PHASE SPLITTERS

Phase Splitters are essential in push-pull audio amplifiers. By them depends  largely the overall performance obtainable by an hifi amplifier.
This page shows a collection of information and explanations concerning the most popular phase splitters taken from some websites. All classic circuits have their problem in order to deliver two signals very very similar and balanced to each other but shifted by 180°.

If you're bored or want to know directly the new phase inverter that is truly balanced go to bottom of this page: you will discover the main features of such new circuit with excellent performances.

NEW PHASE SPLITTER

bullet

From TUBECAD

 

Unless a push-pull power amplifier has a balanced input, at some point the unbalanced signal must be converted into a balanced signal, a push-pull signal. In a tube amplifier, this task is accomplished by a phase splitter, usually a Split-Load or a Long Tail phase splitter. Each circuit has it advantages and disadvantages. These two circuits, however, are not the only possibilities. Transformers have long been used to split the phase of a signal and even a center-tapped choke can be used.
 

 

Split-Load Phase Splitter
    This single triode phase splitter has the best voltage balance, but has dissimilar output impedances per output phase leg.
 

Negatives
-
No gain.
- Limited voltage swing.
- Dissimilar Zo's.   
- Dissimilar PSRR's.

Positives
 

+ Single triode.

 
 

 

Review of Tube Phase Splitters
 
 
 
 

Long Tail Phase Splitter
 

    This two triode phase splitter has the advantage of high gain and large voltage swings, but must be tweaked to provide balanced voltage outputs. 

Negatives
-
Requires tweaking.
- Poor balance.
- Dissimilar PSRR's.

Positives
+
Gain.
+ Large voltage
    swing.
+ Similar Zo's.

 

Cross-Coupled Phase Splitter
 

    This two triode phase splitter has the advantage of high gain and large voltage swings, but must be tweaked to provide balanced voltage outputs.
 


Negatives
-
Four triodes.
- Low input
   overload voltage.
- Sub-optimal use of
   Cathode Followers.
 

Positives
+
Gain.
+ Very low input
    capacitance.
+ Identical Zo's.
+ Identical PSRR's.

 
 

Paraphrase Phase Splitter
 

    This two triode phase splitter has the advantage of high gain and large voltage swings, but must be tweaked to provide balanced voltage outputs.
 

Negatives
-
Requires tweaking.
- Poor high
   frequency response.
- Uses Feedback.

Positives
+
Gain.
+ Large voltage

 

 

  • From BONAVOLTA

  • Phase Splitters

    In a push-pull amp, you need a phase splitter to produce the two identical but in phase opposition signals to drive the power tubes.
    Phase splitters are like religions, you belong to one or another and discussions between fans are often emotional.

    horizontal rule

    Contents

    bulletCathodyne
    bulletLong tailed pair
    bulletSchmidt
    bulletCross coupled
    bulletParaphase
    bulletIsodyne
    bulletTransformer

    horizontal rule

    Cathodyne

    The cathodyne is the most used phase splitter, probably because it's the simplest and yet quite effective.
    Examples are nearly everywhere on this site.

    Pros:

    bulletVery simple
    bulletPerformant when followed by a cathode follower before the power tubes
    bulletCan be directly connected to the previous stage

    Cons:

    bulletLow gain
    bulletNot very linear because of the internal capacitances of the tube
    that are not identical between the grid and the plate or the grid and the cathode,
    but their values are usually so low that their effect is far away from the audio range
    bulletVery sensitive to the load

     

     

    horizontal rule

    Long Tailed Pair

    This is also a common phase splitter. It is also called "cathode coupled".
     

     

     

     

     

     

     

     

     

     

    horizontal rule

    Schmidt

    Two triodes are mounted in a differential. The grid of the second tube is decoupled to the ground through a capacitor.
     

    Pros:

    bulletVery linear
    bulletEasy to adjust and stable
    bulletLow distortion
    bulletCan be directly connected to the previous stage

    Cons:

    bulletQuality of the grid decoupling capacitor critical
    bulletMedium gain
    bulletThe tubes must be paired and from a twin tube

     

    horizontal rule

    Cross coupled

    Also known as Van Scoyoc. It needs a low output impedance from the previous stage, that's the reason for the transformer in this schematic.

    Practically, a transformer is usually substituted by two cathode followers.

     

    horizontal rule

    Paraphase

    This phase-splitter is also called "anode follower" or "see-saw".
    This splitter takes the signal out of the 1st tube, passes it through a voltage divider and injects it into another tube to obtain the inversed signal.
    The second schematic is a simplified version that can be used when the following tubes have no grid current.
     

     

     

    horizontal rule

    Isodyne

    Schematic by E. F. Worthen (August 1958).
     

    Pros:

    bulletVery linear
    bulletEasy to adjust and stable
    bulletLow distortion
    bulletHigh gain
    bulletDirect coupled
    bulletHigh amplitude output
    bulletLow output impedance

    Cons:

    bulletThe tubes must be paired from a twin tube

     

     

    horizontal rule

    Transformer 

    The schematic is self-explanatory, each output connection of the transformer has the same signal but in phase opposition.


    Pros

    bulletVery simple

    Cons

    bulletQuality extremely dependent of the transformer
    bulletNot always easy to find interstage transformers nowadays

     

     

     

     

     
    bullet

    From ANGELFIRE

    The Long Tail Pair Phase Inverter.

    Circuits similar to this were used in amplifiers made by Eico and many others. Note that the asymmetry of distortion has diminished from about 7 to 1 to about 2 to 1.

    The DC plate voltage on the first 12BH7 is applied directly to the grid of the upper triode in the LTP (long tail pair) and to the grid of the lower triode through a low pass filter consisting of the 1 Megohm resistor and 0.22 uf capacitor. This places both grids at the same DC potential while applying signal only to the grid of the upper triode. The 15 k ohm resistor in the cathodes sets the plate current to the correct value for proper operation.

    The AC signal which is applied to the grid of the upper triode causes a variation of plate current as in a normal amplifier. This varying current, which is in phase with the grid voltage, flows through the cathode resistor as well as the plate resistor. The varying current in the cathode resistor causes a voltage variation which is in phase with the grid signal. This voltage is directly coupled to the cathode of the lower triode in the LTP. The grid of the lower LTP triode is grounded for AC through the 0.22 uf capacitor. This makes it a grounded grid amplifier.

