Need help with the Zetex circuit!

Fasti

Newly Enlightened
Joined
Sep 19, 2003
Messages
4
Hello!

I built the Circuit described in the Designnote from Zetex and used the Boardlayout posted on CPF but it doesn't work as it should! It works fine when the Inputvoltage is below 1V (0.9V) but over 1V the output is off and the Circuit draws more and more current with increasing Voltage! At 2 V it draws about 1 A from my Power Supply but the Output is 0V.
I've checked every connection and tested all parts (except the ZXSC310)
BTW I use the ZXSC310 and drilled a hole in the Board where the shutdownpin is! The Datasheet from the ZXSC310 says that its enabled when the STDN-Pin is open or pulled high! I also tried to connect the STDN-Pin to V+ but that doesn't change anything! Please help me I don't know why this thing doesn't work!

Fasti
 

robk

Enlightened
Joined
Mar 11, 2003
Messages
608
Location
Near Daytona Beach, FL
Just a thought... check the polarity on the schottky diode, or remove it from the board and see if you have output on the collector of the 617 transistor. Also, if your "sense resistor" is too short or too low in value, it will draw a lot of current and smoke the transistor.
Rob
 

LED_ASAP

Enlightened
Joined
Dec 13, 2002
Messages
567
Location
British Columbia, Canada
There is nothing wrong except that you are pushing the components beyond their capabilities. Try increasing the value of the sense resistor and everything will be back to normal.
 

Fasti

Newly Enlightened
Joined
Sep 19, 2003
Messages
4
Thanks for the quick replies!

The polarity of the schottky diode is right! I checked the transistor and it seems to be ok! I can only imagine that the sense resistor, I use about 2.5cm of copperwire, is too low but I thought it would be long enough because I only need 22mOhm!

thx

Fasti
 

MrAl

Flashlight Enthusiast
Joined
Sep 9, 2001
Messages
3,144
Location
New Jersey
Hello Fasti,

The sense resistor is rather critical.
If it's too low, the circuit will never start
and the transistor will burn up.

One way to tell if it might be too low is to
turn the circuit on and feel the area near the
transistor (dont put finger ON the transistor!).
If the transistor gets hot, the circuit isnt
starting. Do this quickly though, because
it could take the transistor out within two
seconds or less.

It's easy to adjust really, all you have to do is
double the lenght of the wire that makes up the
resistor, then test again. If the current is too
high, increase the lenght again, if it is too low,
decrease it a little.

If the transistor is already burnt up it will never
start but may continue to draw lots of current no
matter how large you make the sense resistor.
If the transistor is burnt but it's open, then
you wont see much input current but the output
voltage will be almost the same as the input
voltage.

Good luck,
Al
 

koala

Flashlight Enthusiast
Joined
Mar 5, 2003
Messages
2,295
Location
Melbourne, Australia
Take Al's words, he's one of the pioneer /ubbthreads/images/graemlins/grin.gif. Start off with long wire then trim it down.

Vince.
 

Fasti

Newly Enlightened
Joined
Sep 19, 2003
Messages
4
Thx for the quick answer!

I took a longer wire and it works! Trimmed it down until I got a current of 300mA at the Diode! I tried to cut it a little bit more to get 350mA but then the circuit stops working! IMHO this is the most I can expect from the ciruit. Correct me if I'm wrong ;-)!

Thx for the help

Fasti
 

Reboot

Newly Enlightened
Joined
Jul 3, 2003
Messages
28
Location
Aloha, OR
I adjust mine for 400ma thru MrAl's Luxeon simulator ckt (4 diodes & 1ohm resistor), so I know you can get more than 350. What are you using for an inductor? That's the critical part.
 

Fasti

Newly Enlightened
Joined
Sep 19, 2003
Messages
4
The inductor I use is a Coilcraft DO3316P-223 22uH, 2 A!
I adjusted the circuit at 2.7V(fresh pair of NiMH AAs)and get 350mA and at 3V (2 Alcaline AAs) I get 390mA! That should be pretty enough! :)
Thanks for the help!
 

Probedude

Enlightened
Joined
Jan 10, 2008
Messages
456
Location
Ventura CA
Hello Fasti,

The sense resistor is rather critical.
If it's too low, the circuit will never start
and the transistor will burn up.

One way to tell if it might be too low is to
turn the circuit on and feel the area near the
transistor (dont put finger ON the transistor!).
If the transistor gets hot, the circuit isnt
starting. Do this quickly though, because
it could take the transistor out within two
seconds or less.

It's easy to adjust really, all you have to do is
double the lenght of the wire that makes up the
resistor, then test again. If the current is too
high, increase the lenght again, if it is too low,
decrease it a little.

