Discussion in 'Modding' started by Explicit, 27 Sep 2004.
Shipping was $30 plus a $20 documentation fee.
I have put in the keypad routine i was given from my tutor etc.
When i go to build the code, i get the following errors:
Error : undefined symbol:
Error : _Delay (E:\Final Circuit\source.obj)
What does this mean?
Edit Found out why the error was coming up, just had to ensure the delay routine was functional
In the meanwhile I started to develop a little clock based on the russian IW-18 8-digit VFD tube.
The problem with VFD tubes is that they need an AC or DC filament voltage/current between 0.85 and 5V @ 50 to 100mA depending on the model/size, and a positive anode-grid voltage of about 20-30V.
I finished the schematic based on a 16F628, an Allegro A6812SA 20-bit shift register/VFD driver, and a National LM9022M VFD filament driver.
Today I tested the LM9022 (SMD version) VFD power supply on a prototype board:
With an input of +5V/0.18A and connected to the IW-18 I measure 4.84VAC @ 80mA for the filament voltage and 21.65V @13mA for the anode-grid voltage, enough to feed all digits of the IW-18 tube:
Since power supply is working I can start developing the code to address the A6812.
The idea is to make a clock like this one but even smaller.
I will use the SMD version of the 16F88, and the A6812SA will be replaced by the equivalent Maxim MAX6921 in SMD version.
5x7 LED modules are for pussies, for you're a real man you'd buy 5000 LEDs and solder them individually...
Got an update for you guys. Not technically a 7-segment LED clock
Bravo on the effort... Looks great... And your comments had me laughing for a solid 5 minutes (grin)...
Kind regards, Mike
Crazy, really crazy....
Great job Stevey!
I know but I'm lazy...
Inspired by this thread I now have to make a clock
I just got these from ebay:
2.3" Yellow seven seg displays, auction said for 7, but I got 8!
I'm hoping on using a picaxe-28X, a 74HCT138, and a DS1307 to keep time,
I've got pcb which will hold the displays done, so I'll be able to build that then breadboard the driver circuit, before I make the pcb for that,
I'll let you know if it works
stevey, i know how long you have to be working on that and jesus, that has PAID OFF!
Holy hell that is gd...
give me ur address and i will send u that stuff u wanted... led sign junk... for free... you deserve it man!
jesus you show us up... but i just got my own workshop today!!
so stuff is going to get better from my side...
question, does that keep going after unplugging it from pc and does it recover after poweroff? Does it store messages?
Cheers for the comments guys
The display does keep going after it has been disconnected from the PC but the program for the PC is only a test thing to make sure it all works for now. I'll start with the proper program next week. Gotta build a nice case for it too.
Look forward to seeing some of your work bigal, and cheers for the offer of sending the stuff for free, but you don't have to unless you're loaded I'll catch you on msn sometime.
What kind of quantity pricing do you get for 5000 discrete LEDs (grin)?
I think I ended up paying around 2.4p per LED for the 5000 LEDs compared to about 5p for one offs. There were several types that I considered, particularly some of the clear LEDs, but the viewing angle was not good enough in those types.
Sorry for that random comment i don't know where that came from
Well, this thread has been quiet, so here's a bump up again. This isn't technically a 7-segment clock, but I've seen several that aren't (wow SteveyG, I mean wow), so I suppose it's ok.
This is a CD-Clock I did a while ago. Gave a bunch away to family/friends, sold some to a local store, and sold a few on Ebay as kits. The URL shows the actual Ebay information, and it has the rest of the photos, a video, and details on the clock itself.
This thread is indeed quiet, not much time left over at the moment and I'm a little stuck with the ASM-code of my IW-18 VFD clock project.
Wow, your CD Clock is really great!
Nice project and very nice PCB , I'm really jealous.
A few years ago I made the very bad decision to use the demo of Holophase's Circad'98 CAD software to develop all my PCB layouts, therefore I'm not able now to export the Gerber files for industrial manufacturing of my PCB's.