    In normal Class A operation the grid has a negative DC bias on it. In this case the grids are at about 90 volts while the cathodes are at about 100 volts. When the AC signal drives the cathode of the lower triode more positive the grid and cathode are becoming farther apart. This is an increase of the bias which results in a decrease of plate current. We see that the plate current in the lower triode is out of phase with the plate current of the upper triode. If they were exactly equal they would completely cancel and there would be no voltage variation at the cathode to change the bias on the lower triode and there would be no plate current variations. Oops? This contradiction means that the plate current variations in the lower triode are not as large as those in the upper triode. To have the same output voltage from both triodes in the LTP the load resistor of the lower triode must be slightly larger than the load resistor for the upper one.

    Lets look at it from the standpoint of current. The upper tube is producing cathode current. Some of this alternating current flows to ground through the 15 k ohm resistor and the rest enters the cathode of the lower tube. If you write the equation IR + IV2 = IV1 you see that the currents in the two tubes are 180 degrees out of phase. If the resistor could be made infinite for AC while still passing DC the alternating currents in the two tubes would be equal and the load resistors could be made equal. The larger the common cathode resistor, the better the balance.

    Another variation you are likely to see in commercial amplifiers is the simple long tail pair shown below.

     

     Schematic diagram.

     

    As you can see from the legend, the asymmetry of distortion has all but disappeared. The inequality of plate load resistors is also much smaller.

    Because the 12AX7 is a low power tube this circuit can't drive a pair of tubes having 100 k ohm resistors in their grids. This inverter would be found in a cathode biased amplifier or there would be another pair of amplifier tubes between the inverter and the power tubes.

    If a constant current sink is used instead of the resistor in the cathode circuit the circuit will come into perfect balance with equal plate resistors and the distortion will be very low and equal. A constant current sink has an extremely high impedance for AC, nearly infinite, while passing the DC to set the operating point of the tubes. But I have never seen this circuit arrangement in any commercial amplifiers.

     

    A phase inverter circuit often used in high fi and guitar amplifiers in the 1950s and 1960s is the long tail pair. This is in current terminology a differential amplifier. The one shown below is a direct coupled version and is a variation of one used in a commercial amplifier that had cathode loaded output stages. It had to be capable of delivering 50 volts RMS per grid at a reasonable distortion level.

     Schematic diagram.

                                                                                                                Figure 4 Direct Coupled Long Tail Pair Phase Inverter.

     

    This is not a split load phase inverter but the original phase inverter which obtained the inversion by passing the signal through one more stage. This version overcomes the common mode rejection problem by applying feedback to a single ended stage before the inverter part of the circuit.

    Note that the asymmetry of distortion has diminished from about 7 to 1 to about 2 to 1. In spite of the remaining asymmetry it was used by many amplifier manufacturers, particularly Eico, in some amplifiers which sounded very good. Perhaps unbalanced distortion is not as important as I believe it to be. However, I still am not comfortable with this condition.

    The DC plate voltage on the first 12BH7 is applied directly to the grid of the upper triode in the LTP (long tail pair) and to the grid of the lower triode through a low pass filter consisting of the 1 Megohm resistor and 0.22 uf capacitor. This places both grids at the same DC potential while applying signal only to the grid of the upper triode. The 15 k ohm resistor in the cathodes sets the plate current to the correct value for proper operation.

    The AC signal which is applied to the grid of the upper triode causes a variation of plate current as in a normal amplifier. This varying current, which is in phase with the grid voltage, flows through the cathode resistor as well as the plate resistor. The varying current in the cathode resistor causes a voltage variation which is in phase with the grid signal. This voltage is directly coupled to the cathode of the lower triode in the LTP. The grid of the lower LTP triode is grounded for AC through the 0.22 uf capacitor. This makes it a grounded grid amplifier.

    In normal Class A operation the grid has a negative DC bias on it. In this case the grids are at about 90 volts while the cathodes are at about 100 volts. When the AC signal drives the cathode of the lower triode more positive the grid and cathode are becoming farther apart. This is an increase of the bias which results in a decrease of plate current. We see that the plate current in the lower triode is out of phase with the plate current of the upper triode. If they were exactly equal they would completely cancel and there would be no voltage variation at the cathode to change the bias on the lower triode and there would be no plate current variations. Oops? This contradiction means that the plate current variations in the lower triode are not as large as those in the upper triode. To have the same output voltage from both triodes in the LTP the load resistor of the lower triode must be slightly larger than the load resistor for the upper one.

    Lets look at it from the standpoint of current. The upper tube is producing cathode current. Some of this alternating current flows to ground through the 15 k ohm resistor and the rest enters the cathode of the lower tube. If you write the equation IR + IV2 = IV1 you see that the currents in the two tubes are 180 degrees out of phase. If there were a device that had infinite impedance for AC while still passing DC the alternating currents in the two tubes would be equal and the load resistors could be made equal. The larger the common cathode resistor, the better the balance.

    This version also contains two low frequency roll-offs. The two 0.22 uf caps looking into 100 k ohm resistors make up one and the other is the 1 Meg and 0.22 in the grid of the lower tube. As this network rolls off the same signal will begin to be applied to both grids. Although there will still be a considerable amount of in-phase signals applied to the two grids of the outputs, they will be cancelled in the output transformer. This causes a low frequency roll-off and the accompanying phase shift inside the feedback loop. It is possible to design a long tail pair that will do better than this.

    Here's the circuit I came up with.

     

     Schematic diagram.         

                                                                                    Figure 5 Long Tail Pair with Transistor Current Sink. 

    No, I haven't lost my mind. Anything I might say to defend my use of a transistor is going to sound like rationalization. It's there, learn to live with it. It makes a very constant current sink. This is the device mentioned above that has infinite AC impedance while passing DC current. It yields a circuit with a very high common mode rejection ratio, although not infinite, and low distortion.

    Here is how it works. Lets say the collector current tries to increase. The emitter current will also try to increase which makes the drop across the 1.6 k ohm resistor increase. Because the base voltage is held very constant by the voltage divider of the 3 k and 1.1 k ohm resistors the bias on the base-emitter junction will be decreased. This will decrease the base current and cancel out most of the collector current increase. The collector current has to change by a small amount to make the base current change, but the change is very small. The collector current is 3 mA and most 3904s have a current gain of more than 100 so the base current is about, or less than, 30 microamps. The divider current is 4.88 mA. The base current is 0.615% of the divider current.

    As you can see the distortion was very low, much lower than I had expected. Also the distortion at both outputs was very similar. When feedback was simulated the balance remained perfect. The distortion changed though. The plate on the left gave 0.09% while the right hand plate gave 0.175%. This low level of distortion means the distortion imbalance can be disregarded. The transistor can have an overload problem. My first design only used a 12 volt negative power supply and I had to re-engineer it for a -20 volt supply. If I were going to use this circuit I think I would go up to a -30 volt supply.

