Current-Source idea

hugobors · 17th February 2006, 11:12 PM

Hello,

Do you think this can work????
If yes it could be a good adjustable current-source.
Thanks

lineup · 17th February 2006, 11:50 PM

it is a
bit complicated

but, yes
it will work ....as I see it, by a quick look

what actually sets the current is the 10k trimpot
so any power supply ripple or changes
will upset current source
( = low Power Supply Rejection , PSRR )

hugobors · 18th February 2006, 12:13 AM

Quote:

what actually sets the current is the 10k trimpot
so any power supply ripple or changes
will upset current source
( = low Power Supply Rejection , PSRR )

I know, that's why in the definitive design I want to replace the 20K resistor by a jfet current source feeding the trim (surely an other value than 10K)

So you think that the design with the 2 bipolars is ok???

Thanks

lineup · 18th February 2006, 12:34 AM

It will work,
but there has to be a better way to do it.

Now what JFET can run 100 mA ??

The only one I know, is that Power JFET at Pass Labs
they all are making a big fuzz about.

Lets see what new sensation NP have in store for next year,
to make another big fuzz about ....

hugobors · 18th February 2006, 12:39 AM

No I don't want to replace the mosfet by a jfet, I want to replace the 20K resistor by a jfet in cs configuration... so my adjustable tension reference will be power supply noise independant.

Eva · 18th February 2006, 12:43 AM

I wonder how you could get 7V in the left side if you adjust for 8.2V in the right side. Note that the two BE junctions add voltage to the reference rather than substracting.

Also, I think that it could be simplified by replacing the 70 ohm resistor by a potentiometer and connecting the base of the NPN transistor to its wiper and its emitter directly to -30V. This will also improve PSRR a lot, since in your circuit the reference adjustment is directly modulated by power supply ripple, while in the circuit that I propose the modulation happens indirectly through Vcc modulating Ic through the 10k resistor, Ic modulating Vbe.

The best simple current source that I have found is the two-transistor one that forces a Vbe drop in the sense resistor though feedback. It improves a lot with the addition of a zener and a resistor to keep the Ic of the sense transistor more constant and avoid Vbe modulation effects with Ic. The current source transistor is inherently protected from these effects.

The two-diodes-and-a-transistor current source has somewhat worse performance, particularly when the current through the diodes is modulated by supply ripple. It also suffers from Vbe modulation with Vce changes in the current source transistor.

lineup · 18th February 2006, 12:46 AM

Quote:

Originally posted by hugobors
No I don't want to replace the mosfet by a jfet, I want to replace the 20K resistor by a jfet in cs configuration... so my adjustable tension reference will be power supply noise independant.

MOSFET you have.
okay, sorry I took it for a JFET

Eva · 18th February 2006, 01:07 AM

Avoid MOSFETs as current sources. These devices have output capacitances one or two orders of magnitude higher than similar bipolar and JFETs, thus converting your current source into a nice voltage source at HF.

hugobors · 18th February 2006, 07:05 AM

Quote:

I wonder how you could get 7V in the left side if you adjust for 8.2V in the right side. Note that the two BE junctions add voltage to the reference rather than substracting.

Sorry for the mistake......

Quote:

Also, I think that it could be simplified by replacing the 70 ohm resistor by a potentiometer and connecting the base of the NPN transistor to its wiper and its emitter directly to -30V.

Yes I know but I need to drive several mos and so couln'd make all adjustable with all bibolar around each one.

Quote:

t improves a lot with the addition of a zener and a resistor to keep the Ic of the sense transistor more constant and avoid Vbe modulation effects with Ic

Can you explain more or post a shematic???

And so what's the best solution for me???

Eva · 18th February 2006, 08:42 AM

Try this:

I have chosen component values for a 100mA typical current adjustment and 30V supply as you show in your schematic.

The bipolar transistor controls the gate voltage to force the voltage drop across the 220 ohm resistor to be equal to the multiple of Vbe adjusted with the potentiometer.

That working principle is precise, but has a main flaw: The Vbe of the control transistor is not constant, it changes according to its collector current. This means that biasing the control transistor directly from a resistor to ground produces poor PSRR because ripple on the PSU rails will affect the preset current value. To reduce that undesirable effect by an order of magnitude, the collector current through the control transistor is kept more or less independent of PSU voltage by employing a zener and an additional resistor.

