300 to 400 Ma seems excessive.
You shouldn't need more than 100 MA
The second gain stage or VAS only has a bootstrap Current Source.
And would require a large capacitor for low frequency performance.
or possibly increase the current.
likewise MJE type transistor relatively poor transistor
to directly drive mosfets
You shouldn't need more than 100 MA
The second gain stage or VAS only has a bootstrap Current Source.
And would require a large capacitor for low frequency performance.
or possibly increase the current.
likewise MJE type transistor relatively poor transistor
to directly drive mosfets
Last edited:
ESP 101 in class A
Thank you White Dragon,
yes I imagined that it was excessive polarization of 300 or 400 MA. but I wanted to try to hear the influence on the sound of a high polarization of the final mosfets.
Question. If I want to make the esp 101 work in class A it would be enough just to raise the bias of the final mosfets, since already the input differential and mje350 work in class a .
Do not consider for the moment the power supply and heatsinks, but let's just look at the electrical problem.
Is mje350 able to keep the mosfets always on?
I tried to raise the bias up to 500 MA. per rail, no oscillation , the only audible effect , poor and attenuated bass compared to a bias of 200MA. that I usually use with regular bass. 150w on R8 gives me 4.3A current and 34.64V . Since the mosfets have a voltage gain that is worth about 1 it means that mje350 has to supply the volts that are output on the speakers. and the four mosfets should supply about over 1 A each. Is 0.5 A of mje350 enough?
Why does increasing the bias decrease the bass?
I think that increasing the bias increases the heat of the mosfets and thus increases rds on due to the negative temperature coefficient, and mje350 is no longer sufficient to drive the hot mosfets.
If so I should change mje350 with a more powerful one.
Any suggestions?
Giubos
Thank you White Dragon,
yes I imagined that it was excessive polarization of 300 or 400 MA. but I wanted to try to hear the influence on the sound of a high polarization of the final mosfets.
Question. If I want to make the esp 101 work in class A it would be enough just to raise the bias of the final mosfets, since already the input differential and mje350 work in class a .
Do not consider for the moment the power supply and heatsinks, but let's just look at the electrical problem.
Is mje350 able to keep the mosfets always on?
I tried to raise the bias up to 500 MA. per rail, no oscillation , the only audible effect , poor and attenuated bass compared to a bias of 200MA. that I usually use with regular bass. 150w on R8 gives me 4.3A current and 34.64V . Since the mosfets have a voltage gain that is worth about 1 it means that mje350 has to supply the volts that are output on the speakers. and the four mosfets should supply about over 1 A each. Is 0.5 A of mje350 enough?
Why does increasing the bias decrease the bass?
I think that increasing the bias increases the heat of the mosfets and thus increases rds on due to the negative temperature coefficient, and mje350 is no longer sufficient to drive the hot mosfets.
If so I should change mje350 with a more powerful one.
Any suggestions?
Giubos
Mosfets require very little current to drive.
So it is hard to not resist the urge to make a simple amplifier which drives
the mosfets directly from the 2nd gain stage or Vas.
Unfortunately it is not a good idea if using more than 2 mosfets.
Since multiple Mosfets will load down the Vas transistor.
Better design is to use traditionally driver transistors for the mosfets
its not about using a more powerful Vas transistor.
The load on the Vas is reduced by using Mosfet Drivers.
MJE350 is very slow, overbias of the output stage will likely have no audible improvement. Faster transistor will improve high frequency distortion.
The only way to not load down the Vas is to not directly drive the mosfets
and use additional driver transistors for the mosfets.
Likewise attempting to use high bias current, and having a bootstrap current source
keep up with such demands, likely wont happen.
I will only assume the bootstrap is probably set to around a typical 7 to 8ma
with whatever resistor values being used, with a typical 100u capacitor.
Its likely if your driving way to many transistors, and using high bias current.
The Vas current would need to be very high, around 10ma and the bootstrap capacitor
even larger around 150u or larger.
100u is usually large enough, but since bias is so high, and vas current needs
to be larger. 100u likely isnt large enough. Bootstrap capacitor is connected
to amplifier output. So it causes phase shift at low frequency. If its not big enough
low frequency performance will suffer. Again since its driving to many mosfets.
not a very suitable design for driving multiple mosfets. Only suitable
for simple low power amplifiers using 2 output devices
All in all, to do what you want to do, would require different amplifier design.
Again to run so many output devices you need driver transistors, and not load down
the Vas, And use a better transistor based current source.
Common T0126 packages used for better linearity are 2SA1381(PNP) and 2SC3503(NPN)
likely hard to find originals.
On Semi Conductor Currently manufactures them under the Suffix
KSA1381 and KSC3503
So it is hard to not resist the urge to make a simple amplifier which drives
the mosfets directly from the 2nd gain stage or Vas.
Unfortunately it is not a good idea if using more than 2 mosfets.
Since multiple Mosfets will load down the Vas transistor.
Better design is to use traditionally driver transistors for the mosfets
its not about using a more powerful Vas transistor.
The load on the Vas is reduced by using Mosfet Drivers.
MJE350 is very slow, overbias of the output stage will likely have no audible improvement. Faster transistor will improve high frequency distortion.
The only way to not load down the Vas is to not directly drive the mosfets
and use additional driver transistors for the mosfets.
Likewise attempting to use high bias current, and having a bootstrap current source
keep up with such demands, likely wont happen.
I will only assume the bootstrap is probably set to around a typical 7 to 8ma
with whatever resistor values being used, with a typical 100u capacitor.
Its likely if your driving way to many transistors, and using high bias current.
The Vas current would need to be very high, around 10ma and the bootstrap capacitor
even larger around 150u or larger.
100u is usually large enough, but since bias is so high, and vas current needs
to be larger. 100u likely isnt large enough. Bootstrap capacitor is connected
to amplifier output. So it causes phase shift at low frequency. If its not big enough
low frequency performance will suffer. Again since its driving to many mosfets.
not a very suitable design for driving multiple mosfets. Only suitable
for simple low power amplifiers using 2 output devices
All in all, to do what you want to do, would require different amplifier design.
Again to run so many output devices you need driver transistors, and not load down
the Vas, And use a better transistor based current source.
Common T0126 packages used for better linearity are 2SA1381(PNP) and 2SC3503(NPN)
likely hard to find originals.
On Semi Conductor Currently manufactures them under the Suffix
KSA1381 and KSC3503
Last edited: