Preferred JFET in Constant current source?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
D

Dr.H

Member
Joined 2002
Hello all,

From all the wisdom and experience built up over time, is there a preferred JFET to use as a constant current source for say 3-10mA?

What about if the CCS is cascoded, does it mean the CCS quality (sound, PSRR, etc) is less dependant on the JFET?
 
Last edited:
triode_al

triode_al

Member
Joined 2004
Personally, I like the 2SJ170, this little one will cover the whole range.
I have e.g. used it as a CCS for a phase inverter (6SL7); as well as a tube bias for a cascoded 6SN7/ECC82 CCS.
The 2SK170-CCS in a simple form is a little dependent on voltage (say 3-50 volt), which does show up on the work bench, specifically for the higher currents. That said, this is the converted resistance. But that is true of the majority of simple forms.

As well this FET is temperature sensitive; even the 2SK30 is to some extent (which I prefer for the low currents).

I have made complex CCS's (with small mosfets; led gate voltage) but that does make sense only if enough 'headroom' is available, plus the auxiliary current for the CCS bias. Swinging from 5 to 50 volt will even then show a minor increase of current.
 
A

AndrewT

R.I.P.
Joined 2004
if low volts drop across the CCS is not a priority then avoid the expensive k170.
There are lots of much cheaper and more easily available jFETs. bf862, bf244/5, 2n5457/8/9, 2n5460/1/2,

Read Walt Jung on CCSs. He tells us that min Vdrop must be >=2*Vp. This is important.
A lower Idss jFET will have a lower Vp than a higher Idss from the same family. If you need exceptionally low Vdrop then a 2sk170gr is hard to beat.
 
Last edited:
h_a

h_a

Member
Joined 2006
Ideally you would like to use a high input impedance jfet for a constant current source, so a 2SK246 is prefered over the 2SK170. It is, however, a large pinch-off voltage jfet, so voltage drop is larger.

If one's interested in further reading, Erno Borbely discussed this issue nicely.

Hannes
 
h_a

h_a

Member
Joined 2006
Hi,

jeez, sorry I just had a look at Borbely's Jfet-the new frontier articles and these do not exactly contain what I meant.

I think I (correctly ;) ) remember now it was Walt Jung. Sorry for that, but I'll try to dig out a reference. Don't mind reminding me for that as I'm pretty busy currently.

Hannes
 
SCAIFF001

SCAIFF001

Member
Joined 2010
JFET CCS 6SL7 Phase Inverter

triode_al said:
Personally, I like the 2SJ170, this little one will cover the whole range.
I have e.g. used it as a CCS for a phase inverter (6SL7); as well as a tube bias for a cascoded 6SN7/ECC82 CCS.
The 2SK170-CCS in a simple form is a little dependent on voltage (say 3-50 volt), which does show up on the work bench, specifically for the higher currents. That said, this is the converted resistance. But that is true of the majority of simple forms.

As well this FET is temperature sensitive; even the 2SK30 is to some extent (which I prefer for the low currents).

I have made complex CCS's (with small mosfets; led gate voltage) but that does make sense only if enough 'headroom' is available, plus the auxiliary current for the CCS bias. Swinging from 5 to 50 volt will even then show a minor increase of current.
Click to expand...

Hi,

I will require a phase inverter for a ballanced output for a DAC, could you tell me more about your CCS JFET 6SL7, I have not herd anyone talk about this phase inverter yet. Do you have a schematic avaible or can you talk me through all of the components required ?

Thank You

Ian
 
triode_al

triode_al

Member
Joined 2004
SCAIFF001 said:
Hi,

I will require a phase inverter for a ballanced output for a DAC, could you tell me more about your CCS JFET 6SL7, I have not herd anyone talk about this phase inverter yet. Do you have a schematic avaible or can you talk me through all of the components required ?
Thank You
Ian
Click to expand...

Hi Ian,
My phase inverter started off as 6N7 running at 2,5 mA per side with a common cathode having a 2SK170 and a 20 ohm potentiometer on the source to create a CCS. This went to earth, no negative voltage. This allowed to set the current exactly. The CCS operating point was 4 volt at 200 V V-anode.

I later replaced the 6N7 with the 6SL7 with the two cathodes now connected to the same CCS configuration, the CCS operating point (=Vg) lowered a little bit.

It allowed balanced operation within less than a dB. The inverted side of course has a slight lag (phase difference) that could be corrected, e.g. with a small choke but I did not take the trouble.

Now for a DAC you can of course connect both grids to a side of the DAC (+,-) and you can tap one of the outputs like you want.
In this configuration you will not notice the lag I mentioned (and this will show up if I infer correctly) as a HF roll off, which in fact is beneficent.

You must squeeze the 6SL7 to a smaller current, say 0,5 V at Vg (measured) around 3,5-4V.
You must then have a cathode follower, best to follow up both sides with the same Rk as Ra, to balance the drain of current from the power supply. Specifically with a DAC output this helps in the HF noise area!

The configuration is the same as some PASS preamps with fets that would also allow balanced - unbalanced out configurations.

albert
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom