Shop OBEX P1 Docs P2 Docs Learn Events
noisy power affects PLL — Parallax Forums

noisy power affects PLL

Bobb FwedBobb Fwed Posts: 1,119
edited 2008-10-05 22:36 in Propeller 1
I'm sure this has been covered before, and I'm sure it is mentioned in the manual, but I have recently started using a new 3.3V regulator for my prototyping boards and while assembling a new device I wanted to test something on it and couldn't figure out why nothing was working. I eventually figured out that my new regulator (while averaging 3.33V -- according to a voltage meter) was actually quite noisy and consistently alternating between 3.1V and 3.5V (according to an analog oscilloscope). With an external 5MHz crystal, PLLX8 worked fine, while PLLX16 seemed to prevent the propeller from even booting properly. I couldn't get anything to function at the highest PLL. Just by adding a 1µF cap on the power it seemed to operate, but the power was still not smooth. At about 20µF, the power seemed completely flat; just to be safe, I will use at least 100µF (especially once I get more power sucking components on the board), but I have oodles of 470µF caps, so I will use those. Hope this helps someone!

Comments

  • LeonLeon Posts: 7,620
    edited 2008-10-05 22:03
    What regulator was it? It sounds as if it was oscillating because of poor PCB layout or the wrong capacitors on the input and output. Sometimes, attaching a meter probe or a scope probe stops the oscillation, which can be very confusing.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
  • Bobb FwedBobb Fwed Posts: 1,119
    edited 2008-10-05 22:14
    After I discovered it, I tested it under different circumstances. It didn't matter what cap was on the input power, I still saw the oscillation with nothing attached to the regulator at all (in a breadboard). www.nteinc.com/specs/1900to1999/pdf/nte1904.pdf - the datasheet. Looks like it suggests at least 2µF and suggests 10µF on the output. Previous boards I built had a 470µF output cap out of habit, but I was just testing something quick and didn't get around to putting the cap on yet, and thus my test app (which worked fine on another board I had and ran at 80MHz) didn't even execute the first instruction of turning on an LED.
  • LeonLeon Posts: 7,620
    edited 2008-10-05 22:36
    Do they recommend a low-ESR tantalum? It would probably be OK with one of those close to the leads. If you are using a solderless breadboard, that might explain the problem. They should be used on a PCB with short connections.

    I use an LM317 for the 3.3V supply on my (homemade) Propeller PCBs and don't have any problems.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
Sign In or Register to comment.