Published On: Fri, Mar 25th, 2011
Uncategorized | By Dave Peterson

Wired For Sound: The Great PCB Debate

If you use a valve amp, you’ll know how much importance is attached to hardwiring. Hi-tech manufacturing is a must when it comes to keeping costs down… but why is it that our ears can hear the difference? Dave Petersen investigates.

During the 1950s and ’60s, when valve technology was standard in consumer goods, the cost of making a hand-wired amplifier was kept affordable by the low prices of valves and other components needed for the large-volume production of radios and televisions. Since then valves have disappeared from consumer goods, leaving music products particularly guitar amps as the main market.

Suddenly, good valves are expensive. In recent years prices have risen to several times their former level, and although this can be said of many other kinds of product, few can match the steep upward price curve of valves and their supporting components. The truth is this: with labour-intensive assembly procedures factored in, valve amp manufacturers are having difficulty keeping their products within reach of guitarists. There are solutions, and in this article we’re going to examine them while keeping an eye on the most important thing of all the way amps sound.

PCBs: The answer?

One way makers keep costs down is by using the assembly speed advantages of printed circuit boards (PCBs). PCBs are essential to all modern consumer electronics, and a valve amp made with them is significantly lower in price than an equivalent hardwired one (Pic 1). This is because it takes about a quarter of the time to build, and there’s another bonus you get a more consistent product that’s not likely to need much postassembly quality control. For every 10 handwired wiring workers, there’s a skilled technician whose job is to check wiring and correct errors, and this adds further to the cost of hardwired products.

Why then do hardwired amps still have a market, rather than simply leaving it to their more reasonably-priced PCB counterparts? Is there a discernible difference in sound quality? Many guitarists say there is, and this may account for the survival and the recent growth in sales of old-fashioned hardwired amps.

Taste The Difference

From an engineer’s viewpoint, using conventional test methods, there’s no measurable change between PCB and hardwired versions of the same model of amp in standard parameters like output, frequency response, noise and distortion. Induced hum the bugbear of hardwired assemblies is often lower in the PCB version, although this can vary with the PCB layout (indeed, some respected old valve amps used PCBs very early on without noticeable ill-effects). In others that transferred later to this technology the difference is apparent to the player, and that’s useful in helping to determine what causes this impression.

The words ‘to the player’ are significant, because it’s doubtful if it would be noticeable to the average listener. But ask a guitarist to comment on a hardwired amplifier and a later version of it made with PCBs and the answer will probably be that the amplifier ‘feels’ different.

Soft Touch

Touch response or how a guitar’s strings respond to the player through an amp is a difficult aspect of an amp to assess technically. It’s governed by more than one factor, an important one being the amp’s input impedance, which determines how much it loads the guitar’s pickup. Equally important is group delay, the time taken for the magnified incoming signal to appear at the amp’s output. Group delay is the sum of the transient responses (also called rise time or slew rate) of all the stages in the amp’s signal chain. A difference of more than 10 microseconds is perceptible as a change in the touch response. A frequent comment of players checking out early digital modelling amps, whose group delay is inherently big because of the time needed to encode and decode the signal, was that their strings felt a gauge heavier. This is a worst-case example, and this type of amp has greatly improved recently, but it identifies the area we should search to find the causes.

Published On: Fri, Mar 25th, 2011
Uncategorized | By Dave Peterson

The Sound Barrier

If you use a valve amp, you’ll know how much importance is attached to hardwiring. Hi-tech manufacturing is a must when it comes to keeping costs down… but why is it that our ears can hear the difference? Dave Petersen investigates.

Common sense tells us that such delay figures should be swamped by the time it takes for the sound to travel from the loudspeaker to the ear, but we must also consider the ability of the brain to compare the time taken between sending its impulses to the player’s fingers and the time of arrival of the sound pressure wave at the eardrum.

Our brains are sophisticated things. A musician can detect a difference in pitch between two simultaneously played notes of less than 10 cents, although few can name with certainty a single note played in isolation. So it’s easy to factor in constants like the speed of sound when assessing two amps played back to back. As in many other kinds of experience, we seem to be more alert to differences than to common ground.

PCBs OLD and NEW

You might argue that the increased complexity of modern amp circuits might account for this, but that doesn’t explain the perceived difference between the same model built with the two methods. Considering the electrical character of a PCB made using modern techniques may help us find the explanation. In the early days of PCBs, artwork was laboriously laid out by hand using special drafting tape on clear film, a lengthy and expensive process that also required photography. This drove a preference to keep tracking as short as possible, orienting components to bridge the board with their own size and wire length, the tracks serving primarily as a method of making junctions (Pic 3). Larger boards used widely spaced tracks with lots of curvature (Pic 2). From an electronic viewpoint, this wasn’t too different from the tagboards and three-dimensional birds-nest assemblies of hardwired amps. But as circuits became more complex and components smaller it became necessary to line them up in dense rows to squeeze enough of them on boards that got more crowded with added features, and link them with circuitous tracking made using software-driven drafting programmes. Closely spaced parallel tracks, the default preference of software-generated layouts, are a common feature of PCBs drafted like this (Pic 4).