If the transistor is already burnt up it will never
start but may continue to draw lots of current no
matter how large you make the sense resistor.
If the transistor is burnt but it's open, then
you wont see much input current but the output
voltage will be almost the same as the input
voltage.

Good luck,
Al

Thank you Al and thank you archives!
I was running into this tonight. Trying to make a headlamp and couldn't get it to start on 2 cells. "the sense voltage is >> 19mV, why isn't it switching?!!!"

Off to the garage to try a different value sense resistor.


Edit: :mad:
Still not working!
I'm using the "maximum brightness solution" schematic on the Zetex datasheet (page 7) and even raising the sense resistor to 0.8 ohms (first was 0.1 ohms, then 0.18, then 0.39, then 0.78) the circuit still will not start. I originally tried it exactly as the datasheet has it (0.1 ohm sense, same pn transistor and schottky) and it would not work with 2 cells and only dimly with 1 cell.

PCB layout looks great - nice fat traces to minimize voltage drop and inductance contribution. Lines from the sense resistor too are fat. Battery - comes in near the IC so there's no drop there.

Arrgh! I didn't expect to have this much trouble with this chip!
headlamp.gif
board layout
headlamp-scheme.gif


Bottom right pad is the LED +, bottom left is batt- and LED-.
Upper two pads are for an external switch (using that to enable/disable the IC but for now it's just open. Pull down resistor on that pin is not stuffed) Batt+ is right at the ZXSC310 Vcc pin

Transistor was pulled and checked out okay - replaced it anyways.
Swapped out the Zetex 310, same problem.

Looking at my layout above, compared to the datasheet sample layout, I wonder if I need to isolate the 310 more (datasheet layout has REALLY LONG traces from the 310 to the transistor and sense resistor.)



Dave
 
Last edited:

MrAl

Flashlight Enthusiast
Joined
Sep 9, 2001
Messages
3,144
Location
New Jersey
Hi Dave,


PM sent.

I was saying that there are a few reasons why the circuit
may not start up the first time.

1. Sense resistor too small
2. Inductor series resistance too high
3. Transistor burnt out from previous tries to start up

Feel free to PM me again or post here.
 

Probedude

Enlightened
Joined
Jan 10, 2008
Messages
456
Location
Ventura CA
Thanks Al !

I put the schematic up in my previous post - it's straight from the datasheet.

The transistor I pulled checked out okay, but I replaced it anyways. It did not have a short between C and E and the drop from B-C and B-E matched a new part pulled from the tape.

The other thing that tells me the transistor is okay is that I can put the 310 into shutdown mode and with 2 batts the LED will light since the LED is now just direct driven (I was surprised to find the Cree Q5 would light at ~ 2.6V).

When it was working with 1 NiMH cell (but dimly), the voltage on that cell was ~ 1.25V open circuit. The circuit definitely didn't like an alkaline cell of 1.48V - the transistor would start getting warm. With 2 NiMH or 2 alkaline things would get warm quickly.

I've tried various inductors, first this one.
Digikey pn PCD2119CT, 100uH 600mA smd, 0.36 mOhm max resistance.

Also tried using this one Digikey pn M9966, 100uh, 0.28 ohm max, 580mA sat current. Also tried paralleling this one with another since it's a radial inductor and that was easy to do.

LED is a Cree Q5 (also tried a SSC P4), shooting for 50 - 100mA of drive current.

I really appreciate the help.
Dave
 

MrAl

Flashlight Enthusiast
Joined
Sep 9, 2001
Messages
3,144
Location
New Jersey
Hi Dave,

It looks like you have the circuit set up right so i guess there
is something basic that is not working right. From here it is
basically a matter of trying a few things and doing a few tests
to see what isnt working.

Since you are saying that the normal drive current will
only be 100ma, that should be easy to obtain.

For a quick note, the startup current flows through the
inductor, transistor, battery, battery series resistance,
and sense resistor, and ground return. Too high a resistance
in any part of this path can prevent proper start up because
the sense resistor never gets enough current to trip the
sense threshold.

Test 1:

Doing this as quick as possible...
With one cell 1.5v connected, what is the input current
even if it doesnt start?
(This has to be done quickly so the transistor does not
burn out).

Did you use a wire table to determine the length of your sense wire?

You should next try two 0.1 ohm resistors in series, not lengths of wire,
and see if that works better with one cell or two cells.

What is the *gauge* of your sense wire, and how long is it now?
 

