Topic: HashFast BabyJet users thread - page 42. (Read 68982 times)

yes its the average. So anyone do any Overclocking and get good results. Most people seem to be doing 600ish and getting a stable 420-340ghash.

Has anyone tried to extreme overclock it like 900 and see what they can get?

Yes i see the behavior, its seems very "bursty" . Its not luck because mine steadily increases pver 400 then all the way to 600 sometimes. Hopefully the new firmware will let us change this stuff or stabilize somewhat.

Yip. Just saw mine at 1.3TB in fact. I figured it's just lucky streaks, because I also see 200gh/s every now and again. It's the average that counts.

The image is only 2GB. Did HashFast really send you an SD card that's smaller than 2GB? They sent me an 8GB card.

Or did you perhaps just try and copy the image on the first (90mb) partition of the card using a file copy, instead of following Phil's instructions to use Win32Diskimager?

Even if they did send you a < 2GB card (which I find unlikely since they're hard to find these days), you can wait for Hashfast to send you another one by next week, thereby losing $500 in mining in the process, or you can drive over to Wallgreens and pick up an 8gb card for $5. Your choice.

Does anyone else get spurts on the unit where it jumps to like 800? By jump to 800 I mean its not the doubling effect it starts at 400 and ill see it go 500 then 600 and 700 and then touch 800

Yeah, sure Phil. the SD card that came with my BabyJet isn't big enough to hold the image.

Please send me an SD card that is big enough to hold the new HashFast image, or this will be considered breach of contract.

Thanks!

This is in the spirit of being complementary to HF-engineer, additional information and not a challenge....

I feel peoples' pain with underperforming BJs, and I again want to point out that 3.11 on windows has been a good setup for me. Oh, I hate windows too.  My goal is to move this all over to an OS X machine, actually, but the guys that do OS X ports have to resolve some issues, so I'm waiting a few weeks on that.

But on a little mini-desktop running Win 8 and cgminer 3.11, with some modest overclocking, my previously underperforming BJ (390 g/hash out of the box) is running at 430-435, and another machine is running very slightly higher, maybe 435-440.  

With 3.11 I have not experienced stability issues (knock on wood), save the previously mentioned need to hotplug if you restart cgminer.  (3.12 was subject to crashing.)

I've also had BJs horizontally and vertically - made no difference.  

Where I did waste time was trying to devise an extra-cool place for them.  This doesn't seem necessary.  And really cold temps seem to even slow them down. The climate is mild here now, with outside temps ranging from about 5c to 24c.  I've learned to just throw open the windows, ensure adequate (but not crazy) ventilation, and let the things hash.  The procs will go up to about 82c max, but generally hover in the 78-80c range during the day.  Right now it's 11c, and they are in the low 70s, in a room that's probably 13-14c due to the exhaust heat.

My 0.02 cents.  I'll shut up about it now, though.

You will need to install a fresh copy of our custom MinePeon image for your Raspberry Pi:
1. If you are already running MinePeon, first make a backup from the Settings tab (near the bottom).

2. Download this SD card image: Windows: http://setup.hashfast.com/hashfast-minepeon-1.0.zip  Linux: http://setup.hashfast.com/hashfast-minepeon-1.0.img.bz2

3. If running windows, we recommend the Win32DiskImager: http://sourceforge.net/projects/win32diskimager/, If running Linux, you can simply type:
(Where /dev/xxx is the device name of your SD card.  If unsure, type "dmesg" after inserting your card.)

4. Boot your  RaspBerry Pi with the new SD image installed.  Note that it may take up to 4 minutes or so to collect all the current updates from us before it will begin mining.  Listen for the fans to start in your BabyJet.  You can browse here to see when your system comes online and connect to it: http://setup.hashfast.com/.  If for some reason it doesn't come up after five minutes, power-cycle it.  (We've seen some brands of SD card cause it to hang on reboot)

5. Log into the web interface and restore your backed up MinePeon settings (if any).  Note that the RPI will generate a new wallet when you first boot from the new SD image, and will begin mining using that wallet.  It's important to switch your mining settings using the "Pools" tab or restore your backup so that your mining proceeds end up in the right wallet!

If you don't want your RPI checking in for updates, you can disable this in the Settings tab, but I advise you to at least do it manually once in while.  We hope to have some new firmware and software soon!  We also advise you to set a password on MinePeon, especially if exposed to the outside internet.  (not behind a router/firewall)

-Phil

Where can I download the 0.2.4.3hf8" for MinePeon version?


Where and How?

So far, so good, Phil. Thanks much for your time and effort. And please, ignore DoogieHouser. He's obviously a moron and the rest of us appreciate your presence.

I'm very interested in what you have to say and anything that can be done to improve stability. I have Sierras myself and I'm getting the watchdog errors and HF Get Header USB Read err:(-1) and LIBUSB_ERR_IO which results in the unit restarting itself and then hotplugging back over and over again.

This might seem obvious to everybody... but just in case anyone is as silly as me...

I found the machines were able to stay cool easier if:
1) Don't stack them on their sides
2) Take the front face plate off as well

Before this they seemed to be running on the low 80Cs consistently.  Now they are in the high 70Cs.
I also have a fan located underneath them blowing air from bottom to top.

I'm sorry you are not happy.  I do not make policy, nor can I change it.  I did not collect any orders.  I am just an employee, and I wasn't even working for HF until December!

You can "shoot the messenger" if you like, but I'm only trying to help.  If you, or anyone else here, doesn't like my advice, feel free to ignore it.



Feel free to ignore me.  I've only ever suggested people run an image downloaded from http://setup.hashfast.com/, so that should constitute "official" for most people.

You are complaining about reliability, which we are improving, and yet you want to ignore it.  Sounds like you just want to suffer.

I'm working all weekend on my own time, and I'm not paid for support at all, so I'm only here because I want to help.  I could be doing any number of more fun things right now, so save your complaints, because I've got just as many!

-Phil

Okay so here's the latest in version hf8:

1. Fixed bug that was causing a reboot every hour.
2. Created new update mechanism that doesn't need to reboot RPI.
3. Restored cgminer 3.9.0h2 - This puts the hashrate calcs back and seems to be more reliable.
4. Added "Check Now" button to the update section on the settings page.  (This causes an immediate upgrade if one is available)
5. Added "boot_delay=1" to try to help some RPI's hanging at the "rainbow" when rebooting.  (Thanks rMuD!)

Everyone running the HashFast image will get this within the next hour provided your RPI's are online and Automatic Updates are enabled.  You'll be able to tell from the status page.  (Look for "0.2.4.3hf8" for MinePeon version)

Let me know if you have any problems or suggestions.

-Phil

Ok, thanks for the advice.  I added it to the next update (hf8) which is live now.

-Phil

do us all a favour and go spam somewhere else plz

I'm not doing a damn thing until I receive "official" notice from HashFast.

Sorry Phl, you could be any jackass on the Internet.

I've got 8 days left on my warranty, and I'm not going to risk it on somebody I don't know on BitcoinTalk. Ha! What a great way to void the warranties of unsuspecting HashFast customers.

Not this cowpoke.

Sure, and I suppose this voids the warranty, right? Go back to Butterfly Labs!

More downtime. Yay! I hope this is added into the Miner Protection Program calculations. Remember, Batch 1 was locked-in at October 30th, 2012, and we paid DEARLY for these units 47 BTC is now worth over $40,000.

Okay fine! Please add this additional downtime to the calculation for the miner protection program. So far, we're at something like 50% uptime. So., when will I be receiving my next three units?

Quote from: HF-Engineer on February 01, 2014, 05:31:41 PM

Quote from: rMuD on February 01, 2014, 12:16:24 PM

Pushed Where?   I'm still rebooting every 2 minutes

Please follow the re-image instructions I've now posted several times.  Get a fresh start, it seems to fix a lot of issues people are having.  You shouldn't be rebooting.

I am running all my test systems off or RPI's.  Initially they were very unreliable, which is what got us on the path of customizing MinePeon.  Now it's pretty solid.  They have their niggles, and of course all things firmware and software are a work in progress, but they should have a 99.9% uptime now.

If you HAVE an extra PC, and can install Ubuntu on it (Windows is not reliable for mining), this is a bit more more stable, but there is a learning curve and you'll have to get cozy with a command line and shell scripts if you want a reliable system.

We hope to release a pre-configured image at some point similar to what we have on the RPI so it's plug-n-play.  Until then, unless you are comfortable with Linux, I recommend you stick with the RPI.

-Phil