Ren24LSD - Troubleshooting

[michaelc]

I've build my Ren24LSD and I'm trying to test it. I've done the basic tests and the correct voltages are appearing where they are supposed to on the PICS and other chips. Just nothing happening on the output jack. Here are my symptoms.

I've created a test sequence in Vixen using 48 channels. The output is going to an ENTTEC Open DMX USB. The software 'appears' to recognize the dongle because it does not give me a missing error. I confirm this by unplugging it and restarting the program, where it DOES throw an error.

SO I run the Test of all channels and according to Vixen, it looks like it's working but absolutely nothing is happening. My PICs are programmed with all chips set to a DMX default start of 1 and for this test it is the ONLY board I'm connecting. I also have the TERM jumper on.

Is there a way to test for a valid signal coming from the USB dongle. I have an inkling that the issue might be the dongle but since this is a brand new build on the board, I've got to have an open mind. A nice signal status LED would be a nice addition to future versions. Any help would be greatly appreciated.

[dirknerkle]

Do you mean a Ren24LV or a Ren48LSD? I'm confused...

[michaelc]

Ren48LSD

[budude]

Did you enable the DMX Add-in (not just the Enttec Open Plug-in) - if not do so and try your tests again.

[michaelc]

Brian, yes, I forgot to mention that in the description. The DMX addin is checked. I really think it is dongle related. I tried a test late last night connecting one of my LOR boards directly using the dongle and it had a hard time seeing/keeping the signal as well.

I'm going to try a different computer to see if it's a computer/port issue. My current laptop is running Windows Vista64. It all seemed to work last week but now I'm having issues. Troubleshooting 101, remove all accessories and start at the source confirming operation!

[michaelc]

One thing I noticed tonight when trying to troubleshoot the Open DMX. According to the online documentation in the SUPPORT section or ENTTEC, the dongle should show up in the USB section of the device manager as an FTDI device. Mine is popping into the Ports (COM &LPT) section.

Are there any tricks to get this thing to run on a Vista64 machine?

[budude]

Vista - - ugh... I think it's correct to come up as COM port - I haven't used my Enttec Open for quite a while but that sounds right to me anyway. I'll try and find it to double-check. I think my daughter's machine is a Vista/64 (my last Vista box in the house). Is there any way to test on an XP or Win7 machine - or at another PC in general?

[michaelc]

We must think alike. I'm currently trying to configure on another PC.

[mschell]

Just to be sure - you did load the DMX firmware on the Ren48LSD, right? The normal Renard firmware will not work with the DMX output dongle.

[michaelc]

Yes, I am going to re-image the PICS to be sure, but I'm positive that's what I used since I went in and did the DMX specific edits.

Here is the top part of the code that is in the source directory of MPLAB-IDE that was used to burn the PICS

;; filename: renard-fast-20080307.asm
;; Copyright (c) 2006-2008 Phil Short
;;
;; This program is provided free for you to use in any way that you wish,
;; subject to the laws and regulations where you are using it. Due diligence
;; is strongly suggested before using this code.
;;
;; The Author makes no warranty of any kind, express or implied, with regard
;; to this program or the documentation contained in this document. The
;; Author shall not be liable in any event for incidental or consequential
;; damages in connection with, or arising out of, the furnishing, performance
;; or use of these programs.
;;
;; --------------------------------------------------------------------------
;; NOTE:
;; This code is configured for regular use (the firmware generates 30 uS (at
;; 60 Hz) pulses to drive the SSRs, synchronized to the zero-crossing input
;; pin). Here are the DEFINE statements to change for PWM use (generates a
;; variable width pulse sync'd to the zero-crossing signal) or DC use
;; generates a variable-width pulse, not sync'd to anything).

;; PWM_map determines which bits should be 'PWM' (normally 0, for all non-PWM).
;; It is binary-encoded, with bit 0 corresponding to channel 0 through bit 7
;; channel 7. As an example, the value 67 = 64 + 2 + 1 would have three
;; channels (0,1 and 7) as PWM and the rest as non-PWM.

