Simple high performance circuits
There's nothing quite like sitting back and admiring the view of a warmly lit pair of elegant analogue VU meters responding in perfect synchronisation to whatever you're listening to. True, there are many great digital level indicators, even some that attempt to emulate a moving coil meter on-screen. But these can only be used for PC audio, and aesthetics wise (IMHO) don't even come close to the real thing.
VU meters are a way of measuring the level of an audio signal. They measure the an average of the signal they are presented with, not the peak level or the RMS level, which is displayed in decibels in respect to a preset level.
Figure 1. Classic moving-coil VU meter from a Teac cassette deck
Even today analogue VU have some very useful applications for professional and non-professional use alike. They are invaluable for monitoring average levels and give a good indication of the perceived loudness of the source they are connected to.
For example, they can be very effectively deployed to show how 'loud' CDs and other digital audio sources have been mastered when connected to a CD player standard audio interface. One of the especially nice things about this is that a full scale output from the vast majority DACs is around 2V RMS. A VU meter configured for professional line level reads 0dB when the input has an average of 1.125V meaning that a reading of +4dB will be shown when a sine wave of 2V RMS is present, the maximum undistorted level. The same can be done with phono input, once it's past the preamplifier of course, (in this case a consumer level configuration is best with 0VU displayed at 300mV) and the differences in levels between certain labels and formats may surprise you.
There is also purely the aesthetic function of having them on a unit, a perfectly valid reason. After all, the end goal of HiFi is enjoyment!
VU meters have been around for over 75 years now, since Bell Labs and various broadcasting companies in the United States standardised a method of measuring average levels in telephone lines and other signal carrying cables. The standard requires fairly strict tolerances in terms of the mechanics of the moving coil meter itself as well as the electronics. In order to provide any meaningful indication of an AC signal a rectifier must be used to generate a proportional DC voltage across the meter coil. This rectifier consist of fairly esoteric low voltage drop elements so as not to decrease the meter's sensitivity to small signals. All this makes true VU meters rather costly. Even second hand meters can cost upwards of £70 per unit. For a new unit expect to pay perhaps three times this price. The basic requirements of the standard are listed below.
It should be apparent that these targets will not be met in the vast majority of domestic (and sometimes even professional) cases. You can rest assured that any VU meters on domestic recording equipment such as cassette and reel to reel recorders will not conform to the specifications listed above. This isn't too great a worry as all that we really want is to see a reading of 0VU when the signal reaches a level where distortion will be likely and does so without any significant overshoot, insensitivity to low levels, or a delay of over 500ms or so.
Having said that, it's always good to try and get as close to the real thing as possible when working with lower cost non-standard meters. Further reading will show that there are even a couple of improvements that we obtain by using non-standard drive circuitry, sensitivity being a major one as the drive voltage has to push through the rectifier before it can get to the meter itself. A major problem in cheaper non-standard meters.
The VU meters shown in the schematics below will be standard 200uA types, unless specified otherwise.
All VU meters are essentially composed of some sort of rectifier which converts an AC audio signal into an average DC voltage which is then presented across a moving coil meter. The simplest way to do this is with just one diode which forms a half-wave rectifier between the audio source and the meter. This is often found cheap domestic equipment often preceded by a single transistor amplifier stage so that the diode can be made a standard silicon type instead of a more expensive germanium type with a lower voltage drop. This method also has the advantage of somewhat isolating the rectifier, which will generate significant amounts of non-linear voltage drop across it's source impedance, causing gross distortion.
Figure 2. Cheap and nasty 'VU' meter with poor sensitivity and accuracy
There's not much to this circuit, which is the idea, as it's designed to get by on the lowest component count. Due to the half-wave rectification it will be only half as sensitive to an AC input voltage when compared to a full-wave circuit. D1 has to be a germanium or Schottky type for the sensitivity to even approach a reasonable figure, and even then it is very poor. If the circuit is to be connected across the audio path then it must be driven from a very low impedance (preferably the output from an op-amp) and even then it may cause some compromise to the net distortion of the source.
Capacitor C1 provides some damping to the meter in order to prevent overshoot. I've only worked out a ball-park value for R1 as the circuit shown is for display purposes only. It isn't really worth saying any more about this topology, other than that it's greatly inferior to the rest of the circuits shown here. Hopefully after reading this article the reader will be successfully dissuaded from building it.
So far we have seen how a VU meter can be driven using a half-wave rectifier, but for proper use a VU meter should always use a full-wave rectifier. This improves the sensitivity of the meter and allows a more accurate reading of the average levels as it uses all of the waveform. Diodes with low forward voltages such as the 1N34A must be used so that the meter's sensitivity isn't affected. As the rectifier presents a non-linear load to the source it is important that a buffer is placed between the source and the meter to avoid distortion. A common cheap op-amp such as the TL072 can be used for this job. Using a buffer allows the meter to be driven from a high impedance source without affecting accuracy.
