LC: Problems with an external LED Display

[madmax]

Hi!

Back again after a long time. I built a transport section LC-box with an external bigger

LED display for Time Code/BBS and Status Digits as remote control. It has some switches, LEDs and

a LCD display too. It is using a PIC 18F452 and with some help of TK (thanks) it works fine. But…

I’ve still problems with the external bigger LED display. It is not reproducible

dark or bright (see pictures). Even when my box is not connected with my DAW via MIDI.

It is even dark after switching it on or it is bright.

When I push some of the transport keys (play - stop or …) then it changes from

dark to bright and remains bright. The other keys do not work. When I play the DAW

after a while the display changes to dark. Pressing stop or pause or play (can’t

remember exactly) it changes to bright.

I want to have it always bright so I can read the display very well.

Do you have some ideas for me? Hardware or software problem?

Thanks, Max

—extra information—

1. Core MBHP V3,

Necessary hardware modification on core to drive the external display using a ca. 1m long

5-6 Pin shielded cable:

PIC18F452 Pin 19 (SO), Pin 21 (RO) and Pin 22 (SC) are driven by two gates of an 74ALS1004 in

series each.

(74ALS1004: hex inverting drivers, so I’ve to use two in a row for each signal,

http://www.ti.com/lit/gpn/sn74als1004)

2. IO

-Jog wheel

-Switch keys: REW, FF, Play, Stop, Rec; F1/5, F2/6, F3/7, F4/8, shift (->), Scrub/Click, Loop/Pre-Click, Beats/SMPTE, Toggle

-LEDs: 8 pcs.

I use two 74HC165, one HC595

3. External LED Display for Time Code/BBS and Status

-MBHP_DOUT_X4_V2a (made the display connections working)

-for 74HC595 and 12 LED displays

Supply voltage measured at the external display (both situations): ca. 4.95V

4. Firmware:

-MIOS 1.9g and midibox LC v1.6e. I did an update recently. Before i used

LC 1.6a, and MIOS 1.6e I think. No difference.

5. Pictures are shoot with same camera settings.

[sneakthief]

The first thing to always try when a MIDIbox starts misbehaving is a beefier power supply going into the 5v regulator. Measuring the voltage after the regulator won’t always give you the full story about deficiencies in the current being delivered.

[madmax]

Thanks for the information. I have already a good power supply but tried another one, no improvement.

Further tests:

Switched it on and off about 30-40 times - but no operation with the DAW only switching -> all ok.

Switched it on and ran it several hours without DAW connection and operation too (at least with zeros on the display) -> all ok.

It’s mysterious. Is there any HUI/LC command to dim down the display? My Frontier Design Alphatrack dims also

down after a while. Perhaps it comes from the DAW over midi as command???

Max

[technobreath]

Checked soldering if there is some bad joints somewhere? Maybe a ground failure somewhere.

If you had it working earlier and u did no funky changes in software, why would suddenly led display start to fail.

It doesn’t do it by itself that’s for sure.

In my job I have had a lot of training in searching out errors and correcting them. I thought it was sometimes difficult to find them and keep my head straight, but what my experience tells me is that there is always an explanation, often a simple one, but that don’t mean its simple to find. Often u gotta go through lots of stuff to do it right.

Psu isnt tricky. It can still be psu problem, if both psus u tried are somehow inferior…

Have u actually calculated how much amp you need and make sure it is enough?

Linear or switching powersupply?

[SDRman]

:unsure: This might seem silly, but your display itself (pcb etc) doesn’t have an “Auto Dim” function?

Just a suggestion, with 1m display cable, you might have some instability due to parasitic oscillations…maybe more decoupling might help mate…test display, with shortest possible cable?

[technobreath]

Oops I didn’t notice the long cable… May cause trouble as described in above post.

[madmax]

Hi !

The problem occurs only in combination with DAW operation! If I don’t use my DAW with the sequencer,

the display is bright for hours and lots of switch-on and off’s.

I’m using a commercial DC/DC converter Input 9-36V out 5V with 2A.

The long cable may be a problem, but therefore I added the ALS driver gates. Before adding those

it couldn’t work because the uC is not capable to drive the ~100pF line capacitance.

May be I’ll feed the display with an extra power supply just to find out if the PSU has

a problem.

[technobreath]

Ok, to get this factual I made some calculations.

You don’t say how many segments the led display are, but my calcus assumes that there is three 7 seg displays on your external module.

Let’s say each of the segments draw 20ma, wich is pretty standard for a red led. That multiplies to roughly 450ma.

At 5 volt the voltage drop over 1m ribbon cable (estimated to be about 0.16mm2 wich is about average size) is then about 2%. Wich equals like 100mv. That means with display fully on it has like 4.9v to use.

I wouldnt hesitate to assume that if power is the only factor here, it would be enough. So in other words, cable lenght and size shouldn’t be the problem.

How is it wired? In some kinda funky way together in a matrix of some kind, so that when the box does certain things the power available to the display suddenly goes down?

Drivers I don’t know about, but are they logic buffers? If they simply amplify the current at the input it may be a problem, but if schmitt triggeted, they should give u 5v with input of 2.5v and up.

When seeing the calcs, it shouldn’t be a problem at all.

What kinda signal is driving the led? Pwm or linear?

If software is making the leds dim, there would be some kind of way to figure out what the software is doing at that times by looking at the code. The lc only receives MIDI from the daw. Find out what MIDI signals goes to the lc when it happens with a MIDI monitor, then check the source code to see what the lc actually does with the MIDI commands…

But as said earlier, the problem is probably much simpler than u believe… It can be a short somewhere, a bad soldering somewhere etc that makes signals go crazy.

It’s hard to search for faults when you don’t have a complete picture of how the system is wired and works both hardware wise and software wise.

Only solution is to get a complete picture, then isolate the fault as best can be done… It can be many things, but when u find the fault u will notice that it was all very logical. No fault appear by itself. Hehe. Encourageing - I know hehe.

[technobreath]

A word extra. If current draw from led display is 1000ma then voltage drop is about 220mv…

Then again, if the lc and display use more current than the psu can deliver u have a problem. If so, a external psu for the display would solve it. Only take care to not wire the extra 5 volt so it gets to the lc. That is trouble. Tying ground together will be ok.

[madmax]

I’m using 10+2=12 displays, see pictures. No built in dimming.

I did some tests.

-if there is no DAW connection / operation the display remains bright for hours, see picture.

-there is 100mVpp ripple on the 5V supply in the display.

-Voltage drop is not high, I’ve 4.9V there. So I’m within 5V±10%.

-line is 2m long. That was the reason to built this controller. I want to see where I’m in the song.

Therefore I need the LED display separated.

-Improvement: I added 1k to ground to all serial signals to reduce reflexions on those lines.

-during one tests the display went dark again. Now I want to find out which keys bring the display back to

bright (stop and play I known already). I pressed toggle. Only once. DAW was still playing back.

Interesting, what happened.

The display switched off. Restarted, began at all zero: 0000000000 but bright. counting started also at zero,

but not on the small LCD display. After some bars LED and LCD show the same content.

It looked like a reboot on the LED.

Next test: vy. short cable.