LCA TEJAS MK1 & MK1A: News and Discussion

[Dark Sorrow]

@Dark Sorrow Good information. Thanks for education. Do you know if any indigenous effort has been made on developing our own implementation?

I know C-DAC is working on developing implementation for RISC-V ISA.
IIT-Madras is leading an effort to SHAKTI Processor Program.
IIT Bombay is also having similar effort.
I believe DRDO has either developed or purchased implementation for an ISA to be used in seeker, radars, missile's guidance computer etc.

 

[Dark Sorrow]

@Dark Sorrow Good information. Thanks for education. Do you know if any indigenous effort has been made on developing our own implementation?

II would like to add developing or buying implementation is just 50% of battle. Having software support is key to success or failure of chips.
The reason why ARM or x86 chip are successful is because of excellent software support. Without off-the-shelf (i.e. battery included) sofwtare support the chip will fail.

 

[silverghost]

@Dark Sorrow Good information. Thanks for education. Do you know if any indigenous effort has been made on developing our own implementation?

No need to design one from scratch. Join the arm bandwagon & create a chip that is optimized for the aerospace industry. Once we have the design in place there are fabs that can manufacture the chip for us. This will be cost-effective to get our own chip. Most of the people who specialize in this area are of Indian origin. You will have to pay them the equivalent of their US emoluments but it's all doable.

 

[Rajaraja Chola]

What crap! India is by itself a chip design powerhouse. 1 in 5 of the world's best chip designers right here.

Semiconductor Chip Designing and Manufacturing

India is one of the leading centers for global semiconductor design companies. Most global semicondu

www.pib.gov.in

I hope they go all out and try their level best to win it. Be it Argentina or Malaysia or Egypt or elsewhere. Each and every sale means influence, and can lead to a deluge of sales down the road. Provide great service and respond to foreign customer feedback promptly. Indian weapons should have a reputation for being relatively cheap, yet high on quality, reliability and sophistication. Otherwise the present movers and shakers like US & EU have the expensive end covered, while Russia and China corner the not so rich markets.

Look. I know what I am talking about. Chip design is different than designs based on architecture, testing, assembly and packaging.

Yes, India has some of the best R&D centres for chip design based in Bangalore and Chennai almost for all major and mini companies. In fact we have even home grown chip design companies that design and supply to bigwigs. I think in 2020, one chip design Chennai start up sold itself to Reliance Ambani group. So yes this is not to brag, just to imply I have work experience and knowledge in this field though not directly.

90% of chip designs are based on ARM architecture and the remaining on RISC V one. Even Intel designs it’s chip on ARM architecture. The reason being, the manufacturing, testing and packing like has been designed around ARK architecture to save cost and footprint as this field is very very expensive. RISC v is a new one upcoming architecture is open design one. Very few companies use this architecture and even I think IIT Madras Sakthi Processor is based on RISC V. Manufacturing infra for this architecture is even less. Chip is designed around an architecture just like how an fighter is designed around an engine.

 

[Rajaraja Chola]

Please provide a good source for your claim.

I said it’s possible. I read about it sometime long back that some components of the engine are procured from UK from GE UK. The IP still belongs to GE but UK rarely has used that as an sanction point. I do not have time to dig out the source.

 

[MonaLazy]

15 LRUs that are sourced from UK

Indigenous content of Tejas 59.7% by value & 75.5% by numbers

In a written reply to Lok Sabha, Minister of State for Defence Subhash Bhamre said that several parts of Tejas is of foreign origin.

economictimes.indiatimes.com

As of July '18

Out of total 344 LRUs (Line Replaceable Units) involved in the system level of LCA, a total of 210 LRUs have been produced indigenously and balance 134 LRUs are being imported from foreign companies, he said.

"Indigenous content of the LCA Tejas Aircraft is 59.7% by value and 75.5% by numbers," he added.

To reduce dependence on foreign companies, out of the 134 imported LRUs, the indigenisation of 42 units is under progress.

Further efforts are being made by way of indigenisation with the objective to reduce the import content progressively, he said.

Consolidation of technical specification of imported units have been carried out and they are being projected for indigenisation by private companies, DRDO labs and HAL divisions.

15 out of 344 is 4.36% which is not too much. We need to come up with replacement boards for those only.

@Dark Sorrow thanks for that explanation but is the discussion of chip fab or even design relevant to this? These could just be interfaces between pilot/DFCC and the sensor/actuator. For eg if we have to control the landing gear- why do we need to design and fab our own chip for it? Can't we use any available open architecture controllers and burn the controlling software into its ROM just like on a PC motherboard? An LRU is just the fusion of software algorithms & hardware on a single card/board to process I/O from embedded sensors? In any case, we use the Ada language for software creation- which means if a toolchain (editor/IDE/compiler/linker/debugger/simulator) exists for that architecture we should be ok and can code any new chip even with a different architecture so long as it has a toolchain for a high-level language like Ada. Isn't the fact that indigenization of imported LRUs is a task directed at the private sector an indicator that this is a solved problem, only requiring some effort but no major weight lifting in R & D?

 

[Dark Sorrow]

Indigenous content of Tejas 59.7% by value & 75.5% by numbers

In a written reply to Lok Sabha, Minister of State for Defence Subhash Bhamre said that several parts of Tejas is of foreign origin.

economictimes.indiatimes.com

As of July '18



15 out of 344 is 4.36% which is not too much. We need to come up with replacement boards for those only.

@Dark Sorrow thanks for that explanation but is the discussion of chip fab or even design relevant to this? These could just be interfaces between pilot/DFCC and the sensor/actuator. For eg if we have to control the landing gear- why do we need to design and fab our own chip for it?

The point is that countries holding patent or IP can block the semiconductor access to you; as done by west against Russia. In that case the LRU manufacturers also won't sell their LRU, semiconductor vendors will not sell you their chips and this becomes a contiguous problem.
You need to consider this with regard to sell of Tejas to Argentina and UK's weapons embargo. Is the deal worth it to strain our relations with UK and endanger the entire Tejas and other programs.
I am not saying we should design everything ourselves but just acknowledging we are dependent on western counters.
Even US imports subsystems or material from other countries (including PRC).

Can't we use any available open architecture controllers and burn the controlling software into its ROM just like on a PC motherboard? An LRU is just the fusion of software algorithms & hardware on a single card/board to process I/O from embedded sensors? In any case, we use the Ada language for software creation- which means if a toolchain (editor/IDE/compiler/linker/debugger/simulator) exists for that architecture we should be ok and can code any new chip even with a different architecture so long as it has a toolchain for a high-level language like Ada.

On paper one might think we can replace open sourced controllers but that is not true. ISA is just one aspect of a processor. Their are lot of other aspects in a CPU like

1. Different type of peripherals with SoC
2. Memory Controllers
3. Caches
4. System Buses
5. Concurrency support
6. Security features and fault detection
7. Advanced computing hardware support (e.g. SMID, DSP, ML, etc)

and many more
None of these subsystems are open-sourced.

You will also not find drop in replacement of such chips due to their confidentiality and being trade secret.
Another problem you will find is your OS vendor and chip designer have to work in tandem to get system running with new chip this swill be an issue.
Tooling is provided by chip developer hence should not be an issue unless its an indigenous design them we will even have to work on tooling. Interfacing with OS will be an issue and will need to involve OS developer.
In-short it will require massive software redesigning and re-validation effort that everyone wants to avoid i.e. we are vendor locked.

Isn't the fact that indigenization of imported LRUs is a task directed at the private sector an indicator that this is a solved problem, only requiring some effort but no major weight lifting in R & D?

True but if you look from private sector prospective, the orders for such LRUs is very low compared to the risk
private sector will undertake along with capital expenditure. Even after successful complication their is no guarantee that their product will be accepted and even if accented limited numbers, hence the lack of R&D and manufacturing effort.

 

[Blademaster]

True but if you look from private sector prospective, the orders for such LRUs is very low compared to the risk
private sector will undertake along with capital expenditure. Even after successful complication their is no guarantee that their product will be accepted and even if accented limited numbers, hence the lack of R&D and manufacturing effort.

If it comes to the point where we can't even export our weapons abroad because of some foreign IP, there will come a point where we will say fuck the costs we need to export the weapons to offset the R&D and the manufacturing costs, then at that point we will come up with every indigenous product needed to export the weapons without any foreign objection.

 

[karn]

@Dark Sorrow What is used by ISRO ? And what do we use for our nuclear missiles? I'm sure no vendor is sending anything for a nuclear program.

 

[Dark Sorrow]

@Dark Sorrow What is used by ISRO ? And what do we use for our nuclear missiles? I'm sure no vendor is sending anything for a nuclear program.

ISRO uses commercially available MPUs and MCUs.
ISRO is purely a civilian organization and GoI has made a point to keep ISRO a civilian organization for this particular reason so that it doesn't face unnecessary sanctions and can focus on its core field and objectives
As for what is used in nuclear missiles - I don't know. Its a closely guarded secret.
My guess would be earlier we used Altera FPGA and now probably we use in-house developed and manufacturer ASICs, hence long development time..
Interesting fact - We have started using in-house developed ASICs for our conventional missile and smart munition program like Astra BVRAAM, Akash NG, etc. Hence long developing cycle and testing cycle and achieving 95-100% indigenous content.

You are right most vendors will not be willing to sell material that goes in nuclear program (it is a PR nightmare).

 

[Dark Sorrow]

If it comes to the point where we can't even export our weapons abroad because of some foreign IP, there will come a point where we will say fuck the costs we need to export the weapons to offset the R&D and the manufacturing costs, then at that point we will come up with every indigenous product needed to export the weapons without any foreign objection.

