7805 regs under low load conditions

B

bruce varley

Guest
Hi, I have a device that runs on 5V, load current normally about 8mA, but
with periods of up to 30 - 50mA. Power supply is through a 7805 (TO220)
regulator, with a fairly stable, ripple free 13 VDC on the input. I went for
the larger package rather than a 7805L because the equipment is destined for
hot conditions, bit marginal. Anyway, things seem to work fine for a while,
then the micro starts misbehaving. When that occurs I find that the output
of the 7805 has crept up towards 6 volts. This has happened twice with
devices that started off well within spec, Vo between 5.0 and 5.1 volts. The
change seems to be irreversible.

AFAICT, the setup around the regs is OK. 47n ceramic close to the output.
Absolutely no instability showing on the CRO. My fairly old datasheet
doesn't mention anything about low load conditions, but according to the
regulation quoted the output shouldn't get anywhere near as high as I'm
seeing, even with no load at all.

Has anyone seen anything like this, any suggestions?? TIA
 
Hi Bruce,
It sounds like the 7805 is a dud. I'd remove the offending regulator, pop in
another and see what happens. (I haven't seen this before except with one of
the low-power low-dropout 5V regs which required a 10uF cap on the output)
.... Steve
 
"bruce varley"
I went for
the larger package rather than a 7805L because the equipment is destined
for
hot conditions, bit marginal.
Click to expand...

** Please explain.

Glossing over a significant point like this is very suspicious.




....... Phil
 
there should be a cap on the input as well, 10uF should prevent
instability

"Steve Carroll" <scarroll01@optusnet.com.au> wrote in message
news:4465a868$0$25132$afc38c87@news.optusnet.com.au...
: Hi Bruce,
: It sounds like the 7805 is a dud. I'd remove the offending regulator,
pop in
: another and see what happens. (I haven't seen this before except with
one of
: the low-power low-dropout 5V regs which required a 10uF cap on the
output)
: ... Steve
:
:
 
Phil Allison <philallison@tpg.com.au> wrote in message
news:4clntcF16010bU1@individual.net...
"bruce varley"

I went for
the larger package rather than a 7805L because the equipment is destined
for
hot conditions, bit marginal.


** Please explain.

Glossing over a significant point like this is very suspicious.




...... Phil
Click to expand...
Equipment will be running outdoors, ambient temperature up to 40C. In fact,
all the testing so far has been inside, so no temperature extremes have been
encountered. At no time have any of the devices involved been above
moderately warm to the touch.

Sorry, I should have elaborated.
 
I was going to suggest that too.

Bob


On Sat, 13 May 2006 17:50:37 +0800, "E d" <tosser@dfat.gov.au> wrote:

there should be a cap on the input as well, 10uF should prevent
instability

"Steve Carroll" <scarroll01@optusnet.com.au> wrote in message
news:4465a868$0$25132$afc38c87@news.optusnet.com.au...
: Hi Bruce,
: It sounds like the 7805 is a dud. I'd remove the offending regulator,
pop in
: another and see what happens. (I haven't seen this before except with
one of
: the low-power low-dropout 5V regs which required a 10uF cap on the
output)
: ... Steve
:
:
Click to expand...
 
On Sat, 13 May 2006 17:25:12 +0800, "bruce varley"
<bxvarley@weqstnet.com.au> wrote:

Hi, I have a device that runs on 5V, load current normally about 8mA, but
with periods of up to 30 - 50mA. Power supply is through a 7805 (TO220)
regulator, with a fairly stable, ripple free 13 VDC on the input. I went for
the larger package rather than a 7805L because the equipment is destined for
hot conditions, bit marginal. Anyway, things seem to work fine for a while,
then the micro starts misbehaving. When that occurs I find that the output
of the 7805 has crept up towards 6 volts. This has happened twice with
devices that started off well within spec, Vo between 5.0 and 5.1 volts. The
change seems to be irreversible.

AFAICT, the setup around the regs is OK. 47n ceramic close to the output.
Absolutely no instability showing on the CRO. My fairly old datasheet
doesn't mention anything about low load conditions, but according to the
regulation quoted the output shouldn't get anywhere near as high as I'm
seeing, even with no load at all.

Has anyone seen anything like this, any suggestions?? TIA
Click to expand...
It seems that you have tried different devices and get the same
results, so unless the manufacturer is dodgy and all devices from them
are prone to failure, or there is some other explanation. Have you
tried devices by different manufacturers? Also, without knowing the
input/output configuration for the 7805 it is difficult to speculate.
Try the old Vout - Vin protection diode trick and see if this solves
the problem.
 
On Sat, 13 May 2006 10:27:42 GMT, Ross Herbert
<rherber1@bigpond.net.au> wrote:

On Sat, 13 May 2006 17:25:12 +0800, "bruce varley"
bxvarley@weqstnet.com.au> wrote:

Hi, I have a device that runs on 5V, load current normally about 8mA, but
with periods of up to 30 - 50mA. Power supply is through a 7805 (TO220)
regulator, with a fairly stable, ripple free 13 VDC on the input. I went for
the larger package rather than a 7805L because the equipment is destined for
hot conditions, bit marginal. Anyway, things seem to work fine for a while,
then the micro starts misbehaving. When that occurs I find that the output
of the 7805 has crept up towards 6 volts. This has happened twice with
devices that started off well within spec, Vo between 5.0 and 5.1 volts. The
change seems to be irreversible.

