What motivates someone to do-it-themselves? This may at first seem a strange question to ask in a Hi-Fi column, but it has some relevance. In an age when the market is saturated with relatively inexpensive imported equipment it is surprising the number of people that are still prepared to spend plenty of spare time building kits or projects from journals.
Mind you it's much easier for the home constructor now that most projects come complete with their own PCB and the easily obtainable parts, at comparatively low cost. Bearing this in mind I feel that there is another market yet to be exploited in the DIY electronics field. I refer to making equipment completely from ready assembled modules.
At first sight this may seem a strange idea, but consider the attractions. First, the specifications can be guaranteed from unit to unit. Second, the actual amount of hard graft in making, for example, a tuner amplifier is reduced to correctly interwiring PCBs. Since this is so, the constructor whose competence would not stretch far enough to correctly assemble a kit amp would have little difficulty in assembling a complex piece of equipment. Last, but by no means least, this method is sometimes cheaper than actually building everything from scratch.
In any case a saving over buying the shop-built equivalent will be made. This method of construction takes us into system building where the important thing is to correctly interface the modules. Unfortunately the last thing usually considered is the interfacing and a little work on this aspect of design often works wonders!
Getting away from this idea, slightly at least, I thought that I might turn to the problems of interfacing from another angle. In particular, the equalisation required by magnetic pick-up cartridges.
The modern microgroove record poses many problems for the designer which range from getting large amplitude signals onto the vinyl, to the flatness or otherwise of the completed product. Unfortunately this last problem appears to have become insoluble within the last few years.
Since the oil crisis, vinyl has become increasingly more expensive. The record companies' response has been to decrease the thickness of the record. Inevitably the quality has suffered. Modern pick-up cartridges are easily capable of reproducing warp signals. These are subsonic and of large amplitude.
Once they have entered the stereo system they cause havoc. Being subsonic they cannot be directly heard. However the wanted signal is carried 'piggyback' fashion on top of them. The result is premature clipping. Where this occurs on the waveform being reproduced is dependent on the amplitude of the warp.
Large subsonic signals are also bad news for the output stage of the amplifier which wastes power reproducing them. Possibly more important is the intermodulation distortion between these signals and the bass. It must be said though, as a rider to the last observation, that little research has been done in this area. The usual solution to such problems is to use a high-pass filter or a rumble filter in front of the power amp. Unfortunately, these are often badly designed. Either the cut-off frequency is too high, losing valuable bass, or the rate of cut-off is too low resulting in inadequate rejection of unwanted signals.
A simple rumble filter that can be constructed on a piece of Veroboard is shown in Figure 1. A word of explanation is in order.
Most musical signals contain little information below 40Hz. The main exceptions to this rule are organs and bass drums. For example, the bottom note from an organ can be as low as 16Hz. Orchestral bass drum has a fundamental down to 35Hz.
Record warps however occupy the 2-10Hz region. A suitable compromise cut-off frequency for our filter is 20Hz. This coincides with the low frequency limit of human hearing. The rate of cut-off can be 12db/octave allowing the use of the well tried Sallen and Key filter configuration. It is also important that the device used doesn't degrade the performance of the amplifying chain in any way. For this reason a J-fet op-amp (the LF351) is used. This device has a fast slew rate, 12V/us. and used in the 100% feedback mode, has a distortion of less than 0.002% at 2V RMS output across the audio band.
No problems should be encountered as long as the output and input are well separated physically. Power can usually be obtained from the power line in the amplifier or preamp. The unit should be placed between the pre-amp and poweramp.
In order to preserve a good S/N ratio, records are cut with treble pre-emphasis and bass de-emphasis. The latter is required since bass signals generally have a larger amplitude than those in the mid-range and treble.
If the bass were to be recorded at its normal level then the groove spacing would have to be greater to prevent breakthrough between tracks.
In order to reproduce a flat signal from a record, a filter network is required to de-emphasise the top and emphasise the bass. The required curve is standardised with the turnover frequencies being set at 50Hz, 500Hz and 2.1kHz. What happens is that the response is flat from 20-50Hz and then falls at 6dB/octave. At 500Hz the response is 3dB up with reference to 1kHz. Between 500Hz and about 1.5kHz the response is flat but falls away again towards the top end being 3dB down with reference to 1kHz at 2.2kHz.
Since the introduction of linear ICs, manufacturers have aimed to produce high quality devices that would simplify the design of audio front ends. Until recently their efforts have only approached the performance that can be obtained from a good discrete design. Within the last year though, Hitachi have introduced the first of a new generation of front end ICs.
The new device is the HA12017. The distortion generated by this device is extremely small - 0.002% at 20kHz! Combine this with a 10V RMS output capacity and a quoted signal to noise ratio of -72dB giving you a potential world-beater. Moreover the device is not that expensive being in the region of a pound. Unlike the now familiar dual pre-amps, you'll need two ICs for stereo and the circuit is slightly more complex than usual. Even so the chip is certainly value for money.
Whilst we're on the subject of front ends it may be as well to dispel one myth in particular - disc signal to noise ratio. Most people would assume that the lower the noise from the pre-amp the better. In general terms this is sound logic but there are special considerations which apply to disc inputs.
All pick-up cartridges generate their own noise which is proportional to their impedance. Without filling the page with calculations it can be shown that about 1uV is generated by the average moving magnet designs.
This means that the best signal to noise ratio that can be obtained practically is -65dB for a 2mV input. Unfortunately to obtain this level of performance the pre-amplifier must be noiseless!
By an extraordinary set of coincidences I met Simon Bantly recently, the designer of the Black Triangle speaker system. These were recently demonstrated at a London exhibition and showed their unusual performance and excellent bass response.
The speaker system itself is a three drive unit design with infinite baffle loading on the bass unit. Despite the loading and relatively small volume, the speaker can deliver a usable output at 20Hz! Not unnaturally, Simon did not care to divulge how this was accomplished and several patents are pending on the principles involved. Since we live only ten minutes walk from one another, I took the opportunity to go and listen to a pair in domestic surroundings. I was impressed with the overall response and the only fault that I could hear was some slight sibilance on vocals. The mid-range was nicely detailed and the bass was reproduced with its proper weight and without any boominess.
One of the major problems encountered by those who possess speakers with a good bass response is the excitation of room resonances, this problem has been to a large extent overcome in this design by employing a 'suck-out' filter that operates in the 40Hz region. This frequency coincides with the main resonance of many domestic rooms.
As with all things audio, the only way that it can be properly assessed by the potential purchaser is to listen closely and judge by what you hear. As these speakers are expensive (£1,400 per pair), nothing less than a home demonstration should be demanded before parting with your money.
Another interesting characteristic of these speakers is that they are at least partially omnidirectional. Stereo, as we have come to know it, is light years away from the live experience where direction location information is largely absent. How much the image means is dependent on the individual listener. As an interesting footnote to the above Simon uses as his reference source a VMS20E MkII fitted in the Acos Lustre arm. This combination he believes offers all that can reasonably expected from disc. As I use this combination myself I tend to agree!
Feature by Jeff Macaulay
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!