I really don't understand how people's units get so hot. Mine run at 49C with minimum fan rpms (nearly)
how cold is your room in the first place? |
|
|
|
|
is it possible to have a temp log of the unit?
my avalon batch 1 seems to have random reboots where the miner just stops hashing and by the time i realised it, it has cooled down from possible overheating or just cooled down from all hashing, so i can't really tell if it's overheating.
if my miner is set --avalon-temp 54 , what is the cutoff temps for the miner to shutdown?
when it does overheat and shutsdown or goes idle, do i have to manually power down to let it cool or can i just let it run on idle and it will start hashing again once it's cool?
|
|
|
|
|
ok i tried doing some more mods
this time, adding 2 more fans on the outside of the case at the back (batch 1) and removing the fan grills. original fans at the front (actually the seller changed it to a different model, not the NMB types)
added rubber grommets for the rear fans to reduce vibrations-type noise.
couldn't add grommets to the front fans as i did not have bolts and screws that are long enough and the stock self-tapping screws are not long enough.
but strangely, my unit took a longer time to heat up and when it did, the front fans didnt spin at max rpm until it was way past the threshold temps.
i set the threshold temp to 54 degC and normally by when it reaches 51 degC, the fans would spin up loudly but this time round the fans only reached max rpm when temps touched 56 degC.
Separately I also tested using placing the unit vertically and horizontally, both with a box fan blowing at the hotter side (where the heatsinks are mounted).
the unit repeatedly overheated despite placing the hotter surface facing up and the box fan blowing at it, slightly angled towards it.
when placed vertically, it performed far better, maybe because the box fan was blowing directly at the hotter surface. it was able to hash overnight and for hours on end until the aircon was switched off.
Combined test: with 2 fans at the back and unit placed vertically with box fan blowing at the hot side
Results
front fans RPM are significantly lower than before, from 3700-3800rpm to 2600-2700rpm
noise wise (measured using some app on my phone), small changes, a few db lower but at least it's not that loud and high pitched.
temps have stabilised at 51 degC, and 2600-2700rpm.
More mods planned:
adding heatsinks to the hotter side (outside of the case) and mounting fans on it.
adding rubber c-lines to each edge, where the panels meet and adding rubber grommets to all fans.
use epoxy on the casing to secure the nut to make it easier to bolt it in.
the long term plans is to design and maybe get waterblocks milled for this.
|
|
|
|
What's your hashrate?
Batch3 4module 1200wPSU, 110.8 GH/s, 352-353 MHz, 0.3% reject rate, 1.75% HW errors Batch3 4module stock PSU, 111.6 GH/s, 355 MHz, 0.2% reject rate, 0.94% HW errors The new FW fixed a high reject rate problem on the first unit. Was running 2-3% rejects on the 0703 firmware. Mine looks worst. Batch 1 4 module Hashrate swings wildly so it's hard to judge. 1250w psu and around 0.56% rejected. Cant remember how to read the hw error rate though. The fact that I'm in a tropical country doesn't help much with thr temps |
|
|
|
|
I'm a bit concerned with the thermal paste used. The manufacturer seems to use so much of it that it overflows thru the holes and on to the circuit board itself and contacting the chip and other components.
Unless they use non conductive thermal paste.
|
|
|
|
Just watch the freq carefully when using --avalon-auto. New firmware is much more aggressive in overclocking.
After 1.5 hrs on a B3 4M unit with stock PSU, the HW error rate started climbing rapidly while running @360 MHz (6%+ error rate over the runtime, up from 1.7% minutes earlier). Clock speed was not ramping to down to compensate.
Now using --avalon-freq 345-355 to see if that prevents a runaway.
And before anyone says it, yes that unit is getting a PSU swap later this weekend.
What's your hashrate? |
|
|
|
anyone tried replacing the thermal paste between the chips and the heatsink?
or adding thermal paste between the heatsink and the casing (where the heatsinks are mounted onto the case) ?
