EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Chris Wilson on January 04, 2015, 11:39:59 pm
-
In this schematic the 5V regulator has a 1uF input cap, and both a 10uF electrolytic output cap, and a 1uF none polarised cap. My questions are, I have substituted a 2.2uF for the input cap, as I have run out of 1uF ones, will it matter? Secondly, why has the creator also used a none polarised 1uF on the regulator output, as well as a 10uF electrolytic? Thanks. Schematic should be attached.
-
You can use a 2.2uf electrolytic in place of the 1uf electrolytic. No problem there, might even be slightly better. He uses a 1uf non polar cap as well as a 10uf polar cap, because polar caps have poor high frequency performance, while non polar have good high frequency performance.
-
In a practical design you'd want to put the 1uF closer to the IC so that fast transient current demands are supplied by the cap rather than via a long trace all the way back to the regulator. That's why you typically see small distributed decoupling caps located next to (nearly) every IC in a multi-IC design. For example on this board (https://www.eevblog.com/forum/reviews/cable-isp-backend-equipment/) you'll see that there are small MLCCs located right on the power pins of the big power-hungry ICs, with larger MLCCs just a bit further away on the periphery of each IC, and yet larger tantalums providing bulk capacitance yet further out.
-
If I recall the case of standard 7800 series fixed output voltage three terminal linear regulators the datasheet typically call for a 330nf ceramic cap on the input and a I believe a 1uf tantalum cap on the output. In practicality after looking at various circuit designs of actual commercial products I have literally seen all sorts of values used. To be honest even I myself do not know the correct answer if there is such a thing, but it is hard to go wrong with the values listed on the datasheets as a starting point. If your goal is to get the smoothest voltage output possible from the regulator I'm starting to think that it is a shifting target which depends on the internal feedback loop of the regulator, the amount of current being demanded from the regulator, and the amount of not just the capacitance of the caps but also on their impedance.
Essentially the regulator is constantly trying to correct the output voltage to whatever value it is programmed for and there will always be some lag and overshoot by the regulator. Basically that means the darn things are actually oscillating because of the delay in the feedback loop, however what one ends up with is a "window" of safe capacitance values wherein the device will operate within the defined parameters on the datasheet with different capacitors producing different, yet still permissible, amounts of ripple on the output. But if one were to use capacitor values outside of the window the oscillation could potentially become so severe that the regulator output simply goes bonkers. Also from what I understand is that with audio amplifier applications and linear regulators the goal should be to get the smoothest output possible as the ripple could cause distortion in the output but again I don't have much experience with that so I can't really say one way or the other. If true it might just be a case where you have to hook the scope up to a test circuit and try different types and values of caps to see what sort of effects they have.
I'd love to hear what Dave would have to say about this although I fear he might pull a Gordon Ramsey on me. :P
-
Depending on the PCB layout of your circuit I'd slap another non-polarized cap as close as possible to the output transistors, depending on what load they're driving. A 100n ceramic decoupling cap across the 4046 wouldn't hurt either. I usually place decoupling caps close to the parts that draw high current in spikes as well as close to the regulator. For 7805s I've got away with 100n caps at the input/output but then again it was powering only digital stuff.