As I stated before a tutorial on how to upgrade your erupters to 447 Mh/s be advised this is my first tutorial. and I want to thank all the hard work that went into me being able to complete this by everybody here. please send all feedback as a pm as if you find something totally fucked up about my tut I can correct it and reupload a better version. I hope this helps noobs like me do this mod and best of luck mining.

https://docs.google.com/file/d/0BzJZeR-W87vKa2xXc0VkZVhnYnc/edit?usp=sharing

OK here is the chart I have been working on, green areas are possible with stock regulator , yellow I think I have found a replacement but going to have to change inductor also and red is external only. Now this is just for reference this is what worked for me, I did not go to far on the cooling aspect, no soldermask removal or anything other than the stock thermal pad. I am amazed that this chip can even handle this much power.

Well a 17 does not exist most likely I am using a signal generator, however I am going to have a nice chart to post, so far I am at 20 Mhz however were getting up there on voltage and heat density is making me worry. I can tell you so far were looking at 1.50V @ 3.992A for 20 MHz. I started at 12 MHz and made a chart with the common available oscillators and worked with each one until no HW errors in 5~10 min, long term hard to say. Should be done testing by tomorrow, maybe the fan wont quit in the middle of the night and a fire break out  Although so far at this test there are 4% HW errors. Is the maximum that everyone believes 1.8v? I think it is going to take a peltier to go to much farther then that adds another level of something to mess up.

As for the 1.4V at 18 MHz that is right on with what I have at that point.

Well I am not sure, was running it from a lab supply while I was experimenting, I remember it was right around 1.4v although it was over 3A so it wont work the the current regulator, however now that it works I might not need so much voltage before when the software was not working I spent a lot of time "trying everything" finally just left it. Monday I plan to do a better job working on voltage/current/speed parameters. Been running 4 hrs or so.
 
A:567 R:0+0(none) HW:14/.61%

Hi everyone been following this thread quite a lot, so I decided to get one of these devices and of course modify it. I decided to run the clock from a Tektronix function generator, all works as expected at 12 MHz. Although I have not modified the power supply yet I notice at any rate about 12 MHz the green led starts staying on longer than the short blip. If I go to 13 MHz it stays on a little longer and at 14 longer still. While it is hashing it seem to be faster but the delay after it gets done acts like it limits it back to the stock freq. I don't know this for sure yet. I also attempted to remove the regulator IC and feed power directly form the lab supply, while the voltage was correct I could not get it to work at all, green LED went off but cgminer said no devices found, put the IC back in and went back to work, I just pulled the chip and fed where the inductor was. It did draw 1.4A at 1.05V not sure why it didn't work unless there is some kind of power sequence thing. I hope to play with it more tomorrow so we'll see if I make any head way.

why not use silicon labs programmable oscillator?, might need an adapter, but you can programmatically adjust the frequency ?

http://www.silabs.com/products/clocksoscillators/xo/Pages/default.aspx

Did you change both R1 and R2 resistors?

Also, any detailed info you may give about the best way to desolder and solder the new crystals and resistors in would be hugely appreciated. What i really need to know is how much heat can these little crystals take?

No thats just for the 20 and up. he should be good, not sure what wrong. did you use flux?

Here's the latest figures, now running a 20Mhz crystal on ICA 7 which is stable at 406.2 MHash with ASIC powered externally with 1.7V @ 3.96A. ICA 6 is the stock reference running at 333MHash.

bfgminer version 3.1.3 - Started: [2013-08-15 20:58:51] - [  0 days 03:00:24]
-------------------------------------------------------------------------------
5s: 2.66 avg: 3.65 u: 4.37 Gh/s | A:2752 R:30+0(1.1%) HW:92/.82%
ST: 4  GF: 0  NB: 21  AS: 1  RF: 0  E: 9.66  U:15.3/m  BS:36.6k
Block: ...5e0f8f39 #252279  Diff:50.8M (363.7Th/s)  Started: [23:45:10]
-------------------------------------------------------------------------------
ICA 0:       | 960.4/364.6/422.2Mh/s | A:266 R:4+0(1.5%) HW:10/.88%
ICA 1:       | 543.2/363.7/393.6Mh/s | A:248 R:1+0(.40%) HW:13/1.2%
ICA 2:       | 245.2/364.5/477.8Mh/s | A:301 R:1+0(.33%) HW:10/.85%
ICA 3:       | 423.3/364.2/455.5Mh/s | A:287 R:3+0(1.0%) HW:10/.89%
ICA 4:       | 581.1/364.4/447.6Mh/s | A:282 R:6+0(2.1%) HW:10/.91%
ICA 5:       | 348.0/361.7/441.3Mh/s | A:278 R:1+0(.36%) HW: 9/.84%
ICA 6:       | 394.9/333.8/296.8Mh/s | A:187 R:3+0(1.6%) HW: 7/.85%
ICA 7:       | 294.3/406.2/563.5Mh/s | A:355 R:4+0(1.1%) HW:11/.79%
ICA 8:       | 532.0/363.8/460.3Mh/s | A:290 R:5+0(1.7%) HW: 5/.44%
ICA 9:       | 239.9/365.9/411.1Mh/s | A:259 R:2+0(.77%) HW: 7/.61%
-------------------------------------------------------------------------------

