Hi at All,
as I promised in another thread one year ago, i will present my work on a new type of phase splitter, or better to say, a new entire class of phase splitters with a common architecture as in the below attachment.
All circuits are found on differential amplifiers sub-unity.
Due the more type , then I publish the most interesting variant.
All those circuits, i think, make a good job for unbalance to balanced signal converters and they are the natural inverter circuit for simmetrical topologies as Circlotron amplifiers and Brigde Son of Zen type.
I hope this work can help People to build better amplifiers.
All comments are welcome. 😎
Francesco Corrado
as I promised in another thread one year ago, i will present my work on a new type of phase splitter, or better to say, a new entire class of phase splitters with a common architecture as in the below attachment.
All circuits are found on differential amplifiers sub-unity.
Due the more type , then I publish the most interesting variant.
All those circuits, i think, make a good job for unbalance to balanced signal converters and they are the natural inverter circuit for simmetrical topologies as Circlotron amplifiers and Brigde Son of Zen type.
I hope this work can help People to build better amplifiers.
All comments are welcome. 😎
Francesco Corrado
Attachments
Today I updated the page of my website describing the operation of the circuit and inserting additional schematics. 😎
New Phase Splitter/Inverter AMPLIMOS one stage amplifiers
New Phase Splitter/Inverter AMPLIMOS one stage amplifiers
Yes, It could be useful as a balanced line driver, but not only.
As phase splitter at first sight, the circuits may sound complicated, because it uses an excess of transistors (6 or 8), but in reality if you judge it for what it actually does, at the end it is not unnecessarily complicated.
In fact it not only divides and routes the input signal but it also amplifies this last with much voltage gain.
Look to the scheme: the same circuit with the addition of a volume potentiometer could be a great line preamplifier with balanced output. As shown in annex scheme Circlotron, this circuit,inserted into an amp, it simultaneously makes function that would otherwise require three separate stages:
1) Input stage
2) Phase splitter stage
3) VAS stage.
With this phase inverter, to complete an amplifier would be need only a pair of power transistors in EF stage and nothing else.
In the future I'm planning to make up kits of high quality using matched pairs already available (That 300, 320). 😎
As phase splitter at first sight, the circuits may sound complicated, because it uses an excess of transistors (6 or 8), but in reality if you judge it for what it actually does, at the end it is not unnecessarily complicated.
In fact it not only divides and routes the input signal but it also amplifies this last with much voltage gain.
Look to the scheme: the same circuit with the addition of a volume potentiometer could be a great line preamplifier with balanced output. As shown in annex scheme Circlotron, this circuit,inserted into an amp, it simultaneously makes function that would otherwise require three separate stages:
1) Input stage
2) Phase splitter stage
3) VAS stage.
With this phase inverter, to complete an amplifier would be need only a pair of power transistors in EF stage and nothing else.
In the future I'm planning to make up kits of high quality using matched pairs already available (That 300, 320). 😎
Hi Guys
There is nothing new here.
Being able to use both outputs of a diff amp helps to keep the diff balanced. In this age of super-power amps that are internally wired as bridged or balanced-path amplifiers, a true balanced amp designed as dependent signal paths is easily achieved with the above circuits. Built many in the past and they work fine.
Have fun
Kevin O'Connor
There is nothing new here.
Being able to use both outputs of a diff amp helps to keep the diff balanced. In this age of super-power amps that are internally wired as bridged or balanced-path amplifiers, a true balanced amp designed as dependent signal paths is easily achieved with the above circuits. Built many in the past and they work fine.
Have fun
Kevin O'Connor
Nothing new here?
there is nothing new if you NOT have eyes to see.
All other phase splitters that I know do not have at the same time:
1) output signals that are exactly the copy in the mirror of themselves;
2) the same output impedance;
3) the same frequency band;
4) two output signals that are exactly simultaneous;
5) input and output DC level to zero Volt respect the ground;
6) Strong voltage gain.
7) high CMRR and PSRR
......................😎
Hi Guys
mos57, guess I've just seen a lot more circuits than you?
Simpler implementations exhibit the characteristics you strive for. The complexity of the bias control in your other thread (if not yours, then remarkably similar) is not really necessary for this type of balanced output splitter, although it may possibly allow mismatched components. Simpler often requires more care in parts selection. How sensitive is your circuit to mismatches of devices?
