12Hrs LED clock Very accurate timing

[pasanlaksiri]

This is just another clock project. The processor that I am using here is PIC16F628A. Clock timing is very accurate thanks to RomanBlack Zero-error 1 second timer algorithm.

Link1>> Zero-error 1 second timing algorithm

Link2 >> https://www.electro-tech-online.com/threads/zezj-zero-error-zero-jitter-period-algorithm.100299/

Nothing much to say about the project so here it is


**broken link removed** **broken link removed** **broken link removed** **broken link removed**
**broken link removed** **broken link removed** **broken link removed** **broken link removed**
**broken link removed** **broken link removed** **broken link removed** **broken link removed**
**broken link removed** **broken link removed** **broken link removed** **broken link removed**
**broken link removed**

PCB Arts

**broken link removed** ***broken link removed**

Silk Screen Layouts

Controle Board BOTTOM

http://www.4shared.com/file/236695715/97e5185a/Controle_Board_BOTTOM.htmlControle_Board_BOTTOM.html

Controle_Board_TOP

http://www.4shared.com/file/236695744/9d95dc89/Controle_Board_TOP.html

SSD_Board_BOTTOM

http://www.4shared.com/file/236695925/b6566693/SSD_Board_BOTTOM.html

SSD_Board_TOP

http://www.4shared.com/file/236696025/ab32f2f2/SSD_Board_TOP.html

Firmware

http://www.4shared.com/file/236744326/a9b8c28e/GT_mini_Clock_hex.html

 

[Hero999]

Nice work!

What method did you use to produce the PCB?

I've noticed that some of the tracks don't go anywhere: were they drawn by mistake and removed after the PCB was etched?

It looks like an interesting method of doing a silk screen. Have you tried toner transferring the silk screen and covering with a lay of conformal coating to give a black silk screen?

 

[pasanlaksiri]

Hi, Thanks

Silk Screen method is very much time killing job. But the out come is perfect.

to give a black silk screen?

I am little confused, Could you please explain a bit.

You mean like PRP200 Spray ?

 

[Hero999]

I'm talking about using the toner transfer process to put the silk screen directly on to the PCB so you don't have to mess around printing it on paper and gluing it to the board.

Once I've done the toner transfer, I cover it with a layer for conformal coating.

The PCB remains its natural colour and the silk screen is black, see attached.

 

[pasanlaksiri]

Wow its really lovely. I haven't try that method.

All what I am doing it directly print on the PCB using Screen.

 

[blueroomelectronics]

Really nice clock. Just my 2c on clocks I would always use a crystal either 32,768kHz (common watch crystal) or something that divides easily in binary. Another trick is use the mains frequency, in most countries it's an extremely accurate timebase thus no crystal needed.
How hot does that 7805 get when you put 16 volts (full wave rectified 12VAC) into it, I would think it might need a heatsink.

I don't see any digit driving transistors or LED segment current limiting resistors, why did you decide not to include them? The ancient Microchip 16C84 clock app note doubled up some output pins to gain more current. The Mondo Superprobe also avoided resistors but claimed it worked because only one segment is ever driven at any moment.

Looks very professional. Are you planning to sell them as kits?

PS just a sidethought the 16F628A has the very flexible TMR2 special mode, it might be possible to use it as a 1 s or 0.5 s timebase without the fancy Roman Black 1 second fix.

 

[Boncuk]

I've noticed that some of the tracks don't go anywhere: were they drawn by mistake and removed after the PCB was etched?

They're actually no tracks, but "fillers". Making a single sided board the ground fill won't get to any position of the PCB unless you use wire jumps.

Fillers help etching the board evenly. Large copper areas require more time to etch burrying the risk of underetching narrow traces.

Boncuk

 

[pasanlaksiri]

Really nice clock. Just my 2c on clocks I would always use a crystal either 32,768kHz (common watch crystal) or something that divides easily in binary. Another trick is use the mains frequency, in most countries it's an extremely accurate timebase thus no crystal needed.
How hot does that 7805 get when you put 16 volts (full wave rectified 12VAC) into it, I would think it might need a heatsink.

I don't see any digit driving transistors or LED segment current limiting resistors, why did you decide not to include them? The ancient Microchip 16C84 clock app note doubled up some output pins to gain more current. The Mondo Superprobe also avoided resistors but claimed it worked because only one segment is ever driven at any moment.

Looks very professional. Are you planning to sell them as kits?

PS just a sidethought the 16F628A has the very flexible TMR2 special mode, it might be possible to use it as a 1 s or 0.5 s timebase without the fancy Roman Black 1 second fix.

Yes agreed, both of those methods that you mentioned is perfect.
But the reason that I built this clock just to test Roman Blacks algorithms.

Yes, I am selling it locally as a kit.

 

[ericgibbs]

Yes agreed, both of those methods that you mentioned is perfect.
But the reason that I built this clock just to test Roman Blacks algorithms.

Yes, I am selling it locally as a kit.

hi,
As I understand it, the 1second accuracy obtained by using software will not correct for any drift in the frequency of the crystal itself.
Programming tweaking only sets the program timing to precisely 1 second at a particular crystal frequency.

If the crystal frequency drifts due to say ambient temperature changes so will the 1 second period.???

 

[Wond3rboy]

Hi Hero, Bill, Boncuk, Eric and Pasan,

@ Pasan commendable work.

@ Hero, i tried toner transferring the silk screen but the toner doesnt stick to the top like its clearly stuck on your PCB.

 

[Hero999]

@ Hero, i tried toner transferring the silk screen but the toner doesnt stick to the top like its clearly stuck on your PCB.

May be it's too smooth so there's no key for the toner to cling to?

Try sanding the top side of the board before performing the toner transfer.

They're actually no tracks, but "fillers". Making a single sided board the ground fill won't get to any position of the PCB unless you use wire jumps.

Fillers help etching the board evenly. Large copper areas require more time to etch burrying the risk of underetching narrow traces.

It also saves etchant too.

I was talking about the tracks pointed to in the following attachment.

Maybe its the filling algorithm going overboard?

It's no problem on a digital board like this, unless it's part of a crystal oscillator but it could be a problem if the track carries an op-amp's inputs which are a very small signal.

 

[Wond3rboy]

Thanks Hero, about the fillers i only use them for Ground or Power Planes also some times just stand alone copper blobs, keeps etching time to a minimum. I guess they will cause problems when doing wireless but never got the opportunity to do so. Tracks are supposed to left alone.

 

[Hero999]

Another thing, I notice you've used some acute angles which isn't good practise because it can cause problems with etching. You might think I'm being picky and it is true that with the track thickness you've used it's unlikely to cause any problems but it's good to get into the habit of avoiding acute angle, so you won't run into any problems if/when you do use thinner tracks.

I've highlighted a few acute angles and some obtuse angles, which are bad and good respectively. There are a many more acute and obtuse angles on your board, I've just shown a few to give you the general idea.

 

[pasanlaksiri]

Thank you very much Hero for your great help. I am an amateur. Guys like me needs continuous guidance from pros.

Thanks all..**broken link removed**

 

[Mike - K8LH]

Why exactly are you posting these pictures and links? Why post a link for the hex file if you don't include a schematic?

If you're just showing off your nice work then please excuse me (I thought you were posting a project but I guess that's not the case)...

 

[blueroomelectronics]

I would expect the digits to vary in brightness depending on how many segments are lit without the segment resistors. Why would you omit them?

 

[pasanlaksiri]

I would expect the digits to vary in brightness depending on how many segments are lit without the segment resistors. Why would you omit them?

Actually the brightness is even. I tried it with few types of SSDs. I didn't see any brightness varying. But the green colour SSD brightness level is little lower than the red. But not with each segment.

 

[Boncuk]

Why exactly are you posting these pictures and links? Why post a link for the hex file if you don't include a schematic?

If you're just showing off your nice work then please excuse me (I thought you were posting a project but I guess that's not the case)...

You've got the PCB layout and a silk screen. What else do you need?

 

[Mike - K8LH]

You've got the PCB layout and a silk screen. What else do you need?

You're joking, right?

 

[zorbzz]

Well thats what I am gonna make mine from.

Just can you tell me why use a crystal when the pic16f628a has one built in?
Im pretty new to pics so I dont know sorry.