Trouble controlling 7 segment LED display

the delay of an Inverter 4069 should be around 10 ns so with a good system is should be neglectable. But your idea with 2 inverters in the clockline is a good way to compensate the delay in dataline.
How does it work now? do you have stable display but wrong or?
 

AlbertHall

Joined Jun 4, 2014
12,345
I don't understand why it is inverted compared to the original manufacturers diagram. The original used common anode displays.
I can't read the IC part numbers for the shift registers on that original diagram - is there an inverting version of the '595?
Or did the original chip (if it wasn't a '595) operate with the high to low transition of the clocks?
 

Thread Starter

kosalos

Joined Nov 21, 2016
11
No, the way I am inverting the data line by itself is not working. The display always settles to "7." ( a bit pattern that might be decipherable, but not helpful), or flicking nonsense if I fiddle with the resistance of the input to the inverter transistor.
I need to go back to the original schematic I attached to the first post just as a sanity check.
It was working clean as a whistle with nice LED brightness, except the polarity was wrong.

Forgot to mention. Before I joined your club I had tried putting an inverter transistor into the common anode line going to the display.
Didn't work. the display became a flickering mess.
When Albert mentioned inverting the Data line instead I went 'Duh!"

Albert, the schematic shows 74HC595. That's what i used.
I'll look whether there is a inverting companion chip.
thanks for the idea.
 
just 3 points you should check:
1. could one of the connection fail, an open input on CMOS could toggle by himself.
2. the STCP which is the clockpulse to store from shift reg to output reg gets some noise coupled, this can happen if you have long wires parallel. to avoid this put a small cap on this line (max 100pF).
3. Because the circuit switches quite some current Grond Bouncing can happen. Try to lower the GND connection: have a second wire to connect GND with GND.
on 74HC595 and 74HCT595 both clocks are active on low to high transition.
 

Thread Starter

kosalos

Joined Nov 21, 2016
11
reverb.png resolved!
Running the data line through a 4069 inverter, and swapping the digit select lines on U14 worked.
thanks for all the help.
 

RichardO

Joined May 4, 2013
2,270
That only leaves two mysteries.
1. Why was the data inverted in the first place (unless the manufacturers design is flawed - wouldn't be the first time!)?
2. Why didn't the transistor work as an inverter?
1. I would like the answer to that as well. :confused:
2. I am wondering if is because the transistor has too much storage time coming out of saturation. I notice that the Fairchild data sheet for the BD547 calls the device a switching transistor but does not give any switching specifications. The On Semi data sheet calls the BC547 an amplifier.

Unfortunately not knowing the data rate makes it a pure guess.
 
Top