I am  So what did u do? Replace thermal paste or put thermal paste between hs and case? What kind of thermal paste did use? Non conductive? |
|
|
|
|
anyone tried replacing the thermal paste between the chips and the heatsink?
or adding thermal paste between the heatsink and the casing (where the heatsinks are mounted onto the case) ?
|
|
|
|
bump for anyone's help
See my previous post for the link. [img]https://i.imgur.com/oZuHwwH.png[/ img] yeah i am using that right now but some of the holes are not suitable for batch one modules |
|
|
|
does anyone use 4 modules with batch1 avalons? i see the power board PDU is different than batch2 or 3 and i wonder if it is enough to power 4 modules. me... it's enough PSU's the more important part. |
|
|
|
i read from another thread that different batches of avalons have different number of mounting holes to the case?
batch 1, 8
2, 7
3, 5 screws?
Correct. The only things you need to cool are the chips themselves, the rest of the boards are designed passive. That's not to say other bits of the board won't need additional cooling, but for now they're cooled passively. anyway what about the dimensions and position of the holes? |
|
|
|
|
ok
I just did a similar mod for the Batch 1
all i did was tape the open areas between fans and the units. the units in batch 1 are very closely packed together so there's no need to tape it between the modules like dogie did.
Anyway results are rather clear from just 20mins of run time
temps stay the same but the fans work lesser to lower the temps.
they tend to stay at their max rpm for a shorter duration when it reaches the threshold temp
most of the time it's below 3000rpm (the fans are delta afc1212de instead of the nmb ones)
|
|
|
|
too cumbersome. plus not everyone's keen on having a huge water chiller around. waterblock design is too simple, wouldn't have enough cooling power. |
|
|
|
|
i read from another thread that different batches of avalons have different number of mounting holes to the case?
batch 1, 8
2, 7
3, 5 screws?
|
|
|
|
ok basically i am thinking of designing some waterblocks for avalon so i need some information regarding the design, pictures, heat output etc
1. any recommendations for free/cheap cadcam software? for newbie here. must output industry common file types.
2. how much of the energy it uses is given off as heat? eg if it consumes 700w (power meter), how much heat is given off? this is for me to estimate how much radiator space (in terms of 120mm fan) i need.
3. need detailed pics of batch 2 and 3 ( i have a batch 1 so i can look at it again and again)
1. Free/cheap is not analogous to viable for thermal sims. Solidworks is the route to go. Talk to a student friend if necessary, pricing is reasonable for an individual. OpenFOAM is a pretty good flow simulator and free, however you'll have to extrapolate junction temperatures from the flow simulations into your working fluid. There are a number of ways to import DXF's from Solidworks to OpenFOAM. 2. As stated in another post, all of the power goes to heat minus an infinitesimal amount, not worth measuring. Its a lot of work to model (correctly) thermal properties of systems. Perhaps you can heuristically determine parameters by purchasing the 10/$20 chinese waterblocks off of ebay, a thermocouple at the heat source (between chip and block), input and output, and experimenting with flow rates. i dont intend to do thermodynamics on this kind of thing. it's not as if i'm doing this for protein research. i just need to know things like which part of the pcb has alot of heat, eg just the chip's side (the rear of the chip that contacts the heatsink? what about rest of the circuit board (those capacitors and chokes?) |
|
|
|
|
ok basically i am thinking of designing some waterblocks for avalon so i need some information regarding the design, pictures, heat output etc
1. any recommendations for free/cheap cadcam software? for newbie here. must output industry common file types.
2. how much of the energy it uses is given off as heat? eg if it consumes 700w (power meter), how much heat is given off? this is for me to estimate how much radiator space (in terms of 120mm fan) i need.
3. need detailed pics of batch 2 and 3 ( i have a batch 1 so i can look at it again and again)
|
|
|
|
on a serious note... couldnt it be that its metal stretching / creeking against each other from temp variances?
You know, like the tick sounds you get when a car has been shut off from being on for a while?
i always thought it was due to some mechanical temp sensor u know when they layer 2 metals together and 1 shrinks/expands more than the other. anyway it's possible but seller says there's no such noise frm it when he had it. i intend to cut the side panel to add a plastic window so i can monitor it closely. |
|
|
|
to the OP... I think its just you fucking with us, I know you have a quarter in the other hand, just smacking the case =P LOL
haha jk, hope ya figure it out tho...never heard that outta mine. Then again my aftermarket fans are way to loud to hear something like that =P
i am still lost on the 2nd kind of ticking noise. its only now i realise there's 2 kinds of noise. 1 shallower 1 deeper. hope it's from the same source just that the air moving thru the case, acoustic characteristics made it sound different. |
|
|
|
|
Show Posts
