Avalon ASIC users thread - page 174.

senseless

hero member

Activity: 1118

Merit: 541

Woot, My cgminer has now been running for 6h 41m 50s without even so much as a cgminer restart.

Quote from: drakness on May 30, 2013, 12:31:13 PM

Senseless: Sorry for your troubles and these neophyte questions, trying to understand... you are recommending using the ethernet port, but not the wifi?

You suggest that perhaps the TPLink adapter drawing enough power to run both ethernet and wifi, concurrently, might have damaged the chip?

To use ethernet port, is there anything involved besides setting a static IP, other than 192.168.0.100; so you know how to access the web interface?

Correct, just disable the Wifi but leave ethernet in-tact. Wifi is really the only thing you can disable to reduce power.

Yes, that is my current thought that the damage from my chip resulted from the tp-link drawing to much power.

There is nothing else to it other than what you describe.

drakness

newbie

Activity: 33

Merit: 0

Senseless: Sorry for your troubles and these neophyte questions, trying to understand... you are recommending using the ethernet port, but not the wifi?

You suggest that perhaps the TPLink adapter drawing enough power to run both ethernet and wifi, concurrently, might have damaged the chip?

To use ethernet port, is there anything involved besides setting a static IP, other than 192.168.0.100; so you know how to access the web interface?

senseless

hero member

Activity: 1118

Merit: 541

Think I found the problem and the reason my chip fried. Good news is it wasn't damaged (still works fine).

The TP-Link is powered via USB. I reduced the power usage of the tp-link by turning off the wireless modem. The unit has been running for 2 hours now without so much as a fan hiccup. No more -71 or -32 errors in the logs.

I would recommend everyone do the same on the v1.5 controller, disable and delete the wlan interface.

phoenikx

newbie

Activity: 44

Merit: 0

Quote from: senseless on May 29, 2013, 01:52:08 PM

21C atm, unit is sitting on a table humidity is very low

...

I dont think the chip damage is whats causing the instability in my system. Talked with another person who has the same cycling issue with no chip damage. The following crontab script on the miner is a stop gap measure to keep it mining and correct the issue. By cycling the disconnect/connect of the USB controller to the tp-link i can reset my hashes to max.

*/10 * * * * echo "1-1.1" > /sys/bus/usb/drivers/usb/unbind ; sleep 2 ; echo "1-1.1" > /sys/bus/usb/drivers/usb/bind ; /usr/bin/cgminer-monitor

The chip on the left also does not seem to be in such a good condition.

senseless

hero member

Activity: 1118

Merit: 541

Quote from: Kartaiv on May 29, 2013, 01:41:24 PM

Quote from: senseless on May 29, 2013, 01:29:05 PM

Quote from: Kartaiv on May 29, 2013, 01:26:24 PM

Based on the way that 2 pin header melted it looks like the FTDI chip is putting out a lot of heat. The PCB has the ground pad and vias on the PCB top for that chip though. No thermal pad on the bottom. I wonder if you helped dissipate the heat from that chip if it would help...

Unit just started cycling again. Reported sick, then the unit came back at a whopping 14-30Mh/s fans reving up and down.

The FTDI chip isn't warm at all. I'm thinking that usb-a goes to the FTDI and the USB-b goes to whatever chip was destroyed on my board. Unfortunately I don't have a "normal" usb cable to test that theory. I'll give that a try tomorrow.

http://www.mysenselesslife.com/avalon/DSCN5215.JPG

Just curious, what are your ambient conditions around your unit?

21C atm, unit is sitting on a table humidity is very low

...

I dont think the chip damage is whats causing the instability in my system. Talked with another person who has the same cycling issue with no chip damage. The following crontab script on the miner is a stop gap measure to keep it mining and correct the issue. By cycling the disconnect/connect of the USB controller to the tp-link i can reset my hashes to max.

*/10 * * * * echo "1-1.1" > /sys/bus/usb/drivers/usb/unbind ; sleep 2 ; echo "1-1.1" > /sys/bus/usb/drivers/usb/bind ; /usr/bin/cgminer-monitor

Kartaiv

member

Activity: 61

Merit: 10

Quote from: senseless on May 29, 2013, 01:29:05 PM

Quote from: Kartaiv on May 29, 2013, 01:26:24 PM

Based on the way that 2 pin header melted it looks like the FTDI chip is putting out a lot of heat. The PCB has the ground pad and vias on the PCB top for that chip though. No thermal pad on the bottom. I wonder if you helped dissipate the heat from that chip if it would help...

Unit just started cycling again. Reported sick, then the unit came back at a whopping 14-30Mh/s fans reving up and down.

