I have not yet performed a Tian Test as I do not know what is the definition of variable 'u'. I also tried to do google searches to read about Tian testing but did not succeed to find anything. I would like to know what are the objectives of this test and why a theoretical constant current source and a voltage source are connected in series with the feedback resistance chain. As I may understand, the purpose is to inject some sort of frequency to evaluate the rejection effect of the feedback loop and Miller Capacitance.
Headphones used to come equipped with that sort of plug commonly referred to as a Jack abeit the socket on a domestic amplifier was an output.
I assumed that stage equipment had moved on to XLR connectors - I know this style of Jack was used for mixer and amplifier connectors back in the 70's. Stuff that old of passing interest only is not in my instant recall.
Headphones used to come equipped with that sort of plug commonly referred to as a Jack - albeit the socket on a domestic amplifier was an output.
I assumed that stage equipment had moved on to XLR connectors - I know this style of Jack was used for mixer and amplifier connectors back in the 70's. Stuff that old, of passing interest only, is not in my instant recall.
See Plotting Loop Gain Using the Tian Method - Spring 2011
I assumed that stage equipment had moved on to XLR connectors - I know this style of Jack was used for mixer and amplifier connectors back in the 70's. Stuff that old of passing interest only is not in my instant recall.
Headphones used to come equipped with that sort of plug commonly referred to as a Jack - albeit the socket on a domestic amplifier was an output.
I assumed that stage equipment had moved on to XLR connectors - I know this style of Jack was used for mixer and amplifier connectors back in the 70's. Stuff that old, of passing interest only, is not in my instant recall.
See Plotting Loop Gain Using the Tian Method - Spring 2011
Two Grounding Omissions Found! 
a) two large current loops with areas exceeding 100 square centimeters were found. These are formed by the output conductors feeding the output PCB. One of them is almost touching the transformer! Both amplifier channels have these loops with the one having its loop almost touching the transformer exhibiting the loudest hum/buzz/noise.
b) the grounding of the amplifier modules is trough a large aluminium plate at the bottom of the amplifier box. Lifting this plate off the box's base, I found there is paint insulating it from the box. Checking with an Ohmmeter, I found it is completely isolated from ground! This means, the only ground path from the amplifier modules was through the capacitance formed by the amplifier box's base, paint and aluminium plate.
I have the intuitition, which can be wrong, that these omissions may have forced grounding currents to take strange abnormal paths, causing the hum/buzz/noise issues.
My not so efficient memory tells me the output conductors feeding loadspeakers are to be included in the elmination list of loop areas.
a) two large current loops with areas exceeding 100 square centimeters were found. These are formed by the output conductors feeding the output PCB. One of them is almost touching the transformer! Both amplifier channels have these loops with the one having its loop almost touching the transformer exhibiting the loudest hum/buzz/noise.
b) the grounding of the amplifier modules is trough a large aluminium plate at the bottom of the amplifier box. Lifting this plate off the box's base, I found there is paint insulating it from the box. Checking with an Ohmmeter, I found it is completely isolated from ground! This means, the only ground path from the amplifier modules was through the capacitance formed by the amplifier box's base, paint and aluminium plate.
I have the intuitition, which can be wrong, that these omissions may have forced grounding currents to take strange abnormal paths, causing the hum/buzz/noise issues.
My not so efficient memory tells me the output conductors feeding loadspeakers are to be included in the elmination list of loop areas.
Last edited:
So having remedied these anomalies does this indicate final success? The output of your amplifier is also connected to the inverting input.
A radiated electromagnetic field at 50 Hz will induce a current in a conductor if the former is large enough and the latter is critically close.
Seeing the cramped space inside your chassis I suggested containment of the transformer stray field by fitting a copper strap around the transformer.
Whatever the stray field intensity is, it will be amplified. Your best hope is that distance helps mitigate the risk twisting leads also helps. You could re-route your wiring accordingly.
