Monthly Archives: October 2014

The Forums Are Open

The Berkeley Hi Fi School and Music Reference forums are now open. Please feel free to share your ideas, comments and start new discussions related to school’s classes, Music Reference products and or anything else related to high end audio. Simply register and join in the discussions.

Class #2: Notes on Impedance: Part One

Teaching electronics to a group with a widely mixed background is a challenging job. What I have written below, and will in the future, contains material that addresses this wide range of interest and background. Some students may understand all of it and some may be happy to get just a few bits of wisdom. I encourage you to read all of it and not get hung up on the deeper technical points. If you are a beginner all you might learn is that a speaker’s impedance can be all over the place and that some amplifiers don’t like that but some do.

Impedance is just like resistance except it includes a frequency factor. We use the term resistance for DC measurements or for things like resistors that do not change with frequency. Things like light bulbs, space heaters and wire are usually used at DC or low frequencies where the resistance is constant and can be measured with an Ohm Meter. Ohm Meters use a battery to force a current through the resistance, measure the resulting voltage drop and use Ohm’s law to calculate the resistance. This is a very simple thing for an analog or digital meter to do. However to measure impedance we need to provide an AC current. This is not so easy. By now you have probably realized that unlike mechanical things where we can watch them work, move levers and wheels, with electronics there is nothing to see. We use meters, oscilloscopes, generators etc to measure things and some things are easier to measure than others.

DC  measurements are easy. AC measurements are a bit harder and complicated by the fact that  the resistance may vary with frequency, this is when we call it IMPEDANCE. The first problem with measuring impedance should be obvious. At what frequency do we measure it? Once we know that is varies with frequency we have to ask at what frequency or range of frequency are we going to measure?

If the device we are measuring is going to be plugged into the wall we would simply measure it at the line frequency of 60 Hz as that is our power in this country. It would not be too difficult to make a meter that did that. It would be just like a DC ohm meter but have a 60 Hz sine wave generator in place of the battery.

A more interesting example is a speaker. Since we use a speaker over the audio band of 20 Hz to 20 KHz we might want to know its impedance so that we can connect it to an appropriate amplifier or tap on a tube amp. Because the impedance is not constant we might want to know how high and high low the impedance goes. A typical 4 ohm speaker’s impedance might be as high as 50 ohms and as low as 2 ohms.  While an amplifier is not bothered by high impedances, as they draw less current, it may be very unhappy and distort when the impedance gets very low as now it has to provide more current than it can. In your home when you try to draw too much current from the wall it trips a circuit breaker. In an amplifier it may blow a fuse or simply limit the current using its internal electronic protection. Unlike a circuit breaker that has to be reset, electronic protection does not and often does not indicate that it is active, however you will hear it as distortion, or in some designs, a relay disconnects the speaker but usually resets itself automatically in a few seconds. Load line protection just clips the signal and there is no indication other than what you might hear.

Modern Home Theater amplifiers do not like low impedance loads and will usually indicate a fault on their menu screen. Sometimes this fault says “check speaker wires for shorts” because it thinks you load is a short!. These amplifiers are very sensitive to low impedances as they have rather weak power amplifiers that need a lot of protection. Even though they may be rated at 100 watts they may protect way below that level into difficult loads like electrostatic speakers making them useless for that application. When we are driving a device with a voltage we typically call it a load because energy is being expended and work is being done. The work of a speaker is to make sound, the work of a heater is to warm air, the work of a light bulb is to provide light. All these things convert electricity to something else.

Now how do we measure the impedance of a speaker or some other load whose impedance varies with frequency? We have to set up a test circuit using a variable frequency oscillator, an AC volt meter and AC amp meter. We then measure the voltage and current at frequencies of interest and use Ohm’s law to calculate the result. With a speaker we may just want the high and low impedances. With modern automated equipment we usually run a sweep from 20 to 20 Khz. and plot the result. Here is an example of a speaker impedance curve.


The solid dark line is the impedance (read on the left side) which goes off the chart below 100 Hz. The dashed line is the phase (read on the right side) which almost goes to 90 degrees indicating the load is a capacitor at high frequencies. This phase angle looks like a short to electronic limiters and will cause severe distortion with most transistor amps. Tube amps do not have limiters and will play this load. This is one reason that tube amps are preferred with electrostatic speakers.

Note that this Electrostatic speaker goes below 1 ohm at 20 Khz and is only 2 ohms at 8 Khz where trumpet music may cause the amplifier considerable distress. This will be covered in depth in the complete course on power amplifiers.

Part Two will address impedance matching… a very popular and poorly understood topic on the forums.


Class #1: Notes on Inductors

Here is a link to inductors that might help. Don’t fret over the formulas, read the first paragraph and see the pictures.

An inductor will store energy in a magnetic field and later release it as a current. It actually wants to keep the current going through it continuously. Therefore once the field is built up if the current is interrupted the voltage will rise very high trying to restore the current. This will cause a spark of much higher voltage than the applied voltage. If you put a 1.5 volt battery across an inductor and then release the battery while holding the wires you will get quite a shock. Anyone using a traditional ohm meter has likely experienced this when measuring the resistance of an inductor. This same principle is what allows the 12 volt battery in your car to make a good hot spark in your engine. It is interesting to note that the spark-plug fires when the points disconnect the battery, not when it is connected as the collapsing field is much faster (stronger) than the building field.

