The Gridseed ASIC Thread

nightwalker

2[H]4U
Joined
Feb 25, 2007
Messages
2,555
I thought I would start a proper thread on these as we don't have one yet.

I've had quite a few of them hashing away for a while now so I'll share my thoughts.

Buying the actual units:
I've been a lurker on the BTCtalk forums for quite a while now, that place is crawling with scammer's so please be careful who you buy from if you're hunting on there. I bought both of my shipments from Zoomhash and I can say he is 100% legit. But if you are buying more than a couple of units they will ship from China, granted he does use Fedex International Priority so it is very fast. (Took about 3 days to get to me each time).

Even though they supposedly sell 'plug and play' kits now I'm still not convinced it really is.
There are several ways you can power / configure these units. The worse part about getting mine was that the USB cables bundled with them did not even fit the units, the first batch I had to melt down part of the plastic around the mini-USB connector so it would slide into the Gridseed. Then the second batch, they sent micro-USB cables instead *uber facepalm*. You can use your own USB hubs but not all of them will work, at the moment I've tested a 10 port Rosewill one that works and it must be USB 2.0. The Gridseed gear does not like anything USB 3.0.

Powering the units is personal opinion which way you want to go about it, personally I went the way of spending the extra money on getting power bricks for every unit so I didn't have to hack together wires to use the LED / hobby PSU's that most people use ( and also only 1 would fail at a time instead of the whole rack). If you want to go that route then diagrams are available online so you can see how they wire them up. It is not difficult and all you need is the barrel connectors and some wire. I used 18 gauge Pyle speaker cable for testing mine and it worked perfectly.



Configuring them:
As of late, most people are selling bundles with a RasPi as they can easily run 20+ units.
I am using both the Wiibox controller and Windows on different setups.

The Wiibox controller is handy if you don't want to use a computer to run them, you just track down the IP of it in your router settings and you can get into it that way through a browser. It is very straight forward to set up, but you cannot see the stats of each unit. The only thing you see is how many are running and the pool settings. There are firmware upgrades floating around to be able to see more options but I have not messed with any of that as I've heard of people having their pool settings jacked and basically mining for someone else :/

Edit: The last Wiibox i got in a pack refused to work with any settings, so YMMV with these.


CPUminer / CGminer can also be used. There are compiled versions out there specifically for the Gridseeds.

For the Gridseed version of CGminer (Scrypt Only) :

DL Link http://cryptomining-blog.com/wp-content/files/cgminer-3-7-2-gridseed-windows.zip

Download and open Zadig : http://zadig.akeo.ie/

Go to 'Options' and 'List All Devices' then install the driver for the Gridseed, it will probably be named STM32 Virtual COM Port in the device list.

Edit the .bat file with your user / pool info before launching it, once it's connected to your pool then plug the Gridseed unit into a USB 2.0 port then it should start hashing away a few seconds later. If it's not hashing then check the 'hotplug' command in your .bat file, it may need to be changed from 0 to 5. Remember that the miner must be unplugged from the USB port every time you start CGminer.




And lastly, yes, they can mine SHA256 as well as Scrypt, but it really is not worth it profitability wise and they use a lot more power.

That about covers the basics, so really, are they worth it? That depends....

The Good:
They are small, run cool, and mine Scrypt at least 300 khash per unit.
Obviously the best part is the power consumption, the worst case scenario i've seen by plugging a K-A-W meter into a set of 10 units with a controller was 120 watts mining Scrypt.

The Bad:
They will not mine anything other than Scrypt / SHA256, so if you have a big GPU setup don't think about selling it all off to buy these, as we all know, when Scrypt gets ASIC bloated, the next best coins to mine will be something like Scrypt-N.

The price, they are fairly expensive so I only suggest buying these if you don't care about ROI, or if you've got coins burning a hole in your pocket. After all, it's not ALL about making $$$ right? :D

Amazon links for bits and pieces that I use (or have used) that 100% work.

