Re: [Nottinghack] Need some help with Our Raliway Clock

[Daniel S]

Hi Chris,

> So far I/we have built the shield, added an Ethernet
> shield on top of an arduino and powered it through a
> Bench supply which works well and shows the current
> draw below the fuses rating but when I/we connect it
> the clocks PSU it blows the 500ma fuse and it also
> taken out the 160ma one also.
> I have ordered new ones.

Which pins on the PSU out connector were you using?
I've just tried tried powering an arduino off this board,
and it powered up fine, no blown fuses, etc. (up until
now, I'd only powered the arduino over USB).

I've attached a image of the pins I'd suggest using; it
should be ~12v, which is a little on the high side for
an arduino (regulator will get quite warm...), particularly if
you're putting an Ethernet shield on it, but it should
be ok for a test.

> I'm trying to find a schematic for that board and wondering
> if you had one or able to help workout the pin outs on
> that connector?
Sorry, I've not got a schematic for any of it.

Daniel

[Michael Erskine]

On 18 November 2012 05:48, Christopher Hills <oric...@gmail.com> wrote:
> I'm going to contact the lead developer at infotec to see it they have a
> schematic for it, they did send me software for the display but it was for
> the newer signs that the company produce. He did say that he was going look
> for some old documents; but not herd back, which was back in July.

Hi Chris, That's really cool - having worked in the rail industry for
12 years I would've expected cease-and-desist letters or some other
legal action!

[Martin Raynsford]

That must be pester power, when I took a brief look at it all and contacted infotec they said they were happy for me to play with it and make it do things but they were unable to help me.

[Roger Light]

Hi Chris,

> Yes, it did get a bit warm, the regulator is 7805 (5v @1amp). I'm thinking that the amp rating could be a bit higher, which will make it run cooler. We/I also attached a heat sink to it.

Power dissipation in linear/ldo regulators is actually defined by

P = (Vin - Vout)*Iin

Where Iin is usually very close to Iout. High input voltage with low output voltage tends to hotter devices. You can use the power dissipation above along with the thermal specs from the datasheet to calculate the expected junction temperature (and whether it exceeds the maximum). Adding a heatsink drops the junction to ambient thermal resistivity so means you get hotter slower. Thermal design is worth spending a few minutes on and it isn't that complicated. This page gives a good overview:

http://daycounter.com/LabBook/Linear-Regulator-Thermal-Design.phtml

Cheers,

Roger

[Christopher Hills]

Hi,

I have an update on our display.
I have used a Proto Shield in order to get a better build of shield, this has the 20 Pin connector and  a 14 pin connector for the PSU OUT power connector attached also an Ethernet shield. USB power was not enough so using a bench power supply ruining at 7v @200mA (have seen it peek to 400mA. Have an idea on a circuit using a switching regulator (so I can use the 12v from the back lighting for power) which should run cooler than a linear one but more on that later.

So far I'm having trouble getting the clock test (hackspace logo) to show correctly using Daniel's library as you see if you view the video below.
https://www.dropbox.com/sh/ox9ja3i1owvczsy/UcX8yrDgsj

Upon power on the display fills with lit up leds (not all of them) then loads the image. In the wiki http://wiki.nottinghack.org.uk/wiki/BigClocks it states that we are driving just one set of clk/data/strobe pins which is glass 1. I been able to workout that the glass 1 (name on the PCB) is the one we are driving and it's connected to the other one diagonally across from it.
Glass 0 clk/data/strobe pins are not connected on the shield. Is this correct? as the wiki shows the display working well.
What is causing  the leds on glass 0 to stay on?

Thanks for your time

Chris Hills   
      


On Saturday, 17 November 2012 07:09:53 UTC, Christopher Hills wrote:

Hi,

I'm a member of the Brighton Hack space. I'm currently working on a project involving one of the big railway clocks you gave to various hack spaces.

We have removed the outer casing and use one side to show the time to the people in the hack space, keeps good time as well.
I'm working on the other side which is a repeat of the other.
I'm building the arduino shield which was made by one of your members (Daniel)

I'm trying to take it one step further so able to draw power from the PSU board that supplies power to the 2 fluorescent tubes mounted on the rear metal panel. So far I/we have built the shield, added an Ethernet shield on top of an arduino and powered it through a Bench supply which works well and shows the current draw below the fuses rating but when I/we connect it the clocks PSU it blows the 500ma fuse and it also taken out the 160ma one also.

I have ordered new ones.

I'm using the 14 IDC male connector marked PSU OUT which then attaches to PSU IN on the clock controller board so it's able to power it.
The board has these markings  12001P-8  BCP3 Brunel Clock PSU

I'm trying to find a schematic for that board and wondering if you had one or able to help workout the pin outs on that connector?

