|Elliott Sound Products||Project 14|
A stereo power amplifier is limited in its output power by two main factors - the impedance of the load and the internal power supply voltage. To obtain more power, one has limited choices - other than the purchase of a more powerful amp.
The load impedance can be lowered, but if the load happens to be a pair of standard loudspeakers this is not viable, since the impedance is set by the drivers themselves. Increasing the power supply voltage is generally a bad idea, since most commercial amps do not have a wide safety margin with component ratings, and will probably be destroyed if the voltage were to be raised sufficiently to obtain even 50% more power.
The bridging adapter shown in Figure 1 can make an amplifier produce almost 4 times the power for the same impedance - but beware of the pitfalls. Basically, these are:
Figure 1 - Basic (Conventional) Power Amp Bridging Adapter
Construction is not critical, and the adapter has unity gain for each output. Naturally, 1% metal film resistors should be used, and the choice of opamp is not too critical - the TL072 is perfectly acceptable in this configuration, but feel free to use the opamp of your choice. Note that if interconnect leads are to be used from the adapter to the power amp, the 100 Ohm resistors shown must be placed in series with each output to prevent instability - this is important, as an oscillating adapter will inject an AC voltage of perhaps hundreds of kilohertz into the amp's input, with the very real possibility of destruction of the output transistors. Although not shown in either schematic here, bypass capacitors are needed from the opamp's supply pins to earth/ ground - do not leave these out or the opamps will oscillate!
The power supply may be taken from the preamp supply (this should be ±12V to ±15V). The preamp output is connected to the adapter's input, and for the sake of convention, connect the +OUT to the Left power amp's input, and the -OUT to the Right amp's input.
Naturally, for stereo two circuits are needed, as well as a second (preferably identical) stereo power amp. This arrangement is also very useful to convert an otherwise mediocre stereo power amp into a perfectly acceptable sub-woofer amplifier, having plenty of power (depending on the power of the original, of course).
Quality is not so much of an issue for a sub, since only the low frequencies are reproduced, and amplifier distortion is as nothing to the distortion generated by a loudspeaker at low frequencies and high excursions. The disadvantage of the arrangement shown above is that the input impedance is only 50k (R1 || R4), and the noise contribution from U1B will be higher than expected because of the high resistances.
Using a circuit such as the P87B has a number of advantages. The primary advantage is that the input impedance can be a great deal higher because of the input buffer (U1). Although R102 is shown as 100k, it can be reduced to 22k or increased to 1Meg (or more) with no other changes needed. For high input impedance using the basic arrangement of Figure 1, the impedances around the second inverting opamp become excessive, and this causes noise problems. Figure 2 has no such restriction. The input impedance can cause some noise if the source impedance is particularly high, but this is uncommon. Even a typical valve preamp will have an output impedance that's usually less than 47k, so R102 (and R202 - not shown) can be increased to 1Meg with no noise penalty.
Note that if the circuit is used with a valve preamp, you will need to protect the inputs from high voltages - see MOSFET Follower & Circuit Protection From High Voltages for the details of how to protect the circuit from damage.
Figure 2 - P87B used as Bridging Adapter
Using the P87B circuit means that the preamp sees only the impedance set by R102/202 (R202 is in the second channel), and each channel of the power amp sees a source impedance of about 100 ohms. This low impedance means that fairly long interconnects can be used if needed, with no loss of treble.
|The loudspeaker is connected between the amplifier's + outputs only, and neither side of the speaker can be earthed or connected to any other amplifier output - either of these conditions WILL blow up your amplifier.|
I have had a few constructors who have had problems - mainly due to inexperience. I have been doing this stuff for so long that I often forget that many of my readers are novices, and this looks really simple, so off they go and promptly have problems I haven't covered. I shall attempt to remedy the situation forthwith!
When the unit is built, after checking that power is correct (no more than ±15V), some basic tests need to be done.
If all the above tests are OK, you can connect the output of your preamp to the input of the adaptor. Just for safety's sake (and before you connect your power amps), measure the DC output voltages again. If the +OUT terminal now shows a DC voltage where none was evident before, check the -OUT terminal. You will probably find that it has the same voltage, but of opposite polarity. This means that there is DC from the preamp, so use a capacitor (1µF will be ok for either version) in series with the input to get rid of it (or fix the preamp, which may be faulty).
If the power amp has a volume control (or controls), make sure that both channels are set to maximum. Do not connect a speaker until you have verified that the amps' outputs are at zero volts (±100mV or so), and that there are no large voltage swings when the amp (or adaptor) are turned on or off. If possible, the power to the adaptor should be applied first. It is possible to leave it on permanently if powered from an AC adaptor, as current drain is very low.
Once these tests have been completed, you may connect the speaker. Remember that the power will be four times that from a single channel of the amp for the same impedance (a 6dB increase), so overdriving the speakers is quite possible. Use the utmost care, especially with expensive speakers.
For those who have not used bridging or who do not understand the principles, a short explanation of how the adapter is used and how this almost quadruples the output power is called for.
Before you even contemplate using bridging, make absolutely certain that the amplifiers used are capable (and designed for) half the speaker impedance. If you have 8 ohm speakers, the amps must be able to drive 4 ohms. With 4 ohm speakers, the amps must be able to drive 2 ohms - most can't, so you must not attempt to bridge amplifiers into 4 ohm loads. If you happen to have speakers rated at less than 4 ohms, then don't even try - you will blow up your amplifiers! They might survive for a little while, but failure is inevitable.