10 Port USB hub: http://www.amazon.com/gp/product/B004F38WT4
Barrel Connectors: http://www.amazon.com/gp/product/B0058RLD9C
Power Supply: http://www.amazon.com/gp/product/B00D7CWSCG
Wire: http://www.amazon.com/gp/product/B003ZX8NX0
Single Power Bricks (for people who don't want to make their own wires): http://www.amazon.com/gp/product/B003TUMDWG/


I hope this helps some curious people.
 
Last edited:
As an Amazon Associate, HardForum may earn from qualifying purchases.
I've got a unit coming from Zoomhash. It is in the mail as i type this.

I'll update the thread.
 
What is your opinion on the excalibur 3 and it's claimed 200Mh/s mining capability?

https://www.mining-asics-technologies.com/product/excalibur-3-asic-scrypt-miner-200-mhs/

IMO, it's a bit fishy especially when it's essentially gambling 15k for what could be a scam.
Alpha-t may deliver, if they don't it'll make mining asics tech look even worse.

I don't preorder anything, I will only buy it if it's in stock and ready to ship. Until then, it doesn't exist to me :)
 
I don't preorder anything, I will only buy it if it's in stock and ready to ship. Until then, it doesn't exist to me :)

Quite wise and a smart way to not get ripped off, i think i might stick to this mentality. Don't want to get greedy and be burned for it.
 
Quite wise and a smart way to not get ripped off, i think i might stick to this mentality. Don't want to get greedy and be burned for it.

There's a lot of people selling their Scrypt ASIC preorders on forums, it's not the moneymaker like SHA256 ASIC devices were when they first hit the market, even buying big 2TH machines for SHA256 are not worth it these days with the amount of them the chinese are churning out every day and driving the BTC network hashrate into oblivion.
 
Im going to order 40 of these units as I can get them at $175 ea. Going to give them a shot and see.
 
Im going to order 40 of these units as I can get them at $175 ea. Going to give them a shot and see.

I'll need something to replace the hashpower of the two rigs I sold to reduce my heat generation during winter hehe.
 
Im going to order 40 of these units as I can get them at $175 ea. Going to give them a shot and see.

Curious where you got them from because at $175, it's a no brainer. 300kh at 7w, can't argue with that.
 
I very much dislike the packaging on the gridseed units. I'm more attached to aesthetics than I'd like to admit.

Edit:

I would buy the shit out of this though. Just because it looks hilariously awesome.

Rd37FN6.jpg
 
Last edited:
My supplier that I am using that quoted me $175 for 40 units + is MiA lol.

I have emailed him several times over the past few days to try and place the order and haven't heard anything from him yet.

I could possibly organize a group buy if you guys are at all interested, and the supplier decides to ever get back to me.
 
My supplier that I am using that quoted me $175 for 40 units + is MiA lol.

I have emailed him several times over the past few days to try and place the order and haven't heard anything from him yet.

I could possibly organize a group buy if you guys are at all interested, and the supplier decides to ever get back to me.

Would you really be comfortable sending 7-10K dollars to a man that goes MiA every once and a while?
 
The thing holding me back is the ease of use, software wise. I've read through the Gridseed thread at bitcointalk and things still look a tad messy. Don't get me wrong, setting up a GPU rig is not exactly set-it-and-forget-it but this looks like more hoops than I care to jump through for 300 hash. I'll wait for 2nd gen, 5Mh for 10W :)
 
The thing holding me back is the ease of use, software wise. I've read through the Gridseed thread at bitcointalk and things still look a tad messy. Don't get me wrong, setting up a GPU rig is not exactly set-it-and-forget-it but this looks like more hoops than I care to jump through for 300 hash. I'll wait for 2nd gen, 5Mh for 10W :)

They are a pain in the arse, but now there's an image to dump on a RasPi it makes things much easier by the looks of it. Having to unplug / plug the USB hub using CGminer in Win7 is getting annoying. One of my Wiibox controllers was silly easy to set up, the 2nd one i got was completely buggered and refused to do anything. All part of the fun i guess lol.
 
Ok, so I'm just curious, for those who are purchasing Gridseeds,

How many LTC are you expecting to mine in the next 8 months? That puts us around Dec. timeframe, by which the total return will be significantly diminished for these units.

Using the calculator and a 20 pack for $4500 (there's a 10% off coupon for zoomhash "mycode10")
But getting a decent OC, and 380kH on each unit, that's a total of 7600 kH/s.

Using the LTC calculator, with a measly 2% difficulty increase, 2% pool fee/rejects, it shows only about 165 LTC mined through the rest of the year.

But, if you took that $4500 and just bought LTC now ($19.8), then you could buy 227 LTC NOW.

So, why would you buy a pack of Gridseeds, if you can just buy LTC straight up and get more now?

I'm not trying to troll, but as I look back, I would have been MUCH better off if I had just bought LTC for cash when I first bought all my gpu rigs. Also, with the Gridseeds, the resale value on them is likely to be close to zero in 10 months when faster scrypt ASICs are out on the market.
 
You do raise a good point, and yes, if you're looking to ROI then they are kind of a bum deal at the current price. But i'm pretty sure the majority of people buying the big packs are farm owners who's gear is all paid off and using the income of the whole mining operation to make back the coins quickly. Then it's all more profit after that.

Or maybe they just don't care about the price and want new toys to play with :D
 
I'm not trying to troll, but as I look back, I would have been MUCH better off if I had just bought LTC for cash when I first bought all my gpu rigs. Also, with the Gridseeds, the resale value on them is likely to be close to zero in 10 months when faster scrypt ASICs are out on the market.

For me that is an easy choice. Sure, I can make more money faster trading LTC, but I can also lose it just as fast. Buying hardware and having it pay for itself was the clear choice for me. The hardware will always have value, even if the market tanks tomorrow. Now the choice between ASIC and gpu is a bit murkier. I don't have a clear choice there. The power savings of the ASICs are clearly attractive, but if scrypt mining tanks tomorrow, make selling dedicated mining hardware like ASICs hard to sell. I can sell GPUs easy on thEbay.
 
I took a quick look at the BTC/LTC calculators at Bitcoinwisdom.com linked above...
How many of you dedicated miners fully believe (or comprehend maybe) what you see there?

None of the numbers I run there line up even remotely with my "experiences" mining.
And, if you believe what their data "shows", then I do not understand why anyone would spend another dime on HW, of any type. As in, "If it's not already paid for or free you're losing money on every purchase."

Don't get me wrong, as that's not my personal feeling or intention at all, but most of the estimate web sites out there predict a much worse scenario than what I am realizing. Am I just lucky to be making profit and rapidly approaching full ROI?
 
I took a quick look at the BTC/LTC calculators at Bitcoinwisdom.com linked above...
How many of you dedicated miners fully believe (or comprehend maybe) what you see there?

None of the numbers I run there line up even remotely with my "experiences" mining.
And, if you believe what their data "shows", then I do not understand why anyone would spend another dime on HW, of any type. As in, "If it's not already paid for or free you're losing money on every purchase."

It's not making any sort of prediction, it's just math. If the price of BTC and LTC remain stable and difficulty continues to increase, investing in mining hardware is a losing proposition.
 
It's not making any sort of prediction, it's just math. If the price of BTC and LTC remain stable and difficulty continues to increase, investing in mining hardware is a losing proposition.

This is exactly why I am not currently investing anymore. I will keep with my GPU farm and mine till it isn't profitable anymore and who knows perhaps some of the coins who are thinking of changing algorithms will happen and the ASIC equipment that was purchased will no longer work on them while the GPUs will.
 
.......it's just math.......

That's my point. If it is "just math" my experience should more or less reflect the calculations I run. And it does not. Not even close.
I think the flaw in the calculators and why they are almost useless, is that they cannot factor in, at least with any accuracy, the human willingness to gamble or take a risk without the numbers being "on your side".
Just an opinion of course, but based solely on my personal observations.
 
I'm just still kicking myself for not buying more LTC last summer when I was suffering with 4000w of heat in my house from my gpu farm, and LTC was $1.80. I KNEW the price of LTC had to be at a low point, and was seriously considering buying more and just turning off the rigs when the heat was brutal.

Of course, I didn't because I was chicken.

Looking back when I first invested in my hardware last spring, I could have bought around 3000 LTC. Which is slightly more than I've mined so far. I started mining to reduce my risk. It is less risky because the hardware has cash value, I agree.

I'm not speculating on $$/ LTC or any other scrypt, it's just a matter of how many coins you mine.
One of my coworkers paid about 80BTC for one of the first preorders from Cointerra (when BTC was $150 in Oct.) Of course he didn't know a spike was coming.

But, if you look at it from a mining perspective of how many BTC the rig will MINE, it will only produce about 25 BTC (through the rest of the year) since he received it in mid Feb.

Now, had he used CASH to buy BTC instead of preordering a the BTC rig for $12,000 then he could have purchased 80 BTC at the time. Which, is FAR more profitable than mining 25 BTC with the rig.

Over the past year as I see these situations, it seems more profitable to just buy coins instead of preordering.

Of course, now I'm drooling over the prospect of a 100Mh scrypt miner for $10k.
Gee, how many LTC could I buy for $10k?
555 LTC.

I wonder how many LTC that new KnC Titan will mine in 6-8 months... That IS the question.
One thing is for sure in my book, the used gpu market will be flooded with Radeons this coming Fall.
 
Last edited:
One thing is for sure in my book, the used gpu market will be flooded with Radeons this coming Fall.

I doubt it, the same thing was said when BTC mining with GPU's was worthless and then everyone jumped on the Litecoin train, GPU miners will mine coins that use Scrypt-N and the likes instead.
 
That's my point. If it is "just math" my experience should more or less reflect the calculations I run. And it does not. Not even close.
I think the flaw in the calculators and why they are almost useless, is that they cannot factor in, at least with any accuracy, the human willingness to gamble or take a risk without the numbers being "on your side".
Just an opinion of course, but based solely on my personal observations.

I started formulating a well thought out response, but then I realized one word could fully express the point I'd like to get across...

Duh.

Calculators are only going to take into account the factors that are used as inputs. If you don't like the numbers, put together your own model.
 
^^^

yep. I completely agree.

I just used the calculator as a quick run of the numbers; I think even a 2% difficulty increase is not realistic, it will be MUCH higher.
 
I think even a 2% difficulty increase is not realistic, it will be MUCH higher.

And THAT'S the scary part. If we don't get a corresponding bump in BTC and LTC prices, GPU miners are going to have to look for alternatives to scrypt sooner than expected.
 
And THAT'S the scary part. If we don't get a corresponding bump in BTC and LTC prices, GPU miners are going to have to look for alternatives to scrypt sooner than expected.

I see plenty of "casual" miners, small time guys, and people living in places like California with > .30/kWh power costs leaving the game already.

I expect that to accelerate barring a doubling or more of LTC prices. Because scrypt coins of any flavor seem to be dying off except LTC and maybe DOGE.
 
Take a look at the FS forum here and you'll see a lot of casual miners dumping their hardware. A month ago you couldn't find a 280/290 thread for longer than a day.
 
Because the BTC craze that started back in November 2013 from all the positive press it got and which shot the price of it to nearly $1200 (from ~$100) is finally calming down. The alt-coins, which are tied to BTC, have lost a good chunk of their value in the past 2-3 weeks so the noobs that joined in between Nov and March are freaking out, selling the GPUs before they lose their value.
 
Because the BTC craze that started back in November 2013 from all the positive press it got and which shot the price of it to nearly $1200 (from ~$100) is finally calming down. The alt-coins, which are tied to BTC, have lost a good chunk of their value in the past 2-3 weeks so the noobs that joined in between Nov and March are freaking out, selling the GPUs before they lose their value.

In other words... BUY BUY BUY! ;)
 
Back
Top