I have set this special time aside for making a fool of myself 
I simply can't figure if there woulod be a benefit of using the cascoded jfet approach for the X-BOSOZ ??
It should be fairly easy to adapt existing ones to accepting an extra board with the few extra parts, but I would like to understand why or why not to do that.
Cheers
Magura
I simply can't figure if there woulod be a benefit of using the cascoded jfet approach for the X-BOSOZ ??
It should be fairly easy to adapt existing ones to accepting an extra board with the few extra parts, but I would like to understand why or why not to do that.
Cheers
Magura
Is it necessary to cascode the J-fets in a preamp?? I mean, if you adjust the railvoltages to an appropiate level for the Lovoltech J-fets, and maybe a few other component adjustments (I wouldnt know ) Wouldnt that be pretty good for a BosoZ variant? I can understand the cascoding thing for a poweramp, but for a preamp?? I asked something here, so you can keep track:
http://www.diyaudio.com/forums/showthread.php?postid=819111#post819111
Metalman suggested the railvoltage adjustment.
I guess I put this time aside for making a fool of myself, too
Steen
) Wouldnt that be pretty good for a BosoZ variant? I can understand the cascoding thing for a poweramp, but for a preamp?? I asked something here, so you can keep track:http://www.diyaudio.com/forums/showthread.php?postid=819111#post819111
Metalman suggested the railvoltage adjustment.
I guess I put this time aside for making a fool of myself, too
Steen
i think Grey said he had some thoughts on this, but i think we've
used up any free time he may have had to work on projects with
the group buy!
it seems like a lot of us were thinking the same thing at the same time...
BJFETZLS
BXJFETZLS
BXCCSJFETZLS
LOL!
...maybe once the parts get out there and things settle down a bit
we'll see something really cool in the Zen Line Stage bent.
used up any free time he may have had to work on projects with
the group buy!
it seems like a lot of us were thinking the same thing at the same time...
BJFETZLS
BXJFETZLS
BXCCSJFETZLS
LOL!
...maybe once the parts get out there and things settle down a bit
we'll see something really cool in the Zen Line Stage bent.
Hi,
funny how suddenly everyone starts to think about J-Fets
I also had some thoughts about modifying the cs-ccs-x-bosoz:
please correct me if I´m wrong
What you would need is a jfet wich can take around 80mA of current and some 10V (cascoded). The rest should be simple........
With the smaller jfets the output impedance of a bosoz type amp would be very high because of the bigger drain resistors needed to get the same voltage swing with less current.
I´m not shure if the X-feedback can bring it down enough to be of practical use.
When using the lovoltechs you will have to look at the input capacitance so cascoding would still be a nice thing.
William
funny how suddenly everyone starts to think about J-Fets
I also had some thoughts about modifying the cs-ccs-x-bosoz:
please correct me if I´m wrong
What you would need is a jfet wich can take around 80mA of current and some 10V (cascoded). The rest should be simple........
With the smaller jfets the output impedance of a bosoz type amp would be very high because of the bigger drain resistors needed to get the same voltage swing with less current.
I´m not shure if the X-feedback can bring it down enough to be of practical use.
When using the lovoltechs you will have to look at the input capacitance so cascoding would still be a nice thing.
William
I don't know if the Lovoltech's are the best JFET device for the BZLS. I think the 2SK389 would be the better JFET. Cascoded it should be able to handle the current level, and the cascode should also help with achieving a decent output impedance too. I've been mulling over this idea for about a year, but haven't got around to really looking properly at a design.
On the other hand, if the positive rail voltage is lowered to say 20V, the Lovoltech coud be used without needing the cascode.
Now what else can we do to make that acronym even longer ...
On the other hand, if the positive rail voltage is lowered to say 20V, the Lovoltech coud be used without needing the cascode.
Now what else can we do to make that acronym even longer ...
Not a whole lot I guess
I agree with you on the Toshiba J-fet for low level signals! Maybe it would'nt be absolutely necessary to use those Lovoltech's everywhere! They are pretty fancy right now, but maybe not the best for a preamp!!The Japanese J-fets like 2sk170 and dual sk389's has been long regarded as the best in handling low level's. But I guess this is not Nelson Pass country then? Except for RIAA's and the Aleph-J input circuit we didnt see a lot of those japanese J-fets in NP's amp's I think
Still, it would be kind of cool with a BosoZ like J-fet preamp I will built one if I get the opportunity Steen
BTW the J-fet used in the Aleph-J is a 2sj109 device. One of those small dual fet devices in a single house! Pretty cool
Hi,
with the 2sk170 is the the max Idss is 20mA (v-spec). To bias it properly you would need something like 15mA or lower and max. 30-32V of positive rail. The drain resistors would be around 1k for 16V at the ouputs. The max. input voltage swing is then 0,2 - 0,3 volts maximum so you would need a voltage divider or pot at the input.
Just read the Borbely Audio articles (new frontiers part 2). There a BOSOZ with 2sk389 is featured along with the ad-and disadvantages.
william
with the 2sk170 is the the max Idss is 20mA (v-spec). To bias it properly you would need something like 15mA or lower and max. 30-32V of positive rail. The drain resistors would be around 1k for 16V at the ouputs. The max. input voltage swing is then 0,2 - 0,3 volts maximum so you would need a voltage divider or pot at the input.
Just read the Borbely Audio articles (new frontiers part 2). There a BOSOZ with 2sk389 is featured along with the ad-and disadvantages.
william
Magura said:
The bad part is that I still don't understand why ?
I think that the main purpose of the cascode is to provide the common source amp with unity voltage gain. By doing so, we will minimize the miller's input capacitance. Meanwhile, we have the common gate amp above for the necessary voltage gain.
For more info, there is Pass article.
Regards
jh
There are more elegant ways to bias the JFET.
As someone observed earlier in the thread, my clock has run out of hours in a day. (My editor is not pleased with me...running woefully behind on the writing front.) I promise to get back to my bench as soon as possible.
And why not use a power MOSFET in a preamp? Isn't the IRF610 considered a power device? (Hint: Go ahead and run it hard. And, yes, the IRF610 is a pretty good choice for a cascode device as long as you don't exceed the ratings.)
Wasn't the BOSOZ a differential circuit? If so, the preceding schematic is only half a circuit. If not, I plead lack of sleep. If it was a project by Borbely, I plead temporary insanity (just kidding--I'm permanently insane, just ask anyone who knows me).
Grey
As someone observed earlier in the thread, my clock has run out of hours in a day. (My editor is not pleased with me...running woefully behind on the writing front.) I promise to get back to my bench as soon as possible.
And why not use a power MOSFET in a preamp? Isn't the IRF610 considered a power device? (Hint: Go ahead and run it hard. And, yes, the IRF610 is a pretty good choice for a cascode device as long as you don't exceed the ratings.)
Wasn't the BOSOZ a differential circuit? If so, the preceding schematic is only half a circuit. If not, I plead lack of sleep. If it was a project by Borbely, I plead temporary insanity (just kidding--I'm permanently insane, just ask anyone who knows me).
Grey
Yes the scematic is only half a circuit.
I believe this would work, but what do you mean with "more elegant ways to bias the JFET" ?
This could easily be turned into an upgrade board for existing BOSOZ's, as the board would just need a single connection besides the 3 of the original mosfet.
Magura
I believe this would work, but what do you mean with "more elegant ways to bias the JFET" ?
This could easily be turned into an upgrade board for existing BOSOZ's, as the board would just need a single connection besides the 3 of the original mosfet.
Magura
Magura,
"more elegant" would mean a voltage source between the source of the input and the gate of the cascode fets.
This can be done with a resistor connected to the V+ and a 5-10V zener connected to the resistor and the sources.
You could also bias the cascode fet with a voltage divider (two resistors) between plus and ground and set the voltage at 5-10 volts.
The capacitor can be left out first and only put in when oscillation occurs.
Using a fet as cascode device is not possible with these voltages. If you really want this (why?) the voltage should be lowered to 30-40 volts.
When using a current source you could also lower the negative supply voltage to -15V.
I´m not shure the high voltage (positive) has any advantage when building the amp like this. In the original BOSOZ the higher the chosen voltage the higher the bias AND the more the resistors acted as "real" current sources wich lowered distortion.
William
"more elegant" would mean a voltage source between the source of the input and the gate of the cascode fets.
This can be done with a resistor connected to the V+ and a 5-10V zener connected to the resistor and the sources.
You could also bias the cascode fet with a voltage divider (two resistors) between plus and ground and set the voltage at 5-10 volts.
The capacitor can be left out first and only put in when oscillation occurs.
Using a fet as cascode device is not possible with these voltages. If you really want this (why?) the voltage should be lowered to 30-40 volts.
When using a current source you could also lower the negative supply voltage to -15V.
I´m not shure the high voltage (positive) has any advantage when building the amp like this. In the original BOSOZ the higher the chosen voltage the higher the bias AND the more the resistors acted as "real" current sources wich lowered distortion.
William
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.