#define PWM_map 0xFF

;; DC_build should be either 0 (for AC operation) or 1 (for DC operation). It
;; should remain 0 if PWM_map is anything other than 255

#define DC_build 1

#define CLOCKRATE 18432000
#define BAUDRATE 250000

;; CTR_LOCKOUT should be 0 for the DC build, can be a somewhat larger number
;; for non-PWM or non-DC builds.

#define CTR_LOCKOUT 0

;; Bitmap of outputs which sink current (normally 0xFF, for all 8 outputs).
;; This is binary-encoded the same way as the PWM_map 'variable'. A '1' bit
;; means that the output sinks current (is active low), a '0' bit means that
;; output sources current (active high).

#define SINK_map 0x00

#define DEBUG_DMX 0

;; The DMX start address should be between 1 and 505 inclusive.

#define DMX_START_ADDRESS 1

[budude]

Looks right - the code is available in the File Library already compiled also (thanks to Mark) - guess I should update the Wiki to point to that...

[michaelc]

I was playing around with the starting address so I had to compile it anyway. Besides using MPLAB-IDE is so easy, I just copy the source into the source directory, pick my chips and hit the button. I then verify each chip before proceeding on.

[michaelc]

I've pretty much confirmed my cabling and DMX dongle is working, (originally I had the wrong 3pin to RJ45 cable. I was using my LOR one and their wiring is crazy. They send signal down the two unused DMX pins 4/5. I've now rewired one of my adapter cables so that XLR PIN1 goes to RJ Pins 7&8 (Gnd), XLR Pin 2 goes to RJ45 Pin 2 and XLR Pin 3 goes to RJ45 PIN 1

I have additional PIC's on order (should be here Tuesday). Right now I'm running the board with only 5 of the 6 PICS (I dis-membered one early on). Do i need to do anything special to get the first 2 or 3 outputs (Channels 1 thru 12) to work? Do I need to put a jumper in to account for the missing PIC?

Also, other than the basic voltage tests across the PICS and other chips are there any element specific tests I can run to insure my components are working? I've already done some resistance testing on the traces so I know they are connected. I just can't tell if they are working.

[budude]

If you only have 5 pics, just make sure they are in positions 1-5 (they count from upper-left, to the right and then right to left) and leave PIC 6 open (lower left when looking at board with serial interfaces on left). It will work fine w/o it but you won't be able to daisy chain another board from it.

If you need to do that, then tack a wire on the backside of the board from pin 6 on PIC 5 to pin 6 on PIC 6. This will jumper the Tx from PIC 5 straight to the RS485 output normally coming from PIC 6. Of course, this will make this a Ren40LSD so you'll have to account for that as well...

As for other testing, no - not really - if the voltages are 5v across the IC rails, it should be good to do initial testing.

[michaelc]

Brian,

I've got the PICS in slots 1-5, I'm not planning on daisy chaining this yet. I just want to get it working stand-alone first. I passed all the initial testing before I put the chips in. I reconfirmed the voltages across the IC after the chips were installed. The next phase of testing is hooking up and seeing what happens. Unfortunately nothing happens. Thing I need to figure out in order:

I would like to confirm one thing. It WAS my understanding that the REN48LSD inputs were standard DMX. If that is the case then standard Pinouts for a DMX RJ45 call for:

1: Data1+
2: Data1-
3: Data2+
4: Unused
5: Unused
6: Data2-
7: Gnd 1
8: Gnd 2

BUT I just received a second Ren48LSD PCB and I started mapping out and tracing the pins to help me troubleshoot the 1st REN. Here is what I found:

1: Goes Nowhere
2: Goes Nowhere
3: Ties to Pin 3 on the RJ45 OUT
4: Goes to a resistor then the 485 pin 7
5: Goes to a resistor then the 485 pin 6
6: Goes nowhere
7: Ties to Pin 8 and also PIN 7/8 on the RJ45 Out

So, NOW I'm Totally confused. Based on the board analysis, it looks like PINS 4 & 5 are the active ones for the data and the input is not true DMX. Do I have this right? Which pin on the 485 is Data+?