Figure 3. Conventional drive circuit for professional level
Figure 3 shows a fairly simple and straightforward circuit for professional line level that puts what has just been said into practice. The input is buffered U1 and fed into the germanium rectifier and meter via the standard 3.6kΩ resistor. R1 and R2 form a variable attenuator at the input with a range of 0dB to -6dB to allow a generous range of adjustment to compensate for any meter tolerances. C1 is included to supply a bit of damping, but can be omitted if a true VU meter is used.
Figure 4. Conventional drive circuit for consumer level
Figure 4 shows the line level version of Figure 3. In this case amplification, rather than a small amount attenuation, is needed to bring the level up to get the correct reading from the meter. R2 allows this level of amplification to be made adjustable for meter calibration.
In both circuits it is important to note that the rectifier is connected to a separate ground from the rest of the circuit. This is because the non-linear rectifier will cause the ground current to be a distorted version of the audio signal at the input. If this signal were to share the same ground as the rectifier then distorted current will generate a small but significant voltage across the ground impedance which would compromise the distortion performance of the unit that the VU meter is used in.
So buffering the VU meter and it's rectifier with an op-amp is a good idea, but we are still left with the problem of having to go to the trouble of using a separate ground for the rectifier. This can be a bit fiddly as extra wiring is needed, taking up excess PCB space.
Given a way to avoid producing a distorted ground current, the meter could share the same ground as the audio signal, eliminating the fuss of a different ground connection. This is quite straightforward to do simply by modifying the driving amplifier. By placing the rectifier within the op-amp's feedback loop, along with a parallel resistor to set the series resistance that the rectifier sees, distortion in the ground current can be dramatically reduced, bringing it down to a negligible level. Now the meter can share the same ground as the audio input.
Figure 5. Improved drive circuit with clean ground current
*For line level operation R1 and R2 should both be 1kΩ. If the meter is less than 15cm away from the drive circuitry it may be possible to omit R3 and R4 without inducing oscillation, but do so at your own peril!
As can be seen, the rectifier and VU meter now sit squarely between the op-amp's output and inverting input. This turns the op-amp into a current-drive amplifier, in this case the current being directly proportional to the audio input, thus eliminating distortion in the ground current. As the op-amp runs out of open loop gain as frequency increases it will be unable to correct the nonlinearity in the ground current, the harmonics of which will extend well above audio frequency. To solve this problem C1 bypasses the rectifier at supersonic frequencies to keep the ground current clean as frequency increases. Placing R5 across the rectifier decreases the output impedance to 3.3kΩ, simulating voltage drive with a total impedance of 3.6kΩ (R5 in series with R3 and R4). R3 and R4 provide some output impedance to keep the op-amp stable if the meter is located any significant distance away from the driving circuitry. Again, C2 can be omitted if a true VU meter is used.
Using this method a change of topology is not necessary to cater to professional or consumer audio line level. All that needs to be done is to change the values of R1 and R2 which set the gain. Once again, a wide range of adjustment is attainable through R2.
Now that the problem of distorted ground current has been fixed, the main bugbear of the previous circuits in this article has been the need for germanium in order to prevent the forward voltage of the rectifier from degrading the VU meter's sensitivity. These are not particularly cheap devices, and must be soldered fairly gingerly as they are annoyingly sensitive to heat damage. The next step would be to configure the circuit so these could be changed for small, low cost, silicon types such as the ubiquitous 1N4148 without any loss of sensitivity. Or even better, to eliminate the effect of the forward voltage drop completely.
All this can be done surprisingly easily without any increase of components. In fact it can be done with a decrease of components! By simply omitting the resistor connected across the rectifier, the drive circuit can present the rectifier with pure current drive at audio frequencies. As the circuit is now operating in pure current mode, the forward voltage of the rectifier becomes irrelevant as long as the op-amp can supply enough voltage to push past it. With supplies rails as low as ±9V there's still more than enough voltage swing available using the humble TL072.
Figure 6. Improved drive circuit with better sensitivity
*For line level operation R1 and R2 should both be 470Ω. Again, R3 and R4 may be left out if you are sure the meter is close enough to the driving circuitry to avoid stability issues.
