next topic »

[on:]

I can multiplex stuff myself using shift registers.  I can implement MAX7219s up to 20 displays now.  But I still have problems with the TLC5940.  Frankly, I want to give up on them but I still want to drive my 8X8 RGB displays in PWMed color.  What other options are there?  Thanks!

I think TLC5940 and MAX7219 should hook up and have a babby...

[Reply #1 on:]

I still have problems with the TLC5940.

What problems? They seem simple to me...

[Reply #2 on:]

I still have problems with the TLC5940.

What problems? They seem simple to me...

They seem simple to me also, however... The same problem that several people have had.  They burn up.  So what I have on my test setup so far is 3 TLCs (48 channels) and 2 8x8 RGB matrices (48 LEDs) and a reasonable IREF resistor selection (1.8K) and it runs for it bit drawing like 600mA with the matricies being powered and then all of the sudden jumping to about 2A and the matricies not being powered - it's just shorting within the TLC chip.  I have to cut power then or it will burn up.  The TLC controls the anodes and 8 P-channel MOSFETs control the cathodes in this setup with a 1/8 duty cycle.  I set all channels to 1/2 power but I have tried lower power and it still screws up.  I looked at the waves off the MOSFETs and they look fine.  I just can't tell what the hell is going wrong.

[Reply #3 on:]

They seem simple to me also, however... The same problem that several people have had.  They burn up.  So what I have on my test setup so far is 3 TLCs (48 channels) and 2 8x8 RGB matrices (48 LEDs) and a reasonable IREF resistor selection (1.8K) and it runs for it bit drawing like 600mA

600mA, how many volts?

The amps don't matter, it's the watts.

The TLC controls the anodes and 8 P-channel MOSFETs control the cathodes

How does that work? The TLC is a current sink, it has no business being connected to an anode.

[Reply #4 on:]

They seem simple to me also, however... The same problem that several people have had.  They burn up.  So what I have on my test setup so far is 3 TLCs (48 channels) and 2 8x8 RGB matrices (48 LEDs) and a reasonable IREF resistor selection (1.8K) and it runs for it bit drawing like 600mA

600mA, how many volts?

The amps don't matter, it's the watts.

The TLC controls the anodes and 8 P-channel MOSFETs control the cathodes

How does that work? The TLC is a current sink, it has no business being connected to an anode.

It's all run at 5V.

You're right, I am thinking backwards.  The TLC is the sink therefore the negative side therefore the cathodes.  It is connected to the columns.   I scan the rows with the P-channel MOSFETs, the positive side, the anodes.  The matrix:

http://ftp://imall.iteadstudio.com/LED_matrix/Specs_IM120601006.pdf

[Reply #5 on:]

My hexome project uses three TLC5940s to multiplex an 8 by 8 matrix of RGB LEDs.
http://www.thebox.myzen.co.uk/Hardware/Hexome.html

[Reply #6 on:]

Hi!!

I have used some "led pixels", which are RGB leds with an integrated controler.

I used the ones with the WS2801 controler.

It is a 3 channel PWM shift register which can be "cascaded" with as many ws2801's as you want.
They are perfect since they support 3 channels (one for each color) and they support 8 bit color per channel.

Cheers!!

[Reply #7 on:]

My hexome project uses three TLC5940s to multiplex an 8 by 8 matrix of RGB LEDs.
http://www.thebox.myzen.co.uk/Hardware/Hexome.html

Thanks, I am going to check my project against this when I get home but my project is basically the same.  Differences are that I run the lines from VPRG and DSPRG to the Arduino instead of tying them to GND/VCC so I have control of that and I use a different P-channel MOSFET and I think I am using a smaller pullup resistor (about 1K) which I don't think matters other than energy waste.  My decoupling caps are 10uF tantalum.  Would that matter?  But the general idea is the same.  I am looking at this schematic:

http://www.thebox.myzen.co.uk/Hardware/Hexome_files/Hexome%20processor%20schematic.pdf

You have your TPC6108 mislabeled as TCP6108 (probably your fingers were thinking TCP/IP).  What is your value for IREF?

[Reply #8 on:]

My decoupling caps are 10uF tantalum.  Would that matter?

I would say yes, you should have a ceramic cap of 0.1uF as well to get a better high frequency response.

You have your TPC6108 mislabeled as TCP6108

Thanks.
It is actually my dyslexia kicking in, it causes me to swap the order on some letters. Some letters are much more prone to being swapped than others.

[Reply #9 on:]

My decoupling caps are 10uF tantalum.  Would that matter?

I would say yes, you should have a ceramic cap of 0.1uF as well to get a better high frequency response.

You have your TPC6108 mislabeled as TCP6108

Thanks.
It is actually my dyslexia kicking in, it causes me to swap the order on some letters. Some letters are much more prone to being swapped than others.

Thank you.  I will make that a weekend project to add some more decoupling to the TLCs.  Basically this is the way I do decoupling now.  I run an extra 0V line above the power bus which sits above the chips, everything in a row.  The caps are polarized correctly and soldered between +5V which more or less directly goes into the chips and the 0V line.

Code:

OV-------------------------------------------
     |          |          |          |
     C          C          C          C
     |          |          |          |
5V-------------------------------------------
     |          |          |          |
    [TLC5940]  [TLC5940]  [TLC5940]  [TLC5940]
           |          |          |          |
OV-------------------------------------------

I assume I can add the extra .1uF caps in parallel, more or less like this, right?

Code:

OV-------------------------------------------
     ||         ||        ||        ||
     CC         CC        CC        CC
     ||         ||        ||        ||
5V-------------------------------------------
     |          |          |          |
    [TLC5940]  [TLC5940]  [TLC5940]  [TLC5940]
           |          |          |          |
OV-------------------------------------------

[Reply #10 on:]

Not sure about mupltiplexing but I am using a ws2803 (18ch version of ws2801) and no issues with burning up. I had the same trouble with the tlc5940's and hate them now for being so easy to blow up, the ws2803's have been going like champs and don't burn up if you flick fluoros on n off. They're current source, I'm running them through hex inveters to n channel mosfets to drive rgb strips. Just wish they had built in inverting like the ws2801 has.

[Reply #11 on:]

Not sure about mupltiplexing but I am using a ws2803 (18ch version of ws2801) and no issues with burning up. I had the same trouble with the tlc5940's and hate them now for being so easy to blow up, the ws2803's have been going like champs and don't burn up if you flick fluoros on n off. They're current source, I'm running them through hex inveters to n channel mosfets to drive rgb strips. Just wish they had built in inverting like the ws2801 has.

These have been mentioned several times.  I am going to have to buy a few give them a shot.  I will use these for PCB Project #2.

[Reply #12 on:]

a combo like this would be the as1130
offers 132 leds with an 8bit pwm per led
www.ams.com/as1130

[Reply #13 on:]

a combo like this would be the as1130
offers 132 leds with an 8bit pwm per led
www.ams.com/as1130

I've seen that chip mentioned before.  My complaint is that it doesn't drive any commercially available matricies (if I am wrong, please point one out).  Page 15 of the datasheet gives the matrix that it will drive and it is unlike anything avilable.  So unless you want to make your own PCBs and do an enormous amount of soldering, you are SOL.

http://www.ams.com/eng/content/download/185846/834724/105034

Neat demo here:



Obviously it is a capable chip.

[Reply #14 on:]

I wonder if I can daisy chain multiple TLC5940 ICs? I need to drive 25 RGB LEDs.