    Now Let's Do It With a Tube.

     Schematic diagram.

    Figure 6 Long Tail Pair with Pentode Current Sink.

    I selected a pentode rather than a triode because a pentode by its nature is a current sink and the control grid will not have to do as much to make it into one. The 6BH6 was the best tube in this application. The resistor in the cathode works exactly like the resistor in the emitter circuit above to regulate the current and hold it constant. The -70 volt supply MUST be regulated with zener diodes for low noise and stable performance. I often use two or three diodes in series to make up the exact voltage I need for a particular purpose.

    As you can see the distortion figures are much lower than would ever be needed. When feedback was simulated the left hand plate gave 0.15%, and the right hand plate 0.19% distortion. This looks like the one I would recommend and use myself.

    It has one major drawback. It will not drive 100 k ohm grid resistors in the following stage. If fixed bias is to be used in the output tubes they must be driven by direct coupled cathode followers. That is an additional complication along with the negative regulated power supply. If one wants to build the highest quality amplifier with no compromises this circuit might be used along with regulated screen grid voltages as in the 6L6 monoblock amplifier. This phase inverter would, without a doubt, give a good account of itself in a cathode biased amplifier although the negative power supply would still be necessary for the negative voltage for the cathode

    A phase inverter circuit often used in high fi and guitar amplifiers is the long tail pair. This is in current terminology a differential amplifier. The one shown below is a direct coupled version and is a variation of one used in a commercial amplifier that had cathode loaded output stages. It had to be capable of delivering 50 volts RMS per grid.

    Open Loop Phase Inverter.

    So how do you get two signals of equal voltage and opposite phase? The engineer who designed the first push-pull output had to answer that question and there was no one he could ask who had done it before. Well, an ordinary common cathode resistance coupled amplifier inverts the phase of the signal so all we need to do is to run the signal for the second tube through one more stage than the first. But hang on just a minute, the signals have to be equal in voltage and a resistance coupled amplifier amplifies the signal. OK; so let's put in a resistive attenuator to decrease the signal by the same amount that the amplifier amplifies it. You mean like this?

     Schematic diagram.

    Yeh, that's it. You don't want to use a tube that gives a whole bunch of gain like a 12AX7 so a 12AU7 is better. The first stage is just to get up some gain to be later traded off for other improvements by applying feedback. The output of the second stage drives the grid of one of the outputs. Notice how the grid resistor of that output tube is tapped for a small amount of the voltage to be taken off to the grid of the third stage. This inverts the signal for the other output tube.

    A little while ago I picked up an old Philco radio at an antique mall. When I started examining it I found to my surprise it had a push-pull output. It was made in 1937. I don't know when negative feedback found it's way into consumer products but it was after that. The Philco used exactly this same circuit (with different tubes of course, and no feedback). The drive balance of the Philco circuit was within 5%.

    The circuit shown above produced amazingly good balance after I got the resistor values right. I swapped out several tubes and the worst imbalance I could get was about 5 percent. Its chief problem is asymmetrical distortion. When operated at a level sufficient to drive a pair of 6V6s the distortion was 1.9 % from the top output and .25 % from the bottom. At a 6L6 level the distortion was 2.1 % and .3 %. I'll bet some of you are scratching your heads and saying "hmmm, he must have reversed those readings, after all the bottom output passes through one more stage than the top so it should have more distortion." No, I double checked it and then checked it again. What is going on is? The two amplifier stages are operating at the same level and so are acting like a push-pull amplifier and canceling even harmonics. Since most of the distortion is second harmonic most of it gets canceled out. You can also think of it as the third stage is correcting the errors of the second. This asymmetrical distortion seems to be an incurable plague of phase inverters.

    I tried many variations on this circuit in an attempt to clean it up. When I did find something that equalized the distortion it was at the higher level not the lower. Let's look at a complicated circuit with lots of feedback to see if it gets any better.

    The circuit below comes from a commercial P A amplifier that has a 70 volt line output

    Closed Loop Phase Inverter.

      Schematic diagram.                

    I have altered many of the component values in a futile attempt to clean it up. In spite of all the feedback in the circuit the distortion stubbornly remained at .2 % and 1.5 %.

    Cathodyne Phase Inverter.               

    The formal name of this circuit is the "split load phase inverter". That's because the load resistor has been split in half and one half has been moved around the circuit to be in the cathode. This gives two voltage outputs of equal amplitude and opposite phase.

    Here is the circuit used in that little Newcom amplifier which was part of my first high fi system.

                 Schematic diagram.

    The first section of the 12AX7 is operated in a mode called, not quite correctly, zero bias. The ten megohm resistor allows the grid to develop about -0.7 volts of bias. This is because electrons randomly run into the grid wires on their way to the plate. The 100 ohm resistor in the cathode does not contribute any significant amount of bias voltage, it is there so negative feedback can be applied from the output transformer secondary. This stage has a gain of approximately 73. Signal is coupled from the plate of this amplifier by the .01 microfarad capacitor to the grid of the second half of the duo triode. This is a cross between a cathode follower and a common cathode stage. Because the tube is a triode the cathode current is identical to the plate current (not so for a pentode). with identical currents flowing through the two 100 k ohm resistors the voltages across them are identical, if the resistances are identical. Obviously the current flows in the same direction, up or down, in each resistor at any instant of time. Because the signals are taken one off the top end of a 100 k ohm resistor and the other off the bottom of the other 100 k ohm resistor the two signals are opposite in sign or phase. The gain of this circuit from input to either output is approximately 0.9.

    If the grid were placed at DC ground by connecting the 470 k ohm resistor to ground the voltage drop across the 100 k ohm resistors would not be enough to produce the necessary 40 volts peak to peak which is required to drive a pair of 6V6s. By connecting the grid return resistor to the top of the 100 k ohm resistor and then using a 10 k ohm to bias the tube the plate current is large enough to produce the necessary drop, about 25 volts, across each 100 k ohm resistor. Connecting the grid resistor where it is tricks the tube into "believing" that it has 10 k ohms in its cathode instead of the 110 k ohms which is really there. (That's called anthropomorphizing). The grid bias is the DIFFERENCE in voltage between the cathode and the grid not the voltage between the grid and ground.

    Because the lower end is connected to a point which receives signal, the effective resistance of the grid resistor is different from 470 k ohms. Let's say that the voltage at the grid rises by 1 volt. The lower end of the grid resistor will then rise by approximately 0.9 volts. The change in voltage across the resistor is only 0.1 volts. The current through the resistor will change by 0.1 volts / 470 k ohms = 0.213 microamps. The effective resistance of the resistor is change in applied voltage divided by change in current = 1 volt / 0.213 microamps = 4.69 megohms. This allows the use of a smaller coupling capacitor than would otherwise be required.

