectric surrounding the conductors, the material used for the outer jacket and exact crystal structure of the conductor is an influence on the sound as well.
Through subjective evaluation, I've established that when many single strand conductors of the SAME diameter are put together to form a multistrand cable, the subjective frequency response is widened relative to the frequency response of the single strand. This would suggest that multistrand cables are therefore more preferable to single strand cables, However, I've not talked about the subjective ‘time response’ of single or a multistrand cable.
Which means what exactly?
Unfortunately it appears that a single strand cable seems to ‘ring’ at a particular frequency (related to its diameter) and has a fast decay. Whereas multistrand cables on the other hand, again ‘rings’ at a particular frequency (related to the diameter of the strands), but the ringing is much longer in duration.
I believe that neither cable solution is correct, since the fast decay of single strands tends to lead to subjective comments like: tight, fast sound, but not rich or warm. Whereas a multistrand cable can often sound: warm, slow with some smearing.
I must point out that copper is much easier to work with, since its subjective ‘frequency response’ is wider than that of silver for comparable single strand conductors. That is why many silver cables appear to sound not quite right, sometimes they might be a little bright or coloured.
Cable evolution at LFD?
Over the past five years I’ve been experimenting with different diameters of silver, which form a multistrand conductor. Although this method has been tried with copper, e.g. Cardas, Van den Hul, and others, I believe that silver is better, because it possesses better purity and transparency compared to copper.
Essentially the composition and number of different diameters of silver within the multistrand give rise to the overall subjective ‘frequency response’ and ‘time response’. I suggest you think of each different diameter of silver as a bandpass filter, and the overall response is formed from the combination of each of the separate filters. The trick is to make sure that the overlapping between successive filters is correct, which subsequently dictates the exact diameter of each strand of silver chosen.
As the experimentation progressed, it became apparent that the sound became better as the number of diameters increased, which dictated that the difference between each of the diameters reduced.
This is a very labour intensive job, but now we have a formula to determine the size of each strand within the multistrand. If this experiment was only to enable the manufacture of neutral sounding cables. then really it would be a waste of time and money. But what seems to happen subjectively, as the number of strands increases, is that more detail is revealed in a very natural manner, because time smearing seems to be greatly reduced, so giving greater separation of instruments and vocals.
Currently, our best cable uses 23 strands of silver possessing 19 different diameters.
What’s your view on the SACD versus DVD-A debate?
What debate? Are you referring to the advantages of one format over the other regarding sound quality?
Yes
For audiophiles interested in two-channel reproduction, I think there’s no debate, since neither format will have much relevance. This is because the software is the most important part of the equation. I think that SACD software is going to have the largest number of releases and therefore would be the standard that audiophiles would have to adopt.
Unfortunately these releases are really aimed at the 5.1 market, and not at the very small audiophile two-channel market. This would then leave DVD-A, where, unfortunately the number of titles is never going to be significant.
Which means?
There’s not going to be a true high-end two-channel digital standard with any relevance to the two-channel market.
Your preference?
Personally, I prefer the concept of DVD-A, i.e. 24 bit 96 kHz, since it might be brute force, but it’s likely to produce the goods.
Your prediction?
Standard CD format of 44.1kHz, 16-bit will be the high-end digital standard for audiophiles using two-channel software.
Are we likely, in the foreseeable future, to see an LFD SACD and/or DVD-A machine?
No. I’m only interested in the two-channel market, small though it is.
Why?
This uncertainty (format wars) in the market has effectively destroyed the market for very high performing two-channel digital replay, except for components that offer the capability of decoding both formats. What’s more unfortunate, is that most audiophiles don’t have any (or only a small amount of) software to take advantage of either SACD or DVD-A.
Anything else?
Yes. Components that are capable of decoding SACD and/or DVD-A will not have any advantage in replaying conventional CD.
Remember that high quality modern CD players and D/A converters have resolution beyond the 16-bit 44 kHz standard. They possess this extra resolution so that they can decode CD very accurately. I think that 16-bit 44kHz is not that bad sounding, it’s just difficult to produce digital replay equipment that can exploit the potential that CD already possesses.
The increase in resolution that SACD and DVD-A can give is of little use, if designers are only exploiting a small portion of the performance that the humble CD already gives. Remember that D/A converters and CD players require the designer to be skilled in both digital and analogue design, no mean feat.
You’re known to have firm views on balanced inputs and outputs. What’s led you to those conclusions?
LFD believe in simplicity, so over-complication should be avoided. Balanced operation has advantages in environments of high noise or where cables have to be very long. Studios are a good example. The domestic environment normally doesn't require balanced operation from a technical viewpoint. Balanced operation came about, because certain US manufacturers could give some technical bullshit to sell their amplifiers!
Balanced operation invariably adds complication, which can lead to problems, so why make the life of an amplifier more difficult than it needs to be?
It’s difficult to make a simple amplifier sound really good, so doubling the amount of circuitry (as often happens in balanced amplifiers) just makes it more difficult. When LFD amplifiers interface with amplifiers with balanced inputs or outputs, we prefer to use specially designed transformers to convert between single-ended and balanced operation. This is simple, but does put demands on the circuitry and the design of the transformer.