The FTDI chip isn't warm at all. I'm thinking that usb-a goes to the FTDI and the USB-b goes to whatever chip was destroyed on my board. Unfortunately I don't have a "normal" usb cable to test that theory. I'll give that a try tomorrow.

http://www.mysenselesslife.com/avalon/DSCN5215.JPG

Just curious, what are your ambient conditions around your unit?

senseless

hero member

Activity: 1118

Merit: 541

Quote from: Kartaiv on May 29, 2013, 01:26:24 PM

Based on the way that 2 pin header melted it looks like the FTDI chip is putting out a lot of heat. The PCB has the ground pad and vias on the PCB top for that chip though. No thermal pad on the bottom. I wonder if you helped dissipate the heat from that chip if it would help...

Unit just started cycling again. Reported sick, then the unit came back at a whopping 14-30Mh/s fans reving up and down.

The FTDI chip isn't warm at all. I'm thinking that usb-a goes to the FTDI and the USB-b goes to whatever chip was destroyed on my board. Unfortunately I don't have a "normal" usb cable to test that theory. I'll give that a try tomorrow.

Image of my heat sink job:

http://www.mysenselesslife.com/avalon/DSCN5215.JPG

Kartaiv

member

Activity: 61

Merit: 10

Based on the way that 2 pin header melted it looks like the FTDI chip is putting out a lot of heat. The PCB has the ground pad and vias on the PCB top for that chip though. No thermal pad on the bottom. I wonder if you helped dissipate the heat from that chip if it would help...

WinTame2012

sr. member

Activity: 315

Merit: 250

Official sponsor of Microsoft Corp.

Quote from: senseless on May 29, 2013, 01:18:42 PM

Quote from: Kartaiv on May 29, 2013, 01:17:36 PM

Quote from: senseless on May 29, 2013, 12:53:08 PM

Quote from: Kartaiv on May 29, 2013, 12:44:31 PM

Curious to see how the heatsink on top works out. It's probably not how you think it will, but hopefully it helps. That USB chip is not designed to dissipate heat upwards -- It's designed like the ASIC chips to push heat through the board down. Unfortunately, the PCB is not designed to handle that..

Def not going to be as effective; but if it's possible to get some heat out of it, it might just run stable.

Edit:

Just managed to slice a piece of heat sink off small enough to put on it. Been running for 5m and the sink is definitely above ambient. The hash rate seems a little more stable. The sink isn't HOT but it is taking some heat out of the chip.

Are you using some sort of thermal paste/adhesive between the sink and the chip?

Yes, its a thermal tape stuff (it was on the back of the sink i just scraped a bit off and reapplied after i was done butchering it.

I'd try mount on top of this tiny heatsink a 40mm minifan. If this chip is a QFN package than just heatsink wouldn't be enough.

senseless

hero member

Activity: 1118

Merit: 541

Quote from: Kartaiv on May 29, 2013, 01:17:36 PM

Quote from: senseless on May 29, 2013, 12:53:08 PM

Quote from: Kartaiv on May 29, 2013, 12:44:31 PM

Curious to see how the heatsink on top works out. It's probably not how you think it will, but hopefully it helps. That USB chip is not designed to dissipate heat upwards -- It's designed like the ASIC chips to push heat through the board down. Unfortunately, the PCB is not designed to handle that..

Def not going to be as effective; but if it's possible to get some heat out of it, it might just run stable.

Edit:

Just managed to slice a piece of heat sink off small enough to put on it. Been running for 5m and the sink is definitely above ambient. The hash rate seems a little more stable. The sink isn't HOT but it is taking some heat out of the chip.

Are you using some sort of thermal paste/adhesive between the sink and the chip?

Yes, its a thermal tape stuff (it was on the back of the sink i just scraped a bit off and reapplied after i was done butchering it.

Edit: Hit and passed the 15m mark.

Kartaiv

member

Activity: 61

Merit: 10

Quote from: senseless on May 29, 2013, 12:53:08 PM

Quote from: Kartaiv on May 29, 2013, 12:44:31 PM

Curious to see how the heatsink on top works out. It's probably not how you think it will, but hopefully it helps. That USB chip is not designed to dissipate heat upwards -- It's designed like the ASIC chips to push heat through the board down. Unfortunately, the PCB is not designed to handle that..

Def not going to be as effective; but if it's possible to get some heat out of it, it might just run stable.

Edit:

Just managed to slice a piece of heat sink off small enough to put on it. Been running for 5m and the sink is definitely above ambient. The hash rate seems a little more stable. The sink isn't HOT but it is taking some heat out of the chip.

Are you using some sort of thermal paste/adhesive between the sink and the chip?

senseless

hero member

Activity: 1118

Merit: 541

Quote from: Kartaiv on May 29, 2013, 12:44:31 PM

Curious to see how the heatsink on top works out. It's probably not how you think it will, but hopefully it helps. That USB chip is not designed to dissipate heat upwards -- It's designed like the ASIC chips to push heat through the board down. Unfortunately, the PCB is not designed to handle that..

Def not going to be as effective; but if it's possible to get some heat out of it, it might just run stable.