Jumping The Tracks

This kind of tracking is used in computers, but computer circuits have a speed and bandwidth far greater than those of valve amps, so its effects are arguably negligible. Also, this takes no account of the much greater impedance and signal level differences in a valve circuit. In computers, impedances are generally no more than a few hundred ohms, with signal levels confined to 5 or 6 volts at most. In a valve amp they can range over several megohms, with signal levels up to 50 volts. A feasible intertrack capacitance of 10pF (closely spaced parallel tracks measure about 1pF per cm), equivalent to such a capacitor being connected between signal and earth in a valve circuit, would have little effect at the low impedances of digital switching circuitry, amounting to a resistance of 50K ohms. In the grid circuit of a valve long-tail pair driver stage, with working impedances of several megohms and gain about 50 times, it affects much lower frequencies, feasibly from 5Khz upwards. These frequencies are within the range of touch perception and have a measurable effect on transient response in high-gain valve stages (see our techie sidebar).

Published On: Fri, Mar 25th, 2011
Uncategorized | By Dave Peterson

A Workable Solution?

If you use a valve amp, you’ll know how much importance is attached to hardwiring. Hi-tech manufacturing is a must when it comes to keeping costs down… but why is it that our ears can hear the difference? Dave Petersen investigates.

Part of my own work, in response to customer demand, is to return PCB-built versions of classic amps to the sonic character of the hardwired originals. One popular modification is of a famous amp that was re-introduced in an updated version a few years ago. I normally remove its PCB-based preamp assembly and replace it with a hard-wired one, but an equally significant improvement is to disconnect from the power-stage PCB the resistors that couple the driver stage signal to the output valves and mount them on old-style tagstrips with flying wire connections. With this done, the amp ‘comes alive’, in the words of one customer.

Although hardwiring the preamp board makes a tonal difference, the amp only regains the touch response of the hardwired version with this modification. This is the part of the circuit where impedances and signal levels are at their highest and small changes could be expected to have a significant effect.

I wouldn’t suggest that this is the only way of solving the problem. Recently, one or two manufacturers have reverted to an earlier type of PCB layout, using a ‘trackless’ PCB on which the components are fitted between turret tags whose connections are on common printed sections of board, avoiding interconnecting tracks (Pic 5).

While this is only practical with uncomplicated circuitry, it offers a solution for building older designs, which aren’t usually densely populated, on PCBs, so most of their cost savings can be realised without sonic penalties. The claims of guitarists to hear differences that can’t be effectively measured have traditionally been derided by the technical fraternity but now they’ve become a starting point for any amp manufacturer who wants to be taken seriously.


Techie Sidebar: Transient Response Measurements

We were curious about how a subjective phenomenon like the ‘touch response’ of a guitar could be quantified, so we did some benchwork (Pic 6). A possible method of assessing this aspect of amplifiers is the square wave test. A perfect square wave has a practically non-existent rise time it goes from zero level to max in a microsecond. Observing the output of an amp or circuit driven with a square wave reveals a lot about its transient response behaviour, which is an important element, if not the key, to touch response. No valve amp tracks a square wave perfectly, and the slight slope given to its leading edge by most analogue amp circuits is used as a measure of their success or otherwise in following its instantaneous rise-time. The difference between the vertical rise of the test-signal and the slightly less vertical version of the amp’s output is measurable in S (microseconds) on the horizontal ( X-) axis of an oscilloscope.

To test the effect of long parallel PCB tracks on transient response, two such tracks were isolated and a I kHz square wave applied to one end via a resistor, to simulate the high source impedance of a valve amplifying stage, the other track being used as the earth return. This resistor causes the squarewave to slope, even without connecting it to the PCB, because it interacts with the capacitance of the scope’s test probe but the slope increased when the end of the resistor touched the PCB track. The rise time reading the generator direct was measured at 1S, via the resistor on its own 10S, and 15S with the resistor connected to the PCB (Pic 7). The tested track distance of 10cm is not uncommon in larger layouts. Often more than one such circuit would be found in an amp, and three such stages might add 15S to the rise-time over the hardwired version. It’s the difference between a high-quality guitar cable and an average one.

Leave a comment

You must be Logged in to post comment.