    The circuit performs better than I had expected. I had never tested it before because back when I was building my tube circuits I didn't have much in the way of test equipment. The over all gain is 66 based on just one of the outputs, either one. The distortion is 0.3 % at the level required to drive 6V6s to full power and 0.9 % for 6L6s. Not bad when you look at the distortion figures given in a tube manual for these tubes, 6V6 3.5 % distortion at 14 watts and 6L6 2 % distortion at 26.5 watts. In the final installment of this series we will discuss the benefits of negative feedback and the drawbacks of using too much of a good thing.

     
    bullet

    from National Valve Museum

     

    New Phase Splitter

    A R Bailey MSc(Eng) AMIEE, Bradford Institute of Technology. Wireless World, September, 1962.
     

    Improved high-frequency response.

    Circuit of complete EF86 pentode amplifier and ECF82 phase-splitter for use in a high-fidelity amplifier with a large amount of negative feedback.

    In the past many phase-splitters have been investigated and the reader would be quite justified in asking why a further circuit has been developed. The answer is quite simple in that the performance of high-fidelity amplifiers may be no longer limited by the output-transformer but by the response of the phase-inverter. With the best transformers available this is very true, and the improvement in amplifier performance that is obtained by using better circuits is quite startling.

    In order that the reasons for discarding the present circuits may be seen, it is essential that all the requirements of the phase-inverter be first evaluated.

    The first requirement of all phase-inverters is that they should deliver an output that is balanced to within a few per cent and does not alter as the valves age in use. Most of the phase-inverter circuits in use have this property, but the paraphase-inverter (below) does not, as there is no negative feedback to stabilize the gain of the stage.

    Basic circuit of paraphrase phase-splitter.

    The second requirement is that the output-impedances from both halves of the phase-splitter should be approximately equal. The reason for this is nothing like so obvious at first sight. In fact there are two reasons for this requirement. The first is that severe grid-blocking can occur if the amplifier is accidentally over-driven (ref. 1). This can be overcome by using high-value grid stoppers but this gives an additional high-frequency time-constant due to the input capacitance of the following stage. As will be mentioned later, this gives a very undesirable tendency to HF oscillation when high values of negative-feedback are applied. The second disadvantage is that the drive to the two output valves may unbalance severely at high frequencies due to the different time-constants produced. This will have the effect of severely limiting the HF power available and will also increase the HF distortion. The two circuits that suffer from this drawback are the floating paraphase

    Basic circuit of 'floating' paraphrase phase-splitter.

    and the concertina.

                   
    Basic circuit of direct-coupled concertina phase-splitter.

    In both circuits the valve loads that drive the output stage are balanced, but unfortunately the output impedances are not balanced. This is due to the voltage negative-feedback inherent in the circuits. This negative-feedback overcomes the problem of valve ageing mentioned before, but brings the disadvantages that have been just mentioned. In addition, these circuits are not readily DC coupled to the previous stage. This is a severe disadvantage where large amounts of feedback are contemplated, as the additional low-frequency time-constant of a further coupling capacitor can easily cause instability at low frequencies.

    Up to the present it seems to have been generally assumed that these were the only circuits to be avoided, if at all possible. Unfortunately this has proved not to be the case and the cause of the distressing tendency of high-feedback amplifiers to go unstable can often be laid at the door of the phase-splitter. The reason is quite simply one of poor HF response. A poor HF response in the phase-splitter will cause a falling loop gain, but what is more important it will also give a phase shift that tends towards 90 degrees retard. Now, if there is a total of 180 degrees retard at some high frequency the feedback will no longer be negative but positive. If the gain round the amplifier loop exceeds unity at this frequency then the amplifier will oscillate. Even if the loop gain is below unity then the amplifier may go unstable with even quite short leads to the loudspeaker due to the capacitive loading placed on the amplifier. Indeed it has been stated by Crowhurst (ref. 2) that the effect of near instability is quite audible and the amplifier gain margin should be at least eight times if this effect is to be inaudible.

    For this reason it is essential that all phase shifts that can be removed should be removed: either completely, or at least as far out of the way as possible. Here it might be well noted that the use of grid stoppers in feedback amplifiers is to be deplored unless they are absolutely necessary. Many parasitics have been caused rather than stopped by them!

    Basic circuit of Jeffery's high-gain phase-splitter.
    Basic circuit of long-tailed-pair phase-splitter.

    Two circuits that suffer from excessive HF phase shift are shown above. The first diagram shows the circuit due to ]effery (ref. 3) and it is unnecessary to go into the details of the phase shift in this circuit again. The interested reader is referred to the correspondence following the publication of a subsequent article (ref. 4).

    The long-tail circuit shown above is well known and widely used but the fact remains that its HF response is relatively poor. This cannot be due to the second half of the valve as this section is effectively driven as a grounded-grid amplifier. The trouble is due to the first half of the valve and is due to that hardy perennial-Miller effect. The gain of this first valve is effectively halved due to the input impedance at the second valve cathode. Even so the Miller capacitance is quite large and certainly cannot be neglected. In order that some estimated value of frequency response can be obtained, some typical values will be taken. Using the ECC83 as a typical valve, the quoted anode-to-grid capacitance is 1.6 pF so the total value will be certainly as large as 2.0 pF when wiring and valve-base capacitances are taken into account. The gain of each half of the valve can be as much as 60 times, but this would be better reduced to a factor of 50 as there is a supply voltage loss in the common-cathode resistor. The overall gain will therefore be about 25 times when used in this phase-splitter. This will give a reflected Miller capacitance of approximately 50 pF, so the total capacitance loading on the previous stage will be about 60 pF if 10 pF is allowed for all other capacitances. With a 100 kΩ source impedance this will give a -3 dB point at about 25 kHz, and this is clearly not good enough when output transformers with primary resonance's of about 150 kHZ are considered. The phase shift of the amplifier circuits must be as small as possible where the output transformer reaches its first primary resonance; and therefore the bandwidth of this type of phase-splitter can easily degrade the total amplifier performance. The use of a step network across the anode load resistor of the driving valve can help in this matter, but only If the step starts well before the natural fall-off frequency of the circuit itself. In the case just considered this would mean starting the fall-off of amplifier gain by the step network at approximately 2.5 kHZ or lower. This would obviously give excessive reduction in loop gain at the high frequencies, with consequent increase in distortion.

