Connect with us

PCB tips...

Discussion in 'PCB Layout, Design and Manufacture' started by wannabegeek, Jul 4, 2013.

Scroll to continue with content
  1. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    Hi everyone...I hope you're enjoying a day off of work...

    I'm getting ready to bring a project from the proto board to PCB. I haven't done this very many times and I'm always terrified by it.

    I need to write up a schematic for the project, which I'll do by the end of the day.

    However, I'm sitting here doing my morning job hunt and email and felt like starting off with a question anyways....

    I have a Tx circuit that uses a 555 timer to generate a square wave that PWM a laser diode. The source signal is a microphone that gets treated by am op-amp to bring center the output at Vcc/2 and keep the swing close to 1V.

    The 555 generates a crap ton of RF and other noise...Even if I end up switching to a uC in a later version, I suspect that RF and noise is going to be an issue.

    Can I make a ground plane around the PCB section containing the 555, by using some thin copper sheet that I have...? I was thinking of cutting strips of copper sheet, and soldering it to the traces, in effect surrounding the 555 with a copper rectangle. I suppose the thickness of the copper strips is a consideration too.

    On the proto board, it has helped to simply keep some space between the 555 and the other networks. I was reading about star ground and the basic idea of grounding for audio amps and not sure if those ideas apply here.

    I'd love to get your thoughts on this...

    Thanks,
    wbg
     
  2. BobK

    BobK

    7,682
    1,688
    Jan 5, 2010
    You can make a ground plane out of the copper on the PCB. Typical PCB design software will allow you to mark a rectangle as ground plane, and then will separate any traces from it automagically.

    Bob
     
  3. davenn

    davenn Moderator

    13,836
    1,951
    Sep 5, 2009
    a ground plane around the 555 probably wont make a lot of difference ( it wont hurt tho)
    the fact that you have a square wave is the reason for all the noise.
    square waves are very rich in harmonics. So even groundplaning the 555 isnt going to stop all those harmonics
    from its output going into the next stage


    Dave
     
  4. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    What's the power dissipation of your laser diode?

    Chris
     
  5. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    Hi guys...
    I'm familiar with the Fourier Series of the square wave...it's a beast...

    On the proto board I have to use some big caps to keep that crap from starting
    positive feedback in my amps.

    @CDRIVE, the laser diode module is 650nm @5mW...nothing fancy...
    I did find a great lens though at my part time job in an optics lab that I asked to keep...it's for the Rx side.
    I think the name of the game for my laser radio project is the Rx lens.

    My Rx uses a nice and sensitive photo diode but the current to volt op amp can only give it so much gain b/f
    the signal gets all distorted. I thought it was the voltage offset, but the IC is an OP07 which has really small
    offset.

    I'd like to build a simple 2 transistor driver from Sam's Repair Faq and make my own 20mW laser driver...but alas, it's hard to find anything greater than 5mW and when I do there's something fishy about it...

    I guess I'll build the PCB version slowly, testing for signal leakage as I go...

    Cheers,wbg


    EDIT: Hey, how about a harmonic filter like used for radio Tx...? I don't know how to look up the proper toriod material or size and and all that, but I'm sure you guys do..! Or, would a simple discrete choke component help ...?
     
    Last edited: Jul 7, 2013
  6. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    If you're transmitting audio via laser diode why can't you feed it DC and modulate it (AM) conventionally?

    Chris
     
  7. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    I could....in fact I thought about using a mixer IC to encode and decode.

    But, I thought that FM using PWM would be better for long distance experiments.
    If the intensity of a laser drops off roughly linearly with distance, than I figured that
    it would be more sensitive to detect any signal at all and decode the PWM.

    Using the AM principal, I need to be able to detect the full bandwidth of the signal and carrier. I think...

    I chose the PWM method and enjoyed learning about it... :) My first model Rx was a crude Class D amp. I made a push pull MOSFET amp with the LC filter just before the speaker...I had to recreate the square pulse train with a comparator...it was good practice for me...

    Now I just want to bring this project to PCB and that's another lesson for me.

    This is all about practicing basics and sometimes I don't mind doing something in a funny way if it makes me more interested...

    wbg
     
  8. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    I your case (laser diode) light itself is the carrier. It need only be modulated.

    Chris
     
  9. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    The way I think about it Chris...is that to decode the AM signal, my detector needs to be able to sense a BW of voltages. The carrier is DC and at long distance lets just say it's coming in at 1mV ( after current to voltage stage with gain ). Well, I then need to be able to sense
    0 - 1mV because that's where the signal is.

    If I send PWM, my sensor only needs to detect the lowest possible light energy, say it's 1mV. The duration of the 1mV is the signal. Unless I misunderstand, the second method has got to be more sensitive for long range experiments....