[budude]

The Ren48LSD is exactly the same as the RenardSS boards - this is in the Wiki. It does not follow standard DMX wiring but it's very easy to wire up . Pins 4&5 are the active pins as you found out. If you look up the RenSS hookup info (http://www.doityourselfchristmas.com/wiki/index.php?title=File:Wiki_-_Renard_SS_DMX_Data_Cable.jpg) it shows the layout.

[michaelc]

Eureka and Success (sort of). I got the pinouts right (although I attached GND to 7/8 instead of 1/2 - I don't think it matters) and the board is working (mostly). Since I have rainbow floods hooked up, and I only powered 1/2 the board for starters, I could only test the 1st, 2nd and 4th pair for each RJ connector. RJ45 Connectors 3,4,5,6 tested perfect.

I had an issue with the 1st two jacks (ie 1st PIC). Channel 1 and channel 5 only came on very dimly when set to full on. I swapped out the PIC with another one and it as the same problem. What would cause a low voltage on outputs The only thing I can think of is possibly a couple bad transistors. I'll do some add'l testing tomorrow.

[michaelc]

BTW, Have I said Thank you to you this hour? If not, Thank again for all the help. It's a really good design and board if I can stop making rookie mistakes.

[budude]

BTW, Have I said Thank you to you this hour? If not, Thank again for all the help. It's a really good design and board if I can stop making rookie mistakes.

no worries - it's all part of the job! The circuit is pretty simple - I would try retouching up the soldering to the transistors and resistors first - make sure the emitters are grounded and the base has continuity to the PIC through the 470 ohm resistor. The schematic for the v3b is on the Wiki and that part is the same on the v3c. To help the PIC outputs are in the following order channel-wise:

odd connector numbers
1 - Pin 3
2 - Pin 13
3 - Pin 12
4 - Pin 11

even connector numbers
1(5) - Pin 10
2(6) - Pin 9
3(7) - Pin 8
4(8) - Pin 7

The v3c board has a better silkscreen to show the channel numbers in the transistor nomenclature (Qxx). Unfortunately the cost for using super cheap transistors like these instead of costlier MOSFETs for example is that it doesn't take a lot to blow them up...

[michaelc]

Well the beauty of my purpose for using this board (Rainbow Floods) gives me built in spare parts. I will have 12 'extra' (albeit already soldered) transistors to swap around.

BTW, I have this V3c I'm troubleshooting and 2 other V3C PCB's in a box waiting till next years project.

[michaelc]

Ok, after more troubleshooting this morning with a fresh set of eyes and a better mindset, I have done the following:

1) Confirmed continuity of all solder joints
2) Confirmed that the voltage across the PIC from Pin 1(Vdd) to all 8 output pins (3,13,12,11,10,9,8,7) is -4.98 volts when no input signal is present.
3) Confirmed that when all the 'working' ports go full on to a DMX value of 255, the voltage across Pin1 and the output drops to ~4.3volts
4) Here's the kicker, when the 'bad' channels go to full intensity, the voltage drops from -4.98V to -.0007volts across Pin 1 and the output pins.

So my voltage issue seems to be happening BEFORE I even get to the Transistor (unless the symptoms of a bad transistor is that is feeds voltage back up.

Thoughts?

Swapping IC's does not fix the problem.

[budude]

You should see a ~.7v drop across the B-E junction of the transistor when it's on so that matches your "good" ones. I'd have to think about the behaviour with the "bad" ones. Double check that the emitter (pin closest to PIC) is connected to ground and if so - well - get out the solder sucker and swap the transistor out. Besides soldering and the resistor (which is highly unlikely) there isn't anything else there since you've ruled out the PICs. Try one at least - I'm pretty sure that will get 'er done.