    The answer therefore lies in producing a fall-off at the HF end of the spectrum that starts at a much higher frequency. This could be attempted by reducing the value of the output impedance of the previous stage. This could be done by negative feedback with consequent gain loss; or alternatively by using a smaller value of anode load resistor. This latter also gives a severe loss in gain, quite apart from the increased noise that is produced by the increased valve current. The answer was therefore seen to lie in producing a phase-splitter that did not give the large input capacitance of the previous circuit.

    Circuits are known (eg ref. 5) that do give good HF response in phase-splitter service, but they suffer from high cost due to the complexity of the circuitry involved. The circuit that was finally evolved (see below) has a cost that is only slightly more than that of the conventional circuit, but has a greatly improved response at high frequencies.

                  
    Basic circuit of modified long-tailed-pair phase-splitter.

    The operation of the new circuit is just about identical with that of the conventional long-tailed pair except that the first valve is a pentode. This reduces the Miller effect to negligible proportions and increases the total bandwidth by a factor of just under ten times. Even allowing that the gain of the circuit is about 2 dB less than the conventional circuit, this still gives a gain/bandwidth improvement of about seven times. The comparative gain/frequency plots are shown below, where it is seen that the final rate of fall in both cases is identical at 20 dB per decade. This indicates an ultimate phase shift of 90 degrees which was borne out by measurement.

    Relative frequency response curves of original and modified long-tailed-pair phase-splitters.

    Owing to the partition of valve current in the pentode, the anode load resistor is made somewhat greater than that of the triode stage so that a balanced output is obtained. The circuit can be DC coupled to the previous stage as can the usual one, and a complete 'front-end' for driving the output valves of an amplifier is shown at the top of the article.

    This has a built-in 3:1 low-frequency step network that improves LF stability, and a 20:1 HF step network. The HF network may need component values altering for different types of output transformer due to the wide variation in resonant frequencies and other parameters.

    This circuit was originally developed for improving the performance of the Radford MAl2 and MA1S amplifiers and certainly did this to great effect. Equally there is no reason why the circuit should not give a considerable improvement in the stability of other amplifiers using good output transformers. The circuit is simple and stable and has no difficulty in providing drive for the largest output valves.

    A double-pentode having video amplifier characteristics and low input capacitance would enable greater overall gain to be obtained, but so far the author has not been able to find a valve with suitable characteristics. If such a valve were available, then the overall gain could be increased by a factor of about four times. Where sufficient spare stability margin was available this could lead to a further reduction in the distortion of the amplifier in use.

    Some Notes on Phase-Splitters

    J McG Sowerby. Wireless World, December, 1949.
     

    Selections from a Designer's Notebook

    Phase splitters are widely used in audio amplifiers and were treated in some detail in this journal some time ago. The notes that follow are merely disconnected jottings on a few points which, although not original, may be of interest to some readers.

    Fig. 1. 'Concertina' phase splitter: (a) AC coupled, (b) DC coupled.

    One of the most widely used phase splitters is known as the 'concertina' and is shown in Fig. 1(a). In this circuit a triode, V, has equal anode and cathode loads and these are effectively in series. Consequently, any signal current in the valve passes through both resistors and so equal output signal voltages are obtained at anode and cathode. The arrangement gives a gain of little less than one (usually about 0.9) between input and either output. It has the advantage that the balance of the output voltages depends only on the maintenance of equality of anode and cathode resistors. At (a) the bias for the valve is determined by Rb which is small compared with Rk. Consequently the cathode of V is positive to earth by anything up to 100 or more volts. This being so it seems logical to couple the grid of V directly to the previous anode as shown at (b), then the cathode potential of V will be very nearly the same as the anode potential of the previous stage. Besides saving three components the coupling eliminates any phase shift at low frequencies and this is often advantageous if V is within a feedback loop, as in Mr. Williamsons amplifiers. To a sufficiently close approximation the current in V can be taken as ia = Va/R where Va is the anode potential of the previous stage. The anode-cathode voltage of V is nearly Vak = (Vb  2Va) where Vb is the supply HT potential. From this we see that it is reasonable to design for Va = Vb/3 or less. Even so, it is wise to ensure from the valve curves that no grid current flows in V even when Va is 30% more than the design figure, making due allowance for the resultant reduction of Vak. This arises because some variation in Va is to be expected with time and with valve replacement.

    Fig. 2. Cathode-coupled phase splitter with two types of coupling.

    Fig. 2(a) shows a cathode-coupled phase splitter in which the usual positive bias is supplied from a fixed potentiometer across the HT supply. A more economical arrangement is shown at (b) using fewer components, in which the positive bias is the anode potential Va of the previous stage. Both grids are obviously maintained at the same standing potential, but the signal is applied only to the grid of V, as the 'smoothing' circuit CR prevents the signal reaching the grid of V2. Similar time constants are used in the two circuits, and at medium and high frequencies the performances are identical. At very low frequencies, however, CR in circuit (a) merely leads to attenuation of both outputs, whereas (b) reverts to a push-push output of low gain as the frequency approaches zero, because the grids of V1 and V2 both follow the anode potential of the previous stage, and are in phase with one another.

    An advantage of circuit (b) compared with (a) is that a large value of C can easily be used - say 2 μF or more. Such capacitors are usually of the paper-block-in-metal-can type and if used in the circuit at (a) may lead to loss at high frequencies due to the additional stray capacitance at the grid of V1. One such capacitor measured by the writer recently had a capacitor-can capacitance of more than 100 pF, and if such a condenser is fixed to an earthed metal chassis and used as a coupling condenser in (a), that 100 pF will appear between the grid of V1 and earth. In the circuit at (b) this capacitor-can capacitance will only add to C, and so be slightly beneficial. The circuit of Fig. 2(b) has a disadvantage inasmuch as variations in the HT supply potential may lead to a large push-push output from the splitter, and this may have undesirable results on subsequent stages.

    In designing the circuit of Fig. 2(b) the anode current of V1 or V2 may be taken to be Va/2Rc and the anode-cathode potential of each valve will be Vak = Vb  Va(I + Ra/2Rc) nearly. Previous remarks concerning grid current in the circuit of Fig.1(b) apply here too.

    Some designers are disinclined to use the 'concertina' and cathode-coupled circuits as both involve operating the cathode of a valve at a considerable potential to earth. A commonly used alternative is the 'see-saw' circuit. Another circuit which might be regarded as a combination of the cathode-coupled and see-saw arrangements if available, however, and is shown in Fig. 3.