That the thing unless the state is a pariah their is no weapon sanction on them.
Argentina is an exception.
Most countries that we can export to (Malaysia, Egypt, ME, Colombia, etc.) won't have such problems.
However I agree with your instincts. We need to have a local aerospace Eco-system.

 

[MonaLazy]

The point is that countries holding patent or IP can block the semiconductor access to you; as done by west against Russia.

If we design the chip here in Bengaluru and get it fabricated in Taiwan or China then who owns the IP?

In that case the LRU manufacturers also won't sell their LRU, semiconductor vendors will not sell you their chips and this becomes a contiguous problem.

That's nothing out of the ordinary- it's called vendor lock in and is the oldest trick of the trade. That's why open standards exist in the first place. I am sure once the size constraints are known we can re-create the LRUs on our own- chips/PCBs/sensors/actuators. That's how Chinese have been doing their fighters for a while now. This is also the direction of HAL's effort in locating private players that can deliver.

You need to consider this with regard to sell of Tejas to Argentina and UK's weapons embargo. Is the deal worth it to strain our relations with UK and endanger the entire Tejas and other programs.

If geography can tell a tale- it is easy to see who has the more legit claims to Islas Malvinas

It does look like UK has no business being there.

Also, when was the last time UK deferred on anything to accomodate our sensibilities? Like nurturing Khalistanis or debating human rights and other assorted issues in their parliament. What is their stand on Kashmir?

Bottom line- sand box our relationship with each country, as India we don't have to care for UK's sensibilities- leave it to their gov and they are doing a pretty splendid job of it by sanctioning Argentina.

Also, if they have embargoed sales to Argentina we are infact playing by their rules by finding replacements for even the last British component on the plane. UK is saying we wont sell to Argentina. We sell nothing made in UK to Argentina.

Even if the volumes are low, the price per unit will be high enough to justify investment otherwise it wont work.

 

[Dark Sorrow]

If we design the chip here in Bengaluru and get it fabricated in Taiwan or China then who owns the IP?

IC design is not limited to Bengaluru. It is spread evenly along major Indian cities like Pune, Hyderabad, Noida, Mumbai, etc.
Second the answer is very complicated. Even companies like AMD, Intel, Apple, Qualcomm, etc. down own 100% IP for their chips. Some IP or other is generally procured from third parties.
e.g. AMD uses ARM-core (IP) to implement security features, boot-loading, etc. AMD uses IP for Cadence system for audio processor inside their CPU.

People generally use 3rd party IP for sub-systems, peripherals and advanced computing instruction as this provides cost, energy and time effective way to field your chips.

It depends on configuration of chip, but generally IP is owned several different companies and the companies designing and selling chip pays royalties to each individual owner.

On a positive note India has lot of such soft IP providers.

 

[Corvus Splendens]

Korean p**sy was selected

 

[NutCracker]

Korean p**sy was selected
View attachment 171196

Ka-50 , 933/18 = 51M USD ..

Weird to pay that much for a trainer shit.

 

[vishnugupt]

Korean p**sy was selected
View attachment 171196

As predicted

Defense deals are often not only contain mere weapons but a lot more along with it, Defence deals are either thrust down the throat or done to buy an overall security umbrella.

Malaysia like any other Mulla country has the ambition to become the guardian of Ummah so they can't ignore big players like America or Russia because ignoring super-power will be severely detrimental to their dream. ( Because They can't keep quiet without uttering Kashmir )

 

[Super Flanker]

If true then this is another + for Tejas, LCA Tejas offer is cheaper than KAI FA-50 and fulfills all the technical requirements of the RMAF. Lets see what is the final decision of the Malaysians.

 

[Corvus Splendens]

Ka-50 , 933/18 = 51M USD ..

Weird to pay that much for a trainer shit.

 

[jai jaganath]

If true then this is another + for Tejas, LCA Tejas offer is cheaper than KAI FA-50 and fulfills all the technical requirements of the RMAF. Lets see what is the final decision of the Malaysians.

Sorry in very light sense
Hal ka klpd ho gaya
Itna advertise kiya, palm oil keliye mou kiya, cmd media me chilaya
Par fir bhi har gaye
Atleast Koreans were silent and were working rigorously and showed who is more powerful when it comes to dealing and forcing own terms on other
We must learn from them
Now we have to handle humiliation in sm also
Kya bakchodi life hai
Again sorry for writing in this language

 

[jai jaganath]

As predicted

If we can't win this deal where each and every thing was on our side then forget about selling to other nations where competition is complex
This deal was very near to our mouth but Koreans snatched it like crows very tactically
I think it's better to focus in building for ourselves and leave export for a decade atleast let's build the capability to negotiate on our terms
Humse na ho payga
Anyway they tried hard
They must be appreciated