Electronics > Projects, Designs, and Technical Stuff

Custom Nixie/LED/VFD clock

(1/8) > >>

carbon dude oxide:
hello, I have re-started a project now that i have some spare time from my college work :D i do plan on making it fully open source once i have completed it :)

My plan is to make a multi board clock which you can select what display medium according to what parts you have.

I have a main controller board with a micro, time keeping chip and i am also considering adding some other features to allow more flexibility on its features.

Currently i am working on one of the nixie tube display boards (specifically to hold IN-14's but will also be doing IN-12's and a few other footprints).I have finalised the schematic and board layout. The board only contains passive components as the main controller board has all of the active components. I mainly would like to share this board now so that i can get feedback on it and any problems with the schematic that i cant see :) I have done several simulations using LTSpice for the High side control of the tubes (thanks c4757p for the help with that in a thread quite a while ago :) Im trying the project using SMT to save space considering the case i would like to put it in for myself so i have tried to reduce the amount of different parts on the board.

I have attached the schematic and artwork layout of the board if you want to have a look and help me to improve it :) as i would greatly appreciate improvements and feedback :)

sync:
Don't multiplex nixies unless they are specified for this. They got damaged. Every time they ignite at bit of the electrode material will be sputtered.

carbon dude oxide:
I thiugh that IN-12 and IN-14 were able to do multiplexing?

jolshefsky:
Using the leftmost selection subcircuit as a template, I would omit R1 and short the emitter of T1 to ground. I'm assuming the tube selection and digit selections run from 5V logic, so then I'd put R4 at 10K: the MMBTA42 has a hfe of around 40, so Ib = (5V-0.7V)/10K = 0.43mA, then Ic would be around (40)(0.43mA) = 17mA to drive the base of T2. The MMBTA92 has a bit lower gain, so assuming hfe=25, 17mA * 25 = 425mA which should fully saturate T2 right up close to its maximum Ic of 500mA. (Although without a big heat sink pad, you can't expect the device to dissipate more than 50mW or so, and with a Vce(saturation) of 0.5V, that's only 100mA.)

Now you have everything at 10K except the base resistors for the digit selection (the 4.7K resistors R33-R44). If each digit will draw less than 5mA (which I'd expect, given 155V in, even 1mA is 100mW per digit which would be quite bright) you can get away with 10K resistors instead (see calculation above). However, if you'd prefer 4.7K, consider putting 2 10K in parallel instead. If you find you need 5K, you can have both installed.

My point in going with the 10K resistors is something Dave mentioned in a video about getting things made: you need to buy each reel of parts, and to buy half a reel of 4.7K resistors may be more expensive than buying a full extra roll of 10K. It also frees up one more slot in your assembler's machine for a different component in case you need it.

A bypass cap for the high voltage supply would make me happier, but I don't think it's actually necessary. I'd add it to the PCB and if you get some kind of high-voltage switching interference, that might help.

Now, on your board layout, I'd isolate the resistors and transistors of the low-voltage logic to one part of the board near the connectors. Put a silk line around the area to clarify where it might be safe(r) to probe.

And dang: filled copper planes are your friend, man! I'd make both the top and bottom of the board a full ground-potential plane. Set the clearance wide: my ballpark numbers say 1000 volts can arc 100 mils, so say a 50 mil clearance. You can make a cutout 155V plane on the top layer around the PNP transistors.

And then mounting holes. Throw some on there unless you have a case that absolutely doesn't need them (e.g. slots for the board). And if you do have such a case, mark off the mounting areas so you don't put components there by accident.

Maybe a couple fiducial marks too. I'm working on a design that can be wired either through-hole or surface-mount, or a combination of both. For prototyping, I'm probably going to build a reflow oven and do some SMD myself. When I get a solder mask cut, I figure I'll want fiducial marks so I can accurately align the solder mask.

What else ... test points. It might be nice to get into the base and collector of T1 (et al) to check voltage. It might also be nice to be able to get a point to read out the 155 volts and ground. Maybe the collector of T2 (et al) as well to figure out if the transistor is turning on in case your digits don't display. The pins on the tube can probably be probed fairly easily. In some cases, putting the resistor pad close to the transistor pad can make it easy to get a probe between to get a reading.

Now hopefully a real engineer will call me out on anything that's a bad idea. I'm ready to learn, even if I need to be schooled!

sync:

--- Quote from: carbon dude oxide on October 21, 2013, 02:19:33 pm ---I thiugh that IN-12 and IN-14 were able to do multiplexing?

--- End quote ---
Read the data sheet. If it's specified for multiplexing then do it otherwise don't. And no PWM ether. Here is an example of a PWM damaged nixie.


http://www.mikrocontroller.net/topic/308329

Edit: I would use high voltage open drain shift registers. Much easier and no worries about multiplexing. Maybe TPIC6B595 is sufficient. It's only 50V but the 74141 is only 60V too.

Navigation

[0] Message Index

[#] Next page