Maker Pro
Maker Pro

constant current source

B

Ban

Jan 1, 1970
0
kell said:
I have a circuit that I would like to use a pnp transistor such as the
2N3905 for a constant current source, similar to the diagram on page 9
of this datasheet

http://rocky.digikey.com/WebLib/Fairchild/Web Data/LM555_NE555_SA555.pdf

It would have to source about 80 or 100 microamps. Will this work at
such a low current?

The trigger current is 0.5uA, so there is enough margin with this value, but
non-linearities will start popping up with less. If you do not need the
200mA output current capability, but are happy with 10mA source or 50mA
sink, you can use the LMC555, a CMOS version, which has only 10pA input
current on that pin. It will consume much less current and work at
frequencies up to 3MHz.
 
M

Mebart

Jan 1, 1970
0
Yes, it will work fine. But, the 555 uses quite a bit of current (3 ma
to 7.5 ma, depending on the supply voltage) just to light the chip.
And, it requires an extra outboard transistor. I would think you could
do better depending on the degree of regulation you need.

Doesn't Maxim make some precision current sources for low power aps???

Regards,

M
 
R

Robert Baer

Jan 1, 1970
0
kell said:
I have a circuit that I would like to use a pnp transistor such as the
2N3905 for a constant current source, similar to the diagram on page 9
of this datasheet

http://rocky.digikey.com/WebLib/Fairchild/Web Data/LM555_NE555_SA555.pdf

It would have to source about 80 or 100 microamps. Will this work at
such a low current?
One could use a bipolar transistor biased like that for currents down
to the nanoamp region.
However, it is a poor "constant" current source (or sink if one uses
an NPN) due to the Early effect.
Use a JFET (good to about 30V, the max rating i have seen), or a
Depletion MOSFET (good to 500V).
 
W

Winfield Hill

Jan 1, 1970
0
Robert Baer wrote...
One could use a bipolar transistor biased like that for currents down
to the nanoamp region.

Yes. I assume you're talking about figure 13. Just take care, as
Ban said, to use a CMOS version of the 555 with low timing currents.
However, it is a poor "constant" current source (or sink if one uses
an NPN) due to the Early effect.

I disagree, for two reasons. First, the Early effect (most usefully
expressed as a slight change in Vbe vs Vce) is no worse at low
currents than at ordinary currents, and second, the voltage drop
across the emitter degeneration (current-setting) resistor overcomes
the slight Early effect. The transistor current source will work as
well as before, after scaling the design to low currents.

Now, if you were using a simple two-transistor current mirror,
without emitter resistors, that'd be another matter.
Use a JFET (good to about 30V, the max rating i have seen), or a
Depletion MOSFET (good to 500V).

Not necessary, and the large variation of Vgs introduces problems.
 
J

Jon

Jan 1, 1970
0
Kell,
..
Yes. Make sure that the Re is large enough so that the voltage across
it is >>Vbe, the Base-Emitter voltage. This ensures that Vbe and
variations in Vbe with temperature will have minimum effect on the
collector current.
..
Regards,
Jon
 
K

kell

Jan 1, 1970
0
Okay I can use a bigger cap so the chip doesn't load it and affect the
timing. Looking for some accuracy here. Just need to know the
tolerance. I have 1uF dip tantalum caps with K printed on them,
meaning 10%. My 4u7 dip tantalums have an A printed on them. Does
that mean anything about the tolerance?
 
K

kell

Jan 1, 1970
0
And another thing: if anybody wants to suggest an alternative constant
current source, while keeping the headroom down to 3 volts.
 
T

Tam/WB2TT

Jan 1, 1970
0
kell said:
I have a circuit that I would like to use a pnp transistor such as the
2N3905 for a constant current source, similar to the diagram on page 9
of this datasheet

http://rocky.digikey.com/WebLib/Fairchild/Web Data/LM555_NE555_SA555.pdf

It would have to source about 80 or 100 microamps. Will this work at
such a low current?
If you have to buy the transistor, I would get the 2N3906. It has twice the
beta, 60 min at 100 uA. Also, somebody suggested using two transistors in a
current mirror. You might want to check that out; it will give you more
headroom.

For accuracy, at the 1 uF level, I would use some kind of plastic dialectric
cap.

