Does a DMX splitter add extra delay?

[remco_k]

If I connect a active DMX splitter, with some fixtures behind it, does it add extra delay to the DMX data?
I'd say the answer would be yes, and the real delay depends on wich brand and model splitter is used.
Anyone thougts about this?

The reason I ask is that I did a show last saturday and had en noticeable delay (around 200msec) on the strobe (Martin Atomic 3000).
For flashing a strobo on-beat, is not nice...
The company that delivered all rented equipment had connected the DMX splitter also. The one thing I saw it was a showtec or stageline splitter. 19", with (I think) 8 outputs, each with a red and green light.
On the evening itself I did'nt come up with the thougt that that device could be causing the delay, so I did'nt note the brand and model...
For the same reason I did'nt test the Atomic 3000 standalone directly on my DMX interface... (stupid of me is'nt it?)
That would have ruled out or confirmed that the splitter was the source of the problem. Now I'm looking in the darkness, thinking that it could also be my velleman K8062/VM166 interface that is just not fast enough.
FreeStyler was configured to output 256 channels, refresh interval 20 msec, previously I did tests using the same velleman interface and I did'nt have this large delay.

[remco_k]

I agree that a splitter most probably does have a very, very small delay and certainly nothing I have ever noticed by eye.

On the Martin website for the Atomic 3000 strobe under the specification it states for 60Hz power supply it can have a delay from 0-530ms or 0-650ms for 50Hz power supply.
I am assuming this is the flash (strobe) rate. It might be that it does the wait first then the light on.
It might be that it takes time for the bulb to get to temperature, it does in my car with xenon bulbs.

Just a thought, maybe way off track.

The delay you mean is a delay the operator can program through DMX from 0-530 msec (50Hz ps).
I work often with the Martin Atomic 3000, never ever had this problem with this light.
When I press the single flashbutton on the DMX controller, then the atomic 3000 instantly flashes along with no noticeable delay.
Only last time it was a mess with 'long' delays, for some reason I don't know yet.
Another possibility is that the atomic 3000 is defective. Can't test that anymore because it was rented.
I to don't really believe that the splitter adds that much delay, but I just want to rule out all possibilities.

By the way, I had the same delay when using the light in blinder mode. It just goes on about 500ms after I press the blinder button on my keyboard.
But as said, It might be a defective atomic 3000.
I saved my FreeStyler setup from that specific evening, so later this week I'm going to try to reproduce the problem with a simple RGB LED par. If I do get the same kind of delay then I'll know its something with FreeStyler or my DMX interface/driver.
I more and more begin to believe that something odd was going on with my laptop/interface.

[remco_k]

Did get a little step forward with this problem.

It seems that the Velleman VM116/K8062 (or its controller, the K8062E.exe) gets slower for each DMX channel that is above value 0.
So, if al 512 channels are 0 and I use any channel as a flash channel, than there's no delay between the time I press the button and the moment the light goes on.
Then I set 100 other channels fixed to some higher value. Now I try again - and now there is a large delay between the moment I press the flash button and the moment the light goes on.

At least, I think it is that way. I don't have any DMX fixtures at home to test that. I ordered an RGB par for that purpose, should be coming in one of these days.
This blind testing is killing me.

I should tell why I think it is this way:
3 reasons:
-There's a blinking LED on the Velleman interface, it blinks at some speed. If I set the number of channels to 512 it blinks 2 times as slow as when I set the number of channels to 256. If using FS at 256 channel mode and setting all DMX values to 0, the LED blinks 'fast'. Then I set 100 (or more) DMX channels to another value. Now the led blinks noticeable slower. Maybe indicating that it is sending (or getting) the DMX data slower. The Velleman interface gets its data from the K8062E.exe, which is started by the K8062D.dll.
-I looked at the K8062D.dll and K8062E.exe source code. Somewhat strange constructions are used to send the DMX data. Without going into detail I have reason to believe that that could be slowing the whole DMX data transfer down.
- I did test if FreeStyler is'nt the delay. I wrote an K8062D.dll that intercepts the DMX values between FS and my Velleman interface and show it on screen. FS sends the data immediately with no noticeable delay. So that is ruled out.

So what I'm writing here in this post is all guessing and some logical thinking at the moment, soon I'll test it in practice. And if necessary I will rewrite the K8062E.dll *and* K8062E.exe to something better, only the .dll will be needed when that is complete. Ofcourse only if that proved to be (part of) the problem...

[remco_k]

Hi remco_k,

Sounds very logical your testing and makes sense.

Maybe it's time time to get something like an "Enttec DMX USB Pro" or similar, that has an on-board processor and see if that eliminates the problem.

Guess you are right.
Still I have this personal issue. If I have just a little doubt that it could work good with the Velleman I must try. Its a sort of bug in me...
I have that will all kinds of stuff... I buy something, it does'nt work like I expected, I modify it, then it works (or its dead, or something in between)...
So, before buying the Enttec solution, I'll give it a shot.
I'm willing to invest some time in it. Maybe other Velleman VM116 / K8062 users can profit from that. It would be nice to have a cheap USB/DMX converter that actually works good.
New years night I have another light show to do. Ordering the Enttec now is too late to get it in here before that day... So I rather wait.