This is still far from perfect, because the instantaneous Vgs setting for the MOSFET to conduct 100mA depends on its instantaneous Vds, and the Ic of the control transistor is in turn dependent of Vgs, but keeping Vds above 10V minimizes that effect. Trying to compensate for it would require to bias the control transistor with an additional current source (complexity!!).

I have experienced oscillation sometimes in that kind of circuits, so the 47pF capacitor is recommended. Always check with oscilloscope.

17th February 2006, 11:50 PM	#2
lineup diyAudio Member Join Date: Dec 2005 Location: the north	it is a bit complicated but, yes it will work ....as I see it, by a quick look what actually sets the current is the 10k trimpot so any power supply ripple or changes will upset current source ( = low Power Supply Rejection , PSRR ) __________________ lineup

18th February 2006, 12:34 AM	#4
lineup diyAudio Member Join Date: Dec 2005 Location: the north	It will work, but there has to be a better way to do it. Now what JFET can run 100 mA ?? The only one I know, is that Power JFET at Pass Labs they all are making a big fuzz about. Lets see what new sensation NP have in store for next year, to make another big fuzz about .... __________________ lineup

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
point source / omnipole idea... have some questions.	AudioGeek	Multi-Way	2	14th December 2006 10:39 PM
Aleph-type current source, but source follower this time	tschrama	Pass Labs	4	29th July 2005 12:55 PM
Marathon car amp idea current draw	cm961	Car Audio	7	23rd September 2004 05:15 PM
Source follower electret mike -> use with current source?	capslock	Solid State	1	28th February 2003 12:54 PM
R21? and Current Source AC Current Gain	macka	Pass Labs	9	20th May 2002 08:58 PM

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

18th February 2006, 12:39 AM	#5
hugobors diyAudio Member Join Date: Feb 2002 Location: Nantes	No I don't want to replace the mosfet by a jfet, I want to replace the 20K resistor by a jfet in cs configuration... so my adjustable tension reference will be power supply noise independant.

18th February 2006, 12:43 AM	#6
Eva diyAudio Member Join Date: Oct 2003 Location: Near the sea	I wonder how you could get 7V in the left side if you adjust for 8.2V in the right side. Note that the two BE junctions add voltage to the reference rather than substracting. Also, I think that it could be simplified by replacing the 70 ohm resistor by a potentiometer and connecting the base of the NPN transistor to its wiper and its emitter directly to -30V. This will also improve PSRR a lot, since in your circuit the reference adjustment is directly modulated by power supply ripple, while in the circuit that I propose the modulation happens indirectly through Vcc modulating Ic through the 10k resistor, Ic modulating Vbe. The best simple current source that I have found is the two-transistor one that forces a Vbe drop in the sense resistor though feedback. It improves a lot with the addition of a zener and a resistor to keep the Ic of the sense transistor more constant and avoid Vbe modulation effects with Ic. The current source transistor is inherently protected from these effects. The two-diodes-and-a-transistor current source has somewhat worse performance, particularly when the current through the diodes is modulated by supply ripple. It also suffers from Vbe modulation with Vce changes in the current source transistor.

18th February 2006, 01:07 AM	#8
Eva diyAudio Member Join Date: Oct 2003 Location: Near the sea	Avoid MOSFETs as current sources. These devices have output capacitances one or two orders of magnitude higher than similar bipolar and JFETs, thus converting your current source into a nice voltage source at HF.

18th February 2006, 08:42 AM	#10
Eva diyAudio Member Join Date: Oct 2003 Location: Near the sea	Try this: I have chosen component values for a 100mA typical current adjustment and 30V supply as you show in your schematic. The bipolar transistor controls the gate voltage to force the voltage drop across the 220 ohm resistor to be equal to the multiple of Vbe adjusted with the potentiometer. That working principle is precise, but has a main flaw: The Vbe of the control transistor is not constant, it changes according to its collector current. This means that biasing the control transistor directly from a resistor to ground produces poor PSRR because ripple on the PSU rails will affect the preset current value. To reduce that undesirable effect by an order of magnitude, the collector current through the control transistor is kept more or less independent of PSU voltage by employing a zener and an additional resistor. This is still far from perfect, because the instantaneous Vgs setting for the MOSFET to conduct 100mA depends on its instantaneous Vds, and the Ic of the control transistor is in turn dependent of Vgs, but keeping Vds above 10V minimizes that effect. Trying to compensate for it would require to bias the control transistor with an additional current source (complexity!!). I have experienced oscillation sometimes in that kind of circuits, so the 47pF capacitor is recommended. Always check with oscilloscope.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)