A good starting point!
Not at all bad.
I would think it is for Class A operation ???.
Might need some cap parallell to feedback resistor
to cut off very high freq.
And maybe some Zobel filter at output.
The LEDS in the current sources are probably RED.
As they have 1.6V, we can begin to calculate
what currents are running in different Stages.
34V/1k= 30 mA. 5-10 mA is better for the LEDs
As I see it.
The final evaluation and some filter caps in the right places.
A final test and trim.
But as I said: Not bad
But Simple and Good!
Not at all bad.
I would think it is for Class A operation ???.
Might need some cap parallell to feedback resistor
to cut off very high freq.
And maybe some Zobel filter at output.
The LEDS in the current sources are probably RED.
As they have 1.6V, we can begin to calculate
what currents are running in different Stages.
34V/1k= 30 mA. 5-10 mA is better for the LEDs
As I see it.
The final evaluation and some filter caps in the right places.
A final test and trim.
But as I said: Not bad
But Simple and Good!
Quick comment
This design suggests to me that someone knows what they are doing.
There is a mistake: the drain and source are swapped on the VAS gain P-FET.
Is this your own design Lozano?
This design suggests to me that someone knows what they are doing.
There is a mistake: the drain and source are swapped on the VAS gain P-FET.
Is this your own design Lozano?
Because this design uses conventionnal mosfets (not lateral) it is necessary to add some temperature coefficient correction for the finals, otherwise thermal instability and distortion will occur under some circumstances (loud music...)
Regards, Pierre Lacombe.
Regards, Pierre Lacombe.
Lacombe got a point
That means
that when transistors get hotter they will reduce the average current
that runs through them.
and if current is less they will cool off.
Other mosfets and bipolars have positive temp koefficient.
So if they get warm, they conduct more current, getting even warmer.
Called "termal runout".
And you can see that this could end in disaster.
-In this case you have to sense the heat coming
and have something to reduce the currents.
Y
If you use "lateral" mos, they have negative coeffi.
They are very fine transistors for audio.
2SJ218 and 2SKxxx.
The price is higher, but the price is very low
compared to Trafo, big Caps and Heatzink/Enclosure.
In that way, we can say they are A REAL BARGAIN (very low price)
f.bcn - your homepage
I see you have a webspace.
Why not have a MENU-page, as the startpage?
You can copy/download some template (html).
You can use "View source" in Browser's "View"-menu.
Then save it as .html
Here is a link with some FREE template samples:
http://members.fortunecity.se/mattafort/main.html
Some MOSFETs have negative tempkoefficient.P.Lacombe said:
That means
that when transistors get hotter they will reduce the average current
that runs through them.
and if current is less they will cool off.
Other mosfets and bipolars have positive temp koefficient.
So if they get warm, they conduct more current, getting even warmer.
Called "termal runout".
And you can see that this could end in disaster.
-In this case you have to sense the heat coming
and have something to reduce the currents.
Y
If you use "lateral" mos, they have negative coeffi.
They are very fine transistors for audio.
2SJ218 and 2SKxxx.
The price is higher, but the price is very low
compared to Trafo, big Caps and Heatzink/Enclosure.
In that way, we can say they are A REAL BARGAIN (very low price)
f.bcn - your homepage
I see you have a webspace.
Why not have a MENU-page, as the startpage?
You can copy/download some template (html).
You can use "View source" in Browser's "View"-menu.
Then save it as .html
Here is a link with some FREE template samples:
http://members.fortunecity.se/mattafort/main.html
If you have the means to model the performance, check that the
currents in each leg of your diff pair is really the same. You may
find that if you remove the resistor from the drain of the JFET that
you aren't taking the output from that you will achieve a better
match of the two currents.
Erik
currents in each leg of your diff pair is really the same. You may
find that if you remove the resistor from the drain of the JFET that
you aren't taking the output from that you will achieve a better
match of the two currents.
Erik
Re: A good starting point!
Yes there is a mistek in the schematic.
the R value for leds are 3.3 K
halojoy said:
Yes there is a mistek in the schematic.
the R value for leds are 3.3 K
Not sure what you were aiming for, but ...
Looks like about 900uA quiescent per 2SK170.
Seems a little low to me, but it depends on
what you want.
Did you really want a source follower buffer
after the diff amp to drive the output devices?
If so, you probably should
think about reworking the bias points and signal
swing range. Also, your overall negative feedback
is going back to the wrong side of the diff amp.
Does this circuit have enough open loop gain to
even warrant global negative feedback? Maybe you
really wanted common source here, not source follower?
I think you will definately need a bit more sophistication
in the bias generator using the IR output devices. Use
the lateral devices (Hitachi or Toshiba) if you want to
maintain this kind of simplicity. What bias current in the output stages are you looking for, anyway?
Where's the frequency compensation? Are you relying on the
natural semiconductor capacitances to cover this?
Needs a little more work 😉
mlloyd1
Looks like about 900uA quiescent per 2SK170.
Seems a little low to me, but it depends on
what you want.
Did you really want a source follower buffer
after the diff amp to drive the output devices?
If so, you probably should
think about reworking the bias points and signal
swing range. Also, your overall negative feedback
is going back to the wrong side of the diff amp.
Does this circuit have enough open loop gain to
even warrant global negative feedback? Maybe you
really wanted common source here, not source follower?
I think you will definately need a bit more sophistication
in the bias generator using the IR output devices. Use
the lateral devices (Hitachi or Toshiba) if you want to
maintain this kind of simplicity. What bias current in the output stages are you looking for, anyway?
Where's the frequency compensation? Are you relying on the
natural semiconductor capacitances to cover this?
Needs a little more work 😉
mlloyd1
f.bcn said:
SK170 and lateral MOSFET schematics
Here we can find something to study:
Lateral mosfet Schematics
😉 Sorry f.bcn, it is in french 😉
One of the designs:
Here we can find something to study:
Lateral mosfet Schematics
😉 Sorry f.bcn, it is in french 😉
One of the designs:
An externally hosted image should be here but it was not working when we last tested it.
33v is a little bit too high for the 2sk170, the best working voltage for the 2sk170 is about 10-15v, if u want to find a jfet can work at this voltage, 2sk373 is a good choice....
This design seems to be obsolete, now.
Use Mosquito, with cascoded inputs, which has been built many times in Europe. It is essential to use correctly matched transistors in order to obtain specified results : this is the counterpart to pay for the simplicity of the design.
The 10 kohm resistor (between the two 68 ohms) can be susbtantially increased (100 kohm or more, try it) to minimize offset erratic variations, and very low frequency distortion.
Because of the DC response of the amplifier, one must insert A CAPACITOR at the input, if the source is not totally free of offset. 4.7 uF film capacitor, preferabily polyethylene or oil-paper, is adequate.
Resistors (470 ohms) can be inserted in each gate of the input fets, as close as possible of the transistor case, in order to stop HF oscillations, dependig on the wiring.
This amplifier gives really very good results, distortion is quite inaudible (-100 dB ?) at usual listening levels. Bandwidth is in excess of 60 kHz at full power.
Regards, Pierre Lacombe.
Use Mosquito, with cascoded inputs, which has been built many times in Europe. It is essential to use correctly matched transistors in order to obtain specified results : this is the counterpart to pay for the simplicity of the design.
The 10 kohm resistor (between the two 68 ohms) can be susbtantially increased (100 kohm or more, try it) to minimize offset erratic variations, and very low frequency distortion.
Because of the DC response of the amplifier, one must insert A CAPACITOR at the input, if the source is not totally free of offset. 4.7 uF film capacitor, preferabily polyethylene or oil-paper, is adequate.
Resistors (470 ohms) can be inserted in each gate of the input fets, as close as possible of the transistor case, in order to stop HF oscillations, dependig on the wiring.
This amplifier gives really very good results, distortion is quite inaudible (-100 dB ?) at usual listening levels. Bandwidth is in excess of 60 kHz at full power.
Regards, Pierre Lacombe.
- Status
- Not open for further replies.