Pages:
Author

Topic: Becoming Asic Manufacturer (Read 3178 times)

sr. member
Activity: 1470
Merit: 325
September 04, 2018, 02:14:19 PM
#21
-snip-
lol what a question it depends what dimensions you seek to produce and what kind of quality, and the location, the employees, its an od question and can only roughly be answered, you should ask special electric industry engineers not us here, this is ridiculous.
Do you realize that the OP was from 2014 and only brought up by a newbie?
Anyways, the newbie's question seems valid.

HI I want to do the same manufacter asic chips for bitcoin 16nm or 8NM 150GHS each i want to know steps and if u have business plan for this ;; what if that the models exist what s the cost for that please i need business plan of asic production
The user's last login was December 30, 2015, 05:17:36 AM, must be busy somewhere else.
Next time, don't hesitate to open a new topic regarding ASIC manufacturing business, forum members (& scammers) will not hesitate to help you depending on your given info.

Luckily, he left an email address on his profile: [email protected], you might be able to contact him through email.

nope i didnt knew that, the question seemed to general to me so i answered, its not regular occurence in this forum that newbies get dusty topics out of the dust.
legendary
Activity: 2534
Merit: 6080
Self-proclaimed Genius
September 03, 2018, 11:22:29 PM
#20
-snip-
lol what a question it depends what dimensions you seek to produce and what kind of quality, and the location, the employees, its an od question and can only roughly be answered, you should ask special electric industry engineers not us here, this is ridiculous.
Do you realize that the OP was from 2014 and only brought up by a newbie?
Anyways, the newbie's question seems valid.

HI I want to do the same manufacter asic chips for bitcoin 16nm or 8NM 150GHS each i want to know steps and if u have business plan for this ;; what if that the models exist what s the cost for that please i need business plan of asic production
The user's last login was December 30, 2015, 05:17:36 AM, must be busy somewhere else.
Next time, don't hesitate to open a new topic regarding ASIC manufacturing business, forum members (& scammers) will not hesitate to help you depending on your given info.

Luckily, he left an email address on his profile: [email protected], you might be able to contact him through email.
sr. member
Activity: 1470
Merit: 325
September 03, 2018, 05:30:04 PM
#19

 I would like to ask ( btw I have not even a single clue so don't hesitate to eli5 this)

 How much capital would one need to start a business and get asic chips at lowest price from a manufacturer or make one yourself?

 I am guessing (assuming?) getting couple of chips and reverse engineer it and make thousands of the same stuff would be costly if not impossible.

 So lets assume this scenario , I get 10-20 scrypt asic miner chips , pay someone to reverse engineer and find out how its made , go to a chip factory with him , explain the manufacturer how it should be made with him and pay them to make couple thousand at least. How much capital would this take?

 Or any other example you might have in your mind.

 TL;DR , how much would i need to have to become next asicminer,gawminer company.

lol what a question it depends what dimensions you seek to produce and what kind of quality, and the location, the employees, its an od question and can only roughly be answered, you should ask special electric industry engineers not us here, this is ridiculous.
newbie
Activity: 1
Merit: 0
September 03, 2018, 04:26:19 PM
#18
Great questions.

Let me start of first by laying out two scenarios for you.

Scenario 1: Making your own chips

Firstly, the name manufacturer may be a bit of a misnomer.

The process in ANY chip production life cycle is as follows.

Step1: Find a really smart person to make you what is known as an RTL http://en.wikipedia.org/wiki/Register-transfer_level
          This is essentially the ground work for telling the chip what you want it to do. Think of it as sort of the firmware of the chip.
          RTL is generally either written in VHDL or verilog.

Step2: Once you have your RTL completed, you need to compile it to an FPGA. Now I am sure you all remember the FPGA days of bitcoin machines. FPGA's in a general sense are just prototype boards that you can plug in your RTL into in order to make sure it works. So in the case of a crypto RTL, generally folks compile their RTL on something like a XILINX FPGA so they can validate their design. This is why you get many asic companies showing off their FPGA boards at the beginning stages. This illustrates, generally, proof of concept. So if their RTL works, they can partially compile a small number of cores to execute their RTL and in the case of crypto, these RTLs mine coins. So if your fpga mines or even accepts shares, you are golden from an  RTL standpoint.

