Why buy an expensive EM simulator?

[fekete]

This is an issue that bugs me for a long time, and I noticed that I am not the only one bothered by the high prices of EM solvers.
I am not asking why are they so expensive; there are reasons for that, all the promotions, free workshops, advertising in magazines, expensive 3D GUI, research etc.
What I would like to know is what feature or set of characteristics are important for you when deciding to buy an expensive program as opposed to the cheaper ones with a not so advanced GUI and fewer capabilities?
In my experience, I had a very hard time when I tried to convince my boss to buy one of the “state of the art” solvers with the price in the $40K - $50K range. I didn’t mention to him that there are others <$10K because I wanted:
- to easily build complicated 3D structures
- include realistic metal loss in simulations (although this could be done with the cheaper ones too)
But, to be honest, we could have bough a cheaper solver (even two different ones) and pay a decent salary to another designer for a few months with the difference.

So what are your reasons when deciding what solver to buy? How important is price?

 

[flatulent]

This is the same as the decision on any other tool. The usage present and future needs to be estimated and the total cost minimized. This cost includes labor of using the tool and the labor of making and testing prototype products and tweaking them to work properly.

 

[fekete]

Thank you flatulent
I don’t want to … deflate you, but I was looking for more concrete examples not general ideas.
Like somebody might needed to import layouts from Cadence APD so he/she decides to buy HFSS and Ansoft Links for example.
Some of the cheaper solvers have limited import/export features (from mechanical CAD also).
Or the example from my own experience when I needed to be able to geometrically model complicated features.

 

[eirp]

Do you wanna to use the software in company (for making moneys) or in university?

In case of university usage ask for discount/renting which could be say maximum 1/10 of your price per year.

When you work for company you have to calculate the time when invested money into 3D simulator will be back.

I agree with you that "state of art" one is good choice. Our time is too expensive to spent days even with simple tasks (say importing geometry). In fact, with your given conditions the choice is most probably CST or HFSS

good luck!

eirp

 

[fekete]

Thank you eirp
I am talking about commercial versions. The situation I described was from a few years ago, and we did buy MWS and I used HFSS when working with a different company.
Anyway, so you say that import options can be a factor in deciding to buy an expensive software (note that the difference can be over $30K)
This topic started actually from one about designing a coax 2 waveguide transitions where, somebody, as usual, recommended HFSS.


If designing simple transitions is the reason to buy an EM solver then you don’t need an expensive one. You could buy even two cheaper ones and also buy a few very powerful PCs, and still give yourself a raise for reducing expenses. My question is, when you must buy the expensive software because the others cannot solve your problem? (in reasonable time, with reasonable effort, giving reasonable accuracy…)
I guess people working in academia and big companies don’t care so much about this aspect but consultants and people working in smaller companies have to.

 

[eirp]

Hi, fekete

I said importing of different formats but you see that I meant how the soft is user friendly, how much time is needed to make a model. In my case I'm faced with different structure types so I have to prepare them for simulation fastly.

Depends what you're actually gonna to simulate, if you're designing coax-wg transitions for example, then you may need paper and pencil only

Can you please specify your question please, it's not very clear for me...

thx,
e.

Thank you eirp
I am talking about commercial versions. The situation I described was from a few years ago, and we did buy MWS and I used HFSS when working with a different company.
Anyway, so you say that import options can be a factor in deciding to buy an expensive software (note that the difference can be over $30K)
This topic started actually from one about designing a coax 2 waveguide transitions where, somebody, as usual, recommended HFSS.
h**p://

If designing simple transitions is the reason to buy an EM solver then you don’t need an expensive one. You could buy even two cheaper ones and also buy a few very powerful PCs, and still give yourself a raise for reducing expenses. My question is, when you must buy the expensive software because the others cannot solve your problem? (in reasonable time, with reasonable effort, giving reasonable accuracy…)
I guess people working in academia and big companies don’t care so much about this aspect but consultants and people working in smaller companies have to.

Added after 2 minutes:

Hi, fekete

I said importing of different formats but you see that I meant how the soft is user friendly, how much time is needed to make a model. In my case I'm faced with different structure types so I have to prepare them for simulation fastly.

Depends what you're actually gonna to simulate, if you're designing coax-wg transitions for example, then you may need paper and pencil only
Your inputs are not enough to tell more by now I think

Can you please specify your question please, it's not very clear for me...


thx,
e.

Thank you eirp
I am talking about commercial versions. The situation I described was from a few years ago, and we did buy MWS and I used HFSS when working with a different company.
Anyway, so you say that import options can be a factor in deciding to buy an expensive software (note that the difference can be over $30K)
This topic started actually from one about designing a coax 2 waveguide transitions where, somebody, as usual, recommended HFSS.
h**p://

If designing simple transitions is the reason to buy an EM solver then you don’t need an expensive one. You could buy even two cheaper ones and also buy a few very powerful PCs, and still give yourself a raise for reducing expenses. My question is, when you must buy the expensive software because the others cannot solve your problem? (in reasonable time, with reasonable effort, giving reasonable accuracy…)
I guess people working in academia and big companies don’t care so much about this aspect but consultants and people working in smaller companies have to.

 

[fekete]

Depends what you're actually gonna to simulate, if you're designing coax-wg transitions for example, then you may need paper and pencil only

Hi eirp,

Exactly, for certain designs you don’t need an expensive tool. And this is the essence of my enquiry, I would like to hear about the opposite: concrete situations when somebody had to buy the expensive one because there was no cheaper option.
If you browse through the forum you will see that there are many complicated situations described by people who know relatively well how to use a simulation tool. But there are also questions about very simple situations and the person knows almost nothing about the tool.
It seems that some companies spend a lot of money for an EM solver, but they only have to do simple designs and they hire somebody who has problems even with those simple cases while having such a powerful engine in their possession – isn’t that a waste of money?

Coming back, I see that you would pay much more for some software because of the friendly user interface that allows you to work fast. Ok, that is an important reason, that is why so many products are now based on the ACIS GUI (HFSS, MWS, Micro-Stripes …).
How about other cases like the necessity to have the option to impose periodic boundary conditions for antenna arrays …?
I have mentioned recently a situation where there is only one commercial solver available: simulation of metamaterials, and the solver is MEFiSTo (which is actually a low cost one).


Any other concrete examples?

 

[irfan1]

Sometime ago we had to decide between 2 codes. One is a commercial one and the other one is free. All we wanted to do is to calculate the photonic crystal bandstructure. This is a simple +/(+ -) optimization. We could have used the free code. It lacks a user friendly GUI. It lacks easy structure input. It lacks custumor support. But it is free. The major cost of the free software was "time". We give a try for both softwares. Now I am happy that we have decided on the commercial code.

 

[vfone]

Everything is related to the experience of the designer. I know brand new RF engineers, they are saying: cannot design a bloody 3 pole low pass filter if they don’t have the latest version of ADS2004…because this was the version they studied at university.
In my opinion 80% of the applications do not require very sophisticated software, or at least do not require the very latest versions.
Even for EM simulation I’ve seen a lot of freeware/shareware/time-limited or low-cost software’s that can do simple jobs. Anyway you have to spend some time in cut-and-paste tuning. I don’t think exist a designer who simulated for example an embedded antenna, sent direct to production plant, and he went patiently at home :flasingsmile:

 

[khouly]

vfone is right

RF engineer must be very flexible , and he must know how to use hand calculations , not design with software ,
design by hand frist , then verify and optimize with simulator

when i was practicing to design microstrip filter , i have done it with 2 or 3 clicks with @DS ,
but when i designed it manually i figured out that i have learned much more than using simulator directly

 

[fekete]

I agree with vfone and khouly.

I know brand new RF engineers, they are saying: cannot design a bloody 3 pole low pass filter if they don’t have the latest version of ADS2004…because this was the version they studied at university.
In my opinion 80% of the applications do not require very sophisticated software, or at least do not require the very latest versions.
Even for EM simulation I’ve seen a lot of freeware/shareware/time-limited or low-cost software’s that can do simple jobs.

