Topic: ANTMINER S3+ Discussion and Support Thread - page 17. (Read 710164 times)

Take it apart and measure the voltages going into each chips. They should be 0.75V

i have some thermal paste on order going to try that first with reapplication. One board shows ------- on another miner. Hoping the boards aren't dead just need some touch up. Any other suggestions would be appreciated. thanks!

having an issue with my S3+

Just recently purchased through craiglist.

Worked for 3 days without issue now I get XXXXXXXX. I've tried new PCIE connectors, New PSU, updating and loading original firmware to no avail. Any assistance would be great. thanks!

http://imgur.com/cxxfUTL

Also the HWs are high as heck!

Just for a common knowledge: Faulty chip  is U21.Marking on top reads: EB=3L. Short circuits 1v8 and 3v3 lines. Controller seems to boot but after that red and green just keep flashing. Pretty easy to repair, just six legs in that component.

Thank you sir (Y)

looks ok. its normal for it to fluctuate. its the avg you are looking for.

Hi guys,
This is my first time using a dedicated miner, few years ago I used my graphics card.
Do these stats look ok regarding the s3? and what is the best method for making the s3 last? keeping the temps cool? turning it off once in awhile? cleaned out the fans and internals when purchased. Got this at a bargain price in the UK £80 looking to expand on this after a few months.

Current Setup, running at 428.4 Gh/s.
https://i.imgur.com/OrXnuOw.jpg


Iv noticed my miner drops to 350gh/s then back upto 400+ is this common?
https://i.imgur.com/lE64XZv.jpg

not sure if this is in the correct section.

Only Tx line is not what it should. It's that 1.8 before it starts to hash. When it's hashing that voltage drops somewhere between 1.5-1.6 volts. If you have an oscilloscope you can check if there are pulses,or is it a flatline. It's fourth pin from left in upper row. I would suspect pic's fw but if same thing happens with other controller too that can't be the cause. Check if oscillator is running on hashboard.

Here are the voltages coming from the controller to the hashing boards :

hashing board 1
http://i66.tinypic.com/rkvim8.jpg


hashing board 2
http://i65.tinypic.com/2ymuoax.jpg

Thanks for you reply. I will check the voltages on the pins you suggested. I did try swapping the data cables from one unit to the other and that does not change the bad unit(cables from bad unit on good unit work fine).

You have checked that 3v3 from controller? Check if it gets to hash-boards too. Also check 1v8 in the hash-board. And oscillator if you have equipment. Also swap cables from working unit. Connector pin 3 is 3v3 and pin 4 is 1v8.

I have 2 antminer s3+. One that is fine and one that yesterday stopped finding the hashing boards. I can connect to the controller card and have reset and changed the firmware(downgrading to the first firmware for the S3+). I have also swapped the controller card from the working one and the power supplies to see if that fixes the problem but it does not.

When the S3+ boots up both fans turn and controller card and network are up(I can access the web server of the controller card). On the miner status page though it does not show any information under the “ant miner” panel. Pool is alive too by the way. LED lights are blinking both RED and GREEN quickly at the same time. Load seems very low compared to the other miner too. Under the system log it looks like the controller never finds “USB Bitmain asic”.

 I also see +3.3V on the controller and +12V on the hashing boards at the large capacitors on the top of each hashing board. So it seem like power is getting to everything the way it should.

Any help would greatly appreciated. I have a ton of S1’s and they have been happy hashing for a year now so I am surprised this S3+ seemingly not being able to recover.

I have also reset my router and my power supply is a CORSAIR CX series CX750M 750W which is running just the one S3+.

Never have had any issue like that with my S3/S3+

Anyone know if the S3/S3+ suffers from the same "no internet" error where it sets it self on fire thats happening with the S5/S7 ?

Much better idea. Do not use a molex to PCIe if that was what you were talking about.  They just cannot handle that amount of power.



It should be good on the Lightning 1300.  It is their "premium" PSU and top of their line.  So it should be good if you get a bad one with it then it's a dud contact company they will replace.  It should be very good.

But you are right not all are the same.  They make some utter crappy PSU's in lower levels as show here - http://www.rosewill.com/products/21/14/395/0000/series.htm

Its honestly hard to say. I,ve bought a specific PSU twice and got a different sub serial batch that was branded the same... but were made by a different OEM. So one second you have something reliable from say Seasonic, then you're stuck with the same model but fitted with cheap parts.

Happened to me with Corsairs.

Might not be the case with that PSU, but something to keep in mind.

For the Molex to PCI-e. I use 2 molex to one PCI-e 6pin. I think molex safely do 70watts so you could expect safely 140watts on that one pci-e plug, but only if the plugs are quality made, from a company such as EVGA.

Anyways in the best case scenarios, it would probably be best to only run that temporarily. If you do it anyways, touch every connectors regularly to make sure they are cool. If they start warming up, then...


Hence why a wattmeter would really help. The PSU are recommended to safely supply 80% of its rated power, but thats counted at the ASIC, not the wall, so you should be okay. Just barely.

I just went ahead and put it on. I clocked them down to 218 and just used the front 2 connectors. So I didn't have to use any molex adapters.

Seems OK. I'm probably going to have to under volt them soon any way.

That is actually a nice PSU.  The maker does have different quality of PSU's.   The lightning 1300 I have one of and never had an issue it's running right now. The only part that scares me are you saying you will go from Molex to PCIe? Im guessing you mean that as PCIE to molex is not needed for S3. If it is molex to PCIE don't do it. 

Now if you look at the lower price Rosewill has some utter crap psu's.  The one PSU I burnt cord on was a cheap Rosewill the cords were just not up to par.

Do you have a wattmeter / kill-a-watt? At stock speed it should be able to run it fine, but im not sure its a very quality PSU, so it could degrade overtime. I would look into getting more and better PSU's when possible.

Anyways. If thats the question, Your house wont burn down overnight.

you guys think that 3 s3s running at 225 will be safe to run on this PSU..

http://www.newegg.com/Product/Product.aspx?Item=N82E16817182063

i already have the power supply and its happily powering 2 s3s.. i was looking to add another one..
i will have to use PCI-E to MOLEX power adapters as the unit only comes with 6 PCI-E plugs (well it has 8 but 2 have 2 ends and im not sure the wires are beefy enough to handle the power) but that shouldnt be a problem.