Originally Posted by CZ_the_DAY
Many have observed how temperature affects the faulty ONKYO but I can elaborate. While my unit would be on the verge of instability with artifacts in the video, I'd barely cool* the chip SiI9185' (3to1 HDMI switch) and the signal disappeared until it warmed. Same conditions, cooling* neighbouring chip SiI9013' (HDMI receiver) and the signal became flawless temporarily. IOW change a chip's tolerance to noisy power by means of temperature. (Initially I feared faulty chips!) This helps explain the commonly reported wait times after first powering on.
Sorry for dredging this old thread up, but just in case someone was wondering, there is a sort of technical explanation that might explain this weird temperature sensitivity.
I am assuming for the purposes of this discussion that the HDMI receiver IC connects between the HDMI input connector and the HDMI 3-to-1 switching IC on this receiver.
CMOS transistors are sensitive to temperature. When they are cool, they switch faster, causing the output transitions of its internal logic gates or its external output buffers (output pin drivers) to be sharper and the IC draws more current as a result.
If the HDMI switching IC is cooled, it might have faster internal logic edge rates and draw more current from its supply, increasing its sensitivity to local regulator power supply ripple from failing capacitors and causing internal data corruption, particularly at the input buffers where the effective input voltage high and low levels might vary due to power supply ripple shifting the input switching threshold voltages.
If the HDMI receiving IC is cooled, it might also start switching faster and create sharper transitions on its output, increasing the edge rate and improving the apparent signal-to-noise ratio seen at the input pins of the HDMI switching IC that is experiencing excessive power supply ripple and having input threshold issues.
This is just an educated guess, but I have personally debugged similar issues with digital interfaces in the past by using various methods to modify the buffer output switching speeds, or modify the internal switching speeds, to affect the edge rate and power consumption of similar devices.
I currently own 6 various models of used Onkyo receivers that I purchased from private parties for a project. Every single one of them turned out to be in the early stages of HDMI failure or already failed when I got them. They do not qualify under Onkyo's recent recall because I do not have an original receipt and they were not purchased from an authorized reseller, so I am in the process of repairing them myself using information similar to this thread.