AFAICT, the setup around the regs is OK. 47n ceramic close to the output.
Absolutely no instability showing on the CRO. My fairly old datasheet
doesn't mention anything about low load conditions, but according to the
regulation quoted the output shouldn't get anywhere near as high as I'm
seeing, even with no load at all.

Has anyone seen anything like this, any suggestions?? TIA


It seems that you have tried different devices and get the same
results, so unless the manufacturer is dodgy and all devices from them
are prone to failure, or there is some other explanation. Have you
tried devices by different manufacturers? Also, without knowing the
input/output configuration for the 7805 it is difficult to speculate.
Try the old Vout - Vin protection diode trick and see if this solves
the problem.
Click to expand...

Ed and Bob might have both hinted that loop instability might be the
culprit resulting in oscillation, which is usually compensated using
10uF at input and output.

National Semi have this paper on linear regs
http://www.national.com/an/AN/AN-1148.pdf
 
On Sat, 13 May 2006 17:50:50 +0800, "bruce varley" <bxvarley@weqstnet.com.au> wrote:

Phil Allison <philallison@tpg.com.au> wrote in message
news:4clntcF16010bU1@individual.net...

"bruce varley"

I went for
the larger package rather than a 7805L because the equipment is destined
for
hot conditions, bit marginal.


** Please explain.

Glossing over a significant point like this is very suspicious.




...... Phil



Equipment will be running outdoors, ambient temperature up to 40C. In fact,
all the testing so far has been inside, so no temperature extremes have been
encountered. At no time have any of the devices involved been above
moderately warm to the touch.

Sorry, I should have elaborated.
Click to expand...
Sounds like the ground pin may have become disconnected. Replace the regulator.
Since it's happened with 2 regulators so far, perhaps you've struck a bad batch,
which would be extremely unusual. I'd get some regs from another manufacturer.

Since you're using very low current, have you considered using a switching regulator
instead of the 7805? Or are you using the '05 as a voltage reference for an analog to
digital converter?

BTW, how is the ground terminal of the regulator connected to the rest of the circuit?
Is it located on a seperate board and makes connection through terminal pins of a
connector, or a similar arrangement, or is the regulator soldered directly to the board
where the micro is located?
 
Sorry Bruce, I didn't read your post properly and didn't realise that you'd
already tried swapping the reg.
Steve
 
Phil Allison <philallison@tpg.com.au> wrote in message
news:4clrgiF16hfhsU1@individual.net...
"bruce varley"

Has anyone seen anything like this,


** Input side oscillations.




....... Phil

There aren't any there in this case. Steady as a rock.
Click to expand...
 
Do you have a bypass cap (like around 0.33uF ceramic as suggested
by the manufacturers) between ground and the input pin, nice & close
to the regulator? If not, it'd be worth putting one there then seeing
if you still have this problem.

Bob

On Sun, 14 May 2006 19:27:56 +0800, "bruce varley"
<bxvarley@weqstnet.com.au> wrote:

Phil Allison <philallison@tpg.com.au> wrote in message
news:4clrgiF16hfhsU1@individual.net...

"bruce varley"

Has anyone seen anything like this,


** Input side oscillations.




....... Phil

There aren't any there in this case. Steady as a rock.
Click to expand...
 
On Sat, 13 May 2006 17:25:12 +0800, "bruce varley"
<bxvarley@weqstnet.com.au> opined:

Hi, I have a device that runs on 5V, load current normally about 8mA, but
with periods of up to 30 - 50mA. Power supply is through a 7805 (TO220)
regulator, with a fairly stable, ripple free 13 VDC on the input. I went for
the larger package rather than a 7805L because the equipment is destined for
hot conditions, bit marginal. Anyway, things seem to work fine for a while,
then the micro starts misbehaving. When that occurs I find that the output
of the 7805 has crept up towards 6 volts. This has happened twice with
devices that started off well within spec, Vo between 5.0 and 5.1 volts. The
change seems to be irreversible.
Click to expand...
7805s are fine with small loads & are /very/ hard to kill.
It sounds to me like your regulator is oscillating or losing its
ground connection.

To fix oscillation: Solder 100nF ceramic monobloc capacitors (the
small yellow rectangular kind used for digital bypass caps) directly
onto the legs of the 7805. One cap from pin 1 to pin 2 (input to
ground) & a second from pin 3 to pin 2 (output to ground).

If the 7805 loses its ground connection, its output will 'float' up to
almost the *full value of the unregulated input*, & will act almost as
if pin 1 is shorted to pin 3. Make sure that your 7805 has its tag or
middle pin *solidly* connected to both input & output grounds.
--
W
. | ,. w , "Some people are alive only because
\|/ \|/ it is illegal to kill them." Perna condita delenda est
---^----^---------------------------------------------------------------
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top