Topic: GRIDSEED G-BLADE Overclocking 7Mh/s, improvements and repair - page 5. (Read 74130 times)

Hah! I knew you couldn't resist ZiGGy!

Hey, Wolfey2014...

How're doing, buddy...

2.05BTC for 300 chips...

ZiG

Thanks J4...
I just looked and those crooks at Avalon want 2btc per chip? Am I reading that right? What a RIP OFF!
Just corrupt opportunism! Sickening!

Yes you are right as I said many times before, those guys forget one and most logically thing.
If mining is profitable than companies which produce miner will build miners and use it for them self not for selling.
But the only thing profitable here is make miners and sell it, not mining on it, because dificulty raise exponentially every time when some company starts to sell their miners and you can never pay off yours because of that. Because every two months come out 10 times more powerfull miners.

The only profitable way is to build your ovn miners and mine with it.

The second upper gate mosfet should be enough, but 300w for 7.5 to 8MH is useless. We will soon have to undervolt the blade to keep it running and make profit, depending on how much we are paying for electricity. I'm paying around 0.12$/kWh here, all fees included.something like 0.09$ excluding fees.

The best thing to do, on every miner, is over clock until it starts eating more power that what it earns. Then back to stock power draw, and after that, undervolt to keep it working for some time.
Or try to resell before over clocking isn't profitable anymore.

With 43k resistor, with all the extra cooling, I was able to reach 1075MHz. The blade was using 300watt for 7.1-7.2MH/s

The best small heatsinks I've found for the top of the MOSFET are from moddiy. I posted a link somewhere in this thread.
Side fan blowing on the power board should be necessary to cool the MOSFETs and the coil, even with the heatsinks under and on top of the MOSFETs.

Building my own boards for the avalon gen 3 chips.
https://ehash.com/product/a3233-300pcs/
The FPGA are unfortunately too expensive to expect any ROI unless you already own them.
And they are useless once asics are out, since asics are cheaper and more power efficient. So FPGA for X11 and newer algo, if you already own the FPGA could be ok for some time until a manufacturer comes out with the asic chips.

Home made Avalon? What is that? Please describe in detail,,,,
And what about trying to program existing FPGA's?
There are development boards you can purchase and program....
Not speaking as if an authority. Just what I've read and heard here and there along the way.

Thanks!

Nice, new blades going for $200, at X-hash?
2nd mosfet would help, but would that put you over 300watts?
New miners are 28Mhs @ 440watts, lower power draw.
Some point in time very SOON it will not be worth it for the 5Mhs
blade operators to continue mining at stock hash powers.
Might be time soon to crank the bitch up to max hash power!!

I might add a second set of heat sinks on mosfets and glue them
with thermal epoxy fins to fins to try and get some more cooling.
Might be able to crank her up a bit more, running 975Mhz now.
What resistance is needed for over that freq, do I have to go all
the way to 43K?

I guess you could use that thermal switch with a self locking relay and a momentary switch.
You need to connect 1st NO relay contact to Vcc and back to relay coil (after momentary switch)
The thermal will beak the coil power via its connection to GND if it should over temp.
The 2 blade power plugs would be on the 2nd and 3rd relay NO contacts.
If bi-metalic thermal opens then coil power drops on contact 1 and the 2nd and 3rd NO relay
contacts also go from Closed to Open killing blade main powers.

Hi, wolfey, I never tried to overclock to 8MH, the newer blades that are selling for 200$ with the extra upper gate MOSFET solder pads available should be able to reach the 8MH if you solder the extra MOSFET and give all the extra cooling.
I mostly worked on the Zeus blizzard for the last few weeks, but it was disappointing too. Plenty of extra power available, easy overvolting, but a hard limit in the chips and the bypass pins that aren't soldered. And bad cooling that gives high HW errors rate even at stock speed.