Tam
 
W

Walter Harley

Jan 1, 1970
0
kell said:
Okay I can use a bigger cap so the chip doesn't load it and affect the
timing. Looking for some accuracy here. Just need to know the
tolerance. I have 1uF dip tantalum caps with K printed on them,
meaning 10%. My 4u7 dip tantalums have an A printed on them. Does
that mean anything about the tolerance?


You need to worry at least as much about leakage current as tolerance, if
you're trying to make an accurate long-time-constant ramp based on charging
a capacitor with a constant current source.

You might instead consider a microcontroller (or other counter mechanism)
and a D/A converter.

ISTR a thread on this group a few years back discussing the various
alternatives.
 
J

JeffM

Jan 1, 1970
0
kell wrote:
[1 of only 19 Usenet postings (posted from Google Groups)]

No one has mentioned it yet,
but you don't include context when you respond.

Some threads that you may find useful:
http://groups-beta.google.com/group...his-at-Google-don't-see-Usenet-the-way-you-do
Follow the link in that post and follow the link in THAT post.

The **Show options** stuff will prove useful.


Grise puts a fine point on it here:
http://groups-beta.google.com/group...post-from-google-groups+are-universally-hated
He may be a bit strong with his rhetoric--but not by much.
 
K

kell

Jan 1, 1970
0
JeffM said:
kell wrote:
[1 of only 19 Usenet postings (posted from Google Groups)]

No one has mentioned it yet,
but you don't include context when you respond.

Some threads that you may find useful:
http://groups-beta.google.com/group...his-at-Google-don't-see-Usenet-the-way-you-do
Follow the link in that post and follow the link in THAT post.

The **Show options** stuff will prove useful.


Grise puts a fine point on it here:
http://groups-beta.google.com/group...post-from-google-groups+are-universally-hated
He may be a bit strong with his rhetoric--but not by much.

Okay I went to the links and saw the advice to click show options and
then click reply; I'm trying that now to see how it works. Does this
give you the context you need?
I never used a newsgroup until Google started its groups thing. I
actually did not know Google leaves newsreaders in the dark until you
told me.
 
J

JeffM

Jan 1, 1970
0
you don't include context when you respond.
..
Does this give you the context you need?
kell
..
That's better.[1] Actually, I read on Google as well.
I just didn't want you to form a bad habit.

I actually did not know Google leaves newsreaders in the dark
until you told me.
..
Therein lies the problem.
A newbie who posts from Google often does it wrong for weeks
then gets defensive when corrected
(usually by someone in a bad mood.)


[1] You don't have to re-post everything from the previous post;
feel free to snip out as much from the automated blockquote
as you wish .
I only leave enough to show what part I'm referencing ("context").
See above.
 
H

Homer.Simpson

Jan 1, 1970
0
kell said
Okay I went to the links and saw the advice to click show options
and then click reply; I'm trying that now to see how it works.
Does this give you the context you need?
I never used a newsgroup until Google started its groups thing.
I actually did not know Google leaves newsreaders in the dark
until you told me.

If you really want to enjoy USENET you might consider a real
newsreader. Most consider newsgroups-via-Google a pretty poor
interface. YMMV

I'd recommend XNEWS.... but there are many options.
 
J

JeffM

Jan 1, 1970
0
I actually did not know Google leaves newsreaders in the dark
until you told me.
kell

Expanding on what Homer said, if you get a newsreader and use it a bit
you may like it better than Google Groups.
http://www.google.com/search?&q=gravity+forte+xnews+40tude+mozilla

I have Gravity, but only use it rarely
e.g., responding to a poster who cross-posts to more than 5 groups
(and I don't know which one he reads)
or when Google doesn't archive a particular group on the Groups line.

Generally, I prefer reading from my browser.
 
J

JeffM

Jan 1, 1970
0
I actually did not know Google leaves newsreaders in the dark
until you told me.
kell
:
:Expanding on what Homer said, if you get a newsreader and use it a bit
:you may like it better than Google Groups.
:http://www.google.com/search?&q=gravity+forte+xnews+40tude+mozilla
: JeffM

I forgot to include the major point I was trying to make:
Even if you go back to Google Groups after trying a newsreader,
you'll have an appreciation of how others see Usenet.
 
Top