I've used this type of circuit in balanced amp designs, where the two paths are interdependent rather than being two separate amplifiers one for each path. The dependent path approach offers certain possibilities for CMR that the independents don't.
The essence of the balanced outputs is not new, and that is what I was commenting on. The "four" outputs as you describe them is not new either, since they end up being joined to drive two output stages.
The complex biasing scheme is one of many such schemes that have been added in an attempt to match the circuit gains without having to match the devices.
Have fun
Kevin O'Connor
mos57, guess I've just seen a lot more circuits than you?
Simpler implementations exhibit the characteristics you strive for. The complexity of the bias control in your other thread (if not yours, then remarkably similar) is not really necessary for this type of balanced output splitter, although it may possibly allow mismatched components. Simpler often requires more care in parts selection. How sensitive is your circuit to mismatches of devices?
I've used this type of circuit in balanced amp designs, where the two paths are interdependent rather than being two separate amplifiers one for each path. The dependent path approach offers certain possibilities for CMR that the independents don't.
The essence of the balanced outputs is not new, and that is what I was commenting on. The "four" outputs as you describe them is not new either, since they end up being joined to drive two output stages.
The complex biasing scheme is one of many such schemes that have been added in an attempt to match the circuit gains without having to match the devices.
Have fun
Kevin O'Connor
Hi Kevin, yes it's possible that i not kown all circuits as you.
Okay, play your game: i am curious to view one of the circuits you are referring.
Also in my schemes there is an interdependant path. This happens in the 2° stage: i call it signals (a) and (b). Trimmer are placed to correct the mismatich of transistors, elsewhere simple equal resistors can make a good job.
Francesco
Hi Guys
As you know, there are many ways to improve the performance of the comp-diff stage driving dual push-pull VASs. The simplest is to use matched devices that have good linearity. That involves extra labour that some do not wish to expend, and possibly the "rejection" of a good deal of inventory. Another method is to "enhance" the circuit with CSs, CMs, EF-VAS, etc.
Although it is very important to linearise each stage as much as possible prior to adding feedback, the complete amplifier must be considered. Often, an imperfection here can balance an imperfection there to good effect.
Have fun
Kevin O'Connor
As you know, there are many ways to improve the performance of the comp-diff stage driving dual push-pull VASs. The simplest is to use matched devices that have good linearity. That involves extra labour that some do not wish to expend, and possibly the "rejection" of a good deal of inventory. Another method is to "enhance" the circuit with CSs, CMs, EF-VAS, etc.
Although it is very important to linearise each stage as much as possible prior to adding feedback, the complete amplifier must be considered. Often, an imperfection here can balance an imperfection there to good effect.
Have fun
Kevin O'Connor
You should post your better phase splitter that has :
a) the best overall performance
b) same performance but with fewer components.
Only in this way can prove that my work was useless.
Everything else is just "empty words". 😀
Best regards.
a) the best overall performance
b) same performance but with fewer components.
Only in this way can prove that my work was useless.
Everything else is just "empty words". 😀
Best regards.
Hi Guys
Francesco, apologies to you. When I looked at this thread initially, I did not expand the thumbnails, so thought you had complimentary diff amps at the input - which is entirely conventional. That is what I've used for low-THD balanced designs. I use common-mode current sources and loads.
Your upside down rendering of half of the diff amp threw me off. What you have is four transistors sharing the current of one current source. I don't believe that the trimpots alone will assure long-term sharing as you hope. In fact, I really don't see how any expected sharing will happen that can be useful unless the trimpots are very high values.
The folded cascode connection to the VAS is one that I never liked, even though Curl and pass have been using it since the 1970s.
Good luck with the development.
Have fun
Kevin O'Connor
Francesco, apologies to you. When I looked at this thread initially, I did not expand the thumbnails, so thought you had complimentary diff amps at the input - which is entirely conventional. That is what I've used for low-THD balanced designs. I use common-mode current sources and loads.
Your upside down rendering of half of the diff amp threw me off. What you have is four transistors sharing the current of one current source. I don't believe that the trimpots alone will assure long-term sharing as you hope. In fact, I really don't see how any expected sharing will happen that can be useful unless the trimpots are very high values.
The folded cascode connection to the VAS is one that I never liked, even though Curl and pass have been using it since the 1970s.
Good luck with the development.
Have fun
Kevin O'Connor
- Status
- Not open for further replies.