In your experience, do mains leads really make a difference, and are the expensive ones really capable of offering value?
Mains cables do change the sound of audio equipment. Unfortunately there’s a lot of claims made about mains cables that are exaggerated. I suggest that the effect of a mains cable is not as important as interconnect, but is still significant. Value for money judgments are really down to the system, where these relatively small improvements make the difference between a good listening experience and an exceptional experience.
What I do NOT condone is high pricing for cheaply manufactured cables. If a cable must use expensive materials to realise truly high performance, then so be it. But the use of expensive materials for bullshit (leads to high prices) but not for sonic improvements and is not the way forward. It just upsets customers.
I believe the reason why the performance of audio equipment is altered by mains cables, may be due to the influence of RFI on the mains. Generally I’ve found in the development of mains cables that dielectrics with improved high frequency performance possess better sound quality compared to normal PVC insulated cables.
To date, you've chosen not to build digital inputs into your amplification, even though you built CD players and DACs. Why is this?
One has a decision regarding gain control (pot) and input switching, i.e. to perform these functions in the analogue or digital domains. At LFD we believe that these two functions are best carried out in the analogue domain, since a well-designed analogue linestage possesses a much lower noise performance compared to the output of a DAC; and also NO analogue source is compromised by its conversion into the digital domain using A/D converters.
Obviously digital sources must be converted into analogue prior to gain control and switching. Our digital to analogue converter switches between three digital inputs, so multiple high quality digital sources are catered for. The inclusion of cheap D/A converters in an analogue amplifier would compromise the performance of digital sources to an unacceptable level. It takes a lot of effort, time and money, to manufacture excellent sounding D/A converters.
How do you envisage DAC technology and, by implication LFD DACs evolving?
Digital reproduction will always evolve, but sometimes it may go in directions that may seem a little strange. For two-channel reproduction, I feel that the digital side, i.e. over-sampling (upsampling) and D/A conversion is already very good and has only a small way to go now. I admit that the sound varies between the actual implementations of digital processing chips and D/A converters, but most are fundamentally very good (they should be, since CD was born 25 years ago). However, certain aspects of the analogue and power supply design may need to be improved significantly. Experiments over the years have shown me that worthwhile improvements of LFD digital products have come when these two particular areas are addressed very seriously.
Hype, spin, hyperbole and reviewers?
I must state that I have great misgivings about how certain products are described and elevated in status by the reviewers, particularly in the area of digital reproduction. A lot of D/A converters are quite good, but unfortunately many amplifiers don’t allow them to shine sonically. This applies to cheap CD Players in particular. It’s amazing how good some of the old CD players sound when played through very transparent amplifiers and loudspeakers.
Also I find that a lot of so called ‘improvements in sound reproduction’ are not at all, but are tonal or presentational changes that suit the particular reviewer's system.
We must differentiate between synergy attributes and tonal balance etc., and absolute improvements, where the reproduced music starts to imitate real music. This is very difficult, since humans normally identify the more obvious tonal inconsistencies first, but it’s the subtle details, space around instruments, musically and scale that really distinguish the men from the boys.
Support shelves for amplification and DACs. What’s your experience and your advice?
Unfortunately, the mechanical performance of supports for audio equipment do have an influence on the sound. That's why at LFD, we're currently developing mechanical isolation devices. Every material and construction employed in a support, lead to mechanical resonances in the structure, so experimentation is required. The use of spikes can be beneficial, since these devices couple the mechanical vibration at a particular place on a platform to the audio equipment. You just have to choose a location where the mechanical vibration is the lowest. It’s a little hit and miss.
Remember that spikes are NOT mechanical diodes. Mechanical isolation is different to self-resonance of structures and must be designed carefully to have good isolation from low to high frequencies, so that little mechanical energy from the outside environment is allowed to get to the equipment. Also it’s a good idea to make sure that any mechanical vibration produced by the equipment is also absorbed.
Some of our customers have their amplifiers modified, to mechanically damp the electrolytic capacitors with a sticky mastic compound. The complete solution to vibration problem is not easy and is likely to be expensive. Unfortunately, affordable stands are not that good, so a customer normally chooses the sound coloration they prefer.
I generally recommend that the same proportion of the total system cost be paid for mechanical isolation, as you pay for interconnect and loudspeaker cables, around 15% to 20%.
What are your preferred musical tastes?
In a word, varied. It has to be like that or I’ll be in danger of biasing the sound of LFD products to suit one type of music. I normally use Classical, Jazz and Rock music for evaluation. I have to say I only use Pop music a little, since it normally has clearly defined sonic problems.
I use difficult recordings a lot, since all equipment sounds good with simple well recorded music.
It’s only when music gets difficult with many instruments and vocals mixed together, that really good equipment shines through. I also believe that correct sound reproduction will not favour any type of music, it’ll show the music lover the particular strength of each type of music. It’s the emotional connection of the listener with the music that is so important.