Step3: Initiate backend design. Now this backend design can constitute many things. Whether that is making the PCB design, or makeing the mask design.  Essentially it is at this point that  virtually all ASIC manufacturers outsource this part to the experts. Experts being companies that have years of experience doing backend design for ASICs of all kinds. So, the most notable one IMHO would be Innosilicon in China. They do tons of tape outs for many industries inside and outside of crypto.  At this point you would tell them you want a 28NM chip or whatever transistor size you want and give them the RTL. They will validate again, and create the backend and PCB (if not already designed) and then essentially send off what is known as a GDS to a foundry.

Step4:  The foundry will take the GDS and make what is known as the mask. The mask is pretty much what it sounds like. It is a piece of material they will lay on top of the Silicon Wafer and then after various processes including, doping, and photoetching, create the asics.  This mask is very valuable and where most of the money in a tape out comes into play.  With it, you can make that chip.

Step5:  If you use a notable backend company to do your design, then generally there is no need for a so-called engineering batch.  Most well-known companies will merge your RTL with their tried and true backend to make it a sure fire home run. If you do a engineering batch, it is because you are using an untested backend design where making a full scale production run may be a waste of money if you need to make tweaks to the backend. So, the chips are made, the PCB is made, now just assemble it, put a power supply in an enclosure, and sell it right? NOPE!

Step6:  This is where some really important things come for global commerce. Firstly, if you do not have all the regulatory clearings from some big names in safety, some countries will block your equipment in customs. This is generally a must-do although there are some "creative" ways around this. Also, you need to test the chips yourself to make sure it performs as expected. Also, you need to develop your drivers for the machine. So if you are running a raspberry pi as a controller, you have to make a driver so CGminer can recognize it along with maybe making your own fork off cgminer. Then comes the task of making it user friendly with some sort of webGUI.

That is generally the process. Now as far as cost goes here is a break down.
NOTE: the larger the transistor size, the cheaper the manufacturing process. The prices below are for 28NM however if you go 20NM or 16NM these prices will surely go up ten-fold

RTL:  50,000-100,000 USD depending on if you do it or if you hire someone (generally just a salary)
Backend:  300,000 USD
PCB Design: 10,000USD
Mask Making:  1.1M USD
Chip Cost per chip:(greatly varies on many things)
things that affect chip price are:
RAM TYPE:  ECC or Non-ECC
RAM AMOUNT
Number of Cores
Number of Layers
Generally speaking for a crypto chip the price per chip varies. For a SHA-256 chip there is generally no RAM in these chips. so these chips are cheaper. Now scrypt chips and others require alot of ram so they are more expensive. With that being said chips can range from $2-$15 each.

So you must pay for the number of batch1 chips

All in all the total cost comes to around 3M USD give or take..

Scenario 2: Buying Chips

Generally this can be a good idea because of cost savings, however, if will need to proper logistics in order to physically get those chips on a PCB and hook them up to a controller of some kind. If you have those avenues sorted already, it may be a good move. However, if not, it will quickly get hairy.

Hopefully this answered some questions
HI I want to do the same manufacter asic chips for bitcoin 16nm or 8NM 150GHS each i want to know steps and if u have business plan for this ;; what if that the models exist what s the cost for that please i need business plan of asic production
member
Activity: 108
Merit: 10
September 08, 2014, 08:08:16 PM
#17
thanks for re wording my previous post Bobby Grin
newbie
Activity: 13
Merit: 0
September 07, 2014, 05:14:38 AM
#16
You don't need to reverse engineer the chips. You may reverse engineer the software, but you don't need it as well. You need to raise 1M USD at least to start with. It pays for an initial design, floor plan, logic schematic, electric simulation, layout verification and the cheapest multiwafer mask with a very small batch of packaged chips. You also need an engineer to lay out the PCB for your miner, prepare the BOM (Bill Of Materials), order these parts and hire someone for the PCB production and assembly service. You need a Linux programmer to get the software running. If it works out and you have a working miner, advertise it everywhere and raise more capital to run the production with an improved design and better mask set.
legendary
Activity: 812
Merit: 1002
September 06, 2014, 01:40:34 AM
#15
Man this thread has a bunch of newbs that knows wtf they're talking about, unlike most other sections of the forum. Really great post, FinalHash.

