Exact match. Not showing close matches.
'[OT] Voltage regulator'
I just built a negative voltage regulator using a 741 op-amp, a voltage
reference, and an external power transistor. Yes, I know that this is
quite an obsolete way of doing it, considering that there are ICs
available which require much fewer external components. It works quite
well, but it goes into oscillation if the capacitance on the utput is
around 0.1uF and you draw more than a couple mA. If the cap is much
smaller or much larger, it seems to work fine. I am just wonderingwhy it
would oscillate with such a specific capacitance there? Basically, it is
just the transistor's base attached to the output of the op-amp and the
emitter is the output of the regulator, and also is fed back to the
inverting input of the 741. I originally thought that the oscillation
might be because the regulator's pass transistor can only sink current
from the output, not source it, and therefore the capacitance would cause
the output to overshoot and then discharge thru the load back down to the
right value and then happen again in a cycle. However, I find that even
with the large value of cap that causes it to be stable, it still
overshoots under certain circumstances.
The oscillation frequency is around 50kHz.
The actual final regulator will be more complex than what I just
described, but that is just the simples circuit that exhibits the same
Anyone have ideas on why it does this?
>Anyone have ideas on why it does this?
The short version is that you have gain, feedback, and phase shift in the
If you have gain > 1 at the frequency where the phase shift is at 180
degrees, then it will oscillate at that frequency.
Try compensating the opamp, with a cap between - in and out. Small at
first, maybe 50pF, then go up.
Your in for pain on your controls needs for this....
Why do you not try a 7660 Switch capactior inverter? It is made for this.
If you need a regulated negative voltage tha is different than your
positive rail and you rail is much higher then soft clamp it with a
resistor zener combe and pic (chose) capacitors that give the lowest
charge duty for less power waste. I believe you were wanting a negative
bias for a LCD. Here you just need to spilt the - output with a resistor
and a pot for current limit and voltage bais...
Sean Breheny wrote:
> Yes, I know that this is quite an obsolete way of doing it,
> considering that there are ICs available which require much fewer
> external components.
It is in fact called an LM337, negative version of an LM317.
"Slaving" an NPN transistor (MJ3055) to it is even easier than the
> It works quite well, but it goes into oscillation if the capacitance
> on the output is around 0.1uF and you draw more than a couple mA.
I ask two questions: 1} Have you added any "compensation" capacitor
already which may be causing this problem? 2} can you assure me the
breadboard setup is no larger than 2" diameter?
|On Sun, 3 Jan 1999 22:23:56 -0500 Sean Breheny <CORNELL.EDU> writes: shb7
>I just built a negative voltage regulator using a 741 op-amp, a
>reference, and an external power transistor. Yes, I know that this is
>quite an obsolete way of doing it, considering that there are ICs
>available which require much fewer external components. It works quite
>well, but it goes into oscillation if the capacitance on the utput is
>around 0.1uF and you draw more than a couple mA. If the cap is much
>smaller or much larger, it seems to work fine.
Many regulators both commercial and homemade are unstable with zero or
small values of capacitance at the output. Often the best solution is to
add a "large" capacitor to guarantee a large minimum capacitance at the
output, regardless of the load. The capacitor also increases the
performance of the regulator under transient conditions.
I am just wonderingwhy
>would oscillate with such a specific capacitance there? Basically, it
>just the transistor's base attached to the output of the op-amp and
>emitter is the output of the regulator, and also is fed back to the
>inverting input of the 741.
It does that because adding an additional transistor increases the gain
of the 741 and also causes more delay from input to output. The feedback
is not fast enough to keep the output under control. It can get into a
lot of math to figure out exactly what is happening, though there are
simplified analytical methods to tell if a design will be stable or not.
The simplified technique uses the concept of "phase margin". Many
sophomore-level textbooks will present it.
I like to not deal with math whenever possible, so my advice would be to
add a large capacitor (10 uF). Stay at it, you're learning a lot.
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com/getjuno.html
or call Juno at (800) 654-JUNO [654-5866]
Wise words Mike,
one learns alot from simply being aware of these strange happenings in
First, I'd like to thank all those who replied.
Second, I'd like to clarify a bit what I'm trying to do. I'm not trying
to just generate a negative DC bias for an LCD, this is actually part of
a small adjustable bench supply that I am building, and I expect it to
deliver up to about 3 amps from -2V to abt -15V. Right now, I am able to
get that performance out of it with less than 3 mV ripple. The addition
of a large capacitance (>20 uF) cured the oscillation. Thanks Mike!
However, I don't understand how the explanations given apply in my
circuit. I am familiar with Bode plots and the criteria for oscillation.
The 741 (as you all know) is an internally compensated op-amp. SO, as
long as I don't add voltage gain to its feedback loop, and I don't add
lots of delay to the feedback loop(or phase shift), it shoulkdn't
In this case, the pass transistor shouldn't provide voltage gain, only
current gain, because of its copnfiguration. Also, the oscillation varied
in frequency, but often went as low as 50kHz. I wouldn't think that the
transistor would introduce 180 deg of phase shifty (or a significant
proportion of that, enough to de-compensate the 741) at 50kHz!! I tried
adding compensation to the 741 by
placing various value of cap from 100pF to 0.1uF between its output (pin
6) and its inverting input. It didn't help. What I can say is that the
oscillation would only start if the feedback loop contained a power
transistor. I assume that this is due to the much higher cap between the
base and collector.
I know how OT this is, but I would appreciate it if the experts here
could tell me how my thinking is wrong, or advance additional theories to
help me understand this, even though the problem has been solved by the
addition of the cap on the output. I hate not understanding why something
Thanks very much, you don't know how much a resource like this helps an
More... (looser matching)
- Last day of these posts
- In 1999
, 2000 only
- New search...