Probedude

Enlightened
Joined
Jan 10, 2008
Messages
456
Location
Ventura CA
Hi Dave,

It looks like you have the circuit set up right so i guess there
is something basic that is not working right. From here it is
basically a matter of trying a few things and doing a few tests
to see what isnt working.

Test 1:

Doing this as quick as possible...
With one cell 1.5v connected, what is the input current
even if it doesnt start?
(This has to be done quickly so the transistor does not
burn out).

Did you use a wire table to determine the length of your sense wire?

You should next try two 0.1 ohm resistors in series, not lengths of wire,
and see if that works better with one cell or two cells.

What is the *gauge* of your sense wire, and how long is it now?

Hi Al,
I am using resistors for Rsense, not wire. The values I've tried have been 0.05, 0.1, 0.18, 0.36, 0.8 ohms by using combinations of the resistors 0.1, 0.18 and 0.39 ohms that I had on hand.

I'll remeasure the drop across this resistor but from memory it was far in excess of the 19mV threshold per the datasheet (I think it was ~ 270mV).

From memory (I'll remeasure) with a single battery when it would not start up,
current draw from the batt was ~ 60mA when Rsense was 0.8 ohms and batt was at 1.46V. I don't remember what the current draw was with lower resistor values. With the 60mA and 0.8 ohm Rsense, the voltage drop across it would have been 48mV, at least 2x the threshold current.

With 2 batteries current draw from the battery was ~ 1.57A and that's when things would get hot quick.

From the PCB layout you can see that all the connections are to fat traces. Very little inductance or resistance being added to the circuit.
 

MrAl

Flashlight Enthusiast
Joined
Sep 9, 2001
Messages
3,144
Location
New Jersey
Hi again,


Ok, then if the resistor values are correct and the transistor
is working then the chip isnt working. If the chip is really
working then there has to be something wrong with the
circuit board layout.
Perhaps you can double check the data sheet to make sure
all the pins are what you think they are, and that the
circuit board is made right.
It sounds like you are doing everything else right.

A test for the chip itself might be to place a load on the
output pin to ground (say 100 ohms) and measure the
pin output voltage with the sense input shorted to
ground. Apply 50mv to the sense input and verify that
the output pin switches to a much lower voltage.
This would be with no inductor and no diode and no LED,
just a raw test for the chip alone.
 

Probedude

Enlightened
Joined
Jan 10, 2008
Messages
456
Location
Ventura CA
Hi again,


Ok, then if the resistor values are correct and the transistor
is working then the chip isnt working. If the chip is really
working then there has to be something wrong with the
circuit board layout.
Perhaps you can double check the data sheet to make sure
all the pins are what you think they are, and that the
circuit board is made right.
It sounds like you are doing everything else right.

A test for the chip itself might be to place a load on the
output pin to ground (say 100 ohms) and measure the
pin output voltage with the sense input shorted to
ground. Apply 50mv to the sense input and verify that
the output pin switches to a much lower voltage.
This would be with no inductor and no diode and no LED,
just a raw test for the chip alone.

I've checked and double checked but it won't hurt to check again. The circuit is so simple I can't believe the problems I'm having. It works in other ways - less than 1.3V it'll turn on so the IC can drive the transistor. I've even measured the voltage on the Rsense pin and see it's higher than the threshold of 19mV. So it all measures fine, just doesn't work fine.

I sent off an email for tech support at Zetex. I tried my local distributor to talk to their Zetex apps guy but apparently with Diodes Inc buying Zetex (quite a while ago) there's some reorganization going on.

Suppose I'll just fab their demo board from their dwg and see if things get better quickly.
 

Probedude

Enlightened
Joined
Jan 10, 2008
Messages
456
Location
Ventura CA
Received a reply from the Zetex apps guy.

He must have read my note too quickly because his reply was that I need a buck configuration, not boost :confused:

Edit: SOLVED!

:shakehead

:ohgeez:

:whistle:

I had the shutdown pin and gnd swapped. (it's in the schematic above wrong per the pin #'s)

Very excitedly I went off to my datasheet printouts, online datasheet, etc etc, looking for the mistake but in the end the mistake was all mine.

I hate when that happens.

Sorry for wasting your time. Works great on 1 or 2 cells now :eek:

Amazing how it can be wrong and look so right.
 
Last edited:

MrAl

Flashlight Enthusiast
Joined
Sep 9, 2001
Messages
3,144
Location
New Jersey
Hi again,


Hey, no waste of time, we followed a procedure to get to
the bottom of it by checking one thing at a time, and we
found what was wrong and corrected it. Now it works.
That's not a waste of time :)

Happy to hear you got it working now. Good luck with it
in the future.
 
Top