A radiated electromagnetic field at 50 Hz will induce a current in a conductor if the former is large enough and the latter is critically close.
Seeing the cramped space inside your chassis I suggested containment of the transformer stray field by fitting a copper strap around the transformer.
Whatever the stray field intensity is, it will be amplified. Your best hope is that distance helps mitigate the risk twisting leads also helps. You could re-route your wiring accordingly.
The aluminium plate is now earthed with a 2.5 sq mm copper wire connected to earth. At both ends of the wire I used terminal lugs that were fixed using small bolts. I made sure the length of the wire is the shortest possible. The amplifier modules will be fitted with their main heatsinks, which are connected to 0V, in copious contact with the underlying aluminium plate, which is now securely earthed.
The input wiring will be isolated from power earth so as to be fed only from signal earth from the amplifier modules. I will have to make sure signal grounding does not give rise to loops as there are two channels.
Next task is rerouting the output wires as close as possible to the original PCB tracks feeding the output stage.
The stray mangetic flux will be minimized using an aluminium tape wound a couple or 3 times around the transformer's outer edge.
Hopefully, this resolves all these issues, as this journey has now taken far too long. As suggested earlier by Mooly, quick blow fuses can be used for speakers. I will use this amplifier at home, so huge powers are not necessary. Nevertheless, the adventurous journey exploring complexities as a result of allowing for very big powers, was a nice learning experience. I can only say, when I first tried the amplifier with large professional speakers, the sound was beautiful, powerful and detailed. Thanks to this forum, I would have never been capable to achieve this much.
Wrapping an aluminium tape around the toroidal transformer, rerouting the output wires, and hard earthing the main heatsinks had no effect on hum. Signal earth and power earth were split with a 10R resistor. This, increased the hum intensity. It was like staying besides a power distribution transformer like those found in substations.
Earlier in this long, refusing to die thread, I pointed my fingers at the input stage, and claimed it is the cause of these unwanted ingredients. Power ampfifiers and integrated circuits have complex input stages for a reason. One such reason is very high common mode rejection. Literature states, an input stage's constant current source is a strong determinant of common mode rejection. In this project I never explored this sufficiently to make changes to the actual amplifier circuit.
In my amplifier circuit the input constant current source's output current is supposed and assumed to be constant. However, this is not so. For a given base voltage, Ic changes in accordance with the current source's transistor output impedance. This impedance is not infinite, implying, it will sense ANY voltage change of the collector. Power supply imperfections affect this voltage. This results in modulation of the current source's output current. When the differential pair are subjected to signals, the latter have to compensate for changes in the current source's Ic. Complete compensation is difficult to achieve, so the input stage's output ends up soiled with traces of power supply imperfections.
In my case, there is a way of improving the performance of the input stage's current source. This is by using a transistor to steal some base current when the current source's Ie tries to increase.
With the new earth connection fixed to the heatsinks, and removing the circuitry for additional smoothing for the input stage, and also removing the ground splitting 10R resistance, the issue of hum/buzz has been greatly reduced. In case a 10R resistance is necessary, it can be placed between signal source and amplifier.
With the new level of hum/buzz, I have the impression I am trying to improve beyond what is possible with this amplifier. For the power rating, this is extremely small.
The input wiring will be isolated from power earth so as to be fed only from signal earth from the amplifier modules. I will have to make sure signal grounding does not give rise to loops as there are two channels.
Next task is rerouting the output wires as close as possible to the original PCB tracks feeding the output stage.
The stray mangetic flux will be minimized using an aluminium tape wound a couple or 3 times around the transformer's outer edge.
Hopefully, this resolves all these issues, as this journey has now taken far too long. As suggested earlier by Mooly, quick blow fuses can be used for speakers. I will use this amplifier at home, so huge powers are not necessary. Nevertheless, the adventurous journey exploring complexities as a result of allowing for very big powers, was a nice learning experience. I can only say, when I first tried the amplifier with large professional speakers, the sound was beautiful, powerful and detailed. Thanks to this forum, I would have never been capable to achieve this much.
Wrapping an aluminium tape around the toroidal transformer, rerouting the output wires, and hard earthing the main heatsinks had no effect on hum. Signal earth and power earth were split with a 10R resistor. This, increased the hum intensity. It was like staying besides a power distribution transformer like those found in substations.
Earlier in this long, refusing to die thread, I pointed my fingers at the input stage, and claimed it is the cause of these unwanted ingredients. Power ampfifiers and integrated circuits have complex input stages for a reason. One such reason is very high common mode rejection. Literature states, an input stage's constant current source is a strong determinant of common mode rejection. In this project I never explored this sufficiently to make changes to the actual amplifier circuit.
In my amplifier circuit the input constant current source's output current is supposed and assumed to be constant. However, this is not so. For a given base voltage, Ic changes in accordance with the current source's transistor output impedance. This impedance is not infinite, implying, it will sense ANY voltage change of the collector. Power supply imperfections affect this voltage. This results in modulation of the current source's output current. When the differential pair are subjected to signals, the latter have to compensate for changes in the current source's Ic. Complete compensation is difficult to achieve, so the input stage's output ends up soiled with traces of power supply imperfections.
In my case, there is a way of improving the performance of the input stage's current source. This is by using a transistor to steal some base current when the current source's Ie tries to increase.
With the new earth connection fixed to the heatsinks, and removing the circuitry for additional smoothing for the input stage, and also removing the ground splitting 10R resistance, the issue of hum/buzz has been greatly reduced. In case a 10R resistance is necessary, it can be placed between signal source and amplifier.
With the new level of hum/buzz, I have the impression I am trying to improve beyond what is possible with this amplifier. For the power rating, this is extremely small.
The bias resistor R6 for your constant current source is too low in consideration of the voltage rail levels.
The current passed through R6 is more than double what flows through the input LTP which implies if there is ripple on the negative rail you could reduce the impact by increasing the value to 22k.
You could also increase the value of C8 to 100uF to improve positive rail filtering.
If you connected the 10R resistor anywhere on the wire linking the bottom ends of the power supply capacitors that is entering a war zone of current so you need to form a T junction to avoid the damage and station your entry points a safe distance down the track.
The link I provided in a recent post shows how to arrange these. The center tap need more spacing than the supply decoupling capacitors and the 10R isolating the input signal earth references is the last. I drilled hole in my chassis and put a bolt through and used solder eyelets to allow the wires to connect in descending order of current. I used nuts to separate the eyelets. You will need to remove any paint for the bolt to make good contact with the chassis as for your safety earth use star washers under normal load spread ones in accordance with your electrical authority regulations.
The current passed through R6 is more than double what flows through the input LTP which implies if there is ripple on the negative rail you could reduce the impact by increasing the value to 22k.
You could also increase the value of C8 to 100uF to improve positive rail filtering.
If you connected the 10R resistor anywhere on the wire linking the bottom ends of the power supply capacitors that is entering a war zone of current so you need to form a T junction to avoid the damage and station your entry points a safe distance down the track.
The link I provided in a recent post shows how to arrange these. The center tap need more spacing than the supply decoupling capacitors and the 10R isolating the input signal earth references is the last. I drilled hole in my chassis and put a bolt through and used solder eyelets to allow the wires to connect in descending order of current. I used nuts to separate the eyelets. You will need to remove any paint for the bolt to make good contact with the chassis as for your safety earth use star washers under normal load spread ones in accordance with your electrical authority regulations.
The earth connection has been previously secured with an insulated bolt to the bottom of the amplifier box. This has been exchanged with a longer stainless steel bolt. In the original setup, the paint at the contact area has not been removed. The only contact is from the sides of the drilled hole in the metal. To improve the contact I removed some paint, butn found the paint is quite tough to remove. I will uncrew the bolt and remove more paint to obtain a better earth connection to the amplifier box. Aside from regulations, this connection can be a life saver, and I haven't yet given up to living on this planet.
The input circuitry, now shares the same ground as power. Using a separating 10R resistor resulted in more pronouced issues. I remember from reading Douglas Self that an input cascode which is biased from a voltage devider outside the input stage, may suffer hum issues: Self remarks nobody yet understands why.
Simulating for the current source feeding the input stage, I found it is modulated by the input signals and power supply ripple. I am associating the little remaining hum/buzz with a current source that deviates from having an infinite output impedance. I also found the bias supplying the current sources can be further made independent of power supply modulation.
The Huge Toroidal Transformer:
I did several tests to determine whether the huge toroidal transformer, is injecting hum/buzz signals in the input stage via magnetic induction. All tests indicate this is not the case, or if it is, the effect must be very minimal.
Replacing the input stage's current source with an ideal current source, resulted in no modulation and no ripple fed into the VAS. This confirms I am on the right track.
The input circuitry, now shares the same ground as power. Using a separating 10R resistor resulted in more pronouced issues. I remember from reading Douglas Self that an input cascode which is biased from a voltage devider outside the input stage, may suffer hum issues: Self remarks nobody yet understands why.
Simulating for the current source feeding the input stage, I found it is modulated by the input signals and power supply ripple. I am associating the little remaining hum/buzz with a current source that deviates from having an infinite output impedance. I also found the bias supplying the current sources can be further made independent of power supply modulation.
The Huge Toroidal Transformer:
I did several tests to determine whether the huge toroidal transformer, is injecting hum/buzz signals in the input stage via magnetic induction. All tests indicate this is not the case, or if it is, the effect must be very minimal.
Replacing the input stage's current source with an ideal current source, resulted in no modulation and no ripple fed into the VAS. This confirms I am on the right track.
You need to review the values of R9 and R30 in your current mirror.
For more information run a search "comments on audio power amplifier design handbook by douglas self" and read section 3.
There is an example you can look up on this site - the popular Explendid amplifier which demonstrates the point. This does not have a cascode stage in the LTP however there is a Technics high power amplifier that does which uses 680R emitter resistors.
For more information run a search "comments on audio power amplifier design handbook by douglas self" and read section 3.
There is an example you can look up on this site - the popular Explendid amplifier which demonstrates the point. This does not have a cascode stage in the LTP however there is a Technics high power amplifier that does which uses 680R emitter resistors.
Last edited:
Rail decoupling filters for the input stage are being confirmed to improve the performance of the input stage. Removing them for testing purposes, resulted in increased hum issues.
The 10R ground splitting resistor is used to reduce earth loop currents due to different earth potentials. However, I am still mulling where is the best place for this in my amplifier schematic. I would like to keep the input and output with a continous ground, and insert such a resistor just before the input. This way, unwanted ground currents from an external music player should be reduced.
Using a cascode for the input stage current mirror reduced the ripple to 55uV peak to peak. The modulation of the current mirror's output was still modulated by the ripple, but to a much lesser extent.
This may be naive on my part:
I am after ZERO modulation of the input current mirror's output current.
The 10R ground splitting resistor is used to reduce earth loop currents due to different earth potentials. However, I am still mulling where is the best place for this in my amplifier schematic. I would like to keep the input and output with a continous ground, and insert such a resistor just before the input. This way, unwanted ground currents from an external music player should be reduced.
Using a cascode for the input stage current mirror reduced the ripple to 55uV peak to peak. The modulation of the current mirror's output was still modulated by the ripple, but to a much lesser extent.
This may be naive on my part:
I am after ZERO modulation of the input current mirror's output current.
