| ??? 02/08/08 17:49 Read: times |
#150498 - be conservative ... for now ... Responding to: ???'s previous message |
Shakir F.m said:
Richard Erlacher said:
(1) is your oscilloscope probe properly grounded to the ground of your circuit? yes. I don't think there's a problem there. Richard Erlacher said:
(2) Have you got lots (one per power and one per GND of each IC and located within 1 mm of each connection) of small (10nF) bypass capacitors from power to gnd and one larger one ~4.7 uF tanatalum at the power entry to your board? nope! wow thats a good idea. i'll try it out.thanks! i already have a (much) larger 220uF capacitor(aluminium electrolyte. Does it have to be tantalum?) between the regulated side of the voltage regulator IC and gnd. Electrolytics don't have much high-frequency response. Also, if you're using as small a regulator as you say, it's probably better to use a smaller value, say, 20 uF, +100%/-50%. That will have better frequency response than 220. With a circuit that uses that small regulator, I'm inclined to put 22 uF at the power entry, with 2.2 at the regulated output. Richard Erlacher said:
(3) Has your board got a good, low impedance ground plane throughout? erm..no. sorry but I'm a real noob at this..when u say ground plane, u mean like a solid metallic sheet? where does that go on the circuit board? A ground plane is a layer of copper that covers the entire surface of the board and is connected only to Vss. On my hand-wired boards, it covers the wiring side of the board and a power-plane is connected to Vdd and covers the component side of the board. Not all boards have that feature, however. Richard Erlacher said:
(4) Has your power supply got the capacity to provide at least 10x the maximal current your circuit demands? yes.i'm using an 800mA supply Richard Erlacher said:
(5) Have you tried putting a small (10 ohms) resistor in series with the power supply, before the entry capacitor? no. by entry capacitor, u mean the one between the regulator ic and ground ,right? i'll try that. What I mean is the first cap that Vin encounters on entering the board. Yes, it's between the raw supply and GND. Richard Erlacher said:
If your power supply has more than 5 mV of random noise or of ripple with no load, it's too noisy! well, with no load its absolutely clean. no ripples at all. I'm using 220uf capacitors across the voltage regulator ics and the "unregulated" side is a(quite clean by itself) 8v coming from a battery charger. Across? Just what do you mean? Richard Erlacher said:
If your power supply has more than 25 mV of random noise or or ripple with 10x full load, it's too noisy! well, its waaay over 25mv with 1x load, so its noisy as hell! When you say "full load" do you mean with the MCU, or do you have a resistor that draws about the same current ans the running MCU circuit? What I meant by full load was the full load for which the supply is rated. I'm curious what you mean by "it". By that, I mean I'm curious what frequency components are in the "noise". Is there anything at AC line frequency? What about components at the oscillator frequency? Are your wires to the crystal short (< 2 mm)? Richard Erlacher said:
Be sure your power supply wires are large, not as thick as a pencil, but thicker than a pencil lead, (1.5 mm). ok. that makes sense. thanks a ton for the info. I'll try it out and let u know if it works. Could you post a schematic, and perhaps a picture? just to clear up some questions? RE |
