Reply
 
Thread Tools Display Modes
  #1   Report Post  
 
Posts: n/a
Default Need a good mosfet PN for a current sink

I'm wanting to use a FET as a current sink for a differential longtail.
This will be a balanced input front end using a 6922 with about a -60VDC on
the tail.
All the FETS I have around here seem to have the reverse protection diode
built in.
Can anyone recommend a good part number for this? I've been experimenting
with
the MJE350 bipolar part, but I want to try the FET and see if it will be
more stable
with temp and better PSRR. Any ideas appreciated.

Thanks!
RonL


  #2   Report Post  
Phil Allison
 
Posts: n/a
Default




I'm wanting to use a FET as a current sink for a differential longtail.
This will be a balanced input front end using a 6922 with about a -60VDC
on the tail.



** So the FET connects from -60V to the two cathodes - right ?


All the FETS I have around here seem to have the reverse protection diode
built in.



** Every power MOSFET made is like that - so what ?


Can anyone recommend a good part number for this?



** Almost any MOSFET with a Vds of 100 volts.



I've been experimenting with
the MJE350 bipolar part, but I want to try the FET and see if it will be
more stable with temp and better PSRR.



** A MOSFET will be LESS stable with respect to temperature than a bi-bolar
device.

For bi-polars, the Vbe drops by 2mV per degree C rise for a given Ic -
needing one silicon diode in the bias cct to correct.

For a MOSFET , the Vgs drops by 6 mV per degree C rise for a given I ds -
needing 3 silicon diodes in the bias cct to correct.



......... Phil




  #3   Report Post  
 
Posts: n/a
Default

Thanks Phil
That's good info. I've never compared them. But I noticed ARC used FETs
Would it be a noise issue? Or a PSRR or compliance issue? I guess I need to
build both
and actually measure and compare.

RonL


I'm wanting to use a FET as a current sink for a differential longtail.
This will be a balanced input front end using a 6922 with about a -60VDC
on the tail.



** So the FET connects from -60V to the two cathodes - right ?


All the FETS I have around here seem to have the reverse protection diode
built in.



** Every power MOSFET made is like that - so what ?


Can anyone recommend a good part number for this?



** Almost any MOSFET with a Vds of 100 volts.



I've been experimenting with
the MJE350 bipolar part, but I want to try the FET and see if it will be
more stable with temp and better PSRR.



** A MOSFET will be LESS stable with respect to temperature than a
bi-bolar device.

For bi-polars, the Vbe drops by 2mV per degree C rise for a given Ic -
needing one silicon diode in the bias cct to correct.

For a MOSFET , the Vgs drops by 6 mV per degree C rise for a given I ds -
needing 3 silicon diodes in the bias cct to correct.



........ Phil






  #4   Report Post  
Patrick Turner
 
Posts: n/a
Default



wrote:

I'm wanting to use a FET as a current sink for a differential longtail.
This will be a balanced input front end using a 6922 with about a -60VDC on
the tail.
All the FETS I have around here seem to have the reverse protection diode
built in.
Can anyone recommend a good part number for this? I've been experimenting
with
the MJE350 bipolar part, but I want to try the FET and see if it will be
more stable
with temp and better PSRR. Any ideas appreciated.


You don't need a PNP fet.

NPN fet is what you want, but they are fragile and expensive and no
more effective than am MJE340, an npn bjt, as seen in the
schematic at
http://www.turneraudio.com.au/htmlwe...0ulabinteg.htm

The bjt CCS is seen in the middle of the schematic about 3/4 the way down the
image.

There would be zero benefit in using any other device because the
stability of the current flow value etc is all very immune to temp changes
since the current is set by the emitter resistor negative current feedback
developed by the emitter resistor.

And the simple MJE340 CCS is a better CCS than a pentode,
unless the cathode resistor is indeed a large value, since the actual
typical value of the current source resistance is above 20megohms which is
enough for the source to be called a constant current source since the current
changes such a tiny and totally insignificant amount.

Patrick Turner.



Thanks!
RonL


  #5   Report Post  
Patrick Turner
 
Posts: n/a
Default



wrote:

Thanks Phil
That's good info. I've never compared them. But I noticed ARC used FETs
Would it be a noise issue? Or a PSRR or compliance issue? I guess I need to
build both
and actually measure and compare.


The word "" transistor "" stinks much worse amoung tubophiles than Fet,
and there are latereal and enhancement mosfets, j-fets, and so on.

For some very illogical reason ppl have said that fets of one type or another
are the
only type of SS device to use with vacuum tubes.

With all due respect to ARC who have done a stirling job in perpetuating the
myth of the fet.
ARC have dozens of various types of fets in their horridly complex preamps such
as the SP11 etc.
They are monuments to technical cleverness, but nothing sonically clever is
achieved.
The fets can fail rather sooner than the tubes, and the complexity is a
nightmare to service.


All fets are versions of 'field effect transistors' ie, a transistor that has
its
load working current changed by an applied voltage field rather like a pentode.
ie, the gate is a high impedance control point like a grid.
The mosfet is a metal oxide field effect transistor.

I suggest you spend a month studying the basics of all SS device
parameters, the behaviours, the impedances, and so on, and away from a tube
discussion group
because really not many ppl really understand exactly what happens inside any
simple transistor or 'bipolar junction transistor' as they are also called,
or the j-fet or mosfet.