I'd love to look into this kind of stuff, but the initial capital is far out of reach for the typical Joe.
member
Activity: 108
Merit: 10
September 04, 2014, 01:41:43 PM
#14
Ok for 35K i would suggest just buying on a steep discount and then reselling them. Feel free to shoot me an email if you want to talk more about buying in bulk.
sr. member
Activity: 434
Merit: 250
September 04, 2014, 10:28:37 AM
#13
Questions:
What is your electricity rates per KWh?

 Its about 0.1$ , changes from 0.1 to 0.15 but that can be get close to 0 if I can find a big earner , that way I will have a big place that has its own electricy thats already included in the rent.

How Much money do you want to invest?

Well as less as possible but I am assuming 35k for a start of buying some samples and hashing with them and see if they work as advertised , hell I would love to get free samples to check first and go from there , so my big starting capital is barely 35k so I need to spend it as carefully as possible , if I can actually get it going on I have access to 400k worth of line of credit.

Do you need PCB also?

 After reading your stuff I am more intend towards buying go to ready stuff from places like innosilicon (if I didn't spell that wrong Cheesy ) rather than making my own , 35k-50k is not going to get me create new chip and get it manufacture and resell.

Do you have enclosures ready or do you want to buy them?
 Definetly not ready . I have the space and factory under my hand but they have no idea what hash is or what bitcoin is so its almost impossible to adapt
What kind of place do you customers generally want to put their machines in? (big machines that pull alot of power wont work in the average house)

 I definetly have the space but my initial attraction to this was getting a wholesale price and resell these , may it be either just getting chips , putting them on boards and sell them or getting already board at plug and hash readiness and get them wholesale price and resell them , I would be hashing only meanwhile they are under my control just so I wouldn't lose money in while of that time.
member
Activity: 108
Merit: 10
September 04, 2014, 09:30:14 AM
#12
Right on, here are the answers.
How much hash power is there per chip? I assume 15 dollar chip gives more hashing power comparing to 2 $ chip?

This is not a direct correlation. Hashpower:$per chip is not a correlation AT ALL. It really comes down to things like
Is there ram on a chip. Sha-256 chips DO NOT have ram. Meaning that these chips will be cheaper in general.
Secondly, the physical package size matters on cost.
Is your chip 11mmx11mm or is it 30mmx30mm. The bigger the chip the more expensive.
However, a bigger chip does NOT mean a faster chip. Realistically there are several things that come into play here for cost and none of which are correlated to hashpower.

Quote
So per chip how much would it hash?
This is heavily dependent and almost solely dependent on the RTL. If it is a good RTL and well though out, the architectech would have put in alot of thought into how to reduce the RAM package size (in the case of a scrypt chip) so he can fit what he needs on a smaller package size. Also, how the chip operates is important too. Things like BUS size and things of that nature really play more of a roll into how fast a chip is.
Quote
Sha or scrypt doesn't matter as long as you know you can share all of it.
not sure what you are meaning here. If you are saying it does not matter from a cost standpoint, that cannot be the case. However, I think I am just misunderstanding what you are saying here. Elaboration for my mental slowness would be appreciated. Smiley
Quote
Also when you buy where are the 'place to go' ? The best trustworthy known great deals etc etc place that we should definitely check out to buy asic mining chips?
If you are asking where to go to buy physical chips then the answer is... it depends.
Every chip has it's own pros and cons.
Things to consider for a chip are as follows
+Power Consumption
+Speed
+Form Factor
+Heat Output
+PCB Design
+Cost
+Quality**
**These are for chips in a general sense. In crypto Quality is not generally an issue because product life span is very short due to obsolescence


Now what is more important to you is different from person to person.
For instance, if you get free electricity, you do not care how efficient the chip is and how much power draw it uses. However, if you are getting charged a high price on electrical, this should be your paramount goal.

