seems like higher capacitance comes naturally with the higher power rating and transconductance
I dont think there is any way around it
I have source for FQA19N20 they sell minimum of 1000 pieces so no go for me there
If one has the cash to go for it PM.
FQA19N20C is one different kind of beast with more capacitance and more power then there is the FQA19N20 L even more power and again more capacitance.
I use FQP19N20, which is closer match(current wise) for FQA12P20 but has TO220 pack
Having bought 100 each I can spare a few so Pm
They have been running since last August on MY F5.
I quite like the idea to use the 36 and 28 as you mention I am not shure how they perform and my Knolledge is limited so it will have to be a build and see thing the gain of those 2 is quite interesting so try them and report back.
I have seen post about potential buy of the original pair 19 / 12 will need to search for that
For me it is interesting to look at the capacitance and transfer curves of the 12P20 and 19n20C. You could use a transparency and lay them on top of each other and you have jsut about an exact match for same conditions. Now try with 19n20 and you see differences beginning to appear. L version is even worse. Could be wrong, just interesting
You could be right and just posted sometink I did not see.
Tanks that sound realy interesting.
And buzzforb link to Sign Up Sheet is the one I was going to search for
Last edited:
I have some surplus matched FQA19N20C and FQA12P20. What kind of matched (N-N, P-P) sets are you looking for? You can PM me if you are interested in.
Thanks,
Working on a dual purpose board that can do FQA or FQP.
With proper mounting like BKsabath, one should be fine with the smaller package units.
They should promise in one direction, they are still avail via normal distribution.
With proper mounting like BKsabath, one should be fine with the smaller package units.
They should promise in one direction, they are still avail via normal distribution.
That does look suspicious. They should both be the same. This effects V2 as well if you are cascoding the input. My guess is the 10k is right, but it could depend on the rail voltage. I am running 32 volts and need to know for sure.
I sure hope someone smarter than me pipes up.
Rush
R26 should be 10K, assuming 60V supply rails
-60V
________ x R28 = -19.32V the correct voltage reference for the cascode transistor
R26+R28
-60V
________ x R28 = -19.32V the correct voltage reference for the cascode transistor
R26+R28
So for 32 volt rails, R28 should be the same value as R26 (4.75k). That yields about 16 volts for the cascode reference.
Some one say yes, please.
Rush
Last edited:
for 1/3 of rail voltage you will have just above 10V. so still R26=10Kohm
but there is no problem to rise it to 16V. or higher.
but there is no problem to rise it to 16V. or higher.
Last edited:
If you run the jFETs at 70% of Idss and you have 10Vds then each jFET has to dissipate ~70mW. That will be cool to just warm.
For 32V supply rails, I would make R25,R26,R27,R28 10K and leave it at that.
If you want to find the sweet spot, you have the option of putting a resistor in parallel with R27 & R28 to increase the cascode reference voltage, or putting a resistor in parallel with R25 & R26 to decrease the cascode reference voltage.
If you want to find the sweet spot, you have the option of putting a resistor in parallel with R27 & R28 to increase the cascode reference voltage, or putting a resistor in parallel with R25 & R26 to decrease the cascode reference voltage.
These are BLs that I have a bunch of in that range. Papa said they would be too high idss and get hot if you put full voltage on them. So, I am cascoding them and as Andrew pointed out they should be just warm.
Rush
Rush, that's because rail voltage is higher than the jfet breakdown voltage, right?
Perhaps I may have not been specific on my question; actually I'd like to know if replacing blue jfets with violet ones is possible and if it would bring any advantages in terms of gain, noise, sound quality etc.
Many thanks,
Daniel
The J74/K170 are 400mW devices, doesn't mean they're supposed to dissipate the full amount.
Standard practice is 50% Max, read chapter 6 of the Toshiba semiconductor reliability handbook, referenced to in any Toshiba part datasheet.
50% is a bit conservative.
The standard F5 design has J108/K370, 200mW devices.
Vds is a little under 20V, derated Idss is 6mA, makes short of 120mW.
Or ~60% of the 200mW Max Pd.
50% derated J74/K170 handle 200mW, at 10Vds good for 20mA drain current.
High current good : higher Idss, better numbers, e.g. capacitances, lower noise.
High voltage bad : more heat, more gate leakage, etcetera.
Standard practice is 50% Max, read chapter 6 of the Toshiba semiconductor reliability handbook, referenced to in any Toshiba part datasheet.
50% is a bit conservative.
The standard F5 design has J108/K370, 200mW devices.
Vds is a little under 20V, derated Idss is 6mA, makes short of 120mW.
Or ~60% of the 200mW Max Pd.
50% derated J74/K170 handle 200mW, at 10Vds good for 20mA drain current.
High current good : higher Idss, better numbers, e.g. capacitances, lower noise.
High voltage bad : more heat, more gate leakage, etcetera.