Bjts require current to be supplied through their input base control
terminal to make a larger current flow in the load and through the collector to
emitter.
The base current is typically 100 times less than the collector current.
The amount of voltage change at the base to emitter junction needed to
change the base current and hence the collector current is much smaller than
that
needed in a j-fet or mosfet.

I suggest you measure a few devices set up as a one device amplifier on a
breadboard,
and see if you can draw a model of the devices you want to use.

BTW, ARC amps are no quieter than any other well designed amp you might
build at home.

I quite like fets though; I just see no point in using
so many in any preamp as ARC.
The only place I like to use a j-fet in a preamp which I would call a
"tube preamp" is at the input of a phono stage or microphone amp where a
single 2SK369, 2SK147, etc in cascode with a triode will
produce noise that is typically 20dB less than what you get with a 12AX7,
thus allowing MC cartridges to be directly connected to an amp
rather than have to be hooked up via a step up tranny.

The j-fet in such a situation can be trusted with a few millivolts
of signal without causing much distortion.

But where an SS device is used as a CCS, the current *is* constant;
no variation is measurable with say a common cathode voltage of say 5vrms.

Most folks get muddled about constant current sources in LTPs because
gee, it a bloomin transistor, and prejudice sets in about possible non
linearities.

If the CCS does have an actual finite value of 20 megohms,
then 5vrms causes a current change of only 0.00025 mA, rather a tiny amount.
And the change is substantially linear.
See
http://www.turneraudio.com.au/htmlwe...0ulabinteg.htm

But I assure you that although bjts without NFB are attrociously
non linear voltage amps, they make fabulous current sources
when used with the local current FB as in the schematic above.

Probably you need to analyse exactly what dynamic current changes occur in the
LTP.
It took me awhile to understand them.
But the only the tubes and the loads have any significant change.
And the tubes can be very mismatched, and as long as the load R are the same,
there are equal outputs.
Radford used a pentode on one side of the LTP and a triode on the other.

Patrick Turner.






RonL

I'm wanting to use a FET as a current sink for a differential longtail.
This will be a balanced input front end using a 6922 with about a -60VDC
on the tail.



** So the FET connects from -60V to the two cathodes - right ?


All the FETS I have around here seem to have the reverse protection diode
built in.



** Every power MOSFET made is like that - so what ?


Can anyone recommend a good part number for this?



** Almost any MOSFET with a Vds of 100 volts.



I've been experimenting with
the MJE350 bipolar part, but I want to try the FET and see if it will be
more stable with temp and better PSRR.



** A MOSFET will be LESS stable with respect to temperature than a
bi-bolar device.

For bi-polars, the Vbe drops by 2mV per degree C rise for a given Ic -
needing one silicon diode in the bias cct to correct.

For a MOSFET , the Vgs drops by 6 mV per degree C rise for a given I ds -
needing 3 silicon diodes in the bias cct to correct.



........ Phil







  #6   Report Post  
Ian Iveson
 
Posts: n/a
Default

I have used STD1HNC60 but I notice they have been discontinued by
RS. I recall low capacitance with decent transconductance were key
considerations.

The smaller the better, probably. Check out those with the lowest
current capacity and voltage limit you can get away with.

Keep in mind that your CCS must be stable, and work well beyond the
bandwidth of the amp. Patrick's presumably static measurements are
not worth much.

cheers, Ian

wrote in message
...
I'm wanting to use a FET as a current sink for a differential
longtail.
This will be a balanced input front end using a 6922 with about
a -60VDC on the tail.
All the FETS I have around here seem to have the reverse
protection diode built in.
Can anyone recommend a good part number for this? I've been
experimenting with
the MJE350 bipolar part, but I want to try the FET and see if it
will be more stable
with temp and better PSRR. Any ideas appreciated.

Thanks!
RonL



  #7   Report Post  
 
Posts: n/a
Default

Thanks Patrick!
I know I can always get my head on straight aftert reading your post.
I have a MJE340 set up on the bench and it is solid as a rock.
I had noticed that ARC used FET's and thought I might be missing something
there.
I still want to try the FET and make some test befor I commit to a PCB.
Thanks Again
RonL



"Ian Iveson" wrote in message
. uk...
I have used STD1HNC60 but I notice they have been discontinued by RS. I
recall low capacitance with decent transconductance were key
considerations.

The smaller the better, probably. Check out those with the lowest current
capacity and voltage limit you can get away with.

Keep in mind that your CCS must be stable, and work well beyond the
bandwidth of the amp. Patrick's presumably static measurements are not
worth much.

cheers, Ian

wrote in message
...
I'm wanting to use a FET as a current sink for a differential longtail.
This will be a balanced input front end using a 6922 with about a -60VDC
on the tail.
All the FETS I have around here seem to have the reverse protection diode
built in.
Can anyone recommend a good part number for this? I've been experimenting
with
the MJE350 bipolar part, but I want to try the FET and see if it will be
more stable
with temp and better PSRR. Any ideas appreciated.

Thanks!
RonL





Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Artists cut out the record biz [email protected] Pro Audio 64 July 9th 04 10:02 PM


All times are GMT +1. The time now is 05:25 PM.

Powered by: vBulletin
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 AudioBanter.com.
The comments are property of their posters.
 

About Us

"It's about Audio and hi-fi"