It seems that the manufacturers are just looking for profit, and not caring at all for the customers.
The turn over on hardware is so fast they don't really care.
There are a few small changes in the design and in the products that could make them better at almost no cost, but it looks like they aren't interested in listening us.
And when we are asking for the chips to be able to make things better, the price is so high that it will cost us way more to build the hardware than to buy it.
I'll probably switch back on bitcoin mining with a home made Avalon gen 3, since for now, it looks like some ROI is possible, unlike with scrypt hardware.
The limit today for scrypt is around 15$/ MH to expect ROI (and the manufacturers are paying way less than that for the finished product)
For SHA256, the limit is near 1$/GH. I expect that the homemade avalon could cost me something like 0.75$/MH.

I'd love to build our own miners.  I've been working with computer's for roughly 30 years myself.  Originally went to school for electronics, got sidetracked by the pretty pixels, but have always had my hands in the hardware side changing and modifying things.  I look at the hardware that Gridseed and Zeus have put out there and am baffled by some of the design decisions they chose.  The week power components that were used on the blades, and Zeus with wires coming off the boards on two ends....I suppose if you like a messy setup with cables everywhere it would be alright. 

Right! PRE-PAY NOW! LOSE BIG LATER! Or,,is that NOW!!!???
Face it guys, Pre-Sales are for SUCKERS!
Don't play that game!
I don't and I REFUSE to fall for any buy now hype BS!
I can't believe HASHRA,,,,,,,Listen, wait 2 weeks and you can buy the same BIG DEAL 'BUY NOW' miner for 50% less than you can BUY IT NOW, for! No thanks to Scrypt losing value so fast, you won't even lose enough money to pay for shipping ... 2 weeks later!
I have 26 miners making me just under $5/day! Woo Hoo! At least they ROI'd,,, It's all profit but just no where near enough and certainy no where near where most of us predicted they'd be at this stage of the game. But, the market is EXTREMELY unpredictable as it is!
Then again, who is really pulling the strings UP there! Anyone? Or, no one? It's just happening all by itsel,,,yah, right!
STOP THE MADNESS! and let's make our beloved coins PAY US PROFITS for a change!!!!! 

just my .000000000899---jussfivedollahs cents....

Again i asked myself why we have to re-invent this f-----g Blades? For the introduce Price of 3000 $ , they have to offer more! Ok , not for Overclocking , but even if you have some high temps outside , you would get some problems or burned parts under normal conditions. This is more like : "Ok , let us glue some parts together , an look what happens." This mining devices normaly have to checked more than any other casual electronics, because they are designed to run 24/24. And what is about any responsability from manufactor? Nothing! I don´t want to say that Grideed has bad hardware , other firms are surely not really better.But don´t bother , we are all the part of this community , and if YOU are able to sell some parts without any responsabilty and warranty and at high priceses , why not? As long as the "customers" are willing to spend high pre-order prices and they accept , that there is no guarantee, again : Why not? This is what we read everytime : No refunds, restricted warranty , all sales are final.-........
And I contribute to this system too. I received my "overclocked" Blade just 4 weeks ago ( delayed over 5 Weeks), and pay for them 899$ , because i thought : Oh what a beautifull pice of ingeneer!
And now for the same piece of ingeneer you get only 300$ , if you´re lucky, and for mining we have discuss this enough. So Who is the bad guy in this race?Aren´t we , the customers, who encourage those firms , to do things like that? Ok maybe they selling a few hardware to those "big fishes" or "minig farms people" , but they are rare , the "big" part , are every one of us.Only if we supply them, they can continue like this.... And just if they mine by themselves, but they only mine "Srypt" like anyone of us and what happens if their mined coins don´t sell anlyonger?
Sorry, about this , but i only want to encourage you people to fight for your right, because YOU count.... only if those small people with this small antminer U1 or U2  or S1 or GBLADES , goes offline, how do you think this counts? Just an example,In the sum it´s a real big part. But this is the trick of the big players, let them thought they are small...and they are really small!So don´t belive what they say, that you reach you ROI in a copuple of weeks, if one of them offer you a real big deal , catch this, otherwise don´t bother with.
I still belive in cryptocurrencies , because  there is a way out of the controlling of banc systems,but only if we realize decentralized mining we have the benefits, otherwhise is like any other system , thay allready exist. 

