I am not saying that disabling wifi won't help. There is definitely what seems to be a second problem, that is taken care of by disabling wifi, which is a good thing to do in addition to removing the fuse F1. I have a hunch that it is because there is no decoupling capacitor close to USB-B1 (C44 is too distant), and the power consumption of the TP-LINK with wifi enabled is very variable, causing ripple across the internal 5V rail of the control unit...
Thank you for the explanation. Another question that I haven't been able to find an answer for. Why does removing the fuse matter if there is nothing plugged into the port? Is it feeding a 3.3v on the usb-hub or the port itself? I would have thought there would be nothing flowing over that path if there is no completed circuit (usb cable plugged in), is that not correct? (I can't see the PCB designs as I dont have altium viewer installed.. Their site seemed broken on the 3 browsers i tried to register with...) |
|
|
|
The description that the TP-LINK is "drawing too much power" is actually incorrect.
The USB hub chip does not overheat because the TP-LINK draws too much power. The TP-LINK is powered via a 5V line that does not even come from the USB hub chip.
It was just a guess on my part I have no meter or the necessary tools on hand to find out for sure. I couldn't think of any other explanation on why my unit became stable after disabling the wifi. The only thing disabling the wifi is going to do is reduce EMF and power draw. I currently have the F1 in place and my unit has been mining solid for the last 12 hours without a reboot or issue after disabling the Wifi. |
|
|
|
Thanks for the update! One question regarding this line in Wiki Remove the F1(Fuse) for fix the USB-HUB chip overheat, if you are batch2 user and your received Avalon between Apr 26 2013 ~ May 31 2013, your FPGA contoller board still have F1 mounted. you need do this by yourself. How to do this at home properly? You cant/shouldnt just rip it off. You'll need to have a soldering iron to do it properly. Heat up the solder and then pull it off once it's free. Note: I don't know of anyone who has done the F1 fuse removal as of yet. Xiangfu said all new controllers shipped will have this done. I'm scared to do it heh. |
|
|
|
Woot, My cgminer has now been running for 6h 41m 50s without even so much as a cgminer restart. 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. |
|
|
|
|
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.
|
|
|
|
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.JPGJust 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 |
|
|
|
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 |
|
|
|
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. |
|
|
|
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. |
|
|
|
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.JPGEdit: I've uploaded a much clearer photo which shows damage which I was not even able to see with the naked eye. Very concerning. |
|
|
|
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.JPGThe 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. |
|
|
|
And you send the message to the whole channel or to someone?
I cried myself a river in the channel (quite literally). |
|
|
|
I sent them an email, a PM, an irc message and a new support ticket. Apparently someone read one of those as my shipping method was switched to DHL and I received a new updated on my order with a new shipping number; but no direct response "Sir, your request has been processed" sort of thing. So I have no idea what method was successful in contacting them. It appears they're trying to get their CS together. Hope everyone else with issues gets them sorted timely.
senseless, I am wondering on which server channel did you send them IRC message 10X #avalon on freenode |
|
|
|
Got my box today. DHL got their hands on it on May 28 around 1:30PM, it was delivered to my local city on May 29 at 8:41 AM. On may 29 at 10AM I went in and processed customs forms myself was charged roughly 100$ in taxes. Took me 6 hours to navigate the endless gauntlet of minions but was able to get my unit a day early by processing the customs forms myself.
My unit has dust caked as well. I don't think it was from mining so much as china having air so thick you can literally eat it.
Glad to hear you got yours. What happened to the one they shipped by EMS accidentally? Maybe you'll get two for the price of one haha. I sent them an email, a PM, an irc message and a new support ticket. Apparently someone read one of those as my shipping method was switched to DHL and I received a new updated on my order with a new shipping number; but no direct response "Sir, your request has been processed" sort of thing. So I have no idea what method was successful in contacting them. It appears they're trying to get their CS together. Hope everyone else with issues gets them sorted timely. |
|
|
|
Got my box today. DHL got their hands on it on May 28 around 1:30PM, it was delivered to my local city on May 29 at 8:41 AM. On may 29 at 10AM I went in and processed customs forms myself was charged roughly 100$ in taxes. Took me 6 hours to navigate the endless gauntlet of minions but was able to get my unit a day early by processing the customs forms myself.
My unit has dust caked as well. I don't think it was from mining so much as china having air so thick you can literally eat it.
Congratulations!! When did you pay the order? Feb 2 order #1073 not sure the exact time. |
|
|
|
|
Got my box today. DHL got their hands on it on May 28 around 1:30PM, it was delivered to my local city on May 29 at 8:41 AM. On may 29 at 10AM I went in and processed customs forms myself was charged roughly 100$ in taxes. Took me 6 hours to navigate the endless gauntlet of minions but was able to get my unit a day early by processing the customs forms myself.
My unit has dust caked as well. I don't think it was from mining so much as china having air so thick you can literally eat it.
|
|
|
|
|
UPDATE:
I received a response to my email and/or support ticket. My account has a new update that updated my shipping from EMS to a DHL tracking number (Sent as an update to my account/order). They responded to my request in <48 hours or so. So apparently they do read them, but they must just not personally respond to every one of them.
It appears they are indeed trying to improve their customer service.
|
|
|
|
Does the KC705 have realtime PSU monitoring? If so, how many amps is it really using on the vcore 1v rail? I'd really like to know if Vivado's estimates are anywhere near close!
Use the chipscope that comes with the software |
|
|
|
@OP: Don't be born in that country.
Problem solved.
I wasn't, I just happen to live here. Which country is it ?
Some corrupt SEA nation. |
|
|
|
|
I received a response to my email and/or support ticket. My account has a new update that updated my shipping from EMS to a DHL tracking number (Sent as an update to my account/order). They responded to my request in <48 hours or so. So apparently they do read them, but they must just not personally respond to every one of them.
It appears they are indeed trying to improve their customer service.
|
|
|
|
|
Show Posts