Had I known it all, this project would have been completed months ago. But, alas, I cannot get rid of this hum. Regarding increasing the emitter resistors of the current mirror, I simulated increasing the resistance to 330R, but did not get simulation results that suggest this will correct the output hum. The only changes that suggest the hum will be reduced are using rail decoupling and a cascode for the input stage current mirror.
If you run the Tian plot for your circuit you will see the gain at 1Hz is less than 30 dB. That is less than the 35 dB of closed loop gain you are expecting this to deliver.
A simulation I posted last year had 76 dB of gain at 1Hz and 25.7 dB of closed loop gain leaving roughly 50 dB to deal with distortion and hum.
I suggest you go back over your files to the point prior to incorporating the circuit protection where some resistor values in the driver stage might have changes to accommodate that addition - then run a Tian test on one of these.
A simulation I posted last year had 76 dB of gain at 1Hz and 25.7 dB of closed loop gain leaving roughly 50 dB to deal with distortion and hum.
I suggest you go back over your files to the point prior to incorporating the circuit protection where some resistor values in the driver stage might have changes to accommodate that addition - then run a Tian test on one of these.
For the Tian Plot I need the definition of variable u or study the concept behind it. The latter takes time as I am one who seeks to understand down to details. Without such understanding it will be like someone groping in complete darkness.
Very low frequencies, like those below 10Hz, can easily damage speaker coils as the inductive reactance will be almost non-existent. I payed attention to roll down the gain as frequency drops, especially in the region of 0Hz - 10Hz.
I have reinstalled the rail filters and got very good results but I removed the 10 Ohm resistor. I will not be certain until I test the amplifier with better more efficient speakers. Carefully listening into the speakers, I could not hear a hum, the only remaining artifacts is a slight crackle.
I will also increase the current mirror emitter resitors from 20R to 470R. Simulating did not show such a change will negatively affect the amplifier THD or stability. Literature indicates increasing this resistance decreases noise as current mirrors have noisy currents but not noisy output voltages. This means, I should get an additional improvement in lessened noise, and hence, also weaker crackles.
Very low frequencies, like those below 10Hz, can easily damage speaker coils as the inductive reactance will be almost non-existent. I payed attention to roll down the gain as frequency drops, especially in the region of 0Hz - 10Hz.
I have reinstalled the rail filters and got very good results but I removed the 10 Ohm resistor. I will not be certain until I test the amplifier with better more efficient speakers. Carefully listening into the speakers, I could not hear a hum, the only remaining artifacts is a slight crackle.
I will also increase the current mirror emitter resitors from 20R to 470R. Simulating did not show such a change will negatively affect the amplifier THD or stability. Literature indicates increasing this resistance decreases noise as current mirrors have noisy currents but not noisy output voltages. This means, I should get an additional improvement in lessened noise, and hence, also weaker crackles.
I will not stop here; this project has to be completed, and happily tests are indicating this is very probably the case. As insisted for many times, I will not use this amplifier for professional purposes. That would be very naive businesswise. That would be making a bad impression that can easily ruin customer confidence. Present day customers want to be served with the latest, if possible, most expensive equipment from the most reputable manufacturers. I am not interested in setting up such a business. My interest is to set up a home audio system with a flat response. This is why I opted to use professional equipment.
This thread would never have been created had the amplifier not failed. Initially, I was very happy and proud with it. When it failed, after opening it, I became aware this amplifier could be modified to accomodate a more standard amplifier schematic. The original did not have such a circuit. The original was grounded in such a way the supply rails oscillated with the output as the latter was the ground terminal. The input, however, was still ground/earth reference, unlike to what I thought at first.
Some may mistakenly interpret my English writing style indicative of a proud pretentious character. This is not the case, anyone who is proud and pretentious will never acknowledge their mistakes, especially, publicly. The reason behind my seemingly hauty style is I was never exposed to colloquial English, but only to book style and academic English. Since this version of the language is used by scholars, and hence, is by its very nature, more precise, I prefer to use it instead of some other version that is less precise and rigorous.