I know, to much philosophy for a thread like this, i´m sorry , i hope you guys don´t mind

 

Hi ,
Just some suggestions :

I´ve found this thermal fuses on voelkner , just if you want to be really sure and don´t have this on/off effect , they can handle 10 A and react at 93 C° so for one PCB it´s enough , but in worst case they must be replaced. The advantage is , that they cut the complete power off and it remains off. So you can use them direct to power.
----> http://www.voelkner.de/products/3382/Temperatursicherung-93-C.html
I don´t found something similar on mouser , maybe someone else?

Other Idea that i have was to use a circuit breaker like this ----> http://de.mouser.com/ProductDetail/TE-Connectivity-PB/W28-XT1A-7/?qs=sGAEpiMZZMvNllHT6Fj2fjyJLrUxEIJB5lXHI1hSl5Q%3d (There are no pics , but look into datasheet and you know how they look.)

They offer a overload protection of 7A , this is just right for one PCB , and they are very sensitive to ambient temperature , so if the airfow form fan works, they opperate normaly , but when the fan stops they overheat and break the circuit permanently , even there is no overload related to the current. The big advantage is that they are resetable , so they remain off even you push the button

But this might take some time to find the real life conditions and for testing to find the right one. But if it works , it´s sureley a real secure and comfortable sollution , whitout replace any parts or have this switch between on /off. There are also higher ones , with 15 A so its posible to realize one sollution for both PCB. I think this can work , but it´s hard to find the sweet point , so that they dont react under normal condition but quickly react under overtemp.
If you find the right one , its possible to drill a hole in the original cooler and bring them inside like a temp sensor.

This is what i mean : (ok nearly like this  )
http://i59.tinypic.com/hs759c.jpg

Maybe some suggestions from the experts here?

I don´t try with those circuit breakers so far, but i have a bimetal switch at home and at the right place it´s possible to bring them to react only if the fan is off. For those circuit breaker i think you have to operate them to the limits , maybe something near to overload point, so that they remain on if they are cooled from the fan, but without cooling and with the additional heat from the chips , they can switch off quickly.
If someone want to try this, the best thing is to choose one sensor for both PCB , so if there is an overheat both PCB´s goes offline, and to find the sweet point, choose a summer day with high temps, so that you can be sure that it´s not react under "normal" conditions, but also check under lower outside tems if it works properly, then get the fans off and check if they break the circuit. In the datasheet are the reaction times for different situation of overload, so you have to add the ambient temperature correction factor , to calculate wich one to choose.And you want to do real perfect , you can use a 3 Pin Screw terminal for soldering on PCB , one pin without connection to PCB  just directly to the circuit breaker , and the original + pin is conected to PCB and you wire this directly form Sensor.

This is the affect of the ambient temp to the reaction:
http://i57.tinypic.com/vn164l.jpg

What do you guys think?

http://www.ebay.com/itm/12-Pcs-NC-90C-2-Wire-Connector-Bimetal-Thermal-Thermostat-Switch-250V-5A-/231047800117?pt=LH_DefaultDomain_0&hash=item35cb85a535

I didn't figure out what to turn off with this switch because you can't use it on power because it current is too high.
So you should turn of something else. But you must know when temperature drops swich change state and will make blade goes ON again.

Hello --

Blisk made an earlier post in the thread where he attached a picture of a thermal switch.  I am assuming that the switch locks into the "OFF", or open, state when a certain temperature is reached.  If I remember right, his posted example triggers "OFF" at 95 deg Centigrade.

If the switch is installed, it could serve to open the circuit and shut down the Gridseed when triggered.  I would suggest placing the temperature sensitive end nearest the hot-spot components on the external portion of the circuit board.  I think it could be wired in series along the input power line.

Someone just posted data about the thermal behavior of common components.  Performance and specifications began to change at 90 deg Centigrade in the table that was posted.  I would want to shut the Gridseed off before that stage was reached.   -Scryptr

Thanks for this important advice. How do you realize this ?
With 2 Fans its a little bit more secure than with just one , if one fails.But a thermal switch is much better.