mosfetaudio-didik.comHome
JFET – MOSFET Amplifier
Welcome to “mosfet lover” communityIr. F. Didik Wiryawan Adhi
In the era of ’70 until the end of ’80, the audio amplifier manufacturers competed with each other offerring superior technical specifications in terms of the wide frequency response, total harmonic distortion and intermodulation distortion that under 0.01% (ranging 0.01 to 0001%) and many other things that could be regarded as a ‘selling point’ for each of each brand.
Until now, the technical specifications of the ‘ideal condition’ is still used as a reference.
But among the ‘audiophiles’, especially for the DIYer (Do It Your Self), seemingly the technical specifications of the ‘ideal condition’ will not be used as a single standard, although not to be ignored altogether.
Among this community (audiophiles), the resulting sound to reproduce the original sound is as beautiful as preferred.
In the era of ’70 until the end of ’80, the audio amplifier manufacturers competed with each other offerring superior technical specifications in terms of the wide frequency response, total harmonic distortion and intermodulation distortion that under 0.01% (ranging 0.01 to 0001%) and many other things that could be regarded as a ‘selling point’ for each of each brand.
Until now, the technical specifications of the ‘ideal condition’ is still used as a reference.
But among the ‘audiophiles’, especially for the DIYer (Do It Your Self), seemingly the technical specifications of the ‘ideal condition’ will not be used as a single standard, although not to be ignored altogether.
Among this community (audiophiles), the resulting sound to reproduce the original sound is as beautiful as preferred.
It is often found an audio system has a beautiful voice with all the requirements of audiophile, however surprisingly, its common technical specifications are worse than a system of regular ‘HiFi Compo’.
There are 2 different adherents on topology but not necessarily debated:
Adherent 1 : who need very few components, this may be called ‘minimalist’ or “KISS” (stand for Keep It So Simple) with the goal of natural sounded much less of sound coloration, regardless that the technical specifications are beyond the ideal.
Adherent 2 : those are idealist design topology very obedient in the text book and an ideal outcome measurement, which resulted in the last chain becomes very complicated and requires many electronic components.
And if the author may add, one more role models are:
No minimalist but not too ‘text book minded’, but having a main purpose is : to design amplifier having nice and beautiful voice but with the technical requirements which was not too bad.
No minimalist but not too ‘text book minded’, but having a main purpose is : to design amplifier having nice and beautiful voice but with the technical requirements which was not too bad.
AMPLFIER with JFET input – MOSFET output.
Based on the schematic of a famous Swiss made amplifier (Goldmund Mimesis 3), I did some improvement and little modification on each stage.
Based on the schematic of a famous Swiss made amplifier (Goldmund Mimesis 3), I did some improvement and little modification on each stage.
After long listening, testing and did some changes (tweaking) as well as an understanding of the sound character (a little research and trial), then as the end result, I put it in the design of ‘clone’ which now named :
Mosfet Power Amplifier MP50Di.
The Power Amplifier MP50Di, using JFET (2SK170) in the front end and HITACHI Lateral Mosfet (2SK1058/2SJ162) on the amplifier tail end.
The Power Amplifier MP50Di, using JFET (2SK170) in the front end and HITACHI Lateral Mosfet (2SK1058/2SJ162) on the amplifier tail end.
Technically, the comparison of performance design using BJT and FET / MOSFET will not be much different, but the real difference will be felt at the time of listening.
There are nuances unique sound that can not be generated by BJT amplifier.
Design Topology DRAFT.
Before entering into a complete range of the MP50Di, it helps us look at the basic topology of this series starts from the most simple.
In addition, it should be noted that this project (especially MP50Di), is not recommended for beginners. Small mistakes made by one who has not had enough experiencel and do not understand basic circuit work, will result in useless waste of funds.
The most basic circuit topology for this design can be seen in Figure-1.
It is a simple 3 stages amplifier:
Before entering into a complete range of the MP50Di, it helps us look at the basic topology of this series starts from the most simple.
In addition, it should be noted that this project (especially MP50Di), is not recommended for beginners. Small mistakes made by one who has not had enough experiencel and do not understand basic circuit work, will result in useless waste of funds.
The most basic circuit topology for this design can be seen in Figure-1.
It is a simple 3 stages amplifier:
- input, differential amplifier (or LTP = long-tailed pair),
- VAS (voltage amplifier stage) and
- the amplifier end (with mosfet 2SK1058/2SJ162).
Figure-1
The simplicity of the basic design is easily seen on a MOSFET amplifier circuit without including the driver stage.
It is almost impossible for BJT end with this simple design can produce output power of 50W even though half of it.
The output of the differential pair Q1 on Q1 and Q2 (2SC2240), is fed into Q3 (2SA968) which is a VAS with ‘bootstrap’ of R6 and R7.