Thanks for time
Chris Hills
Brighton Hackspace 

[Daniel S]

Hi Chris,

I'm not quite sure why it's not working for you. The only thing I can
think of at the moment is the ethernet shield - have you tried it
without this?

Regarding the clk/data/stobe pins for glass 1+0, it does drive both,
but not simultaneously like the original controller. This is the
reason for tearing on refresh; it updates the top half of the display,
switches to the other board, then updates that.

The key is pin 14 - this is the board select pin. Whether this is high
or low determines which set of the pins the buffer board takes its
clk/data/strobe from. I assume the reason is so both boards could be
identical.
With the original controller, this is static - and inverted by the
first board before being passed to the second board (so one sees high,
the other low, so they pick up data/stobe/clk from different pins).
My firmware toggles this to change which board is output to, not
great, but to does work

Daniel.

> --
> You received this message because you are subscribed to the Google Groups
> "Nottingham Hackspace - Nottinghack" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to nottinghack...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

[Christopher Hills]

Thanks for your answer.

Didn't try it without the Ethernet shield as my plan involves using it, but as it uses SPI needed to change the pins used D10 & D2 (arduino pins) as D10 is the chip select pin for the Ethernet shield. Changed it to 11 and moved 14 (display pin) to D3  so I'm able to use the shield
and make the required changes within BigClock.h and BigClock.ccp

I taken a look at fresh copy of your library and the one I printed out and noticed an error caused by a comment tag // within the BigClock::init() function.
Yours
//digitalWrite(BigClock::latchPin, HIGH);
  digitalWrite(PIN_BOARDSEL, LOW);

Mine is reversed, came about by me hacking the code so able to workout what was going on/wrong.
So I'm going to change it back and upload the code and test tonight.
I was also having hardware issues with the arduino which I believe caused problems with the display showing incorrect data upon it.
Arduino had been replaced before I uploaded that video.

One further question how does one add a new c-source gimp image dump file, so it gets complied in the same way as the hslogo.h ?

Making the change above I see no reason for it not to work.
Thanks
Chris

[Christopher Hills]

Hi Daniel,

Tried the changes suggested but no different. So I decided to swap the boards over, to see it would invert the unwanted pixels.
This is where I found the real cause, if you take a look at the attached images you will see that ribbon cable and the plug which runs from glass 0 to glass 1 has some built resistance and the cable broke when folded.
Also one of the pins came away with the connector.

You wouldn't by any chance have a spare controller board that you send me?
I been able to make a new ribbon cable but might need a better output connection as I have de-soldered the output connector but might not have a good connection so would like to source a working board from another clock display.      

[Daniel S]

Hi Chris,

Glad you've found the likely cause of the problems.

Regarding that board - I'm pretty sure we don't have any spares as such, but we do have several of the clocks.
As they're the hackspaces, not mine, what does everyone else think about sending Chris a board from one of our big clocks?

Daniel.

[Michael Erskine]

Are there are non-functional clocks that can be used for spares? Can
Chris's board be repaired?

[Christopher Hills]

I been able to make a new ribbon cable and de-solder the connector. I did lose some of the pads on the PCB. Might have got away with losing the pads, will need to test. Also I have a new connector on order. If there is a board available, would like to have it.

[Christopher Hills]

Forgot to add, also willing to pay for packaging and postage

[David Hayward]

(I think we should send one).

[Gazz]

i'd second that, let him get his clock working with a known good board, then
he can try and fix the other board at leisure, and not be worrying any
faults are caused by the board with the torn traces.

--------------------------------------------------
From: "David Hayward" <nach...@gmail.com>
Sent: Wednesday, May 08, 2013 5:41 PM
To: <notti...@googlegroups.com>
Subject: Re: [Nottinghack] Re: Need some help with Our Raliway Clock

[Daniel S]

Ok, I've removed one of the two buffer boards from one of the clocks, and will send it to Chris 'soon', unless there are any objections. Chris: Was it just the board you want, or one of the ribbon cables? (if so, which one, the long or short one?)

Daniel.

On Wed, May 8, 2013 at 7:04 PM, Gazz <ga...@kampenwagen.co.uk> wrote:

i'd second that, let him get his clock working with a known good board, then he can try and fix the other board at leisure, and not be worrying  any faults are caused by the board with the torn traces.

--------------------------------------------------
From: "David Hayward" <nach...@gmail.com>
Sent: Wednesday, May 08, 2013 5:41 PM
To: <notti...@googlegroups.com>
Subject: Re: [Nottinghack] Re: Need some help with Our Raliway Clock

(I think we should send one).


