Topic: X6500 Custom FPGA Miner - page 56. (Read 219954 times)

Hang on!   I just realized something (halfway through selling my entire BTC stash). What if the price goes up again and I'm holding USD. I'm not going to have "just the right amount", because I'll be burned for selling low (it's pretty low as it is) and buying high.  

I'm feeling a bit tense right now.    Seems I always manage to screw up when I try to do anything with bitcoin on the exchanges.  

EDIT: Er... Strike that. Of course - I just realized the USD price is fixed so however many BTC I get when I buy back in will cover the price. 

Cablesaurus uses a script that automatically adjusts the BTC prices in terms of the latest Mt. Gox price (or weighted average, I'm not sure). This will give him a predictable amount of money for each product he sells.

I know that can be a little frustrating, but remember, we're paying for these boards in USD. What happens if we set the price in BTC today and the exchange rate goes down next week? We could easily end up giving the boards away at a loss! It's a shame, but Bitcoin really isn't the currency here, only the method of transfer.

If you're really worried about the fluctuations, you can exchange that amount of BTC for USD right now. When it comes time to pay, convert it back to BTC and you'll have just the right amount. Well, except for the exchange fees.

Sorry!

What I think I would have preferred, is a price fixed in BTC at the time of order (at least the maximum), rather than the price being set in USD terms. There's no way to tell how much the order will cost in BTC, between making the commitment and the time to pay the balance. It's a little difficult for anybody to deal with the fluctuating value of BTC, but if the agreed exchange rate is set at the time of commitment, then we'd have a good idea of the most BTC we would have to find to cover the balance. Just a little something to consider, if you preffer payment in bitcoin. 

30w? I was thinking more like 130-140w for 4 X6500's (eight FPGU's). That's 10 or 20w more than the total load. So there's a little room for deterioration. I don't expect more than 1.5% deterioration per year. So 15% over a whole decade. Perhaps I will go as high as 150w. The batteries I have to think about some more. Even if it needs a another $500 or $1000 spent on Absorbed Glass Matt batteries, the whole thing is still going to pay for itself several times over, well within the life of the panel. It would pay for itself twice even without batteries and only running in daylight hours. The extra cost of battery backup so it can run 24/7 would be marginal to the initial setup costs. Even allowing for two or three battery replacements over the decade.


So we need to aim at something more like 85% average over 10 hours. I think that's doable. Also a heliostat might very well be a wise investment. That or solar tracking. As long as the solar panel is reasonably oversize the need for 100% efficiency is averted.


I'm well aware of this. I've been looking into solar systems for quite a while now. I live in South Australia we have an excellent solar yield.  


That's a negligible overhead in terms of running costs. I'm still not convinced that an on site processor is necessary for a network connected unit.  Even still, I cant imagine needing anything as powerful as the processor in a typical palmtop or even an embedded system. If the software can be run over the network, all the better. The point is you don't need a huge power hungry monster that runs on mains power.


And that's why everybody's doing it if they can afford the upfront cost? C'mon now. Haven't you heard? People are putting $5-15,000 dollar units on the family home, that pay for themselves in under 5 years, cut their electricity bill down to a small fraction and sometimes even generate a small profit. Even selling excess power back to the grid is profitable. If buying power from the grid to mine bitcoins (with power hungry GPU's) is profitable, then cutting out the middle man by turning solar into bitcoins without inverters/grid-power at a tenth of the wattage, must easily be profitable.

Or keep going, or sell back the hardware at 25% discount, so it pays off actually in 8 months

Cool, so you will pay it off in 2.5 years.
They still have to pay to withdraw from Mt.Gox/Dwolla to practical currency. Once you factor in conversion costs, it really saves nothing. The only free, major service is AmazonPayments.

You should give a small discount for paying by bitcoins. Basically transfer the paypal savings to the buyer. Plus there's no chance of chargeback. So it's win-win. What do you think?

No problem, Karmicads. Consider it done!

How about this, we'll be much more willing to accept new orders if you're planning to pay with BTC!

Not to rain on your parade, but be realistic:

1-Don't count on a 30w consumer solar panel to still power your devices in 3 years, and your batteries will need to be replaced several times over that stretch. 15w for 14 hours is an expensive battery and will deteriorate quickly from frequent deep cycling.
2-Unless it's extremely well-placed and on a heliostat, your panel will operate at 100% for a grand total of 0% of the time and will only approach 100% for a small part of the day.
3-Regardless of where you live, it is not sunny 100% of the time. Find out your longitude and latitude and google a solar calculator, enter your info and see what kind of realistic production you can expect.
4-You still gotta run a computer.

Small scale solar is cost prohibitive.

I PM'd you, Re: getting an extra 3. Have full payment for total of 4 ready to go. 

Yeah, we're considering it. The difference in price will only be a few dollars, tops, so it might not really be worth it. PM one of us if you really want to get an order in. If the price has come down, we'll announce a new one before we start accepting payments.

Why dont you sell those 20 orders who backed out, im sure you could find 20 people fairly quickly.

After a lot of quotes, spreadsheets, and Digikey scouring, we were finally able to settle on a price point for this round of pre-orders. Unfortunately, 20 of those orders backed out this morning because the buyer thought $50 was the final price, so we're left with 49 X6500s and 2 X6000s. At that quantity, we're able to get the prices down to $610 (X6500) and $410 (X6000).

We learned some hard lessons about estimating prices based on back of the envelope calculations. We also learned how much Paypal charges in fees and that only about 1/3 of the pre-order deposits were paid with BTC!

We still need to get the prototype built, so we won't start accepting full payments until after that's up and running. Once we get that in our hands and everything looks good, we'll set a due date for the final payment and an estimated lead time. Then the orders go out!

Somebody who has an aversion to big power bills.
Somebody who still wants to do cost effective mining after GPU's have run their course.
Somebody who doesn't want to have to use a motherboard requiring video slots, configuration etc.. etc.. Just USB.
Somebody who is smart. 

This is awesome!

lol

Somebody whose mom no longer pays for his electricity.

Well, not "forever" ...the sun will eventually run out of power. 

Trade secret! It is not the original LMZ regulators that were on the first gen, however, you are right about that.

A nice feature would be a IDC connector with some I/Os from FPGA (somwhere on edge of the board) Let's say 16 I/Os plus 2 VCCIO and 2 GND pins. It's shouldn't be a problem to route since you use 4 layer board. It may help to connect some push buttons and some LED's. Just an idea... If someone would like to modify a code and add some functionality like RESET switch. The more I/Os the better but of course if you find sufficient space on board to do that...

Edit: or to connect LCD display, it could show stats, so quick "throw of an eye" and you know that's everything is ok (or not)...