The 'Glassy and Glowy Clock' thread

Discussion in 'Electronics & Electrics' started by Symon, Feb 19, 2013.

  1. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,317
    The live chat guy recommended this for a beginner:
    http://www.pvelectronics.co.uk/index.php?main_page=product_info&cPath=1&products_id=140

    Not as fancy, but a little simpler, and maybe less DIL soldering.
    I would say if you’re right with one, you’re probably right with the other too.

    Another thing he did say if you mention you’re a beginner he’ll throw in a practice PCB.
    That could be very helpful if you solder all of the cheap components first and save the IC’s etc till last (and the nixies being on plug in modules is all good).
    You could afford to make a mess of the first one early.
     
  2. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,317
    I’m pretty keen to get one to see how they are driving the nixies, and if they heat up enough to damage those daughterboards.

    Of course it’s made from scratch or C++ is jibber jabba you don’t need to know.
    I can’t work out what is meaningful about the message in the first picture.
    The second picture I assume you’re using any letter as a pixel in a scrolling display matrix we can’t entirely see.
     
    Last edited: Jan 4, 2015
  3. RyoSaeba

    RyoSaeba Member

    Joined:
    Sep 11, 2001
    Messages:
    12,162
    Location:
    Perth
    I've build that one. It's pretty easy compared to the vertical tube ones.
     
  4. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,317
    Do you know how they get high voltage for the nixies?
    Not that it matters too much, I still need the nixies and the boards they mount them on.
     
  5. Technics

    Technics Member

    Joined:
    Apr 29, 2002
    Messages:
    1,079
    Location:
    Brisbane, AU
    They have the schematics in the instructions that are available. They generate it exactly the same way as all my clocks do. Using a boost converter that is driven by the by one of the PWM outputs of the microcontroller. The PWM output drives a MOSFET to do the switching.

    The upside is that it only requires the one FET, one inductor, one diode and one cap to generate the high voltage. Then a pair of resistors can divide the output so that it can be fed into one of the ADC inputs for closed loop control of the voltage. Usually there would also be a pull down on the gate so that the FET doesn't stay on while the micro is in reset. Otherwise the inductor will saturate and the magic smoke might get let out.

    The obvious downside of the arrangement is that great care needs to be taken to ensure the FET is never switched on for too long. Provided the PWM /ADC code is interrupt driven and the brownout and watchdog featured of the micro are used it is unlikely to happen. I've had a couple of clocks running continuously for a few years and they haven't ever locked up and damaged anything. The gate could be capacitively coupled to ensure it never happened but this would increase the rise and fall times making the design less efficient.

    Because the output to input voltage ratio is so high, these boost converters typically run in discontinuous mode. Duty cycles approaching 100% would be required for continuous mode and therefore it would be very hard to control. The inductor takes some time to charge but will dump the charge very quickly into the output capacitor. Even though the current required by the Nixie tubes is fairly modest the peak currents can be fairly high.

    As an example, an AVR running at 8MHz can run the PWM outputs at 31250Hz with 8 bits of resolution for duty cycle. Based on that there are plenty of online calculators that can be used to work out the required value of the inductor. I.e This one. I would size the inductor so that the calculated on time is somewhere around 70% at the expected input voltage, output voltage and output current. This will allow a reasonable range for regulation and keep the peak inductor current down. Once the value if the inductor value is known, the calculated peak current (plus some margin, say 30-50%) can be used to find a suitably rated part. In my design I needed about 20mA at 180V from a 12V input. So a 330uH 1.2A inductor fits the bill.

    For the (N-channel) MOSFET, it must be able to handle the output voltage (plus a little margin) and peak inductor current (plus a little margin). Lower Rds(on) values will reduce loss and improve efficiency. Other important characteristic are threshold voltage and gate capacitance/charge. The threshold voltage must be low enough to allow the FET to turn on properly with the logic voltage provided by the micro. Many FET's that can handle the Vds aren't designed to be driven by 5/3.3V logic. So when they are the Rds(on) may end up being much higher than expected. The gate capacitance/gate charge affects the turn on/off time. Gate capacitance varies with gate voltage so it's easier to use gate charge as an indication. Ideally the value should be as low as possible. Each time the FET is switched the micro's output has to charge/discharge the gate. As the micro can only source/sink so much current a lower gate charge value allows this to happen faster. Faster switching reduces the time the FET spends in the partially conducting linear region and increases efficiency. You could run the numbers to work out the on/off times and how much dissipation will result. But as we're dealing with a relatively small converter it's probably just easier to compare the FET's available that meet the other requirements and choose one with a comparatively low gate charge value but isn't so esoteric that it commands a high price an it should be fine for this application.

    The rectifier diode must be able to handle the voltage, current and be fast enough operate at the required frequency. Lower forward voltage ratings help with efficiency but it is less critical at this output voltage. There are plenty of suitable parts.

    The output capacitor must be able to handle the voltage and ideally have a low ESR. The capacitance will affect the ripple voltage but the Nixies won't care about some ripple. Based on experience, for the switching frequency and output current in this example any value from about 2.2uF to 22uF should be fine. Larger values are required for lower frequencies/higher output currents and lower values for higher frequencies/lower output currents.
     
    Last edited: Jan 4, 2015
  6. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,317
    Ok, thanks, it sounds like an on-chip hardware PWM would be the go then. That wouldn’t stop because your program crapped out at some stage while you were writing it,
    and the added benefit that it’s adjustable without needing to adjust anything about your program cycle time.
     
