nixie clock, bit of a heating problem

Discussion in 'The Projects Forum' started by magnet18, Jan 5, 2012.

  1. magnet18

    Thread Starter Senior Member

    Dec 22, 2010
    1,232
    124
    OK, it's all working, and the case is most of the way constructed, battery backup and diode gates work like a charm, there's just one little thing left

    when the tubes are off, the entire circuit (all 12 IC's) draws 28mA, not even enough for my power supply to realize power is even being drawn (it wont light up, but still provides power, so i dont care)

    When the tubes are on, it draws ~205-230 mA, quite a bit
    this all leads me to believe that the tubes and their powering circuitry is drawing about 200mA, which is to be expected.
    Apparently, however, when powering all tubes with a single 555 and FET, the chip and fet get quite warm. I ran it for over an hour, and nothing failed, it was just warm enough to feel; probably around 90F

    So, this bothers me a bit. As I understand, my options are-
    add a heatsink
    add a fan
    use bigger parts
    use 2 tube power supplies (was the idea originally but one died, and I never got around to re-doing it)

    or any combination
    are there any other, easier options, or will any one of those work, or should I not worry about it?

    thoughts?
     
  2. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    What is your plate voltage?

    How much current does each "segment"/number draw?

    The nixie drivers I've seen provide 155V@20mA for the nixie tubes, and 5V@200mA for the logic.
     
  3. magnet18

    Thread Starter Senior Member

    Dec 22, 2010
    1,232
    124
    Yea, each tube should draw like 5 ish mA, each neon like 2, so 2*5+5*6=40mA the power supply puts out 180 volts which is dropped to <170 by the current limiting resistors
    during normal operation (earlier numbers were on start-up, 'lytics were charging) the power supply draws like 180mA, so it is very inefficient :(

    [EDIT]
    wait, I forgot to math
    .180A*12V=2.16W
    .04A*180V=7.2W
    ...
    300%efficiency, woo!
    :p
    (someone show me where I went wrong please)

    [EDIT EDIT]
    My bad, 2.5 mA per tube, silly me :rolleyes:
    that makes 10+2.5*6=25mA
    .18*12=2.16W
    .025*180=4.5W

    now were down to 200% efficiency... where else did I go wrong???
     
    Last edited: Jan 6, 2012
  4. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Are the tubes glowing extremely bright? Have you measured the DC plate voltage and current?

    What is your HV supply rated for? Which part is getting hot?
     
  5. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    If the mosfet is so far without heatsink and reaches less than your body temperature I wouldn´t worry about it. I don´t know what frequency is your 555 running at and what is the gate resistor, but the rule of thumb (at least for power devices) is that if you can hold your finger onto it then it is cool enough, higher than that you should start thinking. You should not be alarmed by "slightly warm".
     
    magnet18 likes this.
  6. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Kubeek reads posts better than I do.

    If you can actually hold your finger on the FET but think "this is warm" after it's been running an hour, you don't have a problem.

    I thought it was getting warm instantly. I would hazard a guess that the FET isn't being driven hard enough for switching in the boost supply, hence the losses.

    Quick way to tell: Is the switching frequency or voltage far off from what you calculated when building it?
     
    magnet18 likes this.
  7. magnet18

    Thread Starter Senior Member

    Dec 22, 2010
    1,232
    124
    Oh, I didnt know you guys were posting, I should have refreshed :p
    Check my edits in the above post

    also, I wont worry about it, it warms up quickly, but stabilizes at a temperature I can hold my finger on

    Can anyone help me with the above math?

    and... I didn't actually calculate anything when building it, I just sorta put it together
     
  8. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Then you don't have a problem.

    Is only 1 segment lit at a time and you are calculating for the full load? try labeling what numbers represent which item.

    Then you are doing great!
     
  9. magnet18

    Thread Starter Senior Member

    Dec 22, 2010
    1,232
    124
    ok, apparently I can't remember current numbers :p

    neon-.7mA
    neon-.7mA
    neon-.7mA
    neon-.7mA
    neon-.7mA

    tube-2.5A
    tube-2.5A
    tube-2.5A
    tube-2.5A
    tube-2.5A
    tube-2.5A

    18.5mA*180=3.33W
    220mA*12=2.76W

    BUT!!
    only 3 neons are on at a time!
    so it's 3.07W
    and the tubes are being limited to closer to 2.2mA, so it's 2.754W!!

    It is apparently quite efficient :D
    I'll have to take some real voltage and current ratings of the high side later, see what they actually are
     
  10. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I thought I told you that it would be somewhere around 90% efficient when I first showed you that circuit many moons back.

    At any rate, I don't recall what the gate resistor is, and whether there is a diode in parallel with it.

    You really need to post schematics along with your questions, as all anybody can do unless they are really familiar with your circuit is to make rather general statements. I don't have any of my files with me.
     
  11. magnet18

    Thread Starter Senior Member

    Dec 22, 2010
    1,232
    124
    You're back!!

    How are you?

    and general statements were all I was really going for, as I was wondering how to best keep the chip cool

    (also, I thought 90% was for a different buck-boost you were helping someone with in spice that you suggested)

    [​IMG]
    just one of the 555 circuits
     
  12. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Ah yes. You're using an IRF840 MOSFET. That's rated for 500V, which is maximum overkill.

    You could likely get away with an IRF620 or IRF630, and it would run a good bit cooler, as the gate charge would be much lower as well as the Rds(on) being lower.

    Also, you need a gate resistor of between 20 and 40 Ohms, and a diode in parallel, anode towards the MOSFET. Otherwise, you wind up with a series LC circuit that will ring like the dickens without any snubbing.
     
    magnet18 likes this.
  13. magnet18

    Thread Starter Senior Member

    Dec 22, 2010
    1,232
    124
    wound an 1N4148 work for the diode?
     
Loading...