I've bought a few crystals and heatsinks too many so if anyone wants me to upgrade their units to either 16 or 20 mhz speeds I'd be happy to do it.  

2 day maximum working turnaround and high quality soldering done with an SMD rework station with all components, heatsinks and external power supplies (if necessary) included.  Please PM me if interested.

In regards to powering these things, i was planning to solder a molex plug on the back end of the usb miner so i could simply plug them in and out of a standard pc power supply. Then, for the data, use usb extensions with the power line cut. Hopefully, using that method would allow quite a few overclocked miners at the same time.

Of course, i'm thinking out loud here so if anyone see any trouble with that, please let us know.

Fascinating thread! It's pretty impressive that you're getting almost 200MHs more and the fact that it seems fairly stable is great. I just presumed that they'd be clocked as highly as possible!

Can somebody tell me what types of resistors these are?  What keywords and descriptions to look for?

I'm using 1.5K (in an 0805 package) as R1, and the 16Mhz crystals are from a chinese eBay seller.  I've got some 16Mhz, 20Mhz & 24Mhz crystals coming from Digikey and will replace the current crystals shortly and post the results.  I'm looking to get the same speed Bluestreak66 has reported with a 16Mhz crystal.

Hopefully, bluestreak66 was saying the chips have a theoretical max of 1GH/s and that around 750MH/s would be the limit of these boards.  So hopefully I can get these higher frequencies going with water cooling and decent power supplies. More for fun than profit of course. I'd like to think I can get the BE's to pay for themselves in a year and they'll need all the help they can get.

I've got my 15A external power supply now and it's working with a 16Mhz crystal ok, I'm getting avg. 360MHash/s with this installed, running at 1.35V - is this a normal speed for a 16Mhz crystal?  Little or no hardware errors with this config.

If I put a 20Mhz or 24Mhz crystal in, the light doesn't go off and the BE isn't recognised by the BFGminer software.  Does that mean it's the wrong type of crystal or any suggestions as to how I could make the higher speeds work, even for a short time?

Got 3 new block erupters to play with



Can you guess what one is overclocked

Using a 16.386mhz crystal and replaced r1 with a 1.2kohm resistor. Put a large heatsink on and everything looks stable


dare I try 18.43mhz?

Far better to go for an internally regulated part, trying to get a very accurate external voltage is not easy and usually requires additional voltage regulation chips, plus there is no indication of the conditions a user is running under. Just because a test setup works with a particular clock setup is not really a solid indication that other setups will work the same, or indeed be non destructive.

As long as everyone realizes that you would have no feedback control, that is to say if the component values you choose trashes a certain number of ASICS, there in nothing you can do to control it, other than unplug the power from the device.(unless you have some sort of clock gating circuit)

I'm not particularly pro or con what you are trying to do (hell I've toasted more post 500$ chips than most), but it is important that you take into account variation of the devices.


As regards the inductor, if you say it is cool... then it is cool, but it is a little more complicated than just setting a reference.

Possibly it might be better to use an SPI controlled VCO

There are some really 'cheap' solutions available. That way you can have a bit of software that increases the clock rate until the thing starts producing too many errors, then you just gradually 'slide' the VCO back down a few tens of  KHZ

Take a look at silicon labs, they have a full range of products that can be picked up on ebay or group buys fairly cheaply.

What you will see is that the number of errors will increase as the frequency increases, so with a VCO you can 'trade off', also as the temp is a function of the speed you can slide the VCO Downwards if you start getting too hot, whilst still mining.

Only 'fly' i can see is If there is some sort of clock syncronicity between the USB circuit (UART/USB bridge) and the ASIC

As regards increasing the voltage/current... usually this is a resistor/inductor change......, changing the resistor will only get you so far, and if you get it wrong.. the flyback inductance from the mis-matched resistor inductor, will toast your PSU chip....

What is this "Windows 8" you talk of? Haven't you heard of Linux

I would try it, but I need a little bit more detailed information.. like an "ifixit picture guide" what to desolder where and to solder on where, what tip to use and so on - i am a novice with the soldering iron