I just need to wait for the orded RGB LED par to come in so I can test my theory. At this very moment I just started to rewrite the DLL, so going to try that option anyway.

[remco_k]

Here come's the part I did'nt want to come...

I quickly ported the original velleman code to mine. When testing data transfer to the USB interface I noticed a delay under some circumstances I described earlier.
Turns out that sending one complete DMX512 packet to the Velleman interface takes 600 msec(!!!!) of time when using the device in 512 channel mode, sending all DMX channels higher than 0. And (for example) if sending channels 0...256 value 0 and 256...512 >0 then the time is half.
Thats because the Velleman 'driver' is grouping all zero's together and send them in once, so when sending all zero's, the unit is fast...
I can't believe what I'm seeing here.
Tried the same in 256 channel mode, takes 310 msec... and so on...
This interface is just way to slow, AT HIS INPUT. I'm so, so, so disapointed in this Velleman interface.

Anyway, the good news is that I now know where the delay of my flash button came from... And the source of that delay is really that velleman interface itself.
This interface is unusable for me. Using the interface in 256 channel mode has an delay of 310 msec (depending of the DMX values you send). That is way out of control....
Going to ditch it somewhere now.

Do you think the Enttec Open DMX USB will suite me? I have an adress where I can pick it up in a few days.
I find the DMX USB Pro a little to expensive for those few times I use it.

[kikabyte]

Here come's the part I did'nt want to come...

I quickly ported the original velleman code to mine. When testing data transfer to the USB interface I noticed a delay under some circumstances I described earlier.
Turns out that sending one complete DMX512 packet to the Velleman interface takes 600 msec(!!!!) of time when using the device in 512 channel mode, sending all DMX channels higher than 0. And (for example) if sending channels 0...256 value 0 and 256...512 >0 then the time is half.
Thats because the Velleman 'driver' is grouping all zero's together and send them in once, so when sending all zero's, the unit is fast...
I can't believe what I'm seeing here.
Tried the same in 256 channel mode, takes 310 msec... and so on...
This interface is just way to slow, AT HIS INPUT. I'm so, so, so disapointed in this Velleman interface.

Anyway, the good news is that I now know where the delay of my flash button came from... And the source of that delay is really that velleman interface itself.
This interface is unusable for me. Using the interface in 256 channel mode has an delay of 310 msec (depending of the DMX values you send). That is way out of control....
Going to ditch it somewhere now.

Do you think the Enttec Open DMX USB will suite me? I have an adress where I can pick it up in a few days.
I find the DMX USB Pro a little to expensive for those few times I use it.

Please let me know how u get on with the DMX USB Pro.. I have similar problems and would love to find a quick unit! I will wait for your reply!!

Merry Christmas

[remco_k]

Today I got my christmas present (how you say that in english?) from my boss. So,I decided to order the Enttec USB DMX Pro. But, because of the christmas days I don't want to rely on postal services. Here in the Netherlands packages gets delayed for days because of all the christmas cards and other junk that gets send around these days.
So going to pick it up this saturday at the webshops street address. This way, I am sure I have the unit in on time. (need it on new years night).

I also ordered a little RGB Par, a week ago. That still did'nt arrive (see above story...), I hope that will come in next monday or tuesday so I can do some quick tests with the Enttec DMX Pro. (I don't have any other DMX enabled equipment at home).

I dont expect any problems with it though. We'll see!

[remco_k]

Christmas Bonus is what I think you mean !!!!!
Nice one getting the Pro version, hope you get it all sorted in time. I am sure you will.

Have good Holidays everyone.

Yes indeed. Stupid of me not thinking of the word 'bonus'. Is actually the same word in my native language.

[remco_k]

Just for anyones information:

Last week I went to get my new Enttec DMX pro and newyears night I did a lightshow at a party from 1 to 6 a.m.
I got to choose between the Avolites Pearl Tiger and my own Enttec Interface with FreeStyler.
I choose for FreeStyler, because the Pearl Tiger did'nt have enough programs that could be ran simultaniously.
And; unlike I had with the velleman interface, I did not have a delay at all anymore. yes!
Problem solved. Now when I pressed the single flash button for the martin atomic strobe, it had already flashed before I let loose of the button. Great!
I even made a big VU-meter alike animation on 2 pixeltracks lying next to eachother, so I made one VU-meter that was 2 meters wide.
That sequence ran great at fast speed (<75msec per sequence step, because of the Cue speed was set to 150%), and ran almost all night without any hesitation/hickup/stutter or whatever.
This was not possible with the Velleman interface, I'd say.

So whats the story?
I know for sure now that the problem is IN the velleman interface itself, and the problem can only be solved by replacing the interface by a beter one like I did, the Enttec USB DMX PRO.

[kikabyte]

Brill! Thank you for your review... mine will be on the wishlist!

[djSupport]

Completley unrelated but Can I just say try not to edit too much on the enttec wiki pages at the dmx wiki enttec wrote those pages lol!