Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Can IBIS model simulate a full chip?

Status
Not open for further replies.

ultracalibur

Newbie level 3
Joined
May 28, 2010
Messages
3
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,283
Activity points
1,307
Can IBIS model simulates a full chip?

Hello,

I am simulating a full PCB. But the information of some chips are only provided in IBIS format.

Therefore, I need to simulate the full chip in order to find all the signals in the PCB. According to how IBIS models are generated, as long as I have the parameters for the IBIS input buffer model and IBIS output buffer model, and with the logical diagram for the chip, can I simulate the full chip in the PCB?

Thanks.
 
Last edited:

matrixofdynamism

Advanced Member level 2
Joined
Apr 17, 2011
Messages
550
Helped
24
Reputation
48
Reaction score
23
Trophy points
1,298
Activity points
7,216
What do you mean by simulate full chip? Do you mean, simulate the internal parts like the logic inside?
The purpose of IBIS models is to provide a way to simulate the IO buffers of a device. It is an alternative to SPICE models which have certain drawbacks, among them being that they expose information that may otherwise be proprietery on how the buffers are implemented by a manufacturer. Why don't you post this question on eevblog and electronics.stackexchange also?
 

tggzzz

Advanced Member level 4
Joined
Jan 7, 2015
Messages
114
Helped
17
Reputation
34
Reaction score
16
Trophy points
18
Activity points
807
You are confusing different levels of abstraction, where each level needs a different tool.

IBIS is used to simulate the signal integrity issues, i.e. the analogue behaviour of tracks and packages with signals that will (typically) be interpreted as digital signals. Suitable tools inclde spice and other analogue simulators.

Once the signals have appropriate levels of integrity and will be correctly interpreted, you switch to simulating in the digital domain, using your HLL of choice, typically Verilog or VHDL.

If you try to mix abstraction levels, you will eventually end up trying to boot an o/s kernel using a spoice simulator. The only people that might try that are CPU designers, e.g. Intel.
 

ultracalibur

Newbie level 3
Joined
May 28, 2010
Messages
3
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,283
Activity points
1,307
Thanks for the reply.

The problem I am trying to solve is a PCB on a platform (like aeroplane or missile), and the chip on PCB controls the aeroplane/missile. Now an electromagnetic wave is hitting on that platform. I use electromagnetic code to solve the voltage change on the input of the chip on the PCB.

To understand how the wave affects the aeroplane or missile, I must know the response change at the output of the chip. The problem is, I only have the IBIS models of the chip. So is there a way to do the simulation only with the IBIS models?
 

tggzzz

Advanced Member level 4
Joined
Jan 7, 2015
Messages
114
Helped
17
Reputation
34
Reaction score
16
Trophy points
18
Activity points
807
There are too many variables to answer that question, e.g. frequencies, power, polarisation, angle of arrival, PCB track geometry, time of arrival, and many more.

Your approach should be via "electromagnetic compatability" (EMC), not digital/analogue simulation.
 

ultracalibur

Newbie level 3
Joined
May 28, 2010
Messages
3
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,283
Activity points
1,307
There are too many variables to answer that question, e.g. frequencies, power, polarisation, angle of arrival, PCB track geometry, time of arrival, and many more.

Your approach should be via "electromagnetic compatability" (EMC), not digital/analogue simulation.

There is no need to separate EMC and circuit simulation. It is a hybrid simulation of wave physics and circuit physics. All the things you mentioned, frequency, power, phase, geometry, time delay could be dealt with properly in electromagnetics. I can decompose the problem into electromagetic part and circuit part, simulate them separately and coupled them in a domain tearing and connecting method. Nevertheless, what I am concern about is the proper simulation of the chip.
 

tggzzz

Advanced Member level 4
Joined
Jan 7, 2015
Messages
114
Helped
17
Reputation
34
Reaction score
16
Trophy points
18
Activity points
807
There is no need to separate EMC and circuit simulation. It is a hybrid simulation of wave physics and circuit physics. All the things you mentioned, frequency, power, phase, geometry, time delay could be dealt with properly in electromagnetics. I can decompose the problem into electromagetic part and circuit part, simulate them separately and coupled them in a domain tearing and connecting method.
Go on and do it. Tell us your results when you are finished.

Nevertheless, what I am concern about is the proper simulation of the chip.
Lookup and understand the concept of "combinatorial explosion", particularly when digital circuits are involved.
 

Status
Not open for further replies.
Toggle Sidebar

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top