Q3 is directly driving the final amplifier MOSFETs (K1058 & J162), which can produce up to 100W RMS power output (theoritically), which can not be applied at the BJT end amplifier (because the BJT is ‘current operated devices’ while the Power MOSFET is ‘voltage operated devices’ which does not require large currents to drive it).
It is almost impossible for BJT end with this simple design can produce output power of 50W even though half of it.
The output of the differential pair Q1 on Q1 and Q2 (2SC2240), is fed into Q3 (2SA968) which is a VAS with ‘bootstrap’ of R6 and R7.
Q3 is directly driving the final amplifier MOSFETs (K1058 & J162), which can produce up to 100W RMS power output (theoritically), which can not be applied at the BJT end amplifier (because the BJT is ‘current operated devices’ while the Power MOSFET is ‘voltage operated devices’ which does not require large currents to drive it).
Note: Q3 here using 2SA968, a medium power transistor, even though with a small transistor such as 2SA970 (or MPSA92), it can work well without any difference in sound quality. The use of 2SA968 is juts for job security of the device not too close to allowable heat dissipation and easy installation of the heatsink if needed.
Quality Improvement of Phase-1
Improved quality of phase-1 is done by adding a constant current source formed by the Q3A and QZ (2SC2240) replaced R4 and R5 and ‘current mirrors’ which is formed by Q3B and Q3C (2SA970) replaced R3.
These improvements result in significant sound quality improvement.
Improved quality of phase-1 is done by adding a constant current source formed by the Q3A and QZ (2SC2240) replaced R4 and R5 and ‘current mirrors’ which is formed by Q3B and Q3C (2SA970) replaced R3.
These improvements result in significant sound quality improvement.
The simplest thing that appears in the measurement with the DMM (digital multi meter) is the DC offset of 35mV dropped dramatically to about + / – 3.5mV.
The schematic of this improved quality of phase-1 discribed in Figure-2
Figure-2
Figure-2
This circuit is very similar with Project 101 of Rod Elliot of ESP, Australia, the famous simple mosfet amplifier. (sound.westhost.com/project101.htm)
Theoretically, this quality improvement (Figure-2) were very influential in improving linearity (also increase in ‘open loop gain’), it was proven in comparative listening test.
Ears fatigue which arise at the time ‘listening’ on previous series, had been greatly reduced here. Music reproduction, sounded more detailed than ever before. Sound staging and clarity were improved compared to previously that a little noisy because the sound of several musical instruments more likely fight over each other which caused image of ‘position of the music players’ were not definitively clear.
Quality Improvement Phase-2
Although Figure-2 has been producing a quite sufficient sound and can be said to be ‘superior’ than class ‘HiFi Compo’ (even a well-known branded), however driving final amplifier MOSFETs directly from a simple VAS (Voltage Amplifier Stage), has many weaknesses that result in disruption linearity at power levels above 20W especially in long enough operation.
This is because the VAS current is very small, less able to do a ‘charge’ and ‘discharge’ to the final amplifier MOSFETs quickly.
Amplifier topology without a driver can work very well for the application of class A with power under 20W (example: Musical Fidelity MF A-2) but will be felt weakness on the operation of large power and not a class A (class AB).
Amplifier topology without a driver can work very well for the application of class A with power under 20W (example: Musical Fidelity MF A-2) but will be felt weakness on the operation of large power and not a class A (class AB).
This weakness is one reason of ‘ears fatigue’ as mentioned above, nevertheless this weakness is only detectable by an experienced audiophile, it is not clearly detectable on measurement equipment (oscilosocope).
Improved quality of phase 2 was done by adding one stage of the driver (Q6 and Q7), and VAS quality improvement with a simple change of a series of differential VAS which was formed by Q4 and Q5 as set out in Figure-3.Figure-3The driver mode is common emiter and the final is common-source which both have voltage gain.
With this configuration, the amplifier would give more powerfull and tighter bass (I am a bass lover !)
To be honest, this configuration (tail end topology) was inspired by design of David NJ White, “High Quality 100W MOSFET Power Amplifier” (unfortunately, the link : www. wnaudio.com is no longer exist).
The change of simple VAS into differential VAS should also be supported by changed in configuration mode of differential input circuit (input LTP), by changing it from unbalance mode output (single ended) to balance mode output.
Changed in single VAS transistors into a differential VAS (VAS LTP) that loaded by current mirror Q5a and Q4a, providing a significant linearity improvement compared to the composition of the bootstrap R6 and R7 in Figure-1 and Figure-2.
Changed in single VAS transistors into a differential VAS (VAS LTP) that loaded by current mirror Q5a and Q4a, providing a significant linearity improvement compared to the composition of the bootstrap R6 and R7 in Figure-1 and Figure-2.
With the high charging speed and discharge, it can achieve good linearity especially when dealing with fast music with dominant high notes.
In addition to change of VAS, the topology at the tail end also be changed which is no longer power follower (no voltage gain) but became a voltage gain circuit.
(Tobe continued) @#