Flotherm vs FloEFD
 

Hello everybody

I would be grateful if anybody could clarify to me. I want know that different of Flotherm and FloEFD.

1. Is it OK? Use the FloEFD?

Test sample is semiconductor power module or set(include fluent analysis). Main analysis is thermal.

2. I know that merit of Flotherm is interlocks with T3ster & Fast speed. Is it right?

That's all

I don't speak English well, thank you for asking question.

Please, reply.

use FloEFD, in my opinion Flotherm doesn't have much future and will eventually die.

thank you for your opinion CFDfan. I think you know the FloEFD well. So could you answer my question?



1. Why think about Flotherm wil die?



2. Is it possible to accurately simulate epoxy molding with FloFED?

I worried about what to pruchase flotherm and floefd.

I look forward your reply. thanks & regard.

thank you for the inside info....now I have to run and learn FloEFD. :rolleyes:

Hi, try reading this post:
https://www.cfd-online.com/Forums/fl...vs-floefd.html

Best regards, Gabriel

1. I don't know if Floefd can simulate epoxy molding (I just don't know what is involved there), but Boris, if he reads this email, will most probably answer you.

2. Flotherm, in my opinion will slowly die, because 3-4 years ago Mentor Graphics (being then the owner of Flotherm and Floefd) released a program called FlothermXT. It was done I suppose to overcome the drawbacks of the outdated Flotherm and combine the best features of Flotherm and Floefd. It turned out that the solver and the mesher in this hybrid program, i.e the heart and the head in any CFD code, were taken from Floefd.

Hi everyone,

Sorry, I've been so busy lately with our FloEFD conference as well as with traveling.

No, FloTHERM will not die. There is a clear difference between FloEFD and FloTHERM.
Sure, both are CFD codes and even if FloEFD would have all the capabilities of FloTHERM, the typical user is still different plus the speed of FloTHERM especially for cartesian conform geometry such as server racks or PCBs with chips etc. Most of those parts and assemblies are square shaped but even curved geometry can be handled. FloEFD is simple the CAD embedded CFD solution for the CAD engineers which most FloTHERM users are not. There are a huge amount of companies using FloTHERM for such applications where they either don't want to use a CAD system as you don't design chips or PCBs in a MCAD system.

1. Yes, FloEFD can be used for power module simulations. You will need to model the module as a 3D geometry just like in FloTHERM and apply the corresponding materials.

2. Yes, FloTHERM is very fast as it does not have so many solver capabilities that can slow it down as well as the SmartCells which enable the whole use of complex shaped geometry with an automated meshing and accuracy. FloTHERM XT is a little skimmed down FloEFD solver in terms of capabilities as you don't need cavitation or hypersonic in electronics cooling and therefore a little faster I think as FloEFD, but I have no exact numbers as I never compared them one by one. In about a month there will be more news on FloEFD.

3. It depends on what you mean by epoxy molding? Do you mean the filling of the epoxy over a chip or part of a PCB to cover a chip in an epoxy protecting cover?
FloEFD can simulate the thermal behavior of an epoxy mold such as the heat conducted through it away from the chip into the environment or other parts of the assembly. But FloEFD can also model the injection filling of an epoxy mold and how it flows around the chip for example. V17 comes with free surface capability and for the epoxy flow you would need the Advanced module as it is a non-newtonian flow into an air cavity. Other fluids in free surface can be modeled with the basic FloEFD version, just the non-newtonian fluid requires the advanced module.

So, in the end, it depends on your preference as well as if the tool meets your requirements in terms of features or usability.
If you are working within Creo, Catia, NX or Solid Edge 90% of your time just the application is some electronics device then most likely FloEFD is the better solution for you. You are already familiar with the CAD and can leverage the parametric in optimization studies and an IGBT etc. are mostly just a heat source anyway, so no detailed PCB traces of more detailed chip modeling, although FloEFD with the EDA bridge can do that also pretty good.
It is best to test both and see how you like the workflow and how they fit into your processes and of course, you can also compare speed and accuracy. But consider every aspect of it. So if you measure speed, then also measure the time you need to create the model, setup the simulation, mesh the model, solve it and post process it. Sometimes the model creation might take longer than the solving difference is for example.
This, in general, should be considered in any CFD code comparison. Whatever you buy, it should suit YOUR needs not the one of the sales person or because someone else says so. You have to work with it in future, day in and day out, and need to be able to use it efficiently.