    Fig. 3. Common-anode-coupled phase splitter. Typical values: V1, V2, = 6SN7, Rg1 = Rg2 = 1 mΩ, R2 = 1.8 kΩ, R = 39 kΩ, Ra1 = 80 kΩ, Ra2 = 100 kΩ, C1 = 0.05 μF, C2 = 0.25 μF, Ck = 20-50 μF, HT = 350V. Gain (input to one output) = 17 (approx.).

    It will be remembered that the cathode-coupled circuit of Fig. 2 tends to maintain equality of signal currents in the two valves, and that exact equality is approached as Rc is increased, eventually becoming exact when Rc is infinite. The circuit of Fig. 3 behaves in exactly the same way, provided that we substitute R for the Rc of Fig. 2. However, in this case there must always be some signal voltage across R (to operate the grid of V2) and this signal will be in phase with that obtained from the anode of V1. Consequently, however large R may be made to achieve equality of signal currents in V1 and V2, it will always be necessary to make Ra1 less than Ra2 to obtain equality 0f output voltages. On analysis it turns out that:-

    Ra1 = Ra2 (1 - (ra + Ra2 + 2R(1 + ra/Ra2) / (ra + Ra2 + (μ + 1)R)

    Where μ = amplification factor; ra = AC resistance; for either of the valves.

    for equal signal output voltages. In practice it is convenient to make Rk the common bias resistance for the two valves; this largely controls the direct current through the two valves and hence through R. Next, R is made as large as possible having due regard for the voltage drop across it. Ra2 is then fixed, and Ra1 calculated. The time constant C1 Rg1, is calculated as for a normal amplifier stage, and it is preferable to make C2 Rg2 several times larger than C1 Rg1 to maintain a phase displacement of 1800 between the output signals at low frequencies.

    The gain from input to either output (when Ra1 is properly chosen by means of the preceding equation) is:-

    A = μRa1 / (ra + Ra1 + R (Ra2 - Ra1)/(Ra2 + R))

    Substitution of practical values indicates that this circuit yields a gain approaching twice that for the ordinary cathode-coupled circuit of Fig. 2 when Rc = R, and the other constants are the same in both circuits. In this respect the splitter of Fig. 3 is more like the see-saw from which a correspondingly increased gain can also be obtained, and it also suffers from the disadvantage that any disturbances on the HT line are fed preferentially to one grid, so that the arrangement should be used only with a well-smoothed supply.

    In using the circuit of Fig. 3 and see-saw circuits one point needs watching. A common bias resistor is quite in order as long as it is bypassed. If it is not, then as there is feedback from both anodes to one grid, and feedback from one valve to the other via the cathode resistor, an unfortunate accidental combination of stray capacitances may easily turn either circuit into a cathode-coupled multivibrator at high frequencies. A relatively small cathode bypass capacitor will stop this trouble completely, but may also lead to a non-uniform frequency response. By using a capacitor of more usual value - 10 - 100 μF - these faults are both removed. If this is omitted and 'multivibration' takes place, the amplitude may be quite low and of quite high frequency 100 kHz or more and so difficult to find except with an oscilloscope, of low input capacitance, or a high-impedance valve voltmeter.

     

    and Now From Amplimos

    Experiences with the Cross-Coupled Phase Inverter

    Years ago on a number of a known Italian magazine I read an interesting article about phase inverters  (phase splitter). In reviewing the various pros and cons of each type of drive, eventually he sing the praises of the "Cross-Coupled Phase Inverter", being, according to the author of the article (a well-known and appreciated engineer), the most sophisticated and balanced type that it was possible to achieve.
    The  schemes in  examples were referred to tubes and since I needed for a solid state applicationo , I decided to build one in a hurry with Jfet.

     

    The scheme, visible in Figure 1, was not optimized but still functioning and sufficient to make me the idea if they were confirmed the superior qualities decanted by this circuit.
    As shown in the drawing, I made a series of measurements of the signals with an oscilloscope in various strategic point of the circuit where you see the relative phases and amplitudes.


    I do not deny that he was quite disappointed by the actual operation of this inverter: in spite of the considerable external symmetry, the signals at various points were not at all symmetrical and balanced and in a certain sense, it was similar to the Long Tail Pair or differential amplifier with source coupling.


    The two output signals were quite different in size and almost double that of each other: I was not happy at all.
    Given the almost non-existent contribution to the signal of T2, I decided to suppress it and decouple the second stage from the first by two capacitors to a more orthodox polarization of the output Jfet  . This we see in Figure 2

    The various waveforms and amplitudes of the signals detected in the various points are now visible in figure 2.
    Now the similarities of working with Long Tail Pair are more evident (2° stage), which become virtually identical if you close the switch SW1

    .
    In fact, by doing so, the transistor T2 operates with base to ground while T3 as emitter to ground amp, emulating in practice the long tail pair, preceded by the input buffer realized with T1.
    The amplitudes of the two output signals tend to be more similar than the previous case, but still quite different, as well as different remained the impedances of the input and the output.

    I then concluded it was better to make the phase splitter with a classic transformer with two secondary. I did enveloped it  by a well-known manufacturer with high quality materials  and special wrapping techniques. All this clearly affected the cost of such a jewel of a component that really for the price was comparable to that valuable.


    For a long time I thought of a new solution but clearly the good idea did not come immediately, and several attempts failed, but then finally after several years I get the good idea, so I started to work on the optimization of the circuit.
    They come out out different versions and the most significant are included in the document for the patent application.

    It 'a new type of phase splitter: I was looking for for years, and whose operation is truly symmetrical and balanced even more than a transformer, (not having its bandwidth limitations):  It's born the phase splitter of the 21st Century!!!!
     

       

     

    Description of the new Phase Splitter titled  "Fully differential phase inverter" 

    FIG.1 shows the common basic architecture from which were inspired  all circuits here presented: an input signal is wired with two inverting amplifiers and also with two non inverting amplifiers. Their relative network is made to form also twin simmetric differential amplifiers with a cross-connection of their inputs at the same time. Then their four outputs drive other 2 differential amps to obtain a convergent syntesis in  two output reflecting signals.

    FIG.2  illustrates the basic application circuit in which are transferred in a practical way the ideas expressed in the principle’s schematic of which in FIG.1; being the reference circuit, it will be analyzed in detail in the next.

    The circuit depicted in FIG.3 derives directly from that of FIG.2 where the second stage is also formed by two pairs of symmetric differential amplifiers, rather than asymmetric, as in the previous case (it have been added the resistors R9 and R10), and this allows, for the first time in this kind of circuits, the possibility to quadruple the number of output signals (four signals in pairs with the same phase). As in the previous case, these four signals, apart from the sign, are all synchronous.

