PDA

View Full Version : Issues with Olsen595 56 channel setup



marquisite
11-19-2007, 06:32 AM
I had a friend build a bunch of computer controllers using the Olsen595 method by having a 'brain' which splits the signal up into 8 groups of 7 equalling 56 channels.

Now my problem is that for some reason any lights hooked up to a channel from 29 to 56 seem to be reading the instructions for another channel in the 1 - 28 range.

For example, the roof on channel 34 on the roof could turn on and then channel 6 turns off causing channel 34 to blink off for one 'step' and then turn back on.

The LPT port on the computer is set to EPP mode.

I tried a seperate computer, software, cables, and switching around the controlling boxes without success. The only thing which hasn't changed is the 'brain'.

My question.... is there anything which would be potentially causing the brain to make lights on channels 29-56 (one side of the brain chip) to malfunction such as physical barriers of the LPT port or cables (shouldnt think so as I thought you could have up to 1024 channels if you really wanted to), or is it simply a defective brain?

Last year I had no problems with the hardware as I only had 14 channels

Dan Ross
11-19-2007, 07:00 AM
How are you bringing power to the board?

marquisite
11-19-2007, 07:04 AM
The PCB gets power from inside the controlling computer the same way that a hard drive does. The only difference is the connector has only three pins. Two wires are soldered to the PCB whilst the other one loops back to the connector

RavingLunatic
11-19-2007, 09:21 AM
Since this is your own design, time for some questions:

1) Are you using Vixen to control it? If so:
-- Are you using the Olsen 595 plug-in with it setup for 56 channels?
-- What happens when you use the test channels panel and just turn on 1 channel at a time?

2) Do you have a schematic or board layout to post? Only way to tell if it is a hardware/design issue.

marquisite
11-19-2007, 10:16 AM
This isn't my 'own' design, it is designed by a friend.

I primarily use Vixen with the Simple595 or Olsen595 (with 1 to 56 set on one port) plug in, however I also tried an archived copy of Comet to make sure it wasn't a software issue.

The test channels feature works fine because channels 29-56 seem to be getting 'double' instructions. They'll respond to their own instructions, but also respond to instructions of channels in the 1-28 range.

So when it comes to running a sequence, if the roof is told to light up, it will light up but will turn off for one step (about 50 milliseconds) and turn back on if say a channel in the garden (up to and including #28) turns off.

The designer recommended I get seven replacement 74HC595 8-bit Serial-Parallel IC's in case they are faulty. I have suspicions this is not the cause though. We'll see how it goes I guess.

RavingLunatic
11-19-2007, 06:13 PM
Try using the troubleshooting tips in the wiki at:
http://doityourselfchristmas.com/wiki/index.php?title=64_Channel_Olsen_595_Controller_As sembly_Instructions

Without knowing about the design, I would still be leaning toward bad signals coming from the parallel port. If you had a oscilloscope you could check to see if additional pulses were being sent. Most likely not since you already have the BIOS set for EPP.

One of the most common problems with boards running off of the parallel port is that the port signals are not sufficient to provide reliable data for 595 chips running with a 5 VDC supply.

Try lowering the voltage going into your Brain board for the 595 chips (some have put diodes inline with the voltage to get it to drop), down to under 4.5 VDC should help out. Or use pull-up resistors on the data, strobe and clock lines to bring them up to 5 VDC when HIGH.

marquisite
11-20-2007, 02:16 AM
My friend said he built the board based on a design from Computer Christmas, but I'm not sure which one and he can't remember the link.

He is yet to respond to the DC supply voltage possibility.

I've changed over three of the seven chips without a single bit of improvement.

I think for next year I'll be scrapping the current Olsen595 'brain'... this year its nothing but a headache.

When using Vixen its almost like its teasing me... it looks so 'awesomeness' on the virtual preview but not so... great.... when applied to the physical display because of this issue.

I'm not sure if the problem could cause a fire either as this problem does cause lights on the roof to flicker faster than a single step in Vixen (due to the roof getting its own instructions and another light set instructions too. In Australia, just about all outdoor lighting runs on 24vac adapters, and I noticed an adapter for a light set connected to one of the troublesome channels was hotter than one that was connected to a functioning channel.

jpb
11-21-2007, 06:28 PM
I would be trying "pull up" resistors on the data and strobe pins as mentioned in other threads. I found this fixed my problems when I had channels misreading the on and off commands.

Jon

marquisite
11-28-2007, 09:23 PM
So turns out the problem wasn't anything difficult to fix, just hard to find.

The cable going from the computer to the brain was only 18-gauage with the usual 25 pins and not the full 25-guage obsolete version. After two of these cables were bought and trialed and the problem still existing, a third cable was bought and this time I paid attention to the thickness of it.

Who's big idea was it to sell parallel cables with 7 less wires inside? Weird.

wjohn
11-28-2007, 11:27 PM
did you need those extra? China thought u would not notice!!