The adapter is connected between the preamplifier and the power amps. The power amps must be the same - power rating, minimum impedance rating, etc. Generally, a stereo power amp is used, so when connected in bridge mode we are assured that the amplifiers are more or less identical.
Normally, the speaker is driven from the amplifier output to ground, and the AC swing is limited by the supply voltages in the amp. Consider a 50 Watt per channel power amp - 50W into 8 Ohms requires a signal voltage of 20V RMS:
P = V2 / R
P = 202 / 8
P = 400 / 8 = 50 Watts
To achieve this, the peak voltage is just over ±28V (20 × 1.414), and a power supply voltage of about ±35V will generally be used to allow for losses and mains voltage variations.
The same amplifier into 4 Ohms will deliver close to 100W - provided the power supply does not collapse under the load. For both these examples, only one side of the loudspeaker is driven, and the other is grounded.
Now, if a second amplifier is connected so that its output is exactly 180 degrees out of phase with the first (i.e. inverted), and connected to the normally grounded side of the speaker, as one speaker terminal is driven positive, the other is driven negative by the same amount.
Figure 3 - Voltages Applied to the Loudspeaker Using Bridging
Figure 3 shows this, with the waveforms at each speaker terminal shown. As you can see, as one terminal is driven positive, the other is driven negative by the same amount, and although a sine wave is shown, the principle is not changed by the signal waveform. (Note that both waveforms should be viewed from left to right, otherwise the diagram would indicate zero output from the speakers - which is exactly what you will get if the adapter is not used.)
At maximum power, the 8 Ohm loudspeaker now 'sees' double the voltage that it would receive from one amp alone. Using the formula above, we get:
P = V2 / R
P = 402 / 8
P = 1600 / 8 = 200 Watts
Since the voltage across the speaker is doubled, naturally the current through it is also doubled, and that is the reason that each amplifier must be capable of driving half the normal speaker impedance. This technique is very common in car audio systems, because the nominal 12V (typically around 13.8V when the engine is running) of a car's electrical system is too low to obtain much power except into very low impedances.
Loudspeakers are very difficult to make if the impedance is too low, because there are too few turns of wire in the voice coil, and efficiency is lost. Four Ohms is a reasonable minimum, but even with this impedance a non-bridged car amplifier is still only capable of a maximum of about 5 Watts. By using bridging, close to 20W is now possible, with each amplifier driving the equivalent of 2 Ohms.
This is the reason for all the dire warnings about not grounding either speaker lead of a car audio system - because each lead is the output of an amplifier, shorting it to ground will destroy the power amp because the amplifier outputs almost invariably have a DC voltage of around 6.5V with no signal. If shorted to another speaker lead nothing will happen until signal is applied, and the amp may die as a result.
The same principle applies to the bridged connection shown here - no connection other than to the speakers is possible without damaging the amplifier.
Figure 4 - Power Amplifier Connections
The drawing above shows the speaker connections. Only the amplifier +Ve outputs are used, and you need to be careful with the phasing. If the speaker is connected with the +Ve terminal to the wrong amplifier (Amp2 instead of Amp1), the output will be 180° out-of-phase with the other speaker, assuming it's been wired correctly. This will cause a dramatic loss of bass, because the low frequencies will cancel.
With most power amps, the absolute minimum load (speaker) impedance is 8Ω, because the amplifiers each 'see' only half the connected impedance. Be particularly careful with IC amplifiers such as the LM3886, as their protection circuits are easily triggered with low impedances. I suggest that the supply voltage should not be more than ±30V with these ICs when bridged.
Car amps generally have a single supply voltage, nominally +12V with respect to the chassis. The circuit shown above expects a dual supply, and while this is easy enough to do, it adds complexity for no good reason. If the inverter opamp is biased to half the battery voltage, it can perform the signal inversion, and we only need to capacitively couple the input and output. This simplified version is shown below.
Figure 5 - Single Supply Version For Car Installations
As shown, there is no buffer for the direct signal - it's simply passed through to the output. The inverting stage will always provide a signal that is exactly equal but opposite (in phase). Even if the power amp loads the source signal, the inverter will invert that reduced level to maintain the proper signal level to each power amp. It's less 'elegant' than the versions shown earlier, but it's also simpler. The second channel (for stereo) uses the other half of the opamp, and uses the same ½Supply (Vcc/2) bias voltage to reduce component count.
Because a car's electrical system is rather hostile, I've included R1, C1 and D1. R1 limits the peak current, and should be rated for at least 1W. Zener diode D1 clamps the maximum voltage to 15V, protecting the opamp and C1 from over-voltage. C1 is required for any opamp to bypass the supply. The incoming signal is not buffered - it's used directly to one power amp input, and the inverted output is applied to the other.
The circuit shown in Figure 4 is about as simple as it gets, but it will still work well. You can't use the PCB for this though, because there's no provision for the ½Supply rail to bias the opamps properly to ensure correct operation.
|Copyright Notice. This article, including but not limited to all text and diagrams, is the intellectual property of Rod Elliott, and is © 1999. Reproduction or re-publication by any means whatsoever, whether electronic, mechanical or electro-mechanical, is strictly prohibited under International Copyright laws. The author (Rod Elliott) grants the reader the right to use this information for personal use only, and further allows that one (1) copy may be made for reference while constructing the project. Commercial use is prohibited without express written authorisation from Rod Elliott.|