Shop OBEX P1 Docs P2 Docs Learn Events
Electrical Noise Problem — Parallax Forums

Electrical Noise Problem

kingnebkingneb Posts: 65
edited 2005-10-20 04:36 in General Discussion
I have a circuit that uses three 74LS273 flip flop·chips.· They store a logic state captured from the data bits on the PC's parallel port.· The control bits on the parallel port control the clock for the flip flop.· I have a noise problem with the flourescent light above my workbench plugged into the same outlet.· When I switch it on it somehow sends a clock pulse to the flip flops and captures·the parallal port data when·its not supposed to.· When·you switch the flourescent light off·the flip flop captures garbage data.· What can I do to stop this?· I am using a filtered·voltage regulated power supply to power the flip flops.· The power supply uses a transformer and full wave bridge rectifier before the regulator and capacitors.

thank you···

Comments

  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2005-10-10 19:55
    Do you have a scope so you can verify where the noise is coming from?· A scope would help isolate the problem to the supply running the 72LS273, or to the computer.· In any event, you could try a simply pull-up/down resistor to see if that stabilizes the line.· Also, do you have a .1uF capacitor at the IC across the power pins?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Tech Support
    csavage@parallax.com
  • kingnebkingneb Posts: 65
    edited 2005-10-10 20:03
    I have no scope, they are too expensive.· I do not have caps across the power pins.· I will try .1uf's across the power pins.··Is it safe to put a pull down resistor·across the clock pin?

    Post Edited (kingneb) : 10/10/2005 8:22:15 PM GMT
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2005-10-10 23:30
    Try the caps first.· I forget on the 74LS273 whether the latch line is active high or low.· If memory serves, it's active LOW, and you would use a 4.7K pull-up resistor.· But verify that first.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Tech Support
    csavage@parallax.com
  • kingnebkingneb Posts: 65
    edited 2005-10-11 02:50
    THe clock on the 273 is an active high (I think). The caps do not help, .1 uf's between the vdd and vss pins.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2005-10-11 04:24
    Try a Farady cage.

    If youput the device in a plastic food box [noparse][[/noparse]for electrical insulation] and then wrap the whole box in aluminum foil, you may eliminate the problem.

    Ground the Aluminum foil to your computer's chassis.

    Obviously, this is only a test solution.

    Once you are sure that you have eliminated the problem, you need to provide a similar metal chassis as a permanent solution.

    If the problem still exists, it is in the cabling or power supply transients.

    Try a battery to observe if the power supply is or is not the problelm. A battery inside the box won't create transients. There are several remedies for power supply problems.

    If cables are the problem, you can wrap them in aluminum foil for a trial. Then later go to shielded cables.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    G. Herzog in Taiwan

    Post Edited (Kramer) : 10/12/2005 8:22:07 AM GMT
  • steve_bsteve_b Posts: 1,563
    edited 2005-10-12 12:38
    How about a ferrite magnet around the power cable going to the flourescents? (like those little bulky plastic modules you see on your old CRT monitors power cable).

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·

    Steve

    "Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."
  • FlyingFishFingerFlyingFishFinger Posts: 461
    edited 2005-10-13 23:01
    Instead of starting a new thread I thought I'd put my question here.
    I have a PCB with a large (10A) relay-and-mosfet high current controller I built. Now the board is big and still has a lot of space, so I thought I'd add more stuff like my LCD controller. I also thought to mount a chip socket for my Stamp. What kind of electrical noise problems can I expect if I have diodes across the relay coils, power everything off the same power supply (exept the logic part goes through a 7805) and as yet use no caps anywhere? How can I prevent it? Where would I have to strategically place my caps?
    Please remember I am 15 and may have trouble with elaborate filtering schemes.
    Thanks
    Rafael
  • steve_bsteve_b Posts: 1,563
    edited 2005-10-13 23:31
    flyingfishfinger:

    Sometimes it's best not to worry about everything until you have to!!

    Not that you shouldn't take some precautions when designing circuits....but I don't think you should build too much protection in to something that you may not need to protect from....
    So....polarity protection on power supplies is a great thing....however, putting coupling capacitors and ferrites and such everywhere, although may help, may also just give headaches....

    Anyhow, noise from relays comes in a couple forms. You have the foldback of the electric field around the relays coil....when it's de-energized, it can feed back in to your stamp and play the gremlin...the diode pretty well kills this. You can also have issues with the mechanical motion of the relay....if you are running high currents (hopefully through an appropriately rated relay) you get arcs across the contacts as it's about to make full contact (you sometimes get contact bounce which causes more arcing)....this can give some greif...but you really have to have some sensitive stuff in your circuit to be of worry.

    I'd say, build it first and then go back and deal with noise. Follow datasheets directions if they are asking for capacitors across power leads near their devices....

    If you still want to take the precautions....post your schematic and let us red ink it!

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·

    Steve

    "Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."
  • FlyingFishFingerFlyingFishFinger Posts: 461
    edited 2005-10-14 00:08
    Well I thought I'll keep my design "secret" until its ready unless I run into major problems. It's my personal universal controller board with the stuff your average 15 year old robot builder might use (couple of A/D,D/A, the mentioned LCD driver and the high current controller etc etc). Practically this is my first PCB circuit, not much experience in a) most effective layout and b) practical usage yet (Ive been working with kits and solderless breadboards up to now) Besides I've got a rotten soldering iron, so problems may be quite likely. Also I have veeeery little time to work on it because I'm basically doing homework and stuff 24/7. I did the current controller during my summer break, I hope I'll get to work on it a bit during Thanksgiving break, besides we're moving soon (Nooo! I have to clear my workbench!!)
    Anyway, I'll go on and see what happens.
    Thanks
    Rafael
  • kingnebkingneb Posts: 65
    edited 2005-10-14 02:56
    Someone told me that i should keep my wire lengths short because they act as antennas for radiated electrical noise. The circuit should work if I transfer to a PCB in a metal case.

    Thanks for the help

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    All my exes live in Texas


    Oh yeah, all my exes are six feet under the ground!!!
  • steve_bsteve_b Posts: 1,563
    edited 2005-10-14 03:29
    If you're doing your own boards...a common practice in using 2-sided pcb's is to have the bottom board as a ground plane (basically a giant copper plane that's grounded to your common).
    If you're working with RF....or expect to work in an RF environment....having short leads is important as well as many other things...but depends on what your running (and you're not yet willing to say! tongue.gif )

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·

    Steve

    "Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."
  • Kevin WoodKevin Wood Posts: 1,266
    edited 2005-10-14 04:59
    You might also want to make sure your workbench is properly grounded, ie heavy duty ground strap, with a known good ground.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2005-10-15 11:42
    The ground plane does help. But with flurescent lights, the problem may be creeping in through the power supply lines.

    If you have a single sided board, you can always mount a second board to it as a backing and connect a ground wire to it.
    You might use plastic standoffs or plastic screws to put the two together if you cannot find enough convienent places to drill holes.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

    ······································································ Warm regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan
  • kingnebkingneb Posts: 65
    edited 2005-10-16 21:41
    The Noise problem is affecting my SX circuits now. My project is a pulse width modulation HO scale model railroad throttle. It controls two trains. The pulses are generated by two SX18 microcontrollers. I have an SX28 chip that controls the directional relays and two H bridge drives that control the switch machines (they are not hooked up now, I am just using LEDs as indicators to see which output is turned on and off) and also controls 10 DC prephial switches for accssories. In the final design, three transformers; one will power the IC's, the periphials, and trains. The noise affects the SX28 (makes it go berserk). It seems the noise generated by the locomotive is the culprit. There is also a transformer that powers the locomotive. That may be contrbuting also. I am using a wall adapter to power the SX's for experimentation. There are 100k pull down across the inputs for the SX28. Will lower values help? Should I put pulldowns across the SX18 inputs too? My final design will be in a metal case with printed circuit boards for everything. That might eliminate it completely. However, I want to eliminate the problem on the experimantal solderless breadboards so I can see if my ideas fully work. The flourescent light causes interference too, at the moment it is turned on, and the moment it is turned off.

    Thank you

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    All my exes live in Texas


    Oh yeah, all my exes are six feet under the ground!!!

    One was a lineman who violated the one hand rule.

    The·second put his tongue on 10 car batteries wired in series.

    The third was involved in a tesla coil experiment gone bad.

    Post Edited (kingneb) : 10/17/2005 1:12:01 AM GMT
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2005-10-17 14:17
    I am not sure this will work but it is worth a try and the parts are available.

    On the train transformer [noparse][[/noparse]at the noise source], you can use a computer protection circuit - there are two components a spike arrestor and a low pass noise filter.

    I have one of the low-pass noise filters in a metal enclosue with two lines in, two lines out, and a ground. I am guess that it will block both ways as its schematic appears sysmetrical.

    Try a temporary battery supply. Also, some of the larger breadboards have a ground-able backplane.

    I am very interested in your progress as I am trying to build several projects.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

    ······································································ Warm regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2005-10-17 17:41
    How about a low-pass filter right before the '273 CLOCK pin?· Keep the leads of the RC junction as short as possible and right at the clock pin.· [noparse][[/noparse]Drawing incl.]
    CLK_LPF.jpg
    257 x 167 - 5K
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2005-10-19 10:42
    I was thinking of a CORCOM filter as a low pass on the 110VAC. At 50Mhz, you can expect something like 70db noise rejection.

    It wouldn't hurt to put another one on the SX's power line.

    Both should be part of a shielded case, so the CORCOMs that fit into a hole in the case seem most appealing. [noparse][[/noparse]this assures that the power line going into the case doesn't manage to go around the protection.

    The Art of Electronics confirms that they are excellent to stop noise in AND OUT. The spike arrestors may be unnecessary, but won't hurt.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

    ······································································ Warm regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan
  • kingnebkingneb Posts: 65
    edited 2005-10-19 19:45
    Where can I easily find a shielded case and shielded cables?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    All my exes live in Texas


    Oh yeah, all my exes are six feet under the ground!!!

    One was a lineman who violated the one hand rule.

    The·second put his tongue on 10 car batteries wired in series.

    The third was involved in a tesla coil experiment gone bad.

    Post Edited (kingneb) : 10/19/2005 7:58:49 PM GMT
  • steve_bsteve_b Posts: 1,563
    edited 2005-10-20 04:36
    Coffee can and rf cabling (household cable tv cable -- one cable per wire grounding the shields off)....

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·

    Steve

    "Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."
Sign In or Register to comment.