SPEED IS RELATIVE!
I want to let everyone know that speed is relative. If you have a chip the size of a bitfury chip (small as hell) that runs lets say 1 MH/s and you have a chip 10x the size and it runs at 9.9MH/s these chips still cannot be compared. Just because one is bigger and faster and one is smaller and slower and the bigger one is < 10 smaller chips still means squat.
In order to evaluate speed you need to think about how well the PCB design is. If you have a huge chip, probably going to put 1 or two chips on the board. You can put more small chips on a board, but them cooling may be an issue. So this is where form factor comes into play.

The thing i love about the spondoolies board's are that they are so freaking well thought out. The squeeze a boat load of chips on a big PCB and then overlay the heatsinks accordingly so you can stack alot of stuff on top. This leads to a small form factor but with a higher cost due to their awesome R&D.

So, if you want to answer the below questions (not even necessarily for your use case but just a generalized one) I could point you in the best direction. Generally speaking, when buying alot of hardware, it is always best to do it through a friend of the manufacturing company. Otherwise, those guys do not give anybody the time of day just because folks are hitting them up all the time and they have work to do.


Questions:
What is your electricity rates per KWh?
How Much money do you want to invest?
Do you need PCB also?
Do you have enclosures ready or do you want to buy them?
What kind of place do you customers generally want to put their machines in? (big machines that pull alot of power wont work in the average house)

sr. member
Activity: 434
Merit: 250
September 03, 2014, 09:14:36 PM
#11

 Well it covers most of the stuff but there are very key stuff that are missing.

 How much hash power is there per chip? I assume 15 dollar chip gives more hashing power comparing to 2 $ chip?

 So per chip how much would it hash?

 Sha or scrypt doesn't matter as long as you know you can share all of it.

 Also when you buy where are the 'place to go' ? The best trustworthy known great deals etc etc place that we should definetlly check out to buy asic mining chips?
member
Activity: 108
Merit: 10
September 03, 2014, 04:29:57 PM
#10
Great questions.

Let me start of first by laying out two scenarios for you.

Scenario 1: Making your own chips

Firstly, the name manufacturer may be a bit of a misnomer.

The process in ANY chip production life cycle is as follows.

Step1: Find a really smart person to make you what is known as an RTL http://en.wikipedia.org/wiki/Register-transfer_level
          This is essentially the ground work for telling the chip what you want it to do. Think of it as sort of the firmware of the chip.
          RTL is generally either written in VHDL or verilog.

Step2: Once you have your RTL completed, you need to compile it to an FPGA. Now I am sure you all remember the FPGA days of bitcoin machines. FPGA's in a general sense are just prototype boards that you can plug in your RTL into in order to make sure it works. So in the case of a crypto RTL, generally folks compile their RTL on something like a XILINX FPGA so they can validate their design. This is why you get many asic companies showing off their FPGA boards at the beginning stages. This illustrates, generally, proof of concept. So if their RTL works, they can partially compile a small number of cores to execute their RTL and in the case of crypto, these RTLs mine coins. So if your fpga mines or even accepts shares, you are golden from an  RTL standpoint.

Step3: Initiate backend design. Now this backend design can constitute many things. Whether that is making the PCB design, or makeing the mask design.  Essentially it is at this point that  virtually all ASIC manufacturers outsource this part to the experts. Experts being companies that have years of experience doing backend design for ASICs of all kinds. So, the most notable one IMHO would be Innosilicon in China. They do tons of tape outs for many industries inside and outside of crypto.  At this point you would tell them you want a 28NM chip or whatever transistor size you want and give them the RTL. They will validate again, and create the backend and PCB (if not already designed) and then essentially send off what is known as a GDS to a foundry.

Step4:  The foundry will take the GDS and make what is known as the mask. The mask is pretty much what it sounds like. It is a piece of material they will lay on top of the Silicon Wafer and then after various processes including, doping, and photoetching, create the asics.  This mask is very valuable and where most of the money in a tape out comes into play.  With it, you can make that chip.

