..... also, using some counter-wired bobines or C-L-C pi-greek filters can help too ..... like the one that you already have placed on the output of the head ..... if you have occasion, try to complete it (adding another capacitor also from output and GND, and paralleling them with 100nF ceramic ones) ..... same thing for the one on the dc-dc inputs, and usinf the same on the dc-dc outputs, you probably end taking away 90% or more of these spikes ..... ..... and, also, use good size wires (i know that there's no high currents in the circuit, but is always a good thing to use wires of a decent size for wire-up power sections, like, 0,5mm dia, at least, especially for the input of the dc-dc converters)
Yes, and I was using a linear regulated +/-12vdc Lambda PSU when I was building the prototype of the Kenometer Pro series. That is why the problems didn't appear until later. It worked fine for the prototype but wouldn't allow the meter to be portable and run off batteries and USB.
I did more testing today and found my second drawing to actually work significantly worse than the original.
These converters also emit a strong enough EMF that nearby wires were having noise inducted into them, enough to cause fluctuations on the mcu and visible spikes on my oscilloscope.
I have since removed the circuitry and reverted to my original design. Once again the power is smooth and free from any visible ripple on my scope. I have instructed Kenom to do the same. The ground for the thermopile was also soldered directly to the MCU board instead of the converter stack. This eliminated the small voltage difference created in the inductor.
) ..... and it ended with the IC that had that absorption only without input signals, rising to 80mA with the signals applied, and the 100mA dc-dc converter was messing up all with ripples and spikes (ofcourse ), until i've substituted it with a 250mA converter .....