I should start learn to use Cadsoft's Eagle and redraw the PCB layouts of the 4" LED clock (with silkscreen) before I can think seriously about selling kits.
I bought a few DS32KHZ TCXO's myself to use with the DS1307 RTC.
Is the +/-1minute/year the time precision with the DS32KHZ?
Keep up the good work!
awesome work! very impressive... I have been thinking of doing something similar to that but using a regular HD44780 to control a similar LED display... without using external chips I could only make a 2 line by 8 character display but I would need some way of buffering the lines coming out of the HD44780 since they were meant to drive LCD pixles not LED's... but I never really got around to designing it...
I have been designing my own 6-digit LED clock and came across this website recently. I read all the posts It is nice to see that other people also like jumbo/blue/white LED clocks like I do Some people think I am crazy. This is a great resource for me as well. Thanks for sharing your great works!
Now, I have a question to you guys. This is going to be a little long so please bear with me. I do not know how to explain it really well.
Seems like nobody had this problem before. I did all the programming and everything for the clock but one thing does not seem to be right. It drives me crazy! For some reason, I see the ghost of one display info on the next digit no matter what I do. What I mean is that lets say we are talking about the "seconds" digits. When the seconds change from one number to another, I see the same information on the "minutes" digits ever so slightly(More obvious in full darkness). My first thought was that the transistors (2N3906) I am using might not have been in deep saturation and so instead of being completely on or off like a switch they were acting as resistors, which may not completely turn the digits completely off. Turns out that is not the case. I checked the transistors and they are saturated. Then I thought maybe there was too much stray capacitance that caused digits to act that way. Right now everything is on a proto-board so I do not know if this is the case. Then I tried to use buffers between the uC and the transistors. That did not solve the problem. I played around with the duty cycle and scan frequency of the digits. That did not do anything either. I tried using BCD to 7-segment converter IC to see if driver transistors it has might solve the problem but no luck. The only thing I did not try was to use a shift register instead of manually (in the software) driving the digit selection but I am not too hopeful about that either.
Do you guys have any suggestions? Has anybody had this problem? If so, how did you fix it?
For your reference, I am using 0.56" CA white LED's from LC-LED (for now to test), with PNP transistors for digit selection. Also I am using a PsOC CY8C24966 uC from Cypress Semiconductors. My schematic is pretty much the same as Turboeku's but does not have the drivers and the BCD to 7 segment converter. I convert the numbers myself to simplify my hardware. Also they did not solve my problem so I decided to take them out and directly drive LED's from the uC with resistors. I am suspecting if the uC is not good or something.
If this is successful, I will move onto the 2.3" and 4" designs.
Your help is much appreciated. Thanks in advance!
PS: I apologize for not submitting photos so that you exactly understand what the heck i am talking about but I will soon.
Without seeing your code, my guess is that the ghosting is occuring because you may not be following the correct multiplexing procedure. Assuming you are multiplexing a digit at a time, you must do something similar to:
-Turn off all digits
-Load the correct bit pattern
-Turn on the digit
-Turn off the digit
-Change the bit pattern again
If you're loading the new bit pattern while a digit is active, you'll get ghosting, so blanking the display before loading in the new bit pattern is very important. You may also get ghosting because of the propagation delays. In this case, add a small delay (a few instruction cycles) after turning off the digit before loading in the new bit pattern.
Thanks a lot for your quick reply. I wrote my code in C. I remember putting some delay between the displaying of each digit thinking that propagation delays might cause the problem but I did not completely turn off "ALL" the digits before displaying new information. I use a timer to create my duty cycle (I assume you mean the duty cyle for the delay between turning on and of the digit correct? Or is this extra delay?). I will add delays again strategically to see if it solves the problem. I will let you know. Thanks again!
The reason you'll get ghosting is because you haven't turned off the digits before applying the new data. This means there'll be a small amount of time when the incorrect information is being displayed on another digit. With low refresh rates, the effect is hardly noticable but gets very visible with higher refresh rates.
Just disable the commons to all of the digits before making any changes to the bit pattern and then reapply the common to the correct digit.
Separate names with a comma.