Step5:  If you use a notable backend company to do your design, then generally there is no need for a so-called engineering batch.  Most well-known companies will merge your RTL with their tried and true backend to make it a sure fire home run. If you do a engineering batch, it is because you are using an untested backend design where making a full scale production run may be a waste of money if you need to make tweaks to the backend. So, the chips are made, the PCB is made, now just assemble it, put a power supply in an enclosure, and sell it right? NOPE!

Step6:  This is where some really important things come for global commerce. Firstly, if you do not have all the regulatory clearings from some big names in safety, some countries will block your equipment in customs. This is generally a must-do although there are some "creative" ways around this. Also, you need to test the chips yourself to make sure it performs as expected. Also, you need to develop your drivers for the machine. So if you are running a raspberry pi as a controller, you have to make a driver so CGminer can recognize it along with maybe making your own fork off cgminer. Then comes the task of making it user friendly with some sort of webGUI.

That is generally the process. Now as far as cost goes here is a break down.
NOTE: the larger the transistor size, the cheaper the manufacturing process. The prices below are for 28NM however if you go 20NM or 16NM these prices will surely go up ten-fold

RTL:  50,000-100,000 USD depending on if you do it or if you hire someone (generally just a salary)
Backend:  300,000 USD
PCB Design: 10,000USD
Mask Making:  1.1M USD
Chip Cost per chip:(greatly varies on many things)
things that affect chip price are:
RAM TYPE:  ECC or Non-ECC
RAM AMOUNT
Number of Cores
Number of Layers
Generally speaking for a crypto chip the price per chip varies. For a SHA-256 chip there is generally no RAM in these chips. so these chips are cheaper. Now scrypt chips and others require alot of ram so they are more expensive. With that being said chips can range from $2-$15 each.

So you must pay for the number of batch1 chips

All in all the total cost comes to around 3M USD give or take..

Scenario 2: Buying Chips

Generally this can be a good idea because of cost savings, however, if will need to proper logistics in order to physically get those chips on a PCB and hook them up to a controller of some kind. If you have those avenues sorted already, it may be a good move. However, if not, it will quickly get hairy.

Hopefully this answered some questions
sr. member
Activity: 434
Merit: 250
September 03, 2014, 02:08:07 PM
#9

 Is there any trustworthy asic companies that can mass produce these?

http://www.easic.com/
http://www.umc.com/english/process/c.asp
http://www.esilicon.com/
http://www.xilinx.com/

 These are some examples which I am not sure , are there any trustworthy known sha/scrypt asic manufacturers out there?
newbie
Activity: 4
Merit: 0
September 02, 2014, 10:21:28 PM
#8

 How can we find those companies in china? Also I am sorry about my ignorance but what are the differences between chips? It could be easy for bitcoin but how hard it is for scrypt or even x11 which doesn't exist right now? Cheesy

I know at least two companies who developed bitcoins. If you are interest in bitcoin I can forward your emails to them.

The higher process ( 28nm is higher than 40nm), the cheaper each chip gets and with lower power.

At current bitcoin price, the miners don't make any money, and the asic verdors should make some money ( I guess).

As for  scrypt or  x11 , I can ask around if anyone is making it.

sr. member
Activity: 434
Merit: 250
September 02, 2014, 09:49:24 PM
#7

 How can we find those companies in china? Also I am sorry about my ignorance but what are the differences between chips? It could be easy for bitcoin but how hard it is for scrypt or even x11 which doesn't exist right now? Cheesy
newbie
Activity: 4
Merit: 0
September 02, 2014, 09:32:34 PM
#6
The reverse is almost impossible. it's just too many transisters now.

Bitcoin mining algorithm is already out there, it only take a few engineers to implement the code.
(one for coding, one for backend, one for software)

The big money is in the fab mask, it's about 5M for 28nm, 1.5M for 40nm, 0.5M for 65nm.

If you only want to prove your idea, get 3 engineers to finish the chip (around 120k), and do it in a
40nm MPW(~ 120k )which gives your around 100 chips .

So  the minimal money would be around 250k to prove your idea in 40nm, if your chip is small and low-power,
you can raise 6-10M to do it again in 28nm/20nm/14 nm.


