74hc595 oe pin not working

Ok I thought all was well in my project until I decided to add a fade to the lights
Im simply going thru all the lights and pwming the oe pin down and up
however for some reason one lone pin of the 74hc595 decides to do its own thing and do the opposite of what the oe pin says
so as I fade down and up the brightness one by one of all the leds, this one led is constantly doing the opposite
when I first start I have all the leds fade on, this one stays constant, and when I write them all off it will turn off assuming the oe pin is low enabling the output
but if the oe pin is disabled it turns on
any thought on this? This would be the second 74hc595 to have been found faulty in the batch of 5 from ebay, first time may have been me but this is weird

Have you checked for mis-wiring, shorts between neighbouring pins etc? If the damage isn't static damage (don't wear nylon if handling chips, don't have nylon carpets(!)) then it sounds like the batch are untested factory surplus (which might be false economy).

Do you Have a pull down on that port? If /OE is floating you will have the problem you describe. I would use an ohm meter to verify the connection, Assume Nothing.... You know the definition of Ass U Me?

Doc

No clue on that definition, I don't have a pull down but why would that affect one pin when oe affects them all? And it active low so everything is lit up when its low, the problem is when its high and supposed to be high impedance
when the output is set high, and I pwm the oe pin it simply stays high, and when set low it will do the opposite
think maybe a pulldown on the base of the driving transistor can help? But if its failing to go high impedance then I guess that wouldn't help right?

Put a pull up on it, hen you do so you know that it ALWAYS starts from a known condition... Btw read the word I spelled using the Caps... to define it... If U can.

Doc

Aum? Sorry im not that smart lol don't get it

winner,
Doc is cracking wise, but gets lost in translation.

Ah I get it now, was telling my friends about it and when I said it aloud it clicked lol
And that video helped lol

But then again the startup doesn't really matter because the first thing I do is write that low
and the fault is when I drive it the opposite way, once the program is up
I thought it was a software bug( a really weird one) but then realized its basically impossible with what's happening and other than that oe thing everything else works fine

Why is it so hard just to do it and make a fool out of an old man when you prove him Wrong. Many times I have been wrong, you won't hurt my feelings. I do find it somewhat strange that idea that there can only be one answer to a problem, yours, in your apparently broken sketch or mysteriously broken chips (2 now?) I have been buying that surplus stuff from the internet for years and I have yet to get a bad chip. When I was working I had occasion to deal with a very large variety and number of IC's of all flavors from RTL to ECL to Mos of All kinds and my composite chip failure for research accidents, clumsiness, installation issues, mechanical and electrical damage (static electricity) was about 2% total, yours is 40% was it 2 of 5 74HC595's you bought? that failed?. All from China. For the past 8 years 90% of my ic's were made in and sold from China. What I am really getting to here is that I think either you are mis-handling them electrically or mechanically. Statistically if a chip works it works I rarely if ever saw one that died in circuit... like yours seem to do. IMO

Doc

This is the first time anything has failed me from ebay, I've been doing stuff like this for a year and haven't had problems like this
the fact that one pin isn't cooperating baffles my little knowledge of electronics
If the other 15 pins follow the oe pin, and the oe pin controlls them all, why wouldn't one follow? And actually do the opposite?
I guess im gonna have to double up with another output but still, that's weird result

Well I guess that you are right and you got a bunch of surplus fall out ic's... I only have one real issue with that... Those ic's are surplus... Check the code dates. I might point out (for what it is worth) that although you know what your code was intended to do, you don't have a clue as to what it is really doing except something isn't working right... so please don't follow any of my suggestions... I hate to be told I don't have a clue.

Doc

If you promise to send me the two "bad-out-of-the-box" 74HC595's, I'll tell you how to get the other three to work. For now, I'll predict that the uncooperative LED is driven by the shift register's bit 7. If that's correct, you'll probably see what's wrong right away.

Did you figure it out, or do we have a deal?

haha ^, and if you cant figure that out you can send me the other 3..

winner10920:
This is the first time anything has failed me from ebay

I was unaware that Ebay sold electronic components. I thought that they were just an auction site and the actual seller could be anybody with some of those sellers being much more scrupulous than others. So you only have a point there if it has been the same seller every time, and then the problem is with the seller and not Ebay.

You can get Anything you want... if you know where and in some case's how to ask... Electronics components... Certainly ANYTHING you could find in a catalog... Bare 100W LED's, $26.95, Free Shipping... All Day Long, ( $59.00 with the CC driver )... GREAT Project for an Arduino, a 100 W LED Dimmer, uses a 74HC595 and some small sensing resistors...

Doc

Earlier today, I wrote:

If you promise to send me the two "bad-out-of-the-box" 74HC595's, I'll tell you how to get the other three to work.

Then some other clown said:

patrick4989:
haha, and if you cant figure that out you can send me the other 3 ...

Maybe I'm too sensitive, but I'm thinking that somebody doesn't take me seriously here.

Here's what I think is going on: you're driving the one LED that doesn't work from pin 9 of the74HC595. Pin 9 is called "QH'" in this datasheet - http://www.sparkfun.com/datasheets/IC/SN74HC595.pdf - and its function is "serial data out." The logic and timing diagrams show that it's not affected by OE, and it doesn't have a high impedance state. If you drive an LED from this pin, you can't control it by pulsing OE.

I think that you want to drive that LED with pin 7, "QH," whose function is "parallel data out #7." Pin 7 responds to OE just like the other seven parallel data pins.

Did that fix it? Are you going to send my my shift registers?

Good Call Really good. and more likely than my ideas about pull ups (was the mistake I made the most often, thinking that the processor was in a known state and letting it control the bus line instead of insuring that just in case it is in a known state)

Doc