Hope this helps,
Boris

Thank you. I look forward to your reply. I think that you must be an export.
After read your reply, I can select which simulation to use.

So I have 2 question again.

1. As you say I want see the mold flow that how the mold is filled in the cavity.
But our agent say that It is difficult to get the correct data because FloEFD can not put density or viscosity of mold in floEFD. Is it right??

2. What accessories in FloEFD should I purchase to enable mold flow simulations?

Thank you and look forward reply.

Hi Kurt,

Thank you, yes, I've been using FloEFD for over 10 years now.
Basically correct. It depends on what you want to simulate and which product fits you and your processes best. So FloTHERM and FloTHERM XT are electronics cooling products and cannot do free surface flow of filling a cavity but FloEFD can do that.

So regarding you two questions:

1. Yes, FloEFD can do the mold filling. Its free surface capabilities are working with 2 fluids (air - water, water - oil, air - non-newtonian). The fluids can be entered with all the usual properties for FloEFD fluids (Liquids: density, viscosity, specific heat, thermal conductivity and non-newtonian fluids: density, specific heat, thermal conductivity and multiple models for the viscosity such as Power-law, Herschel-Bulkley, Carreau, Cross-WLF, Second order and viscosity table). So I think more than enough data to provide a detailed material model.
Of course, you will need to have that material property but if you don't have that then often other tools won't be able to help you with that either. Some laboratories can help you to measure the non-newtonian properties in case you don't have them and don't have such capabilities to measure them.

2. For epoxy mold injection you will need the free surface for non-newtonian fluids and that is in the Advanced Module of FloEFD. An add-on module to the basic FloEFD version for your preferred CAD system.
So you will need your CAD integration of FloEFD plus the Advanced Module.


Don't trust agencies, trust the software vendor. Some external company does not know what the software is capable. A distributor might not be 100% up to speed with the latest features but they usually check with the vendor directly to clarify if that is possible. But the mentioned information from your agency does not seem to come from someone who knows FloEFD as this is basic knowledge for any application engineer using FloEFD.
I agree that non-newtonian material properties, especially for the viscosity, are not so easy to get as they are temperature dependent and not as commonly provided by the fluid vendor like for example oil or any Newtonian liquid such as water where you can find it in handbooks. For non-newtonian fluids, you usually need to conduct rheology measurements to determine the viscosity and it changes depending on the mixture of the material and temperature and the shear rate.

Hope this helps,
Boris

Thank you very much your quick reply too.



The purpose of the simulation is to simulate thermal analysis of IPM for a semiconductor module. If possible, I would like to see the mold flow of IPM.


So thanks to boris it became clear which accessary to put in. We decided to buy FloEFD.



Thank you again and I would appreciate your reply.



Be careful not to catch a cold.

Kurt

Hi Kurt,

happy to help. I might not always find the time to look into the forum but there are some people here like CFDfan and Gabriel_C that post here as well every now and then. Otherwise, shoot me a PM to get my attention.
Once you are under support with Mentor, their support team is happy to help with any questions you might have.

Happy holidays,
Boris

Yes I used to work with SWFS (the solidworks embedded version of flowefd) for a long time. I don't have experience in your area of expertise but I could tell you that flotherm is not suitable for you. It was designed explicitly for electronic cooling applications. The reason I think flotherm will eventually die is because the user interface is archaic and they (the developers) tried to modernize it with Flotherm XT (which is embedded in Solidworks version of Flotherm). Later I learned that the solver of Flothrem XT was taken from Floefed. Thus, there is nothing essential left from Flotherm in its future versions (Flotherm XT). This makes me think that the only thing keeping Flotherm alive is its customer base and the revenue coming from it.

