bypass capacitor, long cable and switchmode PSU

Design Concepts
1

the main aggregator about the protodeck’s troubleshooting war is this one

the post where I found and understood my noisy PSU effects is that one

so I can begin..

I read a lot about bypass capacitor.

this article was very good for me to understand well the system.

instead of changing my psu now, I’m  going to put several capacitor at the place where I supply the protodeck modules.

I’m using the smashTV last revision modules so everything gets capacitors near from each IC. even the AINx4 I use, those that gave me headache with my noisy PSU.

I read that with long cable (and unshielded) it could be nice to use bypass capacitor too.

in the previous link, it is writtent:

Another great place for a bypass cap is on power connectors. Anytime you have a power line heading off to another board or long wire, I would recommend putting in a bypass cap. Any long length of wire is going to act like a little antenna. It will pick up electrical noise from any magnetic field. I always put a bypass cap on both ends of such lengths of wire.

so the question is: may my system be improved by adding a couple of capcitors correctly place here and there on my 5V/GND supply of all pots?

here and there mean, for instance, at each rows or something like that…

Do the bypass capacitors at the PSU side of my pots supplying wire enough or not ?

my schematic is this one:

even if it could improve a bit, I can put them to have a more stable system.

any help, ideas, words from you increase my knowledge, so help would be very appreciated :slight_smile:

2

Yes it may improve, just try attaching some capacitors with short cables (crocodile clips ftw) in different places and see how it affects the noise level. Remeber:

workaround != fix, I’d strongly advise trying a different PSU.

3

thanks nILS.

about the PSU, I’d be able to rebuy a new one.

I sized this one because I was afraid about power consumption.

the one I bought is : ECM40

What kind of PSU could you advice me in order to match 5V, around 4A and WITHOUT noise ? ? ?

4

How much does the box need atm? Measure that, add a another 10-20% to be on the safe side and choose your new PSU accordingly. For testing any 7-10V wallwart ~500mA will do. Just disconnect the DOUT chain and/or the display to keep current consumption down. Let’s find out if the PSU is the problem before buying a new one.

5

I’m on this right now..

I insulate better pots wire / PSU & I replaced this freaky deadly 4051… NO PROBLEM OCCURED sinde 10minutes moving the frontpanel, testing etc.

by “insulate”, I only mean: make a better path inside the box.

btw, you’re right. I guess the PSU is noisy. But indeed, I probably don’t have to change it.

6

I can confirm that.

no problem now.

I isolate a bit with a little piece of plastic direclty put over the PSU which doesn’t heat too much… so no prob with that.

the path of cable is better too, but could be improved!

but in order to be safer & more reliable, I guess it could be better to use a linear PSU very less noisy.

I didn’t test with 2 capacitor cause 99% of this problem came from the proximity between cable & PSU (I mean: bypass capacitor doesn’t fix that)

7

in order to be safer & more reliable, I guess it could be better to use a linear PSU

I assume you are presently using a switched mode psu: I am interested as to why do you consider linear psu to be safer?

regards

8

I assume you are presently using a switched mode psu

Topic is: “bypass capacitor, long cable and switchmode PSU”

Because (cheap) switchmode PSUs have a tendency to be extremely noisy (lotsa big spikes in the voltage). That can lead to noise on the analog inputs as well as affect the sound of analog outputs (audio out of the mbSID for instance).

9

one way of removing a great part of voltage ripple from SMPS is to add a so callled Pi Filter to the PSU’s output, that’s a combination of two (preferrably low-ESR) electrolytic capacitors and an inductor (coil) inbetween these caps. The primary side cap should be medium-sized to limit the capacitive load on the PSU (they don’t really like having to drive too much capacity directly, some will have startup problems or even get damaged), so use e.g. a 1000uF one first, then the coil (which must be able to take at least the neccessary current), then another, bigger cap.

distributing some local caps is generally a good plan but won’t really remove the overall noise in the system power. so either smoothe the power like explained above or go for a linear psu - but building a linear PSU that can provide 4 A is a non-trivial task and will generate lots of heat.

S

10

thanks for all advice and ideas!

I think 99% of this problem is the proximity influence effect and not the “noise over the wire”

I have to insulate better the PSU from pots and wire near from it.

or move my psu in the box at another side.. I don’t know.

or build a faraday shield :slight_smile: or ??

11

I found little case for my psu:

I guess it wouldn’t be enough to insulate it.

someone could tell me the correct name for the proximity effect I have between wire and psu?

is it the famous EMI?

117695306-40.jpg
117695306-40.jpg261×200 9.37 KB

12

Yes, it’s electro-magnetic interference. that case would be only partially helpful, the ventilation slots are quite large - EM radiation can escape through any hole that is more than half the wave length in size. so if you wanna keep high frequency noise away from the rest of the box, you should try to have the ventilation slots go to the outside of the box and in all other directions there should be as much uninterrupted metal as possible. This will certainly make the situation better than before but I still guess that a part of your problem is wire-bound noise.

S

13

wire-bound?

14

what you called “noise over the wire” :wink:

15

okay :slight_smile:

for me, it was the same concept  (for me = for me, the cute noobie) :stuck_out_tongue:

the conclusion of all my tests around that is:

when a flat ribbon cable unshielded (with pots on one side & AIN on the other side) is putted VERY near from a part (transfo) of the PSU… false detection flows suddenly. if I keep cable as far enough of this part of the psu, NO problem.

I guess improvements could be done, but now, with the box closed, no problem with that.