    The circuit of FIG.4 is derived from that of FIG.2, when we must use only devices of the same polarity. (NPN or PNP, in this figure all N-channel). It represents a solution when you want to achieve this phase shifter with all vacuum tube devices (triodes, pentodes) or with other devices of which there is no complementary counterpart.

    The circuit of FIG.5 is a direct descendants of FIG.4 that provides up to four outputs with the output phases as shown.

    In the circuit of FIG.6 it have been exchanged some connections compared to those of FIG.4, whose phases output, as shown, reveal that it is perfectly equivalent to this latter.

    The circuit of FIG.7 follows directly from that  of FIG.5, where, with respect to this, it have been exchanged two of the four wires which carry the signal from the first group of differentials (first stage) to the second group (second stage). This operation just swap the order of the output phases but everything else is perfectly equivalent to the former in FIG.6 It is shown by way of example.

    Circuit of FIG.8 derives from that of FIG.3 having different input connections of the first stage: At first sight it not seems to found differential amps, indeed we can view several vertical diff. amps: Q1 with Q2, Q3 with Q4, Q5 with Q6, Q7 with Q8.

    Circuit of FIG.9 descends directly from that of FIG.8 in respect of having cross-connections from first to second stage.

     

    Circuit of FIG.10 is a two output version of that in FIG.9

     

    For circuits of FIG.2,3,4,5,6,7,8,9,10 are valid the following relations (if any):

    ·        Q1 = Q2 = Q3 = Q4;

    ·        Q5 = Q6 = Q7 = Q8;

    ·        R3 = R4 = R5 = R6;

    ·        R7 = R8 = R9 = R10;

    ·        VR1 = VR2;

    ·        VR3= VR4;

    ·        R2=R1//RG

    Referring to FIG.2 and, for the above statement, it is noted that the input circuit is formed by two symmetrical differential amplifiers either coupled by emitters (first bridge circuit). It is noted first that it is equivalent to describe the two differentials as formed by the pairs of transistors Q1,Q3 (high) for the first differential and Q2,Q4 for the second (bottom), rather than formed by pairs Q1,Q2 (left) for the first differential and Q3,Q4 (right) for the second differential. In conclusion the circuit in question can be equivalently decomposed into two differential circuits in vertical, horizontal or cross-connected mode. The important point to note is that, we could breaks down the circuit, still having input of said first differential connected to the hot pole of the signal generator G and the other input of the same differential amp connected to ground (cold), while the second differential amp will always have one input connected to the hot pole of the generator G and the other to ground but in reversed order compared to the first differential, so that the two differential amps have their input cross-connected.

    The second stage is also formed by twin emitter coupling differential amps (second bridge), thus the output impedances seen by the four transistors Q1, Q2, Q3, Q4 are identical.

    Referring to FIG.2, the base of Q1 is connected to the G hot pole (1), through the nodes 8,5,3, while the base of Q3 is connected to the cold pole (2) of the same generator G through nodes 9,7, the resistor R2, then nodes 6,4 and therefore connected  to ground; vice versa, the base of Q2 is connected to the cold pole (2) through said nodes 9,7, the resistor R2 and the remaining nodes 6,4, while the base of Q4 is otherwise connected to the hot pole (1) of signal generator G through the aforesaid nodes 8,5,3; then the collector of Q1 through node 10 is connected to R3 which then through nodes 13,20,27, and terminal 28 reach the positive pole of battery B1; the collector of Q3, arrives via node 14 to the resistor R5 whose other end is connected to the node 13 and from this through the cited nodes 20,27 and terminal 28 reaches the positive of B1 battery; the Q2’s collector through the node 11 is coupled to resistor R4 whose other end  arrives, through nodes 16,27, and the predicted terminal 28, to the positive battery B1; the collector of Q4 via the node 15 is connected to R6 which in turn is connected to the nodes 16 and then 27 and from this to terminal 28, where it is connected to the positive pole of B1 battery; the emitter of Q1 through VR1 is connected to node 12; the emitter of Q3 through VR2 is also connected to node 12; the emitter of Q2 through VR1 is also joined to node 12; the emitter of Q4 through VR2 is connected to the predicted node 12 which is connected to one end of the constant current source I1, the other head of I1 through the node 26 is connected to terminal 30, that is  the negative pole of the B2 battery, the node 10 wired to base of Q5, the node 14 to base of Q7, the node 11 to base of Q6, the node 15 to base of Q8; node 20, carries the positive voltage to one end of the constant current source I2 whose other end is connected to the central contacts of VR3 and VR4 whose external contacts of the two trimmers are connected to the emitters, respectively, of Q5, Q7 and Q6, Q8; collector of Q5, via node 17 reaches, by one side  the output terminal 22 (phase +) and, from the other side, the load resistor R7, which is then connected to the node 18, then it is is connected the other load resistor R8 and by this to the node 19, where there is a first connection to the output terminal 24 (phase -) and a second to the collector of Q6; the collector of Q7 is connected directly to node 25 and then to the collector of Q8; node 18 is connected to node 25, then to 26, and finally to terminal 30 that’s the negative B2 battery pole; the remaining output terminal 23 is connected to terminal 29, the central connection among the two batteries B1 and B2 and from this, connected to the ground system.

    For a common understanding, we suppose in the following to refer as differential amplifiers  formed by horizontal pairs.

    By applying an input signal with the polarity indicated, on the collectors of transistors Q1 and Q4 (10,15 knots), working these BJT as common emitter amps, we get two identical signals both shifted by 180° with respect to the reference of G, which we call signal (A) ; simultaneously on the collectors of Q2 and Q3 (11,14 knots), operating these devices as common base amplifiers, we obtain two equal signals between them and in phase with the reference input signal of G, which we denote as signal (B). The upper node 10 is connected to base of Q5, the node 14 to base of Q7, while on the lower branch, the node 11 is connected to base of Q6 and the node 15 to base of Q8, therefore it is apparent that the pair Q5,Q7 receive a pair of signals (A) + (B), while the pair Q6,Q8 receive a pair of signals identical to the first but in reverse order, hence (B) + (A). In fact, the transistor Q5 receives directly on its base the signal (A) and on emitter, through Q7, the signal (B), while the transistor Q6 receives directly on its base the signal (B) and on emitter through Q8 it receives the signal (A), instead. Forming both pairs Q5,Q7 and Q6,Q8 two differential amplifiers, (whose outputs  are function of these two input signals (A) and (B), on the collectors of Q5 (nodo17) and Q6 (nodo19) we obtain two signals, which, although identical in every other electrical parameter, are however in phase opposition by 180°, which it is the result sought. These signals are withdrawable on terminal 22 (phase +) and 24 (phase-) with terminal 23 bonded to ground act as common reference for both. The outputs are also reported to zero DC voltage and this allows a direct coupling of the invention to the user circuit. It is also noted that, as a result of the symmetry of circuit, the frequency response, the output impedance and any other parameter of the out signals are perfectly coincident. For merit of this, the proposed circuit can also boast of a good external noise rejection and a good behavior against power supply noise, as well as a low intrinsic distortion. Excluding the sign, the mutual phase delay and the time slider between the two supplied signals is zero, ie, they are perfectly synchronous.