Using Iron or some other magnetic material will enhance the effect many times. That factor of enhancement is called permeability. The permeability of air is one. The permeability of transformer iron is many thousand, there for it increases the inductance of a coil of wire by that amount. See the explanation and table of materials in this article.

Of the examples I gave last night the most applicable to audio is the speaker crossover where there is an inductor in series with the woofer. In this application the inductor’s reactance (think resistance) increases with frequency thus reducing the current going to the woofer as we go up in frequency. Not all speakers have this inductor and let the woofer roll off through cone breakup, though this is not the best way to make a speaker. Better speakers roll the woofer off before cone breakup and hand the signal off to the tweeter.

Good speaker inductors, also called chokes, use a lot of copper and are thus expensive. The best ones are air core which uses even more copper. One can calculate the inductance needed for a woofer by knowing the impedance of the woofer and the cut-off frequency desired. This is usually 8 ohms and 1 KHz. In that case the inductor can be found L=8 ohms/6.3 x 1000 hz. (6.3 is 2 pi) or one can use a online calculator.

Using those numbers an inductor of 1.27 millihenries will roll off the woofer at 6 dB/octave starting at 1000 Hz. If the desired crossover point is 2000 Hz the inductor would be half as large, 0.64 mHy. Here is a 1 mHy air core inductor typical for a speaker. It also has 0.48 ohms resistance which will hurt the damping of the speaker somewhat. Thicker copper wire is needed to reduce that.–255-250. Here is the same inductor with heavier wire but still 0.3 ohms.–255-422

Please ask if you have questions or suggestions.

RIAA Preamps

Phono cartridges need both gain and equalization to make their output suitable for an auxiliary input or direct into a power amp.  A typical moving magnet cartridge has an output of 2 to 5 millivolts. Typical Moving Magnet (MM) gain at 1 KHz is 40 dB. The RIAA recording curve adds about 20 dB of boost at low frequencies and 20 dB cut at high frequencies to compensate for the RIAA recording curve and the fact that a magnetic cartridge is a velocity transducer. When displacement transducer is used (strain gauge cartridge) the required EQ is only 12 dB starting at 500 Hz and ending at 2000 Hz for a decrease of 12 db in the high end. This was done to reduce surface noise on records by boosting the high frequencies in the record cutting process so that they could be reduced by the same amount on the playback end.

When Moving Coil (MC) cartridges hit the market in the last 1970s their superiority was quickly recognized. However they required 20 dB more gain in the phono stage to bring them up to standard level as their output is 0.2 to 0.5 millivolts (20 db lower than MM). Early adopters of MC cartridges had two options. Buy an external transformer or a pre-preamp (head amp). Of course there was much discussion about the merits of these two very different ways of solving the low output problem. Both had typical gain of 20 dB and the better ones gave you choice over both gain and cartridge loading which affects the sound of the cartridge. When not loaded many MC cartridges sound bright and edgy because they are undamped. Most cartridge manufacturers give an ideal load resistance (typically 3-10 times the coil resistance is a good place to start if none is given). This load electrically damps the cartridge mechanical system for best response. If a cartridge has too much loading (low resistance) is will start to sound dull and lifeless. There were many high quality pre-preamps made up to the point of the popularity of the Compact DIsc. With the resurgence of vinyl, pre-preamps and complete MM/MC phono stages are coming back. A complete phono preamp will have sufficient gain (60+ dB) and RIAA EQ for MC cartridges along with loading. However it is an excellent, and sometimes better solution, to split the phono stage into a separate head amp and RIAA stage to isolate those functions as the very small, delicate output from a MC cartridge is best amplified in two stages.

The first product from Music Reference ( was the RM-4 pre-preamp which was designed to sit in front of a standard phono preamp and raise the gain of an MC cartridge to suitable level to go into a standard phono stage such as the RM-5, Modulus, CJ, ARC or other standard preamp at the time. It is still a fine solution to implement MC phono systems. What makes the RM-4 special is that is has low noise comparable with solid state units (which is not easy), cartridge loading and gain selection of infinite range through plug-in modules. The gain can be from 0 db to 30 dB. Along with a 40 dB standard phono preamp this combination will give a total of 70 dB gain, more than most MC phono stages and in some cases can go directly to a power amp via a passive preamp (simple volume control).

Although I prefer the active (pre-preamp) solution transformers are also available. Some transformers allow independent choice of gain (step up ratio) and loading, however many do not. In that case the load presented to the cartridge is determined by the chosen step up ratio and may not be the ideal load for that cartridge. With a good pre-preamp one has independent control over gain and loading. While some pre-preamps add noise, a well designed one will have noise below the surface noise of a good record. Transformers can pick up hum and often have problems at the frequency extremes where a good pre-preamp can have bandwidth from a few Hz to 50 KHz. Needless to say, Be aware that some complete phono stages use a transformer to amplify the MC input rather than to do that function electronically. I prefer the all electronic method as long as the design is low noise.