Topic: GRIDSEED G-BLADE Overclocking 7Mh/s, improvements and repair - page 28. (Read 74131 times)

To stay on topic, since I'm back home, I made some improvements to the cooling.
Looks interesting for now.
heatsinks in each chips, forced air flow on the overheating parts with custom cover.
a few pics will be posted soon.

I also tried dual 25V 1000uF capacitors I had at hand, but they don't improve anything.
So 220uF 25V should be enough and a little smaller.

BTW, I managed to melt a power plug from my cables by trying back a 46K resistor. Simply too much current for it to handle if you use a single plug.

I'm so happy that my presence drives you insane!!!!, Zit...

Yes KNC and you have two more companies which you can preorder with 30% paying now and the rest before shipping.
Hashra have to high prices like gridseed have, people forget to calculate when you get money back and forget to calculate that other people get those miners and dificulty will raise like hell.
Nobody learn from Bitcoins?
And you forget to calculate that 250Mhash or 90Mhash machines in factory cost less than 1000$ and I am asking again why do you think they seel them and not use them? Just simple calculation again, because they will never get that much coins from it like when they can sell it fo 10.000$.

I buyed gridseeds because I am interested in ovrclocking it, only that.
And the only but really the only way to have those miner is to build it by yourself.
I have to less knowledge about that but I know that many people have enough of it and we can together build low price miners.
and only this way you will have some profit.

And only if you get on lottery and than you buy those miners and they give you money back and maybe some extra money in 6 months when they pay back itself.
and again, didn't we learned nothing by BTC asisc, just check interent.

I do not compare prices on machines that are "coming soon" or Q/1/2/3/4.
Too many companies with delays.
Right now the A2"s are shipping and they are pricey but the only game in town as of TODAY.

June. 1st might or might not be a different story.
I know of 3 other companies that are "coming soon on or about June 1st." If they are shipping then I will compare.

-Bobby

You state the obvious as though you have or you are going to receive an argument.
Supply and demand is the driving force you speak of.
Please, what is your source for your claim of 250MHs Scrypt for $12,000?
I bet Hashra will match if not beat it if it is legitimate and not just based on Vapor Ware (non-existent on the market if it exists at all) pricing.
Link please?

It still may be viable to overclock the Blades, if not just a fun challenging hobbyists' thing to do on the weekend. The primary question being answered right now is, can it be done? And it seems it can be done, thus far anyway.

hashra is too expencive they offer 90Mhash for 15.000$
You can get 250Mhash for 12.000$
and forget about overclocking and people quickly forget payback time.
If you pay that money for the device when you get money back?
We learn from Bitcoin asics.... almost "never".
also people forget about producing costs of one of that machines it is under 1.000$ and you are paying 15.000$
why do you think it is much more business for them to sell those machines than to use it to mine?

Wow J4! I'm impressed!
From the sounds of it, these things are gonna look interesting when you guys get finished with your modding work!
I hope they deliver what you hope they do. 8+MHs stable would be something! Uhm, I mean, will be something
Sounds like you have everything under control. If I had my own blade, I'd be right there with ya!
I can hardly wait to see your results!   

Just make sure everything in the power system is super beefy overkill and you can then start overclocking without having to be concerned about supply / driver side failures and such. I always do this for my prototypes. They always contain ample current carrying capacity.

I recently took a look and found that Hashra has some new super duper - up to 90MHs Scrypt hashers for sale now. Their own brand with I think, Pioneer chips in them. Bye bye GridSeed ASICs?
Our pod hasher days are now numbered, for sure! So are our blades, but we can still probably get a few to several months of profits out of them before the diff' rises too high for them to be viable anymore. Still, should be plenty of time to ROI on them at least.
That's my guess based on what I and some others think. So we need all the OC / profit increase we can get, now! now! now! Just my non-expanded upon, current point of view.

Just out of curiosity, isn't there a way to simply add solder trace to thickin up power the busses on the boards instead of what ever it is you're doing with the pod pcb's? Or is it just for prototyping?

Go Hackers Go!

OK, share with us what will you do so we can do it too.
I have planned to add some more heatsinks to board, to cool down chips and some thermal switches if something goes wrong it will shut down power supply.
I thought 95 oC is top limit

all the overheating parts will be doubled, including L2.
The up1509 will be powered by a doubled channel because the actual limiting factor is the capacity of the channel.

It should work since the other components aren't stressed that much, even when I overclock @ 7.5MH/s
We might be able to achieve around 8MH/s, maybe a little more.
On my pods, the chips are running at least at 104kh/s, so my math says 8.3Mh/s is doable with the good power board.

So as I see you will add mosfets like there are Q5 to Q8?
Are you sure that other parts like L2 and some resistors can handle that?

Hi,

you are close to what I want to do. I want to double the part of the board that feeds the up1509. this way, the chips will still have the same regulated power supply.
The idea is to try to bridge 2 pods before going on the blade.

I already identified easy connection points on the board, yellow/orange circles

red circles are our hot spots to cool down.
Funny thing is the initial design used 4 vrm (Q5 to Q8) for 8 chips and they decided to cut it down to 3 as on the 5chips pod, but for 40 chips. Should be related to the LTC only mode on the blades.

+5V will be collected after bat54C on R10, 12V at C12 and the return will be at R14, R117, R15, EC3, R31, R36 plus ground. only 9 connection points, should be ok.

 

If you like to use two power supply for powering 40 chips, that is not easy to do with tha low voltage you can quickly jump over 0,5 V or more.

Hey J4bberwock,

Why you don't try to use 1 Gridseed 5-chip board as additional power source...?

Their power section is designed to deliver up to 60W in BTC...which is about 30+ Amps @ 1.75V overvolted... maybe enough to supplement the power section of the 1 blade...plus you could control the Voltage DVDD+ very precisely using a trim pot...so the Blade and the 5 chip would be at the same voltage for parallel operation...

IMHO it will be easier to implement than a custom external source...just an idea to think about...I believe it's doable...  

ZiG

EDIT...: It is already "Cut&Paste" ...almost the same power section...+ it is already manufactured, ready to use...no redesign, PCB, production...etc needed...C42 is a good output point...

I think that is too complicated. I think some extra board for power and you cut existing power chip and replace it with more power.

A simple copy/paste of the actual board with same or upgraded components connected in parallel should work.
That's what I'll try.

so you need about 300W per blade, but I think only way is to made a custom power supply or to get one with 5V and lower to what it takes for OC.

Actually, I would call for a bit more overhead than that. Just so they don't have to work as hard hence, don't get as warm either. Maximum efficiency!

Second this...

5 chips design is consuming 20W+ when volt mod...DVDD to 1.75V-1.80V ,,,so each chip is like 4W /1.75V = 2.30A ...or around 2.5A Max per chip provably will be enough...I guess...40 chips per blade = 100A @ overvolted mode

This is why I suggested you the choice of 100A MOSFETs a few posts above...

ZiG

did not measured it but...
5chip grid with vMod1 drawing 10W without fan, Voltage 1.25-1.35V
assuming 90% efficiency at PSU and from 12V to VRM i would guess:
1.1-1.3A per chip

yes I know it was  just a quick look, it is not even enough powerfull after I check it
how much ampers that 40 chips consume is anybody measured?

I have checked some 5v power supply and to lower voltage to 1,4 or 1,5 V.
But with high amperage it will make it extremly hot.