The Project Status:
From an early point in this thread I suspected the input current source may be the real cause of the hum, buzz and noise issues. Simulations have constantly shown, the theoretically constant current was not constant, but modulated by the input signal. Lately, simulations showed that they were also modulated by the power supply ripple. Temporarily replacing the current source with a theoretical ideal source, resulted in no modulation taking place. This is why I concluded it is the current source that should be mitigated against power supply ripple and other imperfections.
The amplifier is ready, provided it is used for a home environment. Higher powers were not tested for. I am testing it to eliminate remnants of hum and buzz. The noise issues have been reduced, thanks to suggestions by mjona. If I was rude, I would like to humbly apologize. This project has taken too much effort from me.
I am testing for hum elemination by inserting a 3mm aluminium plate and thin iron sheets between the toroidal transformer and the amplifier module sitting next to the transformer. It seems, the little hum issue can be resolved this way, but I have to wait when ambient noises are much lower. Simulating for the buzz issue has confirmed this is generated by the output stage. Disconnecting the output stage from the driver stage did not have any effect on buzz. It remained with almost the same intensity, only the waveform changed. The waveform is a square wave at 100Hz with very short rise and fall times. The duty cycle is very small.
Further improvements to the input stage current source biasing did not affect the output buzz. This means, the input stage is optimal as it is, and needs no more improvements.
I am simulating with the output stage isolated from the driver stage to find ways to bypass this parasitic buzz. With the drivers disconnected, the only way to get an output is through parasitics.
This thread would never have been created had the amplifier not failed. Initially, I was very happy and proud with it. When it failed, after opening it, I became aware this amplifier could be modified to accomodate a more standard amplifier schematic. The original did not have such a circuit. The original was grounded in such a way the supply rails oscillated with the output as the latter was the ground terminal. The input, however, was still ground/earth reference, unlike to what I thought at first.
Some may mistakenly interpret my English writing style indicative of a proud pretentious character. This is not the case, anyone who is proud and pretentious will never acknowledge their mistakes, especially, publicly. The reason behind my seemingly hauty style is I was never exposed to colloquial English, but only to book style and academic English. Since this version of the language is used by scholars, and hence, is by its very nature, more precise, I prefer to use it instead of some other version that is less precise and rigorous.
The Project Status:
From an early point in this thread I suspected the input current source may be the real cause of the hum, buzz and noise issues. Simulations have constantly shown, the theoretically constant current was not constant, but modulated by the input signal. Lately, simulations showed that they were also modulated by the power supply ripple. Temporarily replacing the current source with a theoretical ideal source, resulted in no modulation taking place. This is why I concluded it is the current source that should be mitigated against power supply ripple and other imperfections.
The amplifier is ready, provided it is used for a home environment. Higher powers were not tested for. I am testing it to eliminate remnants of hum and buzz. The noise issues have been reduced, thanks to suggestions by mjona. If I was rude, I would like to humbly apologize. This project has taken too much effort from me.
I am testing for hum elemination by inserting a 3mm aluminium plate and thin iron sheets between the toroidal transformer and the amplifier module sitting next to the transformer. It seems, the little hum issue can be resolved this way, but I have to wait when ambient noises are much lower. Simulating for the buzz issue has confirmed this is generated by the output stage. Disconnecting the output stage from the driver stage did not have any effect on buzz. It remained with almost the same intensity, only the waveform changed. The waveform is a square wave at 100Hz with very short rise and fall times. The duty cycle is very small.
Further improvements to the input stage current source biasing did not affect the output buzz. This means, the input stage is optimal as it is, and needs no more improvements.
I am simulating with the output stage isolated from the driver stage to find ways to bypass this parasitic buzz. With the drivers disconnected, the only way to get an output is through parasitics.
Could you please, explain what you mean?R Dijk said:
I may be able to add more capacitance to the smoothing stages on the power supply PCB, which, unfortunately, has some 20 thick wires connected to it. I have four 1000uF, 100V, capacitors available. I can use one capacitor per rail and insert a 0.1R resistor to form a CRC filters. The CRC filter would be 1000uF, 0.1R, 6600uF with the 1000uF directly connected to the rectifier bridge. I know, it is not precisely what you suggested, but it may be a possible implementation given the hardware limitations that I have with this amplifier.
Hum/Buzz testing with a pair of professional speakers:
Testing during the most quiet time, at midnight, I am pleased to write the hum and buzz issues have been reduced to barely audible.
The amplifier was tested without a signal source connected with both volume knobs set to maximum.
More power supply filtering, as suggested, will further weaken the hum and buzz. Hurrah!
Thanks to all who helped. Mooly, mjona, R Dijk and the rest, rejoice with me.
Testing during the most quiet time, at midnight, I am pleased to write the hum and buzz issues have been reduced to barely audible.
The amplifier was tested without a signal source connected with both volume knobs set to maximum.
More power supply filtering, as suggested, will further weaken the hum and buzz. Hurrah!
Thanks to all who helped.
Mooly, mjona, R Dijk and the rest, rejoice with me.I am posting a Tian test of the amplifier. There are a couple of small changes - the compensation follows the point made by Samuel Groner about reducing the load on the VAS.
If the output standing current is too low the gain at 1Hz is low in comparison to higher standing currents. The file is attached - to run click on the output as usual and copy the expression in blue at the top. You need to click on vout in the header and replace that with the expression copied. There are two cursor options select the two and OK.
To see the margins drag cursor 1 so it aligns close to 0 dB on the plot. You can make fine adjustments using the left and right arrow keys on your keyboard. Drag the second cursor likewise so it aligns with -180 degrees. The margins register in at the bottom of a drop down box. I usually drag this to the left of the screen near the bottom. The gain and phase margins are roughly -11 dB and - 75 degrees.
High power amplifiers are not for everyone - I have drawn up a simulation for a 50 watt version that uses a single pair of MJL21193/MJL21194 outputs The gain margin for this is -24dB the phase -76 degrees.
I have used some Cordell device models in the attachment I went 100% with these for the low power circuit.
If the output standing current is too low the gain at 1Hz is low in comparison to higher standing currents. The file is attached - to run click on the output as usual and copy the expression in blue at the top. You need to click on vout in the header and replace that with the expression copied. There are two cursor options select the two and OK.
To see the margins drag cursor 1 so it aligns close to 0 dB on the plot. You can make fine adjustments using the left and right arrow keys on your keyboard. Drag the second cursor likewise so it aligns with -180 degrees. The margins register in at the bottom of a drop down box. I usually drag this to the left of the screen near the bottom. The gain and phase margins are roughly -11 dB and - 75 degrees.
High power amplifiers are not for everyone - I have drawn up a simulation for a 50 watt version that uses a single pair of MJL21193/MJL21194 outputs The gain margin for this is -24dB the phase -76 degrees.
I have used some Cordell device models in the attachment I went 100% with these for the low power circuit.
The amplifier is nearly humless/buzzless with these being barely audible in a very quiet environment. Connecting a signal cable to the inputs makes the hum return, however.
I would like to ask how unbalanced inputs handle this issue of ground currents? I tried simulations by lifting both inverting and non-inverting input above ground using a 10R. This resistance reduces unwanted signal ground currents but it does not completely cancel it.
I would like to avoid having to use XLR inputs. These require more complexity at the input and the volume control circuitry has to be changed to occomodate them.
I would like to ask how unbalanced inputs handle this issue of ground currents? I tried simulations by lifting both inverting and non-inverting input above ground using a 10R. This resistance reduces unwanted signal ground currents but it does not completely cancel it.
I would like to avoid having to use XLR inputs. These require more complexity at the input and the volume control circuitry has to be changed to occomodate them.