Anyone see anything wrong with this EDN TL431 battery monitor circuit?

[Hammy]

I was looking for a simple circuit for UVLO and came across this:

http://i49.tinypic.com/1sdtef.png

Its from an EDN article.

I didn't think it would work correctly but its so few components I
gave it a try and yep it doesn't work correctly.

The LED starts to turn on way before the threshold voltage and well
below the cathode of the 431 goes low.This in turn turns the PNP on
prematurely, pulling the threshold even higher!

 

[Spehro Pefhany]

I was looking for a simple circuit for UVLO and came across this:

http://i49.tinypic.com/1sdtef.png

Its from an EDN article.

I didn't think it would work correctly but its so few components I
gave it a try and yep it doesn't work correctly.

The LED starts to turn on way before the threshold voltage and well
below the cathode of the 431 goes low.This in turn turns the PNP on
prematurely, pulling the threshold even higher!

It should work with a micropower XX431 that has some tens of uA
minimum cathode current.

 

[Hammy]

But it got in EDN, that's the important part, right?
Since the TL431 can be treated simply as high power zener in many
circuits, I bet you can design something that works.

--- news://freenews.netfront.net/ - complaints: [email protected] ---

I did.

This works and it only consumes less then 800uA at my worst case line
in 25Vdc. I'm using an LMV431.

http://i45.tinypic.com/2vabbbt.png

Simulated results

http://i46.tinypic.com/bjie8j.png

I wont have time to breadboard it until later tonight.

The BJT is used to short out another BJT driving a pfet turning off
the PFET and blocking voltage to the circuit until it reaches about
11.7V. There's about 1V of hysteresis; once the fet closes a 470uf cap
gets charged causing a line sag. I have to play around with the
resistor values but it's a lot more reliable and lower current
consumption then the EDN one.

 

[Hammy]

It should work with a micropower XX431 that has some tens of uA
minimum cathode current.

Tried it if you increases the R it increases the drop turning the BJT
on.

You would kinda think it should work and it may with the right values
of resistors and the right temperature, and the right LED,and 431.

Maybe not work is to harsh it just wouldnt be very consistent.

 

[Spehro Pefhany]

Tried it if you increases the R it increases the drop turning the BJT
on.

A regular 3 cent (or whatever) TL431 needs around 1mA to regulate
worst-case (from memory, check the datasheet), so it won't be
regulating properly at the 60uA where the transistor turns on.

You need a different DESIGN of reference (perhaps with a part number
which looks like xx431) wot is designed to work at very low cathode
currents (usually has a lower reference voltage such as ~1.25V as
well).

Otherwise you'd have to decrease the resistor values by something like
20:1 which would quickly drain the battery it's supposed to be
measuring.

You would kinda think it should work and it may with the right values
of resistors and the right temperature, and the right LED,and 431.

No, wouldn't really work properly. A reference that needs 1mA to
regulate is generally inappropriate as a low battery indicator for a
small battery unless it's pulsed or something.

 

[Hammy]

No, wouldn't really work properly. A reference that needs 1mA to
regulate is generally inappropriate as a low battery indicator for a
small battery unless it's pulsed or something.

My need isn't monitoring a battery it's for UVLO for a flyback smps.
You don't have to convince me the circuit isn't really the best for
its intended purpose far from it.

For monitoring batteries I've used the LTC1440 which is a micropower
comparator with integrated reference.

I'm using a LMV431 see my post above for my UVLO. The LMV431 needs min
80uA for guaranteed regulation and can withstand 30V cathode to anode.
The flyback input is 12V to 25Vdc . The input has a TVS and fuse so
if anyone does go above 25V the TVS will go low impedance and vaporize
the fuse Minimum breakdown is 25.7V max is 27.9. Not overly eloquent
but it works I've done it before.

I just posted the circuit because it doesn't work with the values
shown. I swept the supply from about 5V to 25vdc and the cathode
wouldn't saturate the LED starts to emit faintly though.

 

[Hammy]

My need isn't monitoring a battery it's for UVLO for a flyback smps.
You don't have to convince me the circuit isn't really the best for
its intended purpose far from it.

For monitoring batteries I've used the LTC1440 which is a micropower
comparator with integrated reference.

I'm using a LMV431 see my post above for my UVLO. The LMV431 needs min
80uA for guaranteed regulation and can withstand 30V cathode to anode.
The flyback input is 12V to 25Vdc . The input has a TVS and fuse so
if anyone does go above 25V the TVS will go low impedance and vaporize
the fuse Minimum breakdown is 25.7V max is 27.9. Not overly eloquent
but it works I've done it before.

I just posted the circuit because it doesn't work with the values
shown. I swept the supply from about 5V to 25vdc and the cathode
wouldn't saturate the LED starts to emit faintly though.

This is another bizzare circuit from IR a neg output post regulator
useing the tl431 and a PFET?

http://i45.tinypic.com/20ze4uv.png

 

[Hammy]

I think you typed it into spice incorrectly. When I simulated it, it
appeared to work properly. So, thinking there must be some effect I was
missing, I built it, and it works properly. The LED snaps on/off at about
10.5V.

The main issue I have with it is that it sucks power through the 1k
resistor. I'm guessing they are doing this to prevent any 'premature
elucidation' from the LED.

I included the simulation I did, and another one I hacked up using a couple
of NPNs.

Regards,
Bob Monsen

Your right is does work. I found a dead track on my breadboard which
would have been the ground of the anode on the 431 that might have
been the source of the problem.

All the same I'm using this one it only uses a couple hundred micro
amps. In the breadboarded cct I'm using a LMV431.

http://i47.tinypic.com/23rmfyg.png

Here is a scope shot

http://i46.tinypic.com/xqeyau.jpg

CH2 is the input 120Hz FW rectified,CH1 is the 2N7002 gate. The 2n7002
gate is being discharged through a 1M resistor. The fet still turns
off in around 600uS which is good enough for this. I just wanted to
keep Q1 off until the supply voltage gets around 12V.

For comparison here is the original cct scope shot.

http://i46.tinypic.com/2dw59oo.jpg


CH2 is the input signal and CH1 is the cathode of the 431.

 

[JosephKK]

I was looking for a simple circuit for UVLO and came across this:

http://i49.tinypic.com/1sdtef.png

Its from an EDN article.

I didn't think it would work correctly but its so few components I
gave it a try and yep it doesn't work correctly.

The LED starts to turn on way before the threshold voltage and well
below the cathode of the 431 goes low.This in turn turns the PNP on
prematurely, pulling the threshold even higher!

Yep, like many circuits published in EDN for the past 30 years, it
does not work right. That is part of why i quit my subscription, also
there is the huge jump in advertising pages. Needless to say, the
editorial standards morphed back then and again since.

 

[Hammy]

Yep, like many circuits published in EDN for the past 30 years, it
does not work right. That is part of why i quit my subscription, also
there is the huge jump in advertising pages. Needless to say, the
editorial standards morphed back then and again since.

No it does work. I made a boo boo;-)

There is a dead track in the breadboard I used. I figured that out
when I tried the other cct. I posted and it did the same thing. The
431 wasn't grounded it was floating. But still it uses to much
current.

I have 5 Breadboards and they are all full with works in progress.
Even to do up the little 431 cct I had to cram it in an unused section
of one breadboard. Time to make up another batch of PCB's.

 

[Spehro Pefhany]

Yep, like many circuits published in EDN for the past 30 years, it
does not work right. That is part of why i quit my subscription, also
there is the huge jump in advertising pages. Needless to say, the
editorial standards morphed back then and again since.

It's almost down to a newsletter in size..