I'm sorry, but this is not true.

Flotherm XT was developed to enable complex geometry handling better and getting closer to the needs of importing CAD models such as housings etc. Flotherm is capable of that as well but it is ideal for square shapes models as they exist in server racks and most typical PCBs.
The trend in consumer goods and electronics to more stylish shapes designs required a closer connection to the CAD design tools.
The benefit of having such a close link with Simcenter FLOEFD (formerly named FloEFD) already and the already ideally tuned meshing and solver technology provided the ideal backbone. Similar to what Audi, Porsche and VW do with their modular body kits. This doesn't mean that there is nothing in a Porsche Chayenne anymore because it shares the platform of the VW Touareg and Audi Q7.
The core electronics cooling knowledge from Flotherm is a major part in Flotherm XT, meshing and solver technology won't make up for the knowledge, they are just the tools to make it work. Just like every electronics designer has a lot of knowledge but without the EDA tools he will have a hard time getting anything done in the detail of modern electronics.

Not everyone might need or want Flotherm but rather prefer Flotherm XT and others want to work in their native CAD system and therefore prefer FloEFD. All three userbases are growing.

Regards,
Boris

Boris, working for the Simcenter you know all this much better. What I wrote is my personal opinion, but honestly what is really the core electronics cooling knowledge Flotherm has but say Floefd or other commercial codes in this field don't have. And I am asking about something major. Also Flotherm XT, as I understood, is linked to particular version of SolidWorks and can't read SW files created with later versions which in my opinion is a big drawback. All this makes me (personally) think that Flotherm doesn't have much future
Finally, our company is currently using scStream, that is also well suited for electronic cooling applications, and having dealt over the years with Cfdesign, Floworks (Floefed) and scStream, my personal opinion is that scStream is at present the best CFD code for electronic cooling, especially when very complex geometries are involved.

I know, that's why I want to correct it.
When looking at the marketing of a product you can clearly see which products are actually put on maintenance as it would be a waste of resources to promote it if your intention is not to actively sell it anymore.
Where do you think the electronics cooling knowledge in FLOEFD comes from ;-)
It is the advantage for FLOEFD to have another development team that has this knowledge so FLOEFD can offer this also to some of their customers that need it.
Of course, there are other commercial codes that have similar knowledge too, (getting back to my automotive examples) just like there are other automotive companies that have that knowledge too. However, there are some that have a particular competitive edge in their technology that others don't have. You can implement a lot of equations you can find in textbooks, but that doesn't mean you actually understand how things really work or how such components are structured. With T3STER we gained such a detailed structure understanding which also helps our simulation tool development and there is no other company out there who has that for example, this also is valid for LEDs with our TERALED. Just like Tesla gained a lot of experience from their cars on batteries and e-motors which many other companies still need to pick up. Of course, you can hire experienced developers, but then you only have the state of the art, but what happens in 5 years, you need to hire new ones and you are not really working on the edge but chasing it.
I don't know all the other codes in that detail to make sure I can point out something "major" that makes us unique that would also satisfy specifically you. I know this connection to calibrate your thermal models with the help of T3STER is something some of our customers using T3STER really like. There are probably other capabilities but I'm not an expert in Flotherm to know them all. Also you could argue a car is a car so what makes BMW so different to Audi in the core technology (wheels, engine and chassis for example). There are usually minor differences or single spikes that make one product more attractive than another. Especially with products specializing on an application for some time. There is only so much you can develop and it would need to be a sudden quantum leap to make it a superior product to all others, which is unlikely to happen and these spikes are usually just causing a short delay before another company has that as well as the technology is often out there somewhere.
True, there is a delay in the latest version of SolidWorks in both Flotherm XT and FLOEFD, but there are only few companies who update right away to the latest version of any software they use as it would impact current projects they work on if something is not properly working in the new version or compatible with what they were doing. This is also valid for simulation tools, not just CAD. About 90% of all users we have are behind in the latest version of simulation tools and CAD systems. And Flotherm XT did not choose SolidWorks to provide a solution for SW users specifically but to offer CAD import and modifications when imported. So there will be SW users impacted by that if they use Flotherm XT, but a majority is not. If this is something you really had issues with, then that is fine and your choice. Some things we can fix and others we had to make a compromise.