    But if it's not, that's OK, I just want to learn about PWM and decoding it anyways...

    Cheers,
    wbg
     
  10. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    Is that supposed to be a PW of voltages? If so we're talking about a form of FM transmission and detection?

    Chris
     
  11. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    Yes... :rolleyes: not b/c I typo-ed....I just didn't know the right term...
     
  12. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    OK, I've got your drift now. Since all of the harmonic content in a square wave is generated on the fast rising and falling edge have you considered the following?

    Make the RC ramp linear (constant current) to generate a triangle waveform ramp.
    Tap off the ramp, buffer it and feed it to the laser.

    This will have far less harmonic content and can be filtered much, much easier than a square wave. In fact I don't think it would be difficult to resonate it into a sine wave.

    Chris
     
  13. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    I actually did consider that...I wrote it off b/c I don't know how to PWM the triangle wave form without turning it back into a square pulse ....????

    If I understand correctly, I could shift the phase of the triangle wave...I guess that's what I want right...? And there's probably a voltage controlled phase shifter isn't there... :)


    what do ya think...?
     
  14. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    I tested powering the laser diode module with a triangle wave with a high offset. It looks really good at the Rx side....better than the signal from the
    square pulse...

    I'd guess that the diode will last longer too with a triangle wave...it's just more elegant ...

    The only practical way I can see to implement this is to use a uC.

    f(t) = x(t + phi) for t < T/2
    f(t) = -x(t + phi) for t > T/2

    where T is the period of the wave and phi is the phase shift.

    I can imagine a simple while loop for the triangle wave, the phase change as a function of an input voltage might be tricky for a
    beginner...I haven't even got MPLAB working just yet with SDCC....

    This could be pretty cool though....more impressive than the original 555 powered circuit...
     
    Last edited: Jul 8, 2013
  15. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    You can PWM the 555 by injecting your buffered mic audio into the Control pin. That's the primary function of that pin anyway. ;) Your osc frequency (Carrier) should be a good deal higher than the mod freq.

    Chris
     
  16. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    yes yes...I know... ;) I have a working Tx and Rx using a 555....

    I was testing an offset triangle wave and it produces a clear signal on my Rx.

    I just don't know an easy way to frequency modulate it. The 555 is an easy solution
    for PWM. But to do a similar type of thing with a triangle wave I would need to
    modulate it's phase.
     
  17. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    First off I have to correct something I said. If you inject audio into the control pin it will FM the 555. The FM is not confined to the output pin alone. It will also be present on the sawtooth ramp. As I said before, you can use constant current techniques on the ramp to convert it to a triangle wave. You may find that the sawtooth (as is) will be many degrees cleaner than the square wave without modifying it. I'm not implying that sawtooths are low in harmonic content though. They're just cleaner than square waves. ;)

    Chris
     
  18. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    Oh wow...I didn't understand what you meant before...The saw tooth is the relaxation of the caps right...? It gets sent to the flip-flop which makes the square wave as I recall.

    I was thinking of using an integrator to make my triangle wave and just deal with the noise issue. I handle on the proto board OK , so PCB should be OK too.

    I like the signal of the laser much better when using the Triangle with a healthy forward bias.

    I'll play with it tonight and post back...

    Side note, for some reason I'm the only Linux user who can't get MPLAB X to use the SDCC compiler for their PIC12F675...Eclipse plugin for PIC is next....
     
  19. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    652
    May 8, 2012
    I slapped this together without a lot of forethought but none the less it gives you the basic concept. You can clearly see the output being FM'd and the ramp too but it's not as obvious on the ramp because there's also AM there that follows the modulation. Perhaps we can feed the ramp into a fast AGC amp to remove the AM. It would have to be damn fast though! An alternative would be to pump the signal into a limiter but limiters clip. On the other hand FM receivers employ limiter stages. It's what gives them greater immunity to AM than the discriminator alone.

    Chris

    EDIT: I think it's a better idea to do the AM limiting at the receiver end. This way we're following RF convention.
     

    Attached Files:

    Last edited: Jul 9, 2013
  20. wannabegeek

    wannabegeek

    133
    0
    Aug 17, 2011
    wow...that's a cool simulation...thanks so much for the effort...you know an damn lot about electronics... :)

    I'm thinking about my simple idea, of integrating the 555 output...it's more "my level"....

    And that's not trivial for me, b/c my laser module runs on 3.2 V from a zener source.
    I was just running the diode module using a 2N2222 in a switch mode with the transistor collector connected to the zener output...it didn't matter that the 555 output was between 0 and 8 V. I used a diode on the 2N2222 base to keep current from flowing backwards.

    Now if I have a ramp it needs to be 0 - 3.2 V, ditch the transistor b/c the current is so small anyways...
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-