Displays : yeah... no doubt now, they have to be cool Numitron indeed ! Just looked a bit more closely at the schematic, see close-up below, it says :
" Affichage 7x 3015F "Googled for " 3015F ", lots of hits, it's a Numitron part number !
Google even found me a nice little data sheet for it.
They run on 5V DC at 8mA, so ideal to be driven directly by the TTL logic chips of the day, which this counter uses.
Found some for sale on Ebay, used ones, 30 USD each ?!
On offer, but still...
7 of them in this counter, all in good condition judging by the photos so... 210 USD for the lot, 175 Euros or something, more than twice what I paid for the counter. So I guess it was not such a bad buy after all ?? Not that I internet to resell the displays of course, not my kind !
According to the datasheet they are even more tiny than I though.... 0,36" ?! A third of an inch, so 8+ mm high or something ?! That's like pocket calculator !
Well at least it won't draw much power / cost much in electricity keeping it running, and won't take up much space either on the shelves... hey... would that make it a good candidate to be used as a CLOCK ?! I mean whenever I work on the bench I lose track of time... have no wall clock here, no wrist watch I hate them, and my smartphone is on the desk a few meters away. So having this tiny little counter displaying time on his bright lovely little Minitron smiling at me, could be cool ? I don't know... just an idea. I could hack it to either display time, or be used as a counter of course.. not going to mutilate it to permanently turn it into a clock, I would want it to be switchable at will. That could be a nice little project for me, a step up in design and PCB / CAD design and soldering, from the my first board the other day, that minuscule ALDL/RS232 interface board that fits into a DB9 shell.
I would need I guess 4 8 bit shift registers to drive the Numitron directly, just send the date serially from the micro so I could use a low pin count micro, for compactness and ease of design/soldering/routing. Would need an RTC to keep track of time, I think these come in also small packages with a serial interface as well, don't they ? So that would fit well. Then I would need a way to set date and time, I guess I could just use an UART and stick a DB9 socket at the back of the unit, and/or add a few tactile switches on the PCB, alarm clock style.
Using SMD and serial interfaces for RTC and shift registers, the board could be really low profile and very compact, and could be easily fitted inside the counter despite its very small form factor. The more annoying bit is that driving the displays direclly would require lots of wires, could be show stopper. Would be less wires if I drive the decoders BCD inputs instead, 4 per display, 28 lines. Still a lot of wires, would look ugly and a pain to solder one by one.
I think I could make that better, much tidier and neat looking, if I placed the clock board on the bottom side of the main board of the counter, so that I can access the solder side of the display board/decoder inputs. Then I could shape the clock PCB so that the outputs of my shift registers, line up perfectly with the pins of the BCD decoders. This way I could just run ribbon cables from my board to the display decoder chips. I could maybe make custom flat flex cables ? They would have a much lower profile/thickness than a ribbon cable, and I could also make the spacing of the wires custom, to make it match really perfectly whatever pin and chip spacing / placement of the decoder chips. That would be the best looking, tidiest neatest solution I would think, as well as the easiest to assemble.
Of course I could reduce dramatically the need for wires by not driving the displays directly and instead inject a signal straight to the LF counter input, so that the time would be displayed a frequency... it's tempting but no... it would look crap, with all the leading zeros and the HM digits being all concatenated together; fused... hard to read. With direct control of the displays I can turn off all unused displays, add spacing between HM digits to "format" the displays, make it easier to read and better looking. Hell with 7 displays I don't have enough to display HMS, never mind space them apart... would need 8 displays for that, I am short of one display, bugger ! Hmmm... I don't care about secondes anyway. simply H and M would be good enough, and that leaves me with 3 unused displays to let me format H and M as I want. So, in the end if I drive only 4 displays that only 16 lines to drive, not 28.. much better.
OH WAIT !! I have another idea !!
Let's go back to the idea of injecting a signai into the input of the counter to reduce wire count / avoid driving the displays directly.... now I think of it, even using this method I could STILL be able to "format" the display as I want !! I could just use a few I/O lines of the micro to disable specific displays, or drive specific / individual segment as I wish ! this way I get both a low wire count AND the abiltiy to format the display !
But that means I need a stable time base to make sure I can display what I want... but hey, I have one, it's a counter after all, already has a time base, and it's an OCXO so might be good enough for the purpose !
Say I want to display worse case
"23 H 59" i.e I need a 2,36 or so MHz signal accurate down to one Hz. So I need an accuracy of 0,4 PPM at least. Make that an order of magnitude better, 0.04PPM, or 4.10exp -8 ?. Now can this old OCXO achieve that ?! Don't know, I am not into time nuttery, help please...
So I would have to tap the OCXO then add an external / dedicated digital programmable PLL to generate the appropriate frequency... how easy is it to find / source these, how much do they cost, do they come in a low (enough) pin count package, do they come in a "hand-soldering friendly" SMD package... hmmm... looks like going this route could be a whole can of worms eh ?!
... So I guess I better be reasonable, give up on that and go back to driving the display decoders directly- instead... maybe when I am more clever and experienced I could attempt the PLL route, one day...
Oh ! Or maybe there are some 8 bit micros featuring a basic but good enough PLL built-in ?! Do such micros exist ?!
Anyway, it sounds like a cool little project I could be having fun working on, so think I will be looking into it seriously !
Home
Help
Search
Login
Register