And that is fair, my personal favorite is Audi, but BMW build great cars, I just don't like certain things about them. I've talked to customers who used also a variety of codes during their career and they had various reasons to like or dislike certain tools. Some like ours, some like others. Apparently your need or wish to work inside a CAD system is not that high to favor FLOEFD, so you prefer scStream, I talked to people who used both and found FLOEFD much better.
Choosing a tool is not just done for its capabilities, as there might be other tools that have the same capabilities you need to do your work, but you just liked something better about them which makes them your favorite tool. This might be UI, the nice rendering possibilities but also the support of the product or the reason that you can buy a whole solution portfolio from that company for other simulation types such as FEM or EM etc.

What can I say, this is product and company strategy and it is done by every company, no matter if you are the customer or the vendor. You have to make the choices that are right for your company to be efficient and grow and these choices will not satisfy everyone and it is impossible to make everyone happy.

Regards,
Boris

Hi Boris, thank you for the extensive explanation. I wouldn't necessarily agree with all your points but as I said before you know better. One thing however I think I know better is the difference between Floefd and scStream. Yes, Flowefd may have a bit friendlier user interface, but scStream handles much better very complex assemblies. It has amazing abilities to process hundreds of millions of cells with just 128G of RAM. If one has to deal with such models scStream is unbeatable. If your model is less than say 10-20 millions of cells then Floefd works fine and would probably be the code of choice. One more thing scStream beats the Flowefd, or at least Solidworks flow simulation (if there is a difference between Flowefd and SWFS), is the post processing. Finally scSream has a bunch of solvers suitable for different application areas, while Floefd uses a single solver for all flow applications. It could be that its solver is so advanced that can cover well all applications that's why I am not counting this as a drawback. Again Boris this is only my personal opinion having dealt with the two products above and I am not subscribing anybody to it.

Hi,
Yes, as I said, I don't and cannot know all the codes and their details. Our customers usually know more codes as they test their way through them or used them in various other jobs.
But I've seen FLOEFD handle a few thousand parts native in the CAD tool without to convert into neutral formats to then be imported again in the CFD tool. And it doesn't need hundreds of millions of "cells" if you know about its meshing technology. I put the "cells" this way as I'm not talking about control volumes which you are, "cells" is what the user sees, control volumes can be several times more, depending on the model and the mesh that was generated.
To be honest, post-processing is to get results visualized in a way that our brains can better process them rather than a 3D matrix of numbers as the computer can process it. Sure you can do a lot of amazing stuff with it but in the end, 99% is done with simple plots and not in a wow rendering etc. I know scStream has nice post-processing but you can go ahead with ParaView and EnSight if you want to do more with it (which is not possible with SWFS).
I love cool looking images as well but I've seen plots from customers where they don't even use more than 10 colors, so most likely not the most important feature for everyone.
And FLOEFD is not a single solver, that's just what you see, but only because if you pick the "high Mach number" option for example, it automatically uses a different solver. There are other such options as well. This is key in FLOEFD, don't put the user to the choice of x number of solvers or turbulence models etc.

If someone doesn't like a particular tool, then that is fine. If someone finds the other one better than the other, than that is also fine. Those are opinions and very strong personal and don't apply for everyone else and no one can change them, only that person himself.
Fact is, I haven't heard of a CFD tool out there which is completely rubbish. Every tool has its strengths and weaknesses and some have a very narrow application and others a very broad application but then also not very deep into all of them compared to the ones with a more narrow application range. This is product and company strategy and you have to make these decisions on what market you want to serve. There is no jack of all trades CFD tool.

Most important is to pick the tool that fits your needs the best and provides the best efficiency at high accuracy. To pick a tool because you like it more but spend 3-times the time to get from model to result or 3-times the money is selfish and not in the interest of your company where you want to be "lean" and efficient for a higher margin and better competitiveness.