Topic: X6500 Custom FPGA Miner - page 7. (Read 219829 times)

Or you can do what I have done a few times, get the pins offset by one while trying to plug the connectors together with one hand, short one of the rails to ground, and trigger the safety shutoff on the PS.

I should have at least hit save first...

I was wondering the same...

Has anyone gotten the TLM bitstream to work with a x6500 board? I tried the instructions on the TLM website, but without success.

If anyone figured it out and is willing to share what they did, that would be fantastic.

That is a connection with the Pool error. It happen a lot, it just most of the mining software doesn't log like MPBM does.

Why would that be happening?

Excellent help...power adapter was not strong enough. Currently getting power though a molex connecter, and things seem to be going smoothly!

Thanks guys!

It looks like the cool chip isn't hashing, but it was for a little while (about a minute I guess). Is the DONE LED on? There is one for each FPGA. If you restart the board, does it hash again? What firmware are you using? What hashrate is the pool reporting? Can you share a log file from startup?

To figure this out fast, you might want to join our IRC channel at #fpgamining on Freenode (webchat here).

thanks for the replies, here are the problems...

one chip is running at 34 degrees celcius, the other is at 25. The cool chip has 5 accepted shares with 1 reject, or 16.67%, and the hot chip has 4057 accepted shares with 176 rejected or 4.16%.

So I am wondering

-why such a huge difference in shares?
-am I actually getting the 399.70 mh/sec?
-is there a problem with the board?

I have seen other peoples screen shots and shares on each chip are also way off. I just paid over 700$ for something that I could have paid 500$ for, so I just want to make sure everything is OK.

I can't see your screenshot, but it sounds like Carlton answered it pretty well. Can you upload the image somewhere else?

The temperatures are indeed important, and you do seem to have a fair difference between your FPGA chips. Possible culprits:

1. The way your cooling fans are arranged. Stock coolers? Any extra fans to keep the other PCB components cool? If you have one fan blowing from the side, I've found that the chip that's furthest away from this lateral fan becomes hotter than the chip that is closest to it. Orient any lateral fan so that it's blowing across both heatsinks, and not so that it blows the air from one heatsink onto the other. Also, if you don't even have fans to cool the other PCB components, I'd recommend that (but it doesn't sound like that's the solution to your problem)

2. Is there some other heat source (that's not coming from the x6500) that is closer to one heatsink than the other? Outlet air from a power supply or computer? Encroached by insulating material? (I read about one guy's kids nearly drowning his BFL singles in a beanbag)

3. Check with a torch whether you may have overdone the thermal compound on the heatsink of the chip in question. If you can see any compound that's been squeezed out from under the interface between the heatsink and the surface of the chip package, then this could be the source of the extra heat. You'll have to power the whole board down and apply some appropriate cleaning fluid (the stuff from computer component webstores is expensive for what it is, someone here will be able to tell you which household solvent does the job equally, alcohol probably)

What am i doing wrong here?
 Why is one accepting so many more shares than the other?

glad I haven't tried it yet!

Like I say, having several days of uninterrupted mining with little more than having to restart the MPBM web interface (a memory leak issue that's not responded to any other remedy) is really not so bad.
 

It's generally a bad idea to do this, but I'll admit I do it myself sometimes. The most likely outcome is that you'll burn out the fuse during the fast influx of current (the capacitors will be like a dead short for a moment). Better to turn the PSU off if it's at all possible, really.

I've done this multiple times without any problems.

Is it possible to hot swap these boards when using molex 4-pins to provide them with power?

I've been wanting to try this for a while, but there's very little need as I seem to have them running reliably for several days with little tending. Would still be useful to know that I can safely remove and re-attach individual boards without having to power down the whole PSU.

same here. i have 12 of these little buggers and at least 3 get much higher rejects running the overclocker than they do the 200mhz firmware, even if they are running at a slower clockrate (~190mhz) on the overclocker.

edit: sorry I misread your post, i actually get both lower rejects and invalids, not sure on temps. i run the 200mhz firmware on all of my units without any issues.

I have 2 that wont go above 195mhz with the overclocker firmware. I used the ztexmerge 200 on both, results in lower temps and no invalids, but a little bit more rejects.

I'm not that familiar with them, but that variance does seem strange.  One of mine is running at 198 MH/s at 39 degrees, and the other is 190 MH/s at 40 degrees.  The one with the lower clock has less stales (.81%) over the course of 5 days.  I let it adjust the clocks automatically.  Ambient temp is 15.5 degrees.

mining on slush, doing pretty well so far, and am happy, just one question...

temps are 34.34 and 25.69. the lower temp has no rejects, but the higher temp has 3.07 rejected shares. whats the deal with that and can i lower that number?

I guess I'm lucky that the post office only charge me $2 per box/parcel.