Official Avalon Technical Support Thread

[John Self]

where is the p5 jumper trying to find this on the wiki right now....

on the fpga board?

[JerimiahJohnson]

just found it before you posted that, damn that thing was hidden!

[John Self]

just found it before you posted that, damn that thing was hidden!

I did say next to the USB port

[JerimiahJohnson]

well I have determined the FPGA controllers were broken as well during shipping on both these machines.

*sigh*

so I am withdrawing the bounty but I will give the two gentleman and luke-jr who also helped me determine this 1 bitcoin each sometime tomorrow when my funds come into Gox.


Man they really need to do a better job of shipping these, If I had a working chasis with working fpga controllers and routers, I could build 2 whole 4 module machines, this... really... sucks...

thanks anyways guys please send me your BTC address and Ill make sure you each get a satoshi in the morn

peace -JJ

[JerimiahJohnson]

I mean come on, your using aluminum - an extremely soft metal, and your also using incredibly small screws, like 4 or 5 on each module. its bad news, a couple tumbles and your machine is trashed.

they really need to put some sort of foam bricks that can later be removed between the modules, so it wont be so easy for them to strip the tiny screws break off and smash all the internal components.

Don't get me wrong I love them and I am thank full for the ones I have but I really hate to pay for broken units.

[fpgaminer]

It sounds like you still have multiple, working machines.  Is there a reason you can't take the modules out of the two broken machines and put them in the other machines?  If you've received 2/3's of your order, and still missing 10 machines, that means you ordered 30, and currently have 18 working machines.  Plenty to slot the "extra" 6 modules into.

[JerimiahJohnson]

yeah I could, I am just weary of messing with units that are currently incredibly stable.

I suppose I could use bigger screws for the stripped holes.


not sure how else I could attach the modules unless I just lay them on their side which I already have a couple of those from previous shipments that had 1 or 2 modules which snapped off.

i spent a little over a thousand a piece on these units I feel bad for the people who ordered in batch 3 - paid nearly 6 grand per unit
and they are going to receive messed up units unless they make some changes in the way they ship them.

and yes I have 18 working units right now - two of them I have added an additional module to, but you basically need a broken unit and a high end PSU just to be able to make these 4 module units.


I am told these cables are not just simply IDE cables and where the hell do you get the power cords from?

if you dont have broken units to source from, im not sure how you will make 4 module units
which by the way must have intense cooling or they shut theselves down.

-JJ


btw: fpga miner and john PM me your BTC address - I'll make sure to send you each a coin tomorrow

I think its important to keep this tech support thread going and keep experts around to answer questions and participate
Avalon has stated since the begining that they are not interested in providing tech support.

[cypherdoc]

You're right about expecting them to change their shipping practices.

I predicted this type of problem from the beginning as it happened to me with my gpu miners last year.

The Styrofoam block idea is a good one to prevent snapping the modules off. Or sending the modules separately wrapped.

We even had insurance on my gpu's with FedEx but they literally refused to honor it despite overwhelming evidence it was their fault. I don't ever use them anymore.

[SolarSilver]

The Styrofoam block idea is a good one to prevent snapping the modules off. Or sending the modules separately wrapped.

A styrofoam block would mean every customer has to open the case to remove it or do you suggest a padding
that stays inside the case when in use? What about free airflow?

A much more simple solution would be to have both screws on the bottom of the modules (like now) and screws on the top so the modules are sandwiched in. You'll need some hex spacers too as the modules don't touch the top cover.

So extra holes in the top cover (identical to the bottom one) and extra holes in the heatsink.

I prefer it when the customer does not have to open the case, not everybody will like to fiddle with the captive nuts, it's prone to errors.

[Sitarow]

***Note***
These two units were ordered in Batch 1. However I am led to believe that unit #445 is possibly from Batch 2.

Upgrading Unit 114 power supply reduced power consumption by 70 watts,
Upgraing Unit 445 power supply reduced power consumption by 50 watts at 120 volt input power

I wanted to share my test results after running the updated Firmware released for Avalon on 4.19

I have also included details on the power supply upgrade that was made.

Both units have been upgraded with the Corsair 860i.

Before the power supply upgrade, power consumption on both units running at speed setting of 300 was at 1290 Watts.

After the power supply upgrade both units running at 300 now consume 1160 Watts of power.

Below you will see the results taken at 1, 13, and 26 hour snap shots.

Both Unit 114 and 445 are operating at 300 speed setting with the new firmware loaded and Upgraded Corsair 860i power supply.

Both Unit 114 and 445 combined running at 300 speed setting are pulling 1160 Watts of power.

After 1 Hour

Unit #114 after 1 hour on wifi speed set at 300


Unit #445 after 1 hour on wifi speed set at 300


Screen shot from pool results after 1 hour



After 13 Hours

Unit #114 after 13 hours on wifi speed set at 300


Unit #445 after 13 hours on wifi speed set at 300


Screen shot from pool results after 13 hours


After 26 Hours

Unit #114 after 26 hours on wifi speed set at 300


Unit #445 after 26 hours on wifi speed set at 300


Screen shot from pool results after 26 hours


As you can tell from the images these units are running stable and maintain consistent performance during this time.