Do you still listen to vinyl?
No so much these days. I do feel that the ultimate sound, still comes from the very best analogue systems, however the price is extremely high (e.g. £50,000 plus for front end). Also I think that digital reproduction can be better value for money, since a very well designed D/A converter (and transport) can sell for under £7000 and can give reproduction that analogue front ends would find difficult to match at the same price. We put a lot of effort into the design of our phono stages, so that analogue replay can give digital a good run for its money.
What turntable do you use?
For development, an old Linn with a Zeta and an Ortofon. Overall though, the Pink Triangle Anniversary and the TNT give me more enjoyable sounds. I’ve been influenced by their reproductive qualities. The Linn is there because it’s relatively neutral in the mid band and anyway, it’s a tough design challenge to build a phono stage that can cope with the Linn’s unusual bass characteristics.
You work closely - but not exclusively I believe, with Burton Somervell speakers, but presumably there are other speakers whose quality you respect? Is that correct?
I respect any loudspeaker designer that produces products that are neutral, transparent and possess good imaging. My electronics suit these type of loudspeakers, since there’s an obvious synergy and consequently a very good sound. Remember that slight irregularities in the frequency response are not particularly relevant to me. Generally I find that the following loudspeakers work with our electronics: Burton Somervell, Tannoy, ProAc, AudioPhysic and Neat loudspeakers. Just give me good honest pair of loudspeakers and its amazing what can be achieved sonically.
What sources and speakers do you use when developing your equipment and your cables?
The source is CD, either our DAC3 or Mistral CD player. The use of good CD or Analogue LP is more important than the loudspeakers, since garbage in, garbage out. I only use CD because it’s far more convenient than LP, and fundamentally CD is no better or worse than LP, just slightly different. However, as both mediums get better, their sound gets closer, so that you cannot discern what medium you are listening to.
What makes a good source ‘good’?
Accuracy and musicality. That applies to CD as well as Analogue LP. To achieve the best from either medium requires different skills, so it’s not surprising that when CD or Analogue LP go wrong, they exhibit different compromises.
What speakers do you use during development?
I use a pair of modified Celestion SL600 loudspeakers, rewired and the crossovers outside the box.
These modifications helped immensely, to relieve the rather laid back, and slightly congested sound. Really I don’t think the absolute quality of the loudspeakers is that important when developing our products, since the important characteristics of the system prior to the loudspeakers are obvious with any decent loudspeakers, even £100 loudspeakers.
However, it’s really nice then to use an excellent pair of loudspeakers like Burton-Somervell to get that little bit closer to life-like sound.
Do you offer those modifications to SL600 owners?
No.
But you advocate putting the SL600 crossovers outside the box, right?
Yes, absolutely, and re-wiring too. That way, a lot more of the potential of that pioneering design, the great ingredients, becomes apparent.
Your Battery DiscStage (phono stage) is your most expensive product. Where does the cost go?
I actually don’t get the usual manufacturer’s mark up with the Battery DiscStage, since it would become prohibitive in price. There are very expensive parts in the battery DiscStage, i.e. 26 off sealed lead acid batteries, 1W Tantulum film resistors, switched attenuators, pure silver leaf capacitors and silver wiring.
This is NOT a commercial product, but a statement of design and passive component selection. It also takes one week to manufacture this product.
Some expensive products are not statements like this particular example. Just look inside a product and make an estimate of the parts cost and the labour. This exercise should then produce the end retail price. It appears to me that some products are trying to recoup their R/D costs, and maybe more.
Recouping your R&D costs?
At LFD we don’t take into consideration my time in developing products since I assume that a small amount of money - less than 5% of cost - is used to pay for R/D.
Sometimes this doesn’t work out, but other times it does. I don’t replace products every year, so that over the life span of any particular product, R/D should be paid back. I do this, since I believe that fashions come and go, but quality remains. Also when I improve products, existing customers can upgrade for modest cost, rather than be forced to purchase a new replacement product and have to sell the existing product.
Here at LFD, ours is a long-term strategy, not a way of getting a quick buck. I also feel that a lot of ‘improved’ Mk2 or replacement products are not, but can be less good or just different. This does our industry NO GOOD at all. If a product truly requires expensive parts and high labour costs to produce magical sound, then so be it, but do not over-engineer a standard product, just to push the price up. The best products are never cheap, but the most expensive products are not always the best.
Digital Room Correction is a hot topic in some quarters. Any thoughts on the viability of the technique?
This is not a trivial problem to solve and invariably leads to compromises. It’s very rare to hear any really successful implementations, although one day a good compromise will be found. I actually believe that a more pragmatic solution to the problem is to reduce the problem at source, i.e. prevention not cure.
The placement of loudspeakers within the acoustic environment has a large influence on the resultant sound, so take great care. The application of sound absorbing materials and reflective materials can also be beneficial. From a purely philosophical viewpoint, the application of advanced digital processing may have some benefits, but the side effects may well outweigh those improvements.
Thank you Richard
Thank you Howard.
********************************************************************************************