Clock rate vs Lead-in Cable Length

BradsXmasLights

BradsXmasLights

WiFi Interactive
Joined
Dec 23, 2010
Messages
605
Location
Brisbane, Australia
What clock rate are people running their 6803 pixel strips at?

I have a PIX8AD and the manual says to start with a clock rate of 750. This worked good when directly connected, and with 4core alarm cable of a 1m distance to first pixel. At ~2m, I had mixes results depending on the amount of pixels connected, voltage drop, flickering when dimming, all sorts of strange things like only 70% of the pixels working, etc.

Tonight’s testing was with ~150 pixels connected to a single SPI output. Dropping the clock rate back to 250 and using some scrap CAT6 with 2 pairs joined for clock & 2 pairs for data, i then got perfect results using a 7m length. (7m was tested because that's the amount of offcut I already had :D)

Dropping the clock back even more to 100, didn't really seem to make any real difference to the 100ms interval test sequencing I was using with Vixen.

Also 12VDC Power was from a cheapo Ray Wu unit too. Previously I thought I was getting EMI problems, but it all that weird stuff with flicking when dimming seems to be tied back to to the clock rate/lead in cable length.

I’m keen to know what other people have discovered with cable/lead-in distances!
 
Have you changed the drive chips in your PIXAD8 to the "ACT" versions? Ed changed this recently to make it drive the 180x series chips correctly and it may help with 2801 and 6803 types too.

There is a post in Ed's section of the forum I believe. (maybe search for 180x)
 
Short answer: It may help and it shouldn't hurt.

Long answer: I think Ed tested the new chips and found no problems with them on all protocols, but it's not been absolutely proven that there is a benefit going to the new chips for 2801 and 6803. It did definitely help the 180x types.
 
Hi Driver,
I had similar issues when setting up my 6803 pixel strip.
My setup has a Ray Wu 12V PSU, PIXAD8 and 62 strip segements of 6803 strip (approx 6.3 metre in length).

When connected directly to the PIXAD8 with a short 1 metre lead I had no issue and everything tested great.

I ideally needed a 5 metre extension lead from the PIXAD8 to reach the 6803 strip along the gutter in situation.
When I hooked up the 5 metre extension lead the LED strip segements stayed on with randome colours.
I never actually played with the clock rates at all but ended up having to have add a "dead pixel" half way into the extension lead and also add another V+ and V- supply wire to the first section of the strip which fixed the issue and now performs perfectly.

I wonder if slowing down the clock rate would've prevented me having to add the "dead pixel" length but I think I still would've needed to add the extra power to prevent any 12V power drop over the 6.3 metre strip length.

I hope that is of some use?

Coxie
 
Ed was saying with the new chips all the pixels tended to behave a little better than before. But were mainly for the TM18xx pixels as david said
The old chips still work well with the 6803 and 2801

My leads are all 5m to the first pixel. Running at 750. One of my leads goes to the top of the mega tree for the 2801 star. That one is 4.5m to the floor then a dead pixel and another 4m back to the P12R controller.
One thing you will have to watch is the voltage drop more than anything. Maybe this is causing the problems more than anything else.

Most of my leads are now made from 6 core alarm cable,with 2 wires twisted together for both +ve and -ve then the data and clock lines. Im still getting a small voltage drop issue but most times im injecting power anyway so it doesnt matter.

As for the clock rates, i havent played with them yet. Not really had a need to but I might to just see what reaction I get.
HTH
 
one thing people aren't mentioning, at least directly: what VOLTAGE are the lights. i suspect that the cmos chips in the 12v lights are using 12v to supply the ICs. they may have a voltage divider or something but it will vary with chip and pixel design.


imho the 74LS are putting out adequate signal levels for 5V levels for 2801 and 6803. most of the 180x pixels i have for testing are 12V. we started to see the problem on them. BUT i also think the problem could be more related to voltage than to pixel type. so 12v versions of 6803 may work BETTER with the 74ACT chips instead.


The reason: 74LS outputs TTL level 0 and 1 which are NOT 0v and 5v as you may expect. 74ACT outputs cmos levels which are usually closer to 0v and 5v, called rail-to-rail, signal levels. but input thresholds for 0 vs 1 or usually much lower (3v for 5v cmos). but if the pixel chip is really looking at a 12v, or close, voltage it's threshold might be higher. so the 5v cmos output will usually reach it.


all of this will vary based on pixels, cable gauge, distance, etc. i've tested ALL 5v 2801 pixels with a 5 meter extension cable from MPH. the 12V TM180x i've tested with 5 ft and 10 ft extenders and had very good results.


-ed
 
The pixel chips I've looked at are all 5V. The 12V strip uses a series resistor and zener to generate the 5V for the chip.

I haven't pulled a TM180x pixel apart to see how they power the actual chip.
 
That makes sense Ed - my 6803 strips are indeed 12V based. They also have the resistor & diode that David talks about.

Regarding the clock rate, is there any real advantage of a 750 clock instead of say 250? Even with the lower clock speed, I could see no noticeable difference in response speed.
 
Driver said:
Regarding the clock rate, is there any real advantage of a 750 clock instead of say 250? Even with the lower clock speed, I could see no noticeable difference in response speed.
Click to expand...


Depends on your sequence. When testing SPI clockrates I could see a difference when doing strobing effects, if I turned all the pixels on at once and off at once there was a certain time limit I had to enforce to make sure that all pixels lit, otherwise I would only get part of the string coming on because it wasn't enough time to refresh the whole string. I didn't end up using any of those sequences they were more for my own interest. Most lighting is a gradual transition or "chasing" sequences so there is plenty of time for pixels to sort themselves out.


Cable type and distance are the two main deciding factors in what clock rate you can run. Environmental effects also come into play as well (eg nearby power cabling especially power for your lights since the lights use PWM) so running shielded cable for data is often a good idea.
 
Guys, when you talk about injecting power, is that by running a cable from the V+ on the controller to the next strip, or straight from the PSU itself?


I'm running a PixAd8 with 6803 strip. My PSU's are in the garage and not with the controllers.


Cheers,


Nathan.
 
I run the V+ and the V- straight from the same power supply that is running the strip from the controller You should never mix supplies unless they are designed for it and most are not.

With the pixad8 the output power fuse will be rated to take a certain load so by adding extra strips if the load is below the rated fuse on the pixad8 then i would run the power from there instead or if you have spare outputs on the pixad8 you could use just the power output of that to supply power, this way you give yourself fused protection on your strips
 
You must log in or register to reply here.
Back
Top