There are a lot of company in china which studied the algorithm and implemented in RTL code.
it would be much easier to just select one of those.
sr. member
Activity: 434
Merit: 250
September 02, 2014, 08:26:22 PM
#5

 Yeah Cheesy I don't even have 1000$ invest right now let alone 10m or 5m , its just curiousity (ok maybe also show it to a rich friend who barely started bitcoin recently so there is a little bit of hope as well Cheesy )

 So what would one need to go through to find a whole lot of detail about this to build a solid business plan regarding becoming lets say the next king of X11 asic manufacturer ? X11 seems like the next run away point after scrypt also got hit with ASIC , so I am assuming the first company that makes asics for X11 would probably make a killing.

 Lets say I need all the info I can get to read up on to make a solid business plan that finds the 'cheapest' way to make X11 asic miners?
newbie
Activity: 54
Merit: 0
September 02, 2014, 08:22:12 PM
#4

 I would like to ask ( btw I have not even a single clue so don't hesitate to eli5 this)

 How much capital would one need to start a business and get asic chips at lowest price from a manufacturer or make one yourself?

 I am guessing (assuming?) getting couple of chips and reverse engineer it and make thousands of the same stuff would be costly if not impossible.

 So lets assume this scenario , I get 10-20 scrypt asic miner chips , pay someone to reverse engineer and find out how its made , go to a chip factory with him , explain the manufacturer how it should be made with him and pay them to make couple thousand at least. How much capital would this take?

 Or any other example you might have in your mind.

 TL;DR , how much would i need to have to become next asicminer,gawminer company.

1. Approx $10M from start to finish, but for something that is cryptocurrency related. It's closer to $5M if you have a team already selected and existing IP. This assumes a 28nm process target.
2. You can't reverse engineer ASICs anymore. There are 100 million transistors on there, and nobody is going to figure out what it's doing by looking at a mask layer. Besides, you don't need to reverse engineer anything. The algorithms are well-known already.
3. You can't make just thousands of ASICs. Volumes need to be in the 100k minimum, and it really needs to be in the millions for them to take you seriously. You can't just walk into a fab and expect them to make some for you anyways; you have to know what you are doing and have relationships with the fab companies. Most Bitcoin ASIC people don't have this BTW, and get really shafted on the  wafer prices.

Is this just idle speculation?

sr. member
Activity: 434
Merit: 250
September 02, 2014, 07:49:51 PM
#3

 So about 500k would suffice ? Also I always wonder how that pre-orders got so much trusted , I mean I am sure if I say right now give me 1k for a miner and I will deliver in 6 months people won't believe me Cheesy So they must have some power behind them already to prove they can deliver what they promise.
legendary
Activity: 1512
Merit: 1057
SpacePirate.io
September 02, 2014, 06:58:50 PM
#2

 I would like to ask ( btw I have not even a single clue so don't hesitate to eli5 this)

 How much capital would one need to start a business and get asic chips at lowest price from a manufacturer or make one yourself?

 I am guessing (assuming?) getting couple of chips and reverse engineer it and make thousands of the same stuff would be costly if not impossible.

 So lets assume this scenario , I get 10-20 scrypt asic miner chips , pay someone to reverse engineer and find out how its made , go to a chip factory with him , explain the manufacturer how it should be made with him and pay them to make couple thousand at least. How much capital would this take?

 Or any other example you might have in your mind.

 TL;DR , how much would i need to have to become next asicminer,gawminer company.

I created a component for vending machines and kiosks that used biometric scanners and a custom FPGA that had to be put into a custom chip package onto a board with surface mount components. The electrical engineer (who had 40+ years experience) ran about 80k a year to design the board, and a second guy to write firmware for about $45 a year, the initial development tools and packages ran about 50k, the production runs themselves were about $80k, with an inital order minimum of about $125k. Expect to spend a 6-9 months in development with another 2-3 months for QA and other post production work.  I'd say you would want to hire a team to design and build it for you that already has relationships with small run shops that can do contract manufacturing. Expect to spend $250-400k... now you know how pre-orders got started in the asic arms race Cheesy
Pages:
Jump to: