Quote from: BitMainBuyer on July 21, 2015, 05:55:47 PM

Printed off the 2 ducts and have them mounted with 2x Ultra Kaze's. Here's a comparison:

Old setup
Kaze Pull + Stock push + no ducts = ~53°

New setup
Kaze Pull + Kaze push + 2 ducts = 63° / 59° (there's a few degrees difference between both boards strangely..)

Are the above temps expected with my new setup? I was expecting a bit cooler, especially in this air conditioned environment.

It's possibly worth noting I don't have top or bottom plates due to printing difficulties.

Edit: temps have risen to 65° / 61° now Sad

The temperature readouts are expected to be higher (more accurate) when there is no airflow hitting the temp sensors. It may seem that your miner is hotter, but it isn't, you just get the actual temps. Should you put your own temp sensor on the heatsink, you would see the difference.

Since miners are optimized to have a temperature reading that is 20°C off (from my experience with stock fan, your miner is actually around 100°C when readout reaches 80°C), the bottom duct has small openings to move some air around the temp sensors and lower the temps on them a bit. I have one miner running at 400MHz without a problem in a room that is reaching 45°C during the day. The other one, however, has an older heatsink and I can't push it as hard. Cheesy

It's happily working at 381.25MHz.

Both of my miners regularly have a 5°C difference between boards. I'm unsure if it's the cheap temp sensors or the difference actually exists, I haven't tested.

The best way to see how well is the miner running is to compare HW errors. Up the frequency and let it run for several hours. As long as you're under 0.01% HW errors, you're good.
Here's how's my 400MHz running:
http://www.pohrani.com/f/3h/Do/3LYATpZi/400mhz.png