Saturday, November 9, 2013

How to overclock your Block Erupter And possibly destroy it in the process. « WTFMoogle.com

How to overclock your Block Erupter And possibly destroy it in the process. « WTFMoogle.com:

How to overclock your Block Erupter And possibly destroy it in the process.

by  on Aug.04, 2013, under Electronics

Caution do this at your own risk. I blew mine out so you don’t have to!

That said. Here is how I got my block erupter overclocked to 56o MHz briefly before it died.
Lets start off first with a simple overview of the board
bepinouts2.0
Click for larger size
draw_0719-5
The heart and soul of your Block Erupter. Please note: putting a heat sink on top of this chip will have little if any effect as the heat side of the chip is soldered to the board. You are better off with changing the stock heat sink on the back of the board, or adding on to it as I will show you later. The 4 large 2 pin wide traces are all the main + voltage to the asic.
draw_0719-19
To maximize heat transfer I suggest you carefully remove the solder mask on the back of the board where the asic lies on the other side. I did this with a dull Exacto knife and very little pressure slowly scraping across the surface. If you go this route: USE A NON CONDUCTING THERMAL GREASE!!! or Heat transfer pad!!!! Artic silver may short things out and leave you with a mess to clean up, if not a dead board.
draw_0719_1
This is a Atmel attyiny2313a. This contains the firmware that bridges the usb – serial adapter chip to the asic. You would be stupid to put a heat sink on this as it will never get warm. If it does your doing something wrong. I have seen sales on ebay of heat sink kits for the block erupter that put a heat sink on this chip
facepalm
(Moogle Face Palms)
Next we have the Silicon Labs Uart – usb adapter.
Not only does it do the proper conversion, It coverts the 5v usb to 3.3 that powers the AVR chip
draw_0719-2
Again you would be stupid to put a heat sink on this.
Moving along we get to the fun parts :) This is where your soldering skills will take place.
draw_0719-4
Behold the asic voltage regulator and all its glory. You will learn to hate this chip :D as it will be the source of many of your hardware error issues.  Here is the datasheet, become familiar with it as you will need to know how to properly modify the core asic voltage.

Next we get to the clock generator
draw_0719-13
This operates at 3.3v and 12 MHz. Please note the 3.3v is very important here, as this may be the thing that ended up frying my asic.
The core asic clock speed is this crystal times 28. So 12 MHz * 28 = 336 MHz.
Here is a list of oscillators on Digikey you can experiment with if you dare.
http://goo.gl/nJeCVo< Long Digikey url so I shortened it
draw_0719-14
Lastly the resistors you will need to modify to change the core asic voltage. R1 is conveniently the same as R1 in the datasheet for the Regulator and the same applies to R2 :)
To get a higher voltage you can change R1. Do the math or be lazy like I did :D
Here is the setup I was using.
DSC00478
I used a 10K ohm pot to adjust the core voltage with a volt meter on the pin I soldered to the main asic core voltage line. You should be very careful if you use my method and always have a volt meter on the asic core voltage line. I mostly did changes with the Block Erupter connected to my bench power supply with the current limited to prevent any mishaps.
To the left and right of the Block Erupter are my 2 different clock gen circuits I used. I used a simple machine pin style female header so I could easily swap out the crystal. I found 2 different schematics on Google images.
DSC00483
Closer view of the board. Simple 74f04 clock source circuit.
DSC00484
The other board. This one gave me more trouble and wasn’t too stable.
DSC00485Closer view of my modifications. I put a capacitor on the main usb voltage lines to filter out any noise it might have picked up. Removed R1 and installed a 10k ohm trim pot. Voltage pin probe on the asic core voltage line to monitor with my volt meter. Trick is to set the trim pot to 1.2K ohms before you solder it in place then you can just make slight adjustments to home in a stable voltage. I then removed the clock gen and put it on its side. This way I could just solder a small wire to it to get the default clock back to test if I did any damage at any time. I also soldered in a pin that I could connect my oscillator circuits to.
Lastly I added a mini heat sinked 5v  fan to the back held on with thermal grease and a few small dabs of hot glue on the sides.DSC00488

Results & Things Learned:
  • I ended up destroying my Block Erupter in the end. I pushed it too far and too hard using a 20 MHz crystal, that is 560 MHz. but this lasted 5 minutes tops before the asic became non responsive.
  • If you screw up the clock gen to the asic or have noise on your signal then the Block Erupter will identify its self as 000000 or some random number sequence then what it should be.
  • A 14 MHz crystal is stable. 16mhz is pushing it
  • The Block Erupter Asic may not like a 5 clock source too much.
  • The lowest crystal speed that I was able to use without changing the voltage was 13.5 MHz. There wasn’t much of a speed increase but you will get some satisfaction in getting  a higher hash rate and more accepted work in your mining pool. All though with out changing the voltage resistors you may get more mining hardware errors ad the asic wants to have more voltage.

Lots of people have been wanting to give me bitcoins for my hard work :)
So if you feel generous send them here : 1A7FReBCeFXJR8VUhX4rCm8v5C5mExakHJ
Thank you and be safe.

No comments :