  7. 303-Acid

    303-Acid Member

    Joined:
    Jun 29, 2001
    Messages:
    2,539
    Location:
    Beaumaris, Vic
    I've just received this kit to build over the coming weeks. Wanted one for ages and bit the bullet. :)
     
  8. Physikus

    Physikus Member

    Joined:
    Mar 11, 2015
    Messages:
    2
    PanaPlex Penta Clock

    Once I found in the Internet an interesting Penta Clock with a PanaPlex Display; similar to Union-Jack-Nixie-Tubes. It can display the Time and blink to Italian Five Letter Words chosen by a Randon Generator.;) Here the Weblink: Penta Clock Enjoy..:)
     
  9. Duncan

    Duncan Member

    Joined:
    Dec 21, 2004
    Messages:
    1,462
    Location:
    Sydney
    Bit of a thread resurrection, but I've recently finished these:


    Click to view full size!



    Click to view full size!



    Click to view full size!



    Click to view full size!


    Your inspiration to make the clock - To help share the amazing work of tube maker Dalibor Farny.
    Type of display - Hand-made reproduction of the Z568M nixie tube.
    How do you set the time - Infrared remote control or buttons on the back.
    Microcontroller used - STM8S
    RTC clock chip (if used) - R8025
    Any other features or information - USB powered. Has time, date, alarm, temperature, auto sleep/wake, changeable backlight colours and display modes (fade, crossfade etc). Base is CNC machined from solid aluminium. Currently got it up on Kickstarter.

    :)
     
  10. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,317
  11. roamin

    roamin Member

    Joined:
    Jan 21, 2002
    Messages:
    2,719
    Location:
    melb / Crimebourne
    holy crap i just seen this thread, LOVE these clocks and great effort!

    i have always wanted to make one but the prices of the tubes is killer, $50 - $60ish for a single tube hurts.

    are you still making clocks? got any good sites with good prices for tubes?
     
  12. Tinian

    Tinian Member

    Joined:
    Jan 3, 2009
    Messages:
    15,105
    Location:
    15.0° N, 145.63° E
    Bump. Love looking at everyone's clocks so I've decided to get this one to start with. Given where i plan to put it, I'll add a NTP sync as it won't get a good GPS signal.
     
  13. Technics

    Technics Member

    Joined:
    Apr 29, 2002
    Messages:
    1,079
    Location:
    Brisbane, AU
    It looks like a good kit and the wifi NTP module is a nice touch. It's great to see this thread revived. I haven't been doing much clock making recently but I still have bucket loads or parts for things I'd like to do when life is less busy. I've been slowly (and I mean slowly) restoring some old watchmaking equipment starting with a cleaning machine (for mechanical watches). It's getting a modern electronic facelift. Is anyone else still horologically inclined?
     
  14. OP
    OP
    Symon

    Symon (Plugging your Socket)

    Joined:
    Apr 17, 2002
    Messages:
    4,364
    Location:
    Santiago, Chile
    So last night I finished off this one.

    Your inspiration to make the clock - v8ltd's clock, but I'm not as crazy so I just went with the kit.
    Type of display - 7-segment LED's
    How do you set the time - 2 pushbuttons - 1 for hour, 2 for minutes
    Microcontroller used - none, all transistor logic
    Programming language used - none
    RTC clock chip (if used) - none
    Any other features or information - runs off the mains frequency, there is a minor modification to make the clock run off 50Hz. One thing to note if you are going to buy the kit, is that the PCB traces are very thin and the board itself seems cheaply made. Presumably this is to make everything fit on a board this size and to keep the costs down. I consider myself to be an experienced solderer but I still had two pads lift off the board - expect to do a certain degree of fault finding when building this kit.

    My next step with this is to make a mount for it, but also have a uC automatically set the time via GPS. I've gotten used to all my other clocks setting their own time I don't think I could handle having a manual one in the house.
     

    Attached Files:

    Last edited: Nov 5, 2018
    dakiller and Technics like this.
  15. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,317
    So this is still a thing?
    I’m not real sure if I ever posted my third instalment on this forum. Only my first two are in this thread.
    I’ll have to dig up a photo. I don’t think anything quite like it will happen again in my life.
     
  16. OP
    OP
    Symon

    Symon (Plugging your Socket)

    Joined:
    Apr 17, 2002
    Messages:
    4,364
    Location:
    Santiago, Chile
    I would like to think we can keep this thread going. Sure something like a clock is not a major technical challenge, but some (including me) enjoy building them and using all the different types of displays. Who knows, it might also inspire someone new to this kind of stuff to give it a go.

    So, along those lines, here is another one I did a few months ago. Bit of a camera focus fail with the pics, but you get the idea.

    Your inspiration to make the clock - found some of these displays going cheap on ebay so thought I would try them out
    Type of display - IVL2-7/5 VFD
    How do you set the time - GPS
    Microcontroller used - ATMEGA328-PU not the 'P-PU' version, which is a major pain in the arse if you are using the Arduino IDE, found that out the hard way.
    Programming language used - C
    RTC clock chip (if used) - none
    Any other features or information - I re-used an old case that had an arduino in it, hence the sticker on the top, it is powered by a AVR running Arduino code so it's appropriate :) I've seen a few steampunk designs for this display so I might give that a go next time, I have a few of these displays as they were pretty cheap.
     

    Attached Files:

    echo_ likes this.

Share This Page