Close - I paid in USD and am in the US as well. I requested a refund but did not receive one (yet).

I agree that nobody signed that right away in the HashFast Batch 1 case. They introduced a TOS after the fact etc., so not even a question. However, Batch 2/3/4 may be less clear because there was actually a TOS with specific delivery dates that people agreed to.

I'm just saying that the law doesn't blanketly overrule a specifically signed contract to which both parties agree (again - NOT the case here).

Not quite - you can sign that right away, and you often do when you build a house on spec, buy a high-end car, get something custom manufactured etc.

It ultimately comes down to a contract dispute in such a case.

Take my steps above and change 8332 everywhere to 80.

I don't know if that will work for sure however (never tried) - a corporate firewall generally closely monitors port 80 traffic, if not just doing an HTTP proxy across it. So not sure what your firewall will think about the GetWork protocol.

But worth a try.

If you want prove of receipt of your Refund request, do the following:

If you haven't done so yet, send the refund request to: 'refunds@hashfast.com'
-> You won't get any response.

Using the same email account, send any general question (or the same refund request) to: 'sales@hashfast.com'
-> They will send you an automatic response back from there with a link to the 'case' on fogbugz.com.


Click on that link, and scroll to the bottom. You should see a list of all your 'cases', including your refund request made to 'refunds@hashfast.com'. This proves they received it.  Print out as much as you can and keep.

Has anybody received any indication of when the upgrade kits for the Batch 1 BabyJet's are expected to ship?

I've tried asking HashFast 6 times so far with no response. (I believe they're trying to force me to request a refund by refusing to answer questions.)

I've paid with a wire transfer instead of BTC. As such I have no hope of getting back BTC.

Is there any reason for me NOT to request a refund at this point?

If they ship anywhere within the next 2 months, and ship MPP within 4 months, I can still theoretically ROI, but the chances of that are slim. (Especially if they get sued by the BTC refund crowed and go out of business).

This service has now been shut down.

I've not gotten around to decommissioning yet. I'll decommission on the 27th.

Here are the steps to create your own Azure gateway.

I'll show by example of how I would create westcoastbtc.cloudapp.net, port 8332 which connects to BTCGuild. My values will be within []'s. Substitute with your values when more appropriate.

1) Go to www.azure.com and create an account. You can get a free trial account
2) Enter the azure portal (on www.azure.com)
3) On the left of the portal, select 'Virtual Machines'
4) Click "+" at the bottom to create a new Virtual Machine
5a) Click through: Compute \ Virtual Machine \ Quick Create
5b) DNS Name = [westcoastbtc]
5c) Image = Windows Server 2012  (you can host on Linux as well if you're familiar with it. I'm not listing the instructions for that below.)
5d) Size = Extra Small (enough for 10 blades. You can upgrade later)
5e) User name = [shrek]
5f) Password = [donkey]
5g) Region = [West US]
6) Click 'Create a Virtual' machine and have a [beer] until it's done.

7) Click on your new server, and go to Endpoints
Click 'Add' to create a new endpoint
9) Add a Stand-Alone endpoint
10a) Name = [BTCGuild]
10b) Protocol = 'Tcp'
10c) Public Port = [8332]
10d) Private Port = same as 10c. [8332]
11) Click the checkmark to add it

12) Go back to the Dashboard
13) Click 'Connect'. There will be 2 popups, one from Azure, and one from your browser. Click 'Ok' on the azure one, and open your browser one (this is a Remote Desktop Client shortcut).
14) Enter 5e/5f from above for username & password

15) On the Server itself, go into 'Windows Firewall with Advanced Security'
16) Click 'Inbound Rules'
17) New Rule
18a) Port
18b) TCP
18c) Specific local ports = [8332]
19) Click Next
19a) Allow the connection.
20) Click Next
20a) Applies to 'Domain', 'Private', 'Public'
20b) Click Next
21) Name = 'BTCGuild Port'

It should now be set up. Create a batch file with a shortcut, and run the following:

start "BTCGuild Proxy" mining_proxy.exe -o eu-stratum.btcguild.com -p 3333 -gp 8332

(The 'start "BTCGuild Proxy"' part is just because I have dozens of these and I want to customize the title in the taskbar. You can leave it out).