If you need replacement transistors, you can get them at RadioShack as well (I used those for the first prototype).

[michaelc]

Yeah, the resistor was showing the correct resistance. I guess it's down to the transistor. Not sure how it's sucking and making almost 5v disappear. My REN48, became a ren 36 using Rainbows and now it's potentially down to a REN34 LOL!

[michaelc]

Brian,

Is this a fair replacement from the shack?

http://www.radioshack.com/product/index.jsp?productId=2062586

the pack has 52n2222 transistors.

[budude]

Yep - I think those are the ones I used originally actually. They have a smaller pack of MPS2222A I used also but the ones you pointed to should work fine.

[michaelc]

SUCCESS. After 2 transistor changes, my REN48LSD is fully function. I' mounted it on the hardboard and into the enclosure. Tomorrow I'll mount the enclosure on the outside of the house, and run the cat5.

A HUGE THANK YOU to Brian. After some bumps, my very FIRST full DIY board is assembled and working. (I'm not counting the 3 rainbow floods, they aren't full boards).

[budude]

Awesomeness - the cost for the help is a video of it running now!

[michaelc]

It's only running the 3 rainbow's for now. IT was really a Proof of concept for this year. I wasn't sure it was going to be ready for this year so my show is still running LOR. I've got everything prepped for Vixen so I may play around tomorrow night and switch it. I will post a video though. The musical animation part of the show is only 9 minutes but I'm working on Polar Express to get my show up to 15 minutes. Any more than that, I find people get bored. So I go 15 minutes on, 15 minutes of various fades, shimmers and color sweeps, then the show again.

[michaelc]

I've got a power question related to these boards and I'm not sure if it's part of this thread or another so I'll put it here.

I'm using an old PC Computer Power Supply to supply the board. There are many 'tails' coming out of the box and all have a combination of 5 and 12 volts in various MOLEX configurations. I took a standard 4-pin molex connector (peripheral/hard drive connector) that has two grounds, +12 , +5 volt and connected the 12v and gnd to the first side of the board (my rainbows run off 12V). I want to connect 12v to the other side as well. Here's my question

On a computer power supply are each of the 'tails' coming our isolated or does everything share the same power.

Options:
1) I run another set of wires to my enclosure with another 12 and 5 V supply from one of the other 'tails'
2) I just jump the existing power to the other side of the Ren48LSD.

Electrically, I don't know if there is a difference. Can anyone clarify for me.

BTW, I"m planning on using the +5V to power a 2nd REN48LSD in the same enclosure to run different LED lights at 5V.

[budude]

I'm using an old PC Computer Power Supply to supply the board. There are many 'tails' coming out of the box and all have a combination of 5 and 12 volts in various MOLEX configurations. I took a standard 4-pin molex connector (peripheral/hard drive connector) that has two grounds, +12 , +5 volt and connected the 12v and gnd to the first side of the board (my rainbows run off 12V). I want to connect 12v to the other side as well. Here's my question

On a computer power supply are each of the 'tails' coming our isolated or does everything share the same power.

Options:
1) I run another set of wires to my enclosure with another 12 and 5 V supply from one of the other 'tails'
2) I just jump the existing power to the other side of the Ren48LSD.

Electrically, I don't know if there is a difference. Can anyone clarify for me.

BTW, I"m planning on using the +5V to power a 2nd REN48LSD in the same enclosure to run different LED lights at 5V.

On most PC supplies the 12v leads are all soldered together to one point in the supply. Some better ones have separate 12v leads for the video card power. It is best to power each side of the LSD from a separate lead depending on how much you are planning to draw. Even if using a single lead from the supply, it's best to run separate runs from DC IN 1 and 2 and then tie them together off the board.