Edit:

Just managed to slice a piece of heat sink off small enough to put on it. Been running for 5m and the sink is definitely above ambient. The hash rate seems a little more stable. The sink isn't HOT but it is taking some heat out of the chip.

Kartaiv

member

Activity: 61

Merit: 10

Curious to see how the heatsink on top works out. It's probably not how you think it will, but hopefully it helps. That USB chip is not designed to dissipate heat upwards -- It's designed like the ASIC chips to push heat through the board down. Unfortunately, the PCB is not designed to handle that..

WinTame2012

sr. member

Activity: 315

Merit: 250

Official sponsor of Microsoft Corp.

Quote

I've uploaded a much clearer photo which shows damage which I was not even able to see with the naked eye. Very concerning.

Oh my... Could anybody else check his "being in work" B#2 unit for this issue?

senseless

hero member

Activity: 1118

Merit: 541

Quote from: WinTame2012 on May 29, 2013, 11:19:24 AM

Is this B#2 or B#1 unit?

B#2

I noticed that the B#2 controller has a second USB header on there as well. I'm not sure if that goes to a different comm chip (not the same as the first header?)?

http://www.mysenselesslife.com/avalon/DSCN5212.JPG

Edit:

I've uploaded a much clearer photo which shows damage which I was not even able to see with the naked eye. Very concerning.

WinTame2012

sr. member

Activity: 315

Merit: 250

Official sponsor of Microsoft Corp.

Quote from: senseless on May 29, 2013, 11:05:24 AM

Has anyone noticed on their controller fpga board that the little USB chip deformed due to heat? When i put my finger on it, it's extremely hot. When the unit boots up cold it runs fine for about 15-20 minutes then hash rate goes to zero and it starts cycling running at low hash rates <1Gh/s. It seems like that chip on the controller needs a heat sink. I re-arranged my cabling and went zip tie crazy which seemed to improve the situation somewhat; but the unit still dies after a time. Checking the TP-Link, I was getting usb errors when the unit stops hashing.

https://en.bitcoin.it/w/images/en/a/ab/Avalon-controlunit-v1.2-usb-serail-chip.JPG

The chip is so deformed/warped on my board the lettering is completely gone.

Going to see if i can find a tiny heat sink to affix to it tomorrow. Maybe that will improve things.

Is this B#2 or B#1 unit?

senseless

hero member

Activity: 1118

Merit: 541

Has anyone noticed on their controller fpga board that the little USB chip deformed due to heat? When i put my finger on it, it's extremely hot. When the unit boots up cold it runs fine for about 15-20 minutes then hash rate goes to zero and it starts cycling running at low hash rates <1Gh/s. It seems like that chip on the controller needs a heat sink. I re-arranged my cabling and went zip tie crazy which seemed to improve the situation somewhat; but the unit still dies after a time. Checking the TP-Link, I was getting usb errors when the unit stops hashing.

https://en.bitcoin.it/w/images/en/a/ab/Avalon-controlunit-v1.2-usb-serail-chip.JPG

The chip is so deformed/warped on my board the lettering is completely gone.

Going to see if i can find a tiny heat sink to affix to it tomorrow. Maybe that will improve things.

Bogart

legendary

Activity: 966

Merit: 1000

I ran into a new sort of issue with my (batch 1) Avalon tonight.

I noticed it was no longer mining, and when I logged into the web UI, it loaded, but I was unable to load the cgminer status tab.

I ssh'd in, and saw way too many instances of /usr/bin/cgminer-monitor running:

http://pastie.org/7977320

At that point I rebooted the Avalon using the 'reboot' command, and it resumed working normally.

I'm running the 20130419 firmware.

johnyj

legendary

Activity: 1988

Merit: 1012

Beyond Imagination

I'm still running 0225 firmware, it has been up and running since my last recondition of the heatsink, more than a week. Periodically cgminer would restart after FPGA ran into disorder and reported all strange values from sensors (thus no more submitted shares), but after cgminer restarted, everything would back to normal again, no reboot/hardware reset needed

With this firmware, I can confidently leave the machine running at home and go for vacation

I also tried 0419 version, although the FPGA seldom run into disorder, but once it does, no cgminer restart will solve the problem, a power cycle is needed

uhoh

hero member

Activity: 742

Merit: 500

Circle gets the Square

Quote from: Bicknellski on May 28, 2013, 07:42:37 AM

Quote from: WinTame2012 on May 28, 2013, 07:38:18 AM

Quote from: -ck on May 28, 2013, 05:56:18 AM

Quote from: -ck on May 27, 2013, 07:09:05 AM

~~The direct USB code does not really work at all on the MIPS router that is built into the Avalon device itself~~

Cancel that. Xiangfu set me up with usbmon on his standalone avalon and I was able to track down the problem so direct USB is now working on the avalon itself \o/

Glorious. I gonna tip you from my Avalon's income.

+1

Nice work.

+1

Topic: Avalon ASIC users thread - page 174. (Read 438724 times)