I can confirm I see this behavior with BTCGuild. The difficulty on those mining proxies are all set to 1024 (by BTCGuild - I have no control), so your miners can sometimes have minutes and minutes without finding any blocks. This should in theory not cause problems over the long term, but because of the way PPLNS work, I can imagine it being an issue with statistic reporting.

PPS servers like Eligious work fine with that level of difficulty though.

Interesting. 21m/s as well.

Let me switch some of my blades to BTCGuild and see if there is another issue.

It's at 80% usage currently, so won't be a problem on the server side.

By FAR the biggest performance bottleneck is the speed of your internet between your cube and the server. GetWork is a horribly inefficient protocol (which is why it's not used by pools anymore), so if there is anything slow between you and the server, your performance will suffer. It's a lot better converting from GetWork to Stratum as close as possible to your Blades as possible.

I run my own blades on the proxies, but I have a 40mps connection to the internet. So that is almost as fast as a LAN connection, and I've never seen a slower speed on the cloud proxy vs. my lan proxy. But on a slower connection other people definitely do experience problems.

They're definitely up and running - tons of people connected to it.

Which server? I've been connected via westcoastbtc to Eligius, and has not have any service interruption in the last 7 days.

I've shut down NorthEuropeBTC though (just now), since nobody connected to that over the last 7 days, and in preparation of me decommissioning the farm.

The MPP is based on number of BabyJets:
"In fact, we will give you up to 400% more hashing capacity than the Baby Jet you purchased."

So if HashFast wants to market themselves as having "HALF A TERRAHASH/s (500GH/s) on a single chip.", then "Baby Jet I purchased" becomes a 500ghs machine, and MPP is 2 TH extra. If HashFast wants to insist for the purpose of the MPP that "Baby Jet" is a 400ghs machine, then don't call it a "HALF A TERRAHASH" machine in another place. You can't have it both ways.


But I grant you they have some statements in their that MIGHT have them be able to get away with that. e.g.
"we will increase your mining capacity to up to 2 Terahashes!"

But that would mean that they can run 2 Terahashes on 3 chips, ala 666ghs per chip. Not going to happen.


Besides, it's cheap for them to just ship a 4th chip. You still have to purchase the motherboard to run it, and they'll probably pocket more on those than the manufacturing cost of the chip. They could technically sell their motherboards for $5'000 each and still be fulfilling the MPP - there are no terms anywhere that describes the motherboard cost.

Not quite.

Even if they get to 1 TH per chip (unlikely), and you have the device in hand on December 20th (unlikely), and with modest 10% difficulty adjustments until the end of January (unlikely), you won't ROI by end of January, given 4 times the initial delivery power.

So they'll still need to ship out a full x4 MPP.

So I was doing some calculations, and I don't have a way to make this service run much longer - it's now costing $300 per month and climbing. I will have to pull the plug.

As it stands now, I will have to take approximately 3% of blade computing power to break even. However, because of difficulty changes, in 1 month it will have to be 6% then a month after that 10% etc.  I don't think that's really worth it for anybody.


The thing that's a killer is the data transfer due to the GetWork protocol, which is several TB per month now.

I will post the setup instructions if you want to set up your own private Azure server for this. But as it is now, I'm planning to shut this service down on December 17th.

Which server/port are you connecting to?

Check your internet speed at speedtest.net as well.

Not as far as I know. I've been meaning to modify mining_proxy to give me those kind of stats, but haven't gotten to it yet.

Well, if I can figure out a way to break even with my costs on this proxy, I can keep them running indefinitely. My costs at the moment is around $250 per month so far. I can carry it for a little while longer, but at some point it will be a bit much, since it keeps increasing per month.

However, would everybody be interested to keep connecting if I use a small portion of your blade power (let's say 1%, but maybe 0.5% will be sufficient) to mine under my own account? It will be like a pool fee. Just trying to recover costs - not looking to make money here.

I could also ask for donations, but I can't scale with that. With a %, if I get 100 times the number of clients connecting, I can easily just configure 100 times the amount of processing power to match (got to love Azure).

Any other ideas?