As an extra I have also included the before and after images of the power supply upgrade.

The following are before and after power readings for each unit, followed by before and after upgrade images.

Unit #445 Before and After Power Readings

Initial Load both units combined was at 1280
Amps, Watts, Volts


Power Consumption after single Unit Upgrade


Unit #445 Before and After Power Supply Install

Unit #445 Power Supply Old vs New


Unit #445 Before


Unit #445 After

  


*************************************

Unit #114 Before and After Power Readings

Note that these readings are for two Units. First unit "445" has already been upgraded.

Power Readings Before upgraded Unit 114



Power Readings After Unit #445 and Unit #114 power supply have been upgraded
These photos have been taken after 27 hours of operation

Amps, Watts, Volts


Unit #114 Before and After Power Supply Install

Unit #114 Power Supply Side by Side


Unit #114 Before


Unit #114 After

[mitak64]

What else, besides the relatively lower temps of unit 445, makes you think it's a Batch 2 ?

[cypherdoc]

Well done sitarow.

Did you consider externalizing the PSU?

Someone has demonstrated much lower temps.

[Sitarow]

What else, besides the relatively lower temps of unit 445, makes you think it's a Batch 2 ?

I suspect the voltage is lower on Unit 445 as it is slower then the first unit.

Also the power supply was 700 watt on unit 445. The temps are lower on unit 445 and the fans are set for lower speeds.

So Batch 1 is 67000 at 282
and Batch 2 is 65000 at 282

Or Batch 1 is 72000 at 300
and Batch 2 is 70000 at 300

I think that I can push these units over 300 but I would wait for advice on that.

I also noticed that the fan speed is selected from the first temp sensor that is located on the controller.

I am tempted to put the fans at 100% speed as they are hardly being used in the present environment.

These units with this setup are Stable even at 26 temp 1 and 49 temp 2

Unit 445 is overall lower temps and I wish I could set the fan speed higher then the software 30% without modding.

[Sitarow]

Well done sitarow.

Did you consider externalizing the PSU?

Someone has demonstrated much lower temps.

The heat on the new power supply is non issue. I can set that power supply fan to be at 100%  but there is no need for that.

What I wish to do is increase the fan speeds of the two black fans. However I would like to do that via software.

[Sitarow]

I have recieved requests for the location of the unit number.

They are located near the location of the 4th module close to the back inside the unit.

Unit #114


Unit #445



And an update.

Unit #114 is crunching away. However into its 40th hour unit #445 stalled however restarted itself.
After 52 hrs into this experiment Unit #445 stalled however cgminer did restart but would not process work.

I left it in that state over 4 hours and nothing. I had to do a hard boot by killing power. I will keep an eye out and post a log of the event.
However I suspect that the main controller may be having the issue.

Question I have if anyone is able to answer

How do we increase fan speed without modding?

Here are screen shots of the standings.

Unit # 114 After 56 Hours of operation





Unit# 445 After 1 Hour of Operation 56 Hours into tests

Screen shot below is the result after 1 hour. Unit itself never lost power or lost wifi


Details from pool for Unit# 114 and Unit# 445 after 56 Hours into Test




***** Note *****

Unit # 445 after 56 Hours into test.

Unit # 445 stalled at 44 hours into test and restarts time to restart unknown. Unit #445 stalled at 51 hours into test again.
Cgminer restarted but would not process work but would restart in an attempt to process work.
Never self corrected from this state for 4.5 hours. Hard reboot issued at 55 hours solved condition.

[cypherdoc]

Sitarow,

please consider making your images smaller by using:

Code:

[img width=900]***[/img]

[Sitarow]

Unit #445 has been up for 24 hrs with out any problems after I powered it off and on.

I also blocked one of the rear holes so it forces air out of the fan hole in the middle.

Seems to have helped however fan speed is IMO still to slow. I will be getting a manual fan control unit for the fans today and test if that helps.

http://www.newegg.com/Product/Product.aspx?Item=N82E16835118217

Edit: Have not yet done the manual fan control option. Will update when I move in that direction.

Still holding out for software fan control.

[Sitarow]

Unit #445 has been up for 39 hrs with out any problems after I powered it off and on.

However Unit #114 has stalled 92.5 hours into the test. Software reboot command did not solve the problem.

On examination the fans were running at 100% but not registering after auto cgminer restart nor software initiated system reboot.

Seems that there is a problem with the controller hanging.
Edit: or the usb communication to the controller. If this happens again I will open the unit up and unplug the usb and plug it back in. Unless that is a bad idea

Note Unit # 114 stalled after 92.5 hours of operation. Waited 4 hours to see if it would fix itself but it just kept restarting cgminer.
Like Unit # 445 it only started working after a powering the unit down.

[kaerf]

I've been getting a lot more failures lately that require a hard reboot. Don't know if it's the 419 firmware or something else...

[PuertoLibre]

It is a good thing we have beta testers that allow folks to wait until these are "stable builds". The Firmware seems to be improving quite a bit.

@ BitSyncom

Any chance the Corsair power supply can be added to the Wiki as a viable solution?