Validity of IPOSIM

Announcements

From sunburn to sun earn – we’ve got the power! Watch our #poweringgreen videos now.

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
zhenqihui
Level 1
Level 1
5 replies posted 5 sign-ins First reply posted

1. As calculated loss of FF225R12MS4 and FF300R17ME4 by IPOSIM, FF225R12MS4 has a lower loss, but my platform test results show FF225R12MS4 loss is more then 100W higher than FF300R17ME4, which is completely different from the calculation results. 

 

2. I changed the Ron of the FF300R17ME4 from 4.7ohms to 3.3 ohms. The recalculated results show that the  Eon+Eoff, and Erec, which determine the loss, are exactly the same as the original results, however,  the switching loss has increased obviously. How to explain this calculation result? Is the calculation model of IPOSIM reasonable? 

0 Likes
1 Solution
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello Zhen 

Please let me know what is switching frequency you are operating in experimentation. 

FF300R17ME4 can operate lower switching frequencies and FF225R12MS4 can operate at higher frequencies.

FF225R12MS4 device is having higher VCE so if you operate at lower frequency the conduction losses will be more. kindly increase your switching frequency and check the losses and compare it.

Thanks

Guru

View solution in original post

0 Likes
13 Replies
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello @zhenqihui 

Thanks for posting the question in the Infineon community.

1)I understand that you have done IPOSM for FF225R12MS4 and FF300R17ME4. The experiment you have conducted for FF300R17ME4 and you are comparing the this experiment results with IPOSIM of  FF225R12MS4, From this data it is difficult to comment. Please provide the IPOSIM conditions which you have simulated and experiment you have conducted?

2)Yes the IPOSIM Results are reasonable if your hardware conditions(Rg,Temp,PWM) are same. 

The attached images are not accessible. Please properly attach and send us.

Thanks 

Guru

 

0 Likes
lock attach
Attachments are accessible only for community members.
zhenqihui
Level 1
Level 1
5 replies posted 5 sign-ins First reply posted

Thank you very much for your reply.

First, let me introduce my purpose: I want to choose an IGBT module to build an H-bridge inverter and drive a transformer to make a DC power supply. I selected the analogous model DC/AC Application-Single Phase Module in IPOSIM for loss calculation. The parameter settings are shown in Figure 1:

zhenqihui_0-1645597686355.png

Figure 1 The parameter settings on IPOSIM

I selected two IGBT modules, FF225R12MS4 and FF300R17ME4, for comparison and calculation. The calculation results are shown in Figure 2. Obviously the loss of MS4 is much lower than that of ME4, so I want to use MS4 instead of ME4 to reduce the loss.

zhenqihui_1-1645597728868.png

Figure 2 calculation results of IPOSIM for two IGBT modules

Then, I built the platform shown in Figure 3 to carry out the test experiment. I used FF225R12MS4 instead the FF300R17ME4 which is I used originally on the platform. The test results found that the loss of MS4 is almost 200W higher than that of ME4, which is completely opposite to the calculation result of IPOSIM. I am very confused as to where the problem is.

zhenqihui_3-1645597825610.png

 Figure 3  Schematic diagram of experimental platform

Further, I adjusted the Rg_on of FF300R17ME4 from 4.7 ohms to 3.3 ohms in IPOSIM. The calculation results found that the parameters related to loss did not change, but the loss results changed significantly, so I doubted the rationality of IPOSIM. The results are shown in Figure 4.

zhenqihui_4-1645597861272.png

Figure 4  the rationality of IPOSIM

0 Likes
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello @zhenqihui 

Thanks for sharing the data.

We analyzed your data. Please summarized your results following two cases

Simulation

Looks the simulation results are correct in fact we verified internally also look the same results we got. The IPOSIM  provides different power losses for different Gate resistors you could see in your fig.4 results.

may know which case you have seen that loss are change? 

Hardware

The FF300R17ME4 is not using in your present hardware because higher loss that's is fine. what about your gate resistance values  did you changed it or keeping same for FF225R12MS4? .

The FF225R12MS4 experiment you have conducted at right may I know what is Gate resistor value , PWM strategy and heat sink parameters?

while you are comparing the simulation data with hardware make sure your simulation parameter's should be same

Thanks 

Guru

0 Likes
lock attach
Attachments are accessible only for community members.
zhenqihui
Level 1
Level 1
5 replies posted 5 sign-ins First reply posted

In fact, my purpose is not to make the results of the experimental platform consistent with the simulation results of IPSIM, but to compare the loss of FF225R12MS4 and FF300R17ME4. Our existing equipment is using FF300R17ME4. If an IGBT module with lower loss can replace the FF300R17ME4, the volume and weight of the equipment can be further reduced, which is our desired result. So we used IPSIM calculation to find the target device FF225R12MS4 with lower loss. We thought it could complete our task, but the actual test result was not satisfactory. This is the reason why we asked for help.

In order to further clarify the problems we communicated, the following content is organized. firstly, I will introduce the IGBT driver, then, the experimental platform and test method are discussed, finally, I will summarize my work.

1. IGBT Driver

The IGBT driver on the experimental platform adopts 2SP0115T2A0-12, and selects Rg,on and Rg,off according to the parameters calculated and set by IPOSIM for experimental testing. There is a problem here, when IPOSIM sets the Rg,on and Rg,off of FF225R12MS4, the resistance value can only be greater than 4.7ohms, but the resistance value recommended in the 2SP0115T2A0-12 data sheet is smaller, as shown in Figure 1.

 
 

zhenqihui_2-1645757347125.png

Figure 1 Different recommended values for Rg,on and Rg,off

In fact, the loss calculation results of IPSIM vary greatly depending on the selection of Rg,on and Rg,off. However, according to the set value of IPSIM and the recommended value of the 2SP0115T2A0-12 data sheet, the platform experiment was repeated, and both of the loss measurement results were almost the same. There is no discernible difference, which is why I doubt the validity of IPSIM.

2. Experimental Platform
In order to verify the IPSIM calculation result that the loss of MS4 is much lower than that of ME4, an experimental platform was built, and the input and output power of the inverter bridge were measured simultaneously to measure the loss, as shown in Figure 2. The aluminum radiator is used, the length, width and height are 240mm*370mm*100mm, and the thickness of the bottom plate is 15mm.

zhenqihui_3-1645757420118.png

Figure 2 Experimental Platform and Loss Measurement Method

On this experimental platform, I used two IGBT modules, FF225R12MS4 and FF300R17ME4, to compare their losses. The experimental results show that the loss of FF225R12MS4 is much higher than that of FF300R17ME4, which is completely opposite to the calculation result of IPSIM.

3. Summarize

Through the above work, I have the following questions.

Question 1: Why IPSIM has restrictions on the Rg,on and Rg,off settings of different IGBT modules, as shown in the screenshot on the left of Figure 2.

Question 2: Why the loss of FF225R12MS4 is more than that of FF300R17ME4, which is completely opposite to the calculation result of IPSIM.

Question 3: Different choices of Rg,on and Rg,off, the loss calculation results of IPSIM are very different, but according to the set value of IPSIM and the recommended value of the 2SP0115T2A0-12 data sheet, the platform experiment was repeated, and the loss measurement results were consistent. There is no obvious difference on the loss with different Rg,on and Rg,off, what is the reason for this?

It is necessary to supplement and modify the description. In the case of 25kW output power, the loss of FF225R12MS4 is about 80W larger than that of FF300R17ME4, instead of the previous 200W

thanks

Zhen

 

 

0 Likes
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello @zhenqihui 

Thanks for providing your system details 

I have answered your questions please follow the below points 

Answer-1

The Gate resistor (4.7 ohm ) was provided by IPOSIM is basically taken based IGBT

characterization, The IGBT characterization done at 4.7 ohm so that is the IPOSIM tool was given minimum

limitation.

 

Please let me know how you are measuring the losses in your system?. Based on your loss estimation

process  we can comment on Q2&Q3.

Based on 

Thanks 

Guru

 

  

0 Likes
lock attach
Attachments are accessible only for community members.
zhenqihui
Level 1
Level 1
5 replies posted 5 sign-ins First reply posted

Thank you for your support and help. I will introduce the loss measurement method in detail below. I use the PW3390 power analyzer produced by HIOKI to measure the loss, using the wiring mode shown in Figure 1, The input power measurement in the yellow box in the lower half of the measurement schematic diagram corresponds to the 3P4W wiring method, and the output power measurement corresponds to the 1P2W wiring method.

zhenqihui_0-1645858390749.png

Figure 1 Experimental Instrument and Loss Measurement Method

The readings of the measurement results are shown in Figure 2. Where P123 is the input power, P4 is the output power, and Loss1 is the loss measurement result.

zhenqihui_1-1645858429375.png

Figure 2 Measurement result reading

The measured losses are the sum of the losses of the three-phase input rectifier and the inverter bridge. The final measurement results show that under the condition of 25kW output power, the loss of FF225R12MS4 is about 80W larger than that of FF300R17ME4.

My purpose is not to make the results of the experimental platform consistent with the simulation results of IPOSIM , but to compare the loss of FF225R12MS4 and FF300R17ME4. Our existing equipment is using FF300R17ME4. If an IGBT module with lower loss can replace the FF300R17ME4, the volume and weight of the equipment can be further reduced, which is our desired result. So we used IPOSIM calculation to find the target device FF225R12MS4 with lower loss. We thought it could complete our task, but the actual test result was not satisfactory. This is the reason why we asked for help.

I think, at least the result should be that FF225R12MS4 has less loss than FF300R17ME4, not the other way around.

Thanks

Zhen

0 Likes
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello @zhenqihui 

Thanks for your detailed explanation.

Looks your power loss calculation method is correct.

same power you are applying in both cases ?

while you are comparing the experimentation are you using same gate resistor for both devices?

below recommendations to be followed

1)The gate resistor to be tuned separately for FF225R12MS4, same gate resistor cannot be used for both devices.

2)Make sure your thermal resistance between junction to ambient should be calculated and try to keep as minimum as possible of both cases .

Thanks

Guru 

 

0 Likes
zhenqihui
Level 1
Level 1
5 replies posted 5 sign-ins First reply posted

Thank you very much for your reply.

In fact, the experimental results show that the loss of the FF225R12MS4 is almost unchanged during the adjustment of Rg_on and Rg_off from 4.7 ohms to 1 ohm, and is not as sensitive as Rg_on and Rg_off in the IPOSIM calculation.

Thanks

Zhen

0 Likes
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello @zhenqihui 

The power loss calculation mentioned in the below image is not correct. 

Vaddemani_0-1646305876134.png

In above mentioned system  it will give additional losses from both sides (AC  side and DC side)

1.Bus bar losses 

2.DC link capacitor losses 

Generally we used double pulse test for device power loss calculation. Please conduct  double pulse test and calculate your device losses.  Please click here 

I have one more doubt how you estimated the  FF225R12MS4 loss  is 80W  from below fig?

Vaddemani_1-1646306346858.png

 

Thanks 

Guru

0 Likes

Thank you very much for your reply.

The situation you mentioned is correct, but in fact, these additional losses can be basically considered to be the same in the comparison of the two IGBT module inverters. The main loss difference is the loss of the IGBT module inverter. What we are concerned about is the relative magnitude of the inverter losses of the two IGBT modules. Based on the above purposes, the loss measurement method will not be problematic.

The loss difference of 80W for the two IGBT modules under the condition of 25kW output is calculated based on the actual measurement readings, as shown in the table below.

zhenqihui_0-1646480028769.png

 

Thanks

Zhen

0 Likes
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello Zhen 

Please let me know what is switching frequency you are operating in experimentation. 

FF300R17ME4 can operate lower switching frequencies and FF225R12MS4 can operate at higher frequencies.

FF225R12MS4 device is having higher VCE so if you operate at lower frequency the conduction losses will be more. kindly increase your switching frequency and check the losses and compare it.

Thanks

Guru

0 Likes

Zhen uses 10kHz frequency in his system.. By the way, could you please help to answer my question?

Econo PIM 3 module with IGBT 7 in it - Infineon Developer Community

0 Likes
Guru_Prasad
Moderator
Moderator
Moderator
250 replies posted 100 solutions authored 25 likes received

Hello @Alan1 

The FF225R12MS4 will give good performance at a higher frequency so kindly test at >15kHz and check the losses.

Our AE is working on your question. you can see the response for your question on that particular page.

Thanks

Guru 

0 Likes