This is the reason why I started the discussion. 90% of the postings recommend expensive solvers that I won’t mention again. They are very good products, no doubt, but I cannot believe that I am the only one that has a limited budget available and has to be careful with the spending. In fact, some post are looking for free software, some are writing their own codes etc. In between these and the lucky guys who can afford to spend the price of a Mercedes on one piece of software, there must be an army of “cost conscious” users. Sure I would like to drive a luxury car, but I have other expenses too and I have to divide my budget.
I am looking for concrete examples, in both ways – companies that can do well with a low cost products (like if they only design simple waveguide filters, or only planar antennas etc.) and real life examples when they had to go with the luxury CAD product.

RF engineer must be very flexible , and he must know how to use hand calculations , not design with software ,
design by hand frist , then verify and optimize with simulator

when i was practicing to design microstrip filter , i have done it with 2 or 3 clicks with @DS ,
but when i designed it manually i figured out that i have learned much more than using simulator directly

That is another interesting issue, managers should appreciate better good designers; and if the designer takes the effort to do the job with a cheaper software, some of that price difference should go to the designer …

 

[plasma]

Hi feket

The rich become richer and the poor stay poor.
If you look for free codes you will be limited and stay with simple
Products. If you will invest wisely in good code you will
be in the front of technology. you will be able do much more advanced things
Then you can do now. It will help you to decrease costs and make more.

Think about it. In term of new expensive technology like the
Computers in the beginning they were so expensive see what it did to the
Computer users and were the people that don't use computers only use
"Hand calculations"
Because you are using your mind you need the best tool to do the best
in your profession

THX
PL

 

[fekete]

The rich become richer and the poor stay poor.

It is true in general, but not always. Many of the big companies of today started low.
Money is important factor but if you are smart in your field and also with the way you spend, you can overcome the disadvantage of a lower budget.

If you will invest wisely in good code you will be in the front of technology …

Yes, invest wisely; ie do not always buy the most expensive code if you don’t need that. There are good codes at lower prices, depends on the applications.

I compiled a list of approximate prices for a perpetual commercial license with most of the capabilities included (the one year licenses are usually ~ one third of that):

HFSS, MWS, Micro-Stripes - (FEM, FDTD, TLM) $40-50K

Ansys, Analyst (FEM) $30K

Fidelity, XFDTD (FDTD) $20-25K

Gmacs, EMC Studio (3D MOM) $20-25K

IE3D, Sonnet (2.5 D MOM) $15-20K

MEFiSTo (3D TLM) $10K

EMS3D (2.5D MOM) $8K

These are prices on the US market. I might be wrong, please correct me if you have fresh information. You can see anyway, that there are some important price differences.

And there are many other EM simulation products that I don’t have pricing information on, like:

Empire, SEMCAD, QW-3D, Concerto (FDTD)
FullWave Infolytica (FEM?)
Feko, Singula, Wipl-D, PAM-CEM (3D MOM)
WASP-NET (Hybrid)
Aseris (FDTD, MOM, etc.)
If you have pricing information on these and others, please enlighten us.

Then there are the system simulation tools with a planar EM solver included:
Aplac, AWR Microwave Office, ADS, Ansoft Designer, CST DS, Genesys


So, again, if you invest wisely you will notice that some of the lower priced products might be right for you, this way the smart poor can become richer.

 

[XuQing]

fekete, u are right. We don't need expensive EDA software to do anything.

But I think the cheaper one maybe not so popular like the expensive one. If this is right, whether we need buy several cheaper softwares instead of one expensive software.

 

[irfan1]

If you ask me, I would say that the biggest cost is time, not money. Poorly written software may give you a lot of headache and cost you incredible loss of time. Note that I am not on the side of expensive software but on the side of well coded, well documented software. On the other hand, quality comes with a cost. I am not aware of any quality software that's reasonably cheap. Well, for some simple tasks, simple geometries, simple materials and simple boundary conditions it may be reasonable to use a low cost or even a free EM simulator.

 

[fekete]

Every tool is as good as the person using it. I remember of a semi-joke: A big expensive machine stopped working and no one from the big company could fix it with all the fancy computerized monitoring systems. They called a consultant – he arrived and his only tool was a hammer. He analyzed carefully the machine, listened to the noises and then went to a certain part and hit the machine with the hammer. The story goes that the machine started working. When he asked for $2K they said: what for one hammer hit?
- No, the hit was 5 bucks, $1995 for knowing where to hit. (Or something like that …)

 

[vfone]

"To invent, you need a good imagination and a pile of junk"
Thomas Alva Edison

 

[fekete]

Nihao XuQing,

We don't need expensive EDA software to do anything

? I never said that. In fact I gave a few examples at the beginning for when you have to go with the more expensive software.
Until now the reasons I collected from answers would be:

1. Advanced graphical preprocessing capabilities -> reduce setup time
2. Capacity to import geometry from layout or mechanical CAD software -> reduce setup time
3. Good support and documentation -> reduce learning time

Nobody seems to care if the software has some advanced EM features included. All the EDA software companies have to do is to provide a nice user interface – graphical and human.

But I think the cheaper one maybe not so popular like the expensive one

The more expensive ones have bigger marketing budgets, so everybody gets to know of them.
They also have very low academic prices, this way many graduates will choose the software they already know from school. Sometimes it seems more like a fashion issue. But fashion changes …

irfan1

Poorly written software may give you a lot of headache and cost you incredible loss of time

Poorly written software?!! This is an insult to the hundreds of researchers and engineers behind the products that I have listed. Some might not have the fanciest GUI (which, by the way it is not too much “written” by Ansoft or CST, they bought the core of it) but they are well written codes with the results of years of research included. None of the products listed above are poorly written software!
A few examples:

XFDTD use to have a not so advanced user interface for a long time. However, behind this product stands Raymond J. Luebbers, who is a well-known FDTD expert. I think he wrote the first FDTD book (among other things) from which some of the people working for the competitors learned a few things. I don’t think somebody of his caliber would be involved in “poorly written software”.

Sonnet, while a not so expensive product is one of the first commercial EM software on the market. Behind Sonnet stands James Rautio (who worked with R. Harrington) and he is a well-known MOM expert. He is writing EM software “poorly” since 1983.

MEFiSTo does not have an ACIS user interface but an in-house developed 3D one. Behind MEFiSTo stands Prof. W. Hoefer, again a well-known specialist with numerous books and papers on TLM and other methods. Surely someone of his scientific stature would not supervise over poorly written software.

EM3DS while being the product with the lowest price on that list is a well-known tool in the RF MEMS field and not only. It is also the only commercial product using volumetric MOM.
Even in this board it was recognized that it is the best RF MEMs tool

in fact, EM3DS is included in other MEMS products as the EM solver, like in Convertorware

Empire like XFDTD used to have a not so nice user interface. However, it is backed by reputable researchers from Duisburg University like Prof. Ingo Wolf.

SEMCAD – don’t know to much about it, however in a recent post

it seems that this is a software with certain advantages when used for modeling EM – human body interaction


And so on! There is no poorly written software on that list!
In one of your previous messages you were complaining about noisy advertising for a certain EM software product (Microstripes):


But how about people who want to discredit products other than their favorites?
Since you are a researcher, I would have expected a little more respect towards the hard work of other researchers.

I am not aware of any quality software that's reasonably cheap. Well, for some simple tasks, simple geometries, simple materials and simple boundary conditions it may be reasonable to use a low cost or even a free EM simulator.

All of the product listed above are quality software and can do much more than that. It is very easy to launch unfounded innuendos. How are you better than the loudmouth Microstripes guy? :sm31:

 

[irfan1]

I don't remember addressing any of the software that's on your list and I don't remember insulting any researchers. Surely there is poorly coded and poorly documented software even among very expensive software. If you read my post throughly, you'll see what I am against. I am against poorly coded and poorly documented software. I hope you will be more carefull next time when you accuse someone. And I am expecting an apology from you!

 

[fekete]

Your comment appeared almost immediately after my message where I posted the list of software products. I should believe that there is absolutely no connection? You are posting without reading the previous messages? I think you are the one who should be more careful.

Instead of saying that maybe you used the wrong term, you are insulting them again by using the term poorly coded. No software is perfect but to accuse so many hard working people of doing poor coding is something that deserves a harsh response.