Electronics > Projects, Designs, and Technical Stuff
Numitron keep-warm testing for latched parallel design with OC output
(1/1)
leonababy:
I would like to make some numitron clocks, and I have a stock of vintage clock chips that I can get HRS:MINS:SECS from.  Unfortunately, it is hard to source BCD versions - mine are 7-segment, and I figured out a way to get the SECS, and in the process I created latch signals.  The clock chip can't drive a numitron, or course.  I need about 25mA max for each segment.

Most people go serial - I may do that after building my share of direct drive clocks.  Looking around, I found the TLC59212, an 8-bit latch with the needed drive, and (inverted) OC outputs.  In researching numitrons, I found it is recommended to keep the filaments warm.  An early datasheet shows a CD4511B (active high) driving a numitron segments, with a 400 ohm pull-up to keep current in the segments.  My thinking was to place the keep-alive resistor from each output to ground - the reverse.  When the segment is "off", the resistor flows current to ground.  Since I am planning a circuit with VCC=5V, the TLC59212 has about VOL of 0.17V, and I want about 4V across each segment, I used a 1N400X to drop 0.7V to the common connection of the numitron.

I calculated that I would need about 334 ohms, so I ran some tests with pulldowns (keepalive) resistors of 330 and 390 ohms.  I am fairly pleased with the results, and am posting to see if anyone has comments from their experience or knowledge of driving filament devices.

Inrush current:
Using the 330 ohm resistor, I have 1V across the filament when off, and it is mostly visible on close inspection but dim.  Inrush current 27mA when turned on.
Using the 390 ohm resistor, I have 0.83V across the filament when off, and it is visible only in the center 25% on close inspection but dim.  Inrush current 29mA when turned on.
Eliminating the pulldown, the inrush current with a cold filament is 87mA.

Turn-on time:
It tames 1us to turn on to 4V across the filament with 330 or 390 pulldown.  With no pulldown, the filament voltage goes from 0 to 3V in 1us, but it takes 25ms to rise to 4V.  In the graphs below, note that the filament voltage starts at 1V with the 330 ohm resistor, and 0.8V with the 390 ohm resistor.

Any comments or observations are welcome.  The reduction from 87mA to 27 or 29mA in inrush current looks good but I am a NOOB. Should I take the hit on power when "off" and use the 330 ohm?

Don

ps the graphs are of voltage across the segment filament
james_s:
I've never understood why people want to muck with these old clock chips beyond using them to repair vintage clocks. It's trivial to do the same thing with a microcontroller and you get a great deal more flexibility, the output can be any format you can dream up.
Cyberdragon:

--- Quote from: james_s on September 23, 2019, 04:39:34 am ---I've never understood why people want to muck with these old clock chips beyond using them to repair vintage clocks. It's trivial to do the same thing with a microcontroller and you get a great deal more flexibility, the output can be any format you can dream up.

--- End quote ---

No way could any micro drive that much current either, it would still need a segment driver system.
james_s:
Well sure, but that's trivial too. You can get power shift registers that can drive them directly. Incandescent is a bit of a pain to multiplex because you need a diode for each segment.
Navigation
Message Index
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod