The hype about ASICs last year pretty much killed the market for FPGAs, everybody went and preordered ASICs so the custom FPGA boards dropped off the market, also helped by the fact that Avalon (Lancelot by ngzhang)  and BFL were the major suppliers of FPGA at the time and dropped their FPGA products. It just did not make sense to buy them at $10 to the BTC when ASIC was promised so very soon. Fast forward to $100 plus BTC and just a trickle of ASIC and everybody newbies want in again.

If you can't get a Lancelot or Enterpoint. then perhaps one of the (expensive) Kintex dev boards they are talking about on https://bitcointalk.org/index.php?topic=9047.0 but I can't vouch for the ROI.

The custom hardware thread is a good place to look, but its mostly ASIC these days https://bitcointalk.org/index.php?board=76.0

The fpgaminer thread is the place to go if you want to play with these development boards https://bitcointalk.org/index.php?topic=9047.0

The cheaper ones are little more than toys (fun toys though) ... 5MHash/s on the LX9, 10MHash/sec on the DE0-Nano (though it can be pushed to 35MHash/sec if you hack the power supply), but neither will give a positive ROI.

PaperClip

True, nothing can grow forever in the physical world,

You may enjoy the threads in this subforum https://bitcointalk.org/index.php?board=81.0
Optimator is predicting 250M by December https://bitcointalk.org/index.php?topic=216358.0  (though he has his critics)

The direct chip orders may well be the decider. If the third-party board manufacturers get their act together, then we may see a huge spike in chip orders, which the foundries will be glad to supply with their 10 week leadtime (provided Avalon and BFL don't decide to choke off the supply ... its their Intellectual Property after all).

And there are others planning ASIC designs too (some will be scammers, but not all). There is money to be made here with the examples of Avalon and ASCIMINER (and perhaps BFL too, who are sitting on a pile of preorders they are failing to supply).

Good luck with your investments, at least you are putting your money where your convictions lie.

This way doth madness lie ... https://en.wikipedia.org/wiki/Solipsism

Yes, good idea, but you started it

Anyway I think I've said all I want to say about this, I just got a bit of a bee in my bonnet about ASICs the last couple of days, no hard feelings. Have a good one. 

Agreed, caveat emptor is always a good rule. Have a nice day now 

Nice try, but you'll need to find a buyer who's a bit weak on the math too, otherwise you'll only get its value for its remaining mining life, which won't make up the difference between what you paid and the sum of your mining earnings.

I've gone over my math with the help of http://coinish.com/calc and I think I've been wildly over-optimistic. 209BTC would be the breakeven price, assuming the rate of change of difficulty remains unchanged. 26% per fortnight would breakeven at 165BTC
https://bitcointalk.org/index.php?topic=225285.msg2369876#msg2369876

But kudos on getting such a good sale price 

Nice try, but you'll need to find a buyer who's a bit weak on the math too, otherwise you'll only get its value for its remaining mining life, which won't make up the difference between what you paid and the sum of your mining earnings.

I've gone over my math with the help of http://coinish.com/calc and I think I've been wildly over-optimistic. 209BTC would be the breakeven price, assuming the rate of change of difficulty remains unchanged. 26% per fortnight would breakeven at 165BTC
https://bitcointalk.org/index.php?topic=225285.msg2369876#msg2369876

But kudos on getting such a good sale price 

Running the numbers for a batch 1 Avalon through http://coinish.com/calc (NB use expert mode), gives a maximum price of 209BTC before your expected breakeven goes to never.

 Of course this assumes the rate of change of difficulty remains constant, which it won't.

If I plug in my 26% per fortnight (14'th root gives 1.6645% per day) that I used on https://bitcointalk.org/index.php?topic=221318.0;all  its drops to 165BTC, so I was being optimistic (my prediction was for a 250BTC to break even at 26% per fortnight, it clearly won't, it actually needs 1.096% per day or 16.5% per fortnight).

Nice toy (http://coinish.com/calc I mean, but an Avalon would be nice too, but not at that price!)

Cool calculator. I was waiting for someone to call me out on my math, but this is just perfect. Use the expert mode to see the return taking into account difficulty increase. One possible request is to enter the increase per fortnight rather than per day (saves having to use a calculator to calculate the 14'th root of the factor).

Irrelevant, you're buying it with BTC, so if it does not return your investment in BTC then the exchange rate matters not one whit. If you spend USD to buy it, then you're effectively investing its equivalent value in BTC at the time of purchase. And again if it never returns your investment, you lose.

On your previous post, delivery time is of the essence in calculating ROI. Do your math carefully.

280BTC is way too much for an Avalon, you'll never make a ROI on that.

Take a look at this thread, lots of good info on expected ROI.  https://bitcointalk.org/index.php?topic=221318.0;all

I posted this in the newbie thread, but I think it could do with repeating here.


^^^^ ... this. But let's try working backward. 80 days ROI at current difficulty, so 40 days to recover half the cost. Use that to calculate the maximum acceptable rate of difficulty increase (so the other half of the cost is recovered by mining for eternity). 40 days is almost 6 weeks or 3 difficulty increases. Cube root of 2 is 1.26, so as long as the difficulty increase is less than 26% per fortnight, you'll recover your costs. But factor in delivery time, and the exact point in the block difficulty cycle to get an accurate answer, and perhaps you won't make any profit at all. Do your math folks.

^^^^ ... this. But let's try working backward. 80 days ROI at current difficulty, so 40 days to recover half the cost. Use that to calculate the maximum acceptable rate of difficulty increase (so the other half of the cost is recovered by mining for eternity). 40 days is almost 6 weeks or 3 difficulty increases. Cube root of 2 is 1.26, so as long as the difficulty increase is less than 26% per fortnight, you'll recover your costs. But factor in delivery time, and the exact point in the block difficulty cycle to get an accurate answer, and perhaps you won't make any profit at all. Do your math folks.

FPGA scrypt is tricky as you need to attach high bandwidth RAM, and what you end up building is not that much different from a GPU card (albeit with a much more flexible processor core). ASIC would take this further by building the RAM on-die, which will give much faster access times. Its doable, but will anyone make the up-front investment to make it happen?

Edit: cross posted with your's Spectrome, but it looks like we're in agreement. Welcome!

But the ROI on ASIC's ordered today is negative(*)! So if you try to keep up you will actually fall even further behind than where you started from!

(*) With the possible exception of the DIY chip purchase schemes.

Which is why I chose 20% to illustrate my example, it may seem high now but once the next batch of Avalons and all those DIY chip miners come on stream it can ony go upwards. Perhaps the chip purchases will be profitable. Avalon is roughly 0.08BTC for 300MHash/s plus assembly costs, but you're still looking at up to 3 months delivery on new orders - 10 weeks for the chips, then all the business of getting them built onto mining boards. I'd have taken a punt if I had the BTC in hand, but the minimum order is typically 10 chips, and I'm a bit light on coin at the moment.

Lots of these freebies popping up these days, but whatever ... my sig wallet's public anyway so why not join the fun. If I win we can all play chase the BTC together   [edit] OOPS wrong smiley, let's have this one instead 

Of course you could just mine your coins. Pools don't insist on any contact details, and provided they don't log your IP address, your coin is completely untraceable. (Tinfoil hat alert: Of course the NSA could be sniffing all your packets for bitcoin transactions, but they're not going to admit to that so you'll probably get away with it providing you don't do anything really, really bad).

I have a little theory about why mining is so popular even though the ROI on eg new ASIC kit is negative ... its just a coin laundromat: USD -> MtGox -> BTC -> buy ASICs/FPGAs/GPUs -> mine -> shiny clean untraceable BTC

Just my 0.02BTC

Yes, but do the math on the ROI and be sure to allow for the delivery time until you can start mining. ASIC's as currently priced will never pay for themselves.

A little worked example. The Blockfury Erupter is priced at 2BTC and mines at 300MHash/sec. Right now that earns you 0.01BTC/day and would pay for itself in 200 days. But difficulty is spiralling upwards. So let's take the figure of 20% every two weeks (a little high perhaps, but it makes the math easier and you can always plug in your own guess as to the future difficulty). After just 8 weeks the difficuty will have increased by 1.2 * 1.2 * 1.2 * 1.2 = 2.1 times, so your return from mining is halved. And here's the sting, if you mined for the rest of eternity you will only make the same return as you got in those first eight weeks (checkout https://en.wikipedia.org/wiki/Zeno's_paradoxes for nice read). Total reward less than 1BTC, you lost money.

Now the Erupter is the ASIC that perhaps you could get your hands on soonest. With BFL or Avalon you'll be waiting months, and all the while the difficulty spiral is eating into your earnings.