Figure 6 shows the current mode VU meter driver, configured for professional line level. There's not much difference between this circuit and the one described in Figure 5 except that R5 is now gone, turning the driving circuit into a pure current output amplifier, and some values have been changed in order to make the circuit suitable for use with a non-standard 500uA meter. D1, D2, D3 and D4 are now thankfully 1N4148 small signal diodes. C1 is still included to improve linearity at high frequencies along with R3 and R4 to stop the capacitance of any cabling connecting the meter to the driving circuit causing problems with stability. Meter damping is provided by C2, which is now mandatory so there aren't any voltage spikes generated by the meter's inductive reactance to the drive current which could impair operation. Gain adjustment is again provided via R2, with R1 preventing a short between the inverting input and ground if R2 is adjusted to zero. This could, however unlikely, potentially damage the meter so it's better to have the resistor, just in case.
At this point it's worth considering how important it is for the VU meter to conform to the set out at the beginning of this article. True VU meters have their scale compensated to adjust for the rectifier's forward voltage and may not give an accurate reading when used with the better sensitivity circuit shown in Figure 6. However, as previously mentioned, the vast majority of VU meters made do not conform to these standards and almost always lack an adjusted scale, making them much more suitable for such a circuit. Many of these non-standard VU meters are quite excellent devices, with well behaved needle movement and solid construction which is all that is required in almost all cases. It is the opinion of the author that a high quality non-standard VU meter used with the topology shown in Figure 6 is actually a superior alternative to a true VU meter with more conventional drive circuitry, having much better sensitivity than the latter.
Figure 7. A very popular, but non-standard VU meter
Figure 7 shows the non-standard miniature VU meter that the driving circuitry in Figure 6 is optimised for. If you've spent any time online looking for VU meters then the chances are that you have come across this device. I have successfully deployed this meter in a couple of preamps and effects units and can say with confidence that it is a very good part indeed. I managed to obtain a pair for less than £10 on eBay, although it is important to make sure they are the Taiwanese made devices and not the Chinese ones, which are apparently useless, judging from the reviews of other buyers. They are also very easy to mount on unit fascias.
Using the current drive method, it is possible to make the meter's coil current a linear function of the input voltage, whether AC or DC. This makes it perfect for a simplistic, yet effective universal Volt-meter that can measure both AC and DC voltages. By setting the gain of the gain fairly high and placing a high impedance stepped attenuator across the input an accurate and very versatile Volt-meter is made with a wide range of input voltages.
Figure 8. Universal Volt-meter with selectable ranges of 500mV, 5V, and 50V
Figure 8 shows a practical implementation of a Volt-meter using the current drive method. S1, along with R1 to R4, form a stepped attenuator which provides 3 different sensitivity settings, in this case 500mV, 5V and 50V, while D1 and D2 form a protective clamp to make sure that the voltage at the non-inverting input of U1 does not exceed the supply voltage, which could damage the op-amp. R6 provides the calibration adjustment for the 500uA meter, giving a wide adjustment range of roughly ±50%, more than enough for all but the very shoddiest of meters. C1 prevents the meter's inductance from causing any problems. The meter itself should be a 500uA, type with a full reading of 500uA being given when 500mV, 5V, or 50V is present at the meter's input.
As the drive circuitry requires a split supply, the second op-amp in the TL072 is utilised along with R8 and R9 to provide a virtual ground across the 9V battery. R7, C2 and C3 provide some AC decoupling for U2, improving stability and slightly increasing the power efficiency when measuring AC voltages (by rather dubious amounts, though). C4 is shown, as it's necessary to bypass the op-amp's supply rails for stability reasons.
The great thing about this circuit is that it's just so easy to use, due to the rectification there's no need to bother with test probe polarity, while the simplicity allows a small footprint so that a fairly small plastic project box can easily be used. It measures DC and average AC voltages without the hassle of having to select between the two while still giving fairly accurate readings with a low component count. I strongly recommend building this circuit if you currently only have one multimeter.
Many of the circuits listed on the web have numerous flaws including distorted input current, ground current, poor sensitivity, inaccurate half-wave operation and many other deficiencies. While this article contains a total of 6 different driving circuits for VU meters, the author considers only Figures 5 and 6 to be worth building in comparison to the other examples. They are both easily adapted for single supply operation and stand-alone use. If the meter is a true type, then it is important to use the circuit shown in Figure 5 so as not to upset the accuracy towards the lower end of the scale. For all other meters, Figure 6 is the best choice as it avoids the use of expensive and fiddly germanium diodes, has a lower parts count and also sensitivity theoretically all the way down to zero (although the offset voltage in the driving op-amp will mean this is never fully the case in practice).
Ultimately it all boils down to cost, space and simplicity. However, considering the cost of the meter in comparison to the driving circuitry, which is simple enough to take up little room and can easily be powered from the range of supply rails present in most audio equipment, it is obvious that going the extra mile with some of the active circuitry shown in this article is well worth it.
If you have any questions or comments relating to this article then feel free to contact the author.
Michael Fearnley 2015