Blue starts blinking over time? (UPDATE: RJ fixed my thermal issue)

[CaptKirk]

Hi,

Any ideas why both my units (one more than the other) would develop a blue blink over time?

I set a Vixen sequence to ramp up blue to full for 30 seconds, fade down and ramp up yellow for thirty, ramp down and repeat (loop).  After a little while (1/2 hour at the most) the blue will start to blink when it is ramped up.  It will start with a single blink then progressively get worse.  Once the blue starts blinking, the blinking exists even from DMX Deck.  

Then it gets weirder.  If you touch the CC chip for blue the blinking stops.  However, you seem to need to run a program to cause the blink- If I set to manual (white) mode, the LEDs stay white and if the CC circuit was the probem I should see white blink yellow if the CC chip is thermal cycling.

It seems to be a control thing but I am not sure.  I have touched up the CC circuit and it is solid.  The PIC was touched up when I was trying to figure out why the diag firmware would not work.  And the second one has the same issue just less severe, although with DMX Deck after the Vixen loop above I had a really interesting alternating blink going for a while until the really bad one decided to speed up the blinking sequence with full blue from DMX Deck.

Again, yes thermal is suspect (why??) but why not in manual mode with full white at the hotest???

THX, Kirk

[CaptKirk]

Tried this morning with a static full blue using DMX Deck and after 1/2 hour it did not start blinking.  That may mean it needs longer to start (real thermal issue), or that the changing Vixen pattern needs to be used to recreate the issue (logic problem).  I will try longer tonight, maybe use a hair dryer (on low) to try to recreate with a static color.

Anyone have any other thoughts?


THX

[vairmoose]

On the blinking...      sounds like either an intermittent power (or ground) ...  in the power side to the LEDs (connectivity lost when leg heats up  (or cools off)...   or similar connectivity issue in the control side for the blue...    either at the  pic or at the led control chip?    if it is a connectivity issue,  just  touching  a probe to the offending pic leg would cause connectivity... invalidating the test results....         

(just a random gathering of thoughts that have zero beneficial value) 

Larry. 

[CaptKirk]

The touching I was talking about is literally putting a finger on the board in the area of the CC.  I would not think the CC circuit should be that touchy that a luke warm temp would cause the flashing.

I had not thought of pushing on IC legs with a probe so I will give that a try also after the thermal tests.  Very interesting that both units would have the same bad connection though- just the shear randomness of the universe says it should not occur on both.  The only common thing between the two is the DMX signal from a single dongle and computer/Vixen (or DMX Deck??)

[RJ]

Is it possible the driver chip is not making good contact with the pcb so its not dumping heat to the copper under it? When you push on it you pull heat and maybe make it touch.

Try squeezing a little heatsink compound under the chip to transfer heat and see if it stops.

RJ

[CaptKirk]

That is brilliant- I wondered about that copper near the CC chips.  I had some CPU termal paste I injected as much under as I could and put on the edges of the CC chips.  That seems to have been the issue- odd it did not occur in MAN mode.  Thank you for your time again RJ.

Now I only have a fading of the red channel over a long period of run time.  The yellow in the sequence (100% red 75% green) gets a shade of green to it over time.  The red CC has been thermal pasted to the copper along with the green and blue channels.  It is not bad but it is noticable- that just may be the way it is...

Kirk

[egenoup]

What number level does this happen (i.e. 10 or 158 or 255)?  Does it happen across the range? does it start solid on and then starts blinking after a while? 30 secs or 30 mins?

Thanks,

Charles S.

[CaptKirk]

What number level does this happen (i.e. 10 or 158 or 255)?  Does it happen across the range? does it start solid on and then starts blinking after a while? 30 secs or 30 mins?

Thanks,

Charles S.

The apparently temperature issue was that the color would start out solid at full intensity (255) and then after a while (maybe a half hour) I would get one blink of the blue during a 30 second on period (see sequence info above) then a little while latter I would see two blinks during the 30 second blue period then more and more until it was blinking on and off at about a half second beat rate.  The blue never blinked when on White in MAN mode so I was worried it might be a logic issue (DMX data getting trashed somehow) but it really looks like it was temp.  Maybe when it is white, there is less current the blue channel had available to absorb so it would not go into thermal protect mode??


Adding the compound to the CC chips has eliminated that issue (ran over night no blinks) BUT I did start noticing a red fade during the yellow on that part of the sequence (the yellow started out yellow and slowly got more green).  That is the issue I am looking at now but it is not very pronounced so it does not kill my use of the lights at a Cub Scout Blue and Gold celebration for our Pack on Thursday.  This is the reason I was in a hurry to try to get these units and get them running (thank you Jim and everyone who had ideas for getting these going).

[CaptKirk]

Charles,

And I forgot to add: I believe I have seen the "flicker" issue you are having.  If you stop about 3/4 of the way up the slider on any of the colrs (not sure how to see the actual value)  I can see a slight dimming and brightening of that color.  It is not super noticable but it does seem to be there.

Kirk

[egenoup]

After a long talk with RJ and a good explanation as to why there would be more thermal discipation in the middle of the range versus the top (full on), I did exactly what RJ told me and it solved the problem (for the most part).  It is definitely a thermal issue.  I placed a piece of angle aluminum on the chip using the artic alumina epoxy and it took 10 minutes for it to flicker a few times.  I blew on the "heat sink" and it instantly stopped blinking.  I then decided to take a reading  WOW!  140 degrees!  At that temperature is where it begins to flicker. 

10 minutes of constant on is more that any one color will ever be on (in my case), so I will just epoxy heat sinks on each of the chips, or try to figure out how to just put one length of angle aluminum.  I need to clear the diode on one side and the resistor on the other.  At least now I know what needs to happen in order to prevent the Aether from flickering randomly at midrange.

Thanks RJ!!!

Egenoup/CharlesS.

[RJ]

140 deg's  F or C ?

If 140 F that is nothing and should not be putting it in shutdown.  The chip might be bad. if what you did didn't fix it then I think we have an issue with the chip.

RJ

[egenoup]

What is the part number of the "better" chip?

DOH! now that I think of it, it was 140F, which is like nothing for that chip...

[RJ]

It sounds like a bad chip. It's max operating junction temp is 150 deg C which is 302 deg F!

It's a little pricey but if you want to use the heavy duty version of this driver it is this :

You are not allowed to view links. Register or Login

If you do make sure to put a small amount of heat sink compound on the bottom of it as it's heat comes out through a pad on the bottom.

RJ