     

    In FIG.11 is proposed the simplified version of the circuit object of application, which expect dual power supplied by batteries (B1, B2) or equivalent, six active devices of amplification (Q1, Q2, Q3, Q4, Q5, Q6) including bipolar transistors (BJT) or unipolar transistors (JFET, VFET, MOSFET, SIT and the like), one constant current source (I), two trimmers (VR1, VR2) two input resistors (R1, R2), two capacitors (C1, C2), six load resistors (R3,R4,R5,R6,R7,R8), two signal input terminals (1-2), three output terminals (25-26-27), three power supply terminals (21-22-23). A signal generator (G) is applied with the polarity shown between the terminals 1 and 2 while the signal supplied to the utilizer is taken between terminals 25-26 for the positive phase and 26-27 for the negative phase, being 26 the common ground terminal.

     

     

    FIG.11 shows the simplified version of those type of circuits that, also maintaining the same FIG.1 principle’s structure, have a minimal number of components. Since this circuit is quite important, it will be treated in more detail in the following .

    In FIG.12 is presented a version circuit derives from that in FIG.11 utilizing unipolar JFET devices, instead of transistors, as example

    FIG.13 is instead a variant of FIG.11, in which they were exchanged collector’s connections between transistors Q3,Q4 to  transistors Q5,Q6 in order to decrease the value of the total amplification rate. In fact it not should forgotten that the circuit of FIG.11 has a high coefficient of amplification in that the pair Q1,Q3 connected to Q5, and Q2,Q4 connected to Q6, on the lower branch, form two identical push-pull folded cascode circuits.

     

    For FIG.11,12,13, we define the following features and limitations:

    1.       transistor Q1 = Q2 = Q3 = Q4 ;

    2.       transistor Q5 = Q6;

    3.        resistor R3 = R4;

    4.       resistor R5 = R6;

    5.       resistor R7 = R8;

    6.       resistor R2 = R1// RG

    Referring to FIG.11, it is observed that the bases of Q1 and Q4 are connected, through nodes 8,5,3 and terminal 1, to the positive phase of signal generator G, while the bases of Q2 and Q3, through nodes 9, 7, resistor R2 and node 6, to the ground signal. At the same ground, through terminal 2 and nodes 4 and 6, is connected the other end of the signal generator G (the negative phase). Consequently, we can observe that the network in which are inserted the aforesaid transistors Q1, Q2, Q3,Q4 form a multiple differential amplifier. In particular we can see two vertical differential amps of which the first formed by the pair Q1,Q2 and the second formed by the pair Q3,Q4. At the same time with the  same network and the same four devices we can see also two horizontal differential amplifiers, of which the upper formed by Q1,Q3 and the bottom formed by the pair Q2,Q4. In addition, for the created network, there are two further simple differential amplifiers the first of which formats by transistor Q5 (for the upper branch) and the other by Q6 (for the lower branch); Q1 collector through node 10 is connected to base of Q5, Q3 collector to emitter of Q5 through node 14, and, on the lower branch, collector of Q2 through node 11 is connected to base of Q6, while collector of Q4 is connected to emitter of Q6 through node 18; finally, collector of Q5 through node 15 is wired to output terminal 25 and collector of Q6 via node 17 is connected to output terminal 27; terminal 26 connected to ground through node 24 is also the reference for the output signals present on terminals 25 and 27.

    Applying a signal between input terminals 1, 2, with the polarities as indicated we obtain two signals equally amplified in voltage on collectors of transistors Q1 and Q4 (nodes 10, 18), with negative phase, which we will call signal (A), while on both collectors of transistors Q2 and Q3 we get two equal signals also amplified in voltage (generally a bit different from the first), with positive phase, which we will call signal (B) (nodes 11, 14). It follows that base of Q5 receives a signal (A) identical to that received from emitter of Q6, while emitter of Q5 receive a signal (B) identical to that received from base of Q6. Forming each of said transistors Q5 and Q6 a differential circuits where the amplification coefficient is proportional to the sum of the signals (A) and (B) present on the respective bases and emitters, it follows that the signals on the collectors of transistors Q5 and Q6 (and then on the outputs 25 and 27) will be amplified at the same level of amplitude but with opposite phases. It is also noted, as a result of the circuit’s symmetry, that the frequency response, the output impedance and any other parameter of the two signals provided are very coincident. For the same reason, this proposed circuit can also boast of a good external noise rejection and a good behavior against power supply noise, as well as a low intrinsic distortion. A further advantage, compared to existing phase-inverters, is the not negligible property of being able to be free of DC voltage (offset) at its outputs, which obviates use of capacitors between its outputs and user equipments.

     

     

             In the following, some main features:

     
    bullet

    Easiness and economy cost of realization of accurate phase inverters (phase splitters);

    bullet

    Generation of two or four phase-shifted signals perfectly specular , which are maintained exactly balanced even at very high frequencies;

    bullet

    Absence of DC voltage on the outputs;

    bullet

    Direct connection of the user without the interposition of coupling transformers or capacitors thanks to the absence of offset voltage versus ground on its outputs;

    bullet

    Strong voltage amplification of the input signal from an audio source or other equipment;

    bullet

    Lower harmonic distortion;

    bullet

    High CMRR and PSRR due to the high symmetry of circuit;

    bullet

    Flexibility of operation due to the fact that the circuit can be realized with two or four outputs signals, as well as with low or high voltage gain;

    bullet

    Naturally predisposed for multi-amplification;

    bullet

    High stability and reliability of the realized circuits;

    bullet

    High Insensitivity to external disturbances and induced noise;

    bullet

    Ease of connection with other electronic devices, due to absence of external offset voltages;

    bullet

    High bandwidth especially when referring to transformers phase splitter.                                                                                                                                                        

     

     

     

     

     

     

     

     

     

     

     

    These circuits are know overall in audioapplication (preamplifier and push pull amplifiers, unbalanced to balanced signals in professional audio gear),but also in Switched Mode Power Supply, in instrumentation and other fields of science and physics.

    If you or your company have interest in this last application, please contact me:     E-mail:  info@amplimos.it