On 8 May 2013, at 17:37, Christopher Hills wrote:

Forgot to add, also willing to pay for packaging and postage

On Wednesday, 8 May 2013 13:00:56 UTC+1, Christopher Hills wrote:
I been able to make a new ribbon cable and de-solder  the connector. I did lose some of the pads on the PCB. Might have got away with losing the pads, will need to test. Also I have a new connector on order. If there is a board available, would like to have it.

--
You received this message because you are subscribed to the Google Groups "Nottingham Hackspace - Nottinghack" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nottinghack+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "Nottingham Hackspace - Nottinghack" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nottinghack+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "Nottingham Hackspace - Nottinghack" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nottinghack+unsubscribe@googlegroups.com.

[Christopher Hills]

Many thanks guys
I only need the top mounted board. It's the first buffer board, gets it's connection from the controller, also they look the same.
If you send an email to oric...@gmail.com I will reply with delivery details.

On Wednesday, 8 May 2013 19:18:58 UTC+1, Daniel wrote:

Ok, I've removed one of the two buffer boards from one of the clocks, and will send it to Chris 'soon', unless there are any objections. Chris: Was it just the board you want, or one of the ribbon cables? (if so, which one, the long or short one?)

Daniel.

On Wed, May 8, 2013 at 7:04 PM, Gazz <ga...@kampenwagen.co.uk> wrote:

i'd second that, let him get his clock working with a known good board, then he can try and fix the other board at leisure, and not be worrying  any faults are caused by the board with the torn traces.

--------------------------------------------------
From: "David Hayward" <nach...@gmail.com>
Sent: Wednesday, May 08, 2013 5:41 PM
To: <notti...@googlegroups.com>
Subject: Re: [Nottinghack] Re: Need some help with Our Raliway Clock

(I think we should send one).


On 8 May 2013, at 17:37, Christopher Hills wrote:

Forgot to add, also willing to pay for packaging and postage

On Wednesday, 8 May 2013 13:00:56 UTC+1, Christopher Hills wrote:
I been able to make a new ribbon cable and de-solder  the connector. I did lose some of the pads on the PCB. Might have got away with losing the pads, will need to test. Also I have a new connector on order. If there is a board available, would like to have it.

--
You received this message because you are subscribed to the Google Groups "Nottingham Hackspace - Nottinghack" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nottinghack...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "Nottingham Hackspace - Nottinghack" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nottinghack...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "Nottingham Hackspace - Nottinghack" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nottinghack...@googlegroups.com.

[Christopher Hills]

Hi Daniel and all,

Received the board, thanks again for sending it to me. Placed the old one in the post today recorded, so you should get it soon.

At first it didn't show any change, so I put 17 & 9 display pins to gnd, it became a  little less flickery but still the same.
Daniel, where it's sharing the constant square wave to pins 1 & 3 going to D6.
I removed the connection to board 0 which has produced the following effect

https://dl.dropboxusercontent.com/u/15542210/2013-05-22%2000.33.02%20Ck3.mp4

As you see it's almost black on the left hand side and the contrast when adjusted shows the led's on that side updating. It's not quite there yet but close.
By any chance are you using an old version of the Arduino Software?
https://code.google.com/p/arduino-timerone/issues/detail?id=7
As I have found this issue report when using the SPI and Timer1 libraries. In which the two don't get on and when using SPI the Timer1 is not working, so I believe that driving the display and reason for the non-update on boards zero.

You stated in the Wiki it's needs a constant square wave which was not dependent upon frequency.
So I'm thinking a 555 timer will be able to provide that requirement. Going to check the original board for a crystal to see how it provided the signal required before. Also will need to remove the references in the code to the Timer1 library as I believe it won't be needed.    
Do you think, all that is needed with the code is to change D6 from an output to an input so the 555 or crystal replaces the need for the timer1 library.

Chris Hills      

 

On Wednesday, 8 May 2013 23:21:30 UTC+1, Christopher Hills wrote:

Many thanks guys
I only need the top mounted board. It's the first buffer board, gets it's connection from the controller, also they look the same.

If you send an email to <removed> I will reply with delivery details.

[Daniel S]

Hi Chris,

Thanks for sending the old board back, I'm not quite sure what's going on, or why it's not working.

Just before I sent you the board, I tested it with a sketch compiled with Arduino 1.0.1, and Timer1-v9 (I did have to change the type of the image in the example sketch from "unsigned char" to "const unsigned char" to get it to compile), and that worked fine.

Regarding the constant square wave - I honesty can't remember what frequency it required, only that being approximatively right was enough to make it work. If you want to be sure, I guess the best way would be to measure what's coming out of a working clock controller, and match that.
I'm sure a 555 timer would be able to produce the required signal, but it shouldn't be necessary.

One thought: what arduino are you using? I've tested it with an Uno (R1?). I don't see why it wouldn't work with others, but...

Daniel.