• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 62 subscribers
  • Views 322 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CSD18540Q5B: Variation of CSD18540Q5B

John Fecht
Prodigy 60 points

Part Number: CSD18540Q5B
Other Parts Discussed in Thread: CSD18512Q5B, LM5145, CSD18534Q5A, CSD18531Q5A

Hello,

I am working on making a solution that uses the CSD18540Q5B. After running my calculations I've noticed that when using the typical parameters my design will not overheat. However, when using only the worst case parameters, the design comes close to overheating. 

I'm curious where variation occurs. For example, if I order a reel of 1000, how much will the operating parameters change in that set of 1000? And then, if I order another reel of 1000 half a year later, is the variation from Lot A to Lot B much greater than the overall variation of Lot A?

If the majority of variation occurs chip-to-chip, we can then do testing to remove the boards that overheat and end up with a percentage that passes. However, if the majority of variation occurs lot-to-lot, we could end up with 1000 boards that don't work.

 

  • 0
    TI__Mastermind 40610 points

    Hello John,

    Thanks for your interest in TI FETs. Can you tell me more about your application, the requirements and how the FET is being used and operated? There will be part-to-part and lot-to-lot variation. I'm checking with the product engineer to see what data we have available to try to help you resolve this. One of the biggest contributors is typically Rds(on) but switching loss can also contribute significantly in a switch mode application where the FET is switched at high frequency (100s of kHz). I see your company is based in Ames, Iowa. I am originally from Iowa and attended Iowa State University. I look forward to your response.

    Best Regards,

    John Wallace

    TI FET Applications

  • 0 in reply to John Wallace1
    Prodigy 60 points

    Hey John,

    Currently we're using the CSD18540 as the high side switching FET and the CSD18512Q5B as the low side fet in tandem with the LM5145 buck converter. The low side loses a small amount of power compared to the high side fet, according to the LM5145 quickstart calculator. Since our application is to design a solar charge controller, we've modified the control loop control the input voltage instead of the output voltage. Here's the main parameters we're using to configure the LM5145: 

    Vin range: 20-36 V

    Vout typ: 15V

    I_in Max: 6.1A

    I_out Max: I_in_max * Vin / Vout

    Fsw = 280kHz

    We currently attach these to panels that have a set point voltage (V_in) of around 27V. At this input voltage the output voltage is around 11A. According to the calculator, the CSD18540 should lose around 1.5 W of power if every parameter in the calculator is worst case. 

    However, when we push the input voltage up to around 36V and assume worst case parameters, the loss increases to nearly 2.5W, which is mostly dominated by switching losses. Running the "typical" parameters, we see that loss is decreased to 1.5W of power.

    Since we can use the same circuitry to run a 27V set point charge controller as a 36V set point charge controller, I thought we might be able to run a test and match the boards that run cooler with the 36V configuration, and then the ones that run warmer can be used for the 27V configuration.

    By knowing what the variation is between ordering, it will help us understand if the ratio of "grade A" boards to "grade B" boards will differ lot to lot, or if it will remain fairly stable every time we order.

  • 0 in reply to John Fecht
    TI__Mastermind 40610 points

    Hi John,

    Thanks for the updated information. Can you share the LM5145 quick start calculator with your values? I'd like to see the distribution of losses. Typically, the parametric distributions in production are tighter than the datasheet specs and there is usually a guard band on top of the datasheet limits. TI can only guarantee what is specified in the datasheet. TI has a number of Excel-based FET selection tools that allow the user to compare up to 3 different TI FET solutions including a sync buck tool. Please let me know the inductor value. I can get the rest of the information from the LM5145 datasheet for the gate drive. Link to the tool is below. 

    https://www.ti.com/tool/SYNC-BUCK-FET-LOSS-CALC

    Thanks,
    John

  • 0 in reply to John Wallace1
    TI__Mastermind 40610 points

    Hi John,

    One more thing. The CSD18512Q5B is a 40V FET and there's not much margin for transients and switching spikes when used for 36V input. It may be possible that the voltage at the SW node is exceeding BVDSS from the bottom FET. For 36V input, I would recommend a 60V FET.

    Thanks,

    John

  • 0 in reply to John Wallace1
    Prodigy 60 points

    To answer your first reply, we're using a 5.6uH inductor. I'll attach images of the values I'm using since I can't attach the quickstart calculator:

    The values I'm using for the FET might be slightly off since I was trying to account for the the gate drive voltage being 7.5V. I've updated the low side fet to be a CSD18540 in the calculator as per your suggestion.

    After some internal discussion, we may add a thermistor located nearby the high side. Then we can throttle the output current down if the temperature is too high. That being said, knowing the lot to lot variation will assist us in providing a better product to our customers.

    Thanks!

  • 0 in reply to John Fecht
    TI__Mastermind 40610 points

    Hi John,

    Thanks for the update. I will review and get back to you shortly.

    Best Regards,

    John

  • +1 in reply to John Wallace1
    TI__Mastermind 40610 points

    Hi John,

    I filled out the LM5145 quick start calculator and I'm not sure how it does the loss calculations. I used the FET selection tool I developed which includes the effects of common source inductance as detailed in the application report at the link below. My tool uses max Rds(on) at the closest value of VGS where on resistance is specified in the datasheet (usually at 4.5V or 10V) and it takes into account the temperature coefficient of Rds(on). It uses typical charge and capacitance values from the datasheet. Similarly, the tool uses typical QG at the closest value of VGS (usually at 4.5V or 10V). 

    Using the CSD18540Q5B for both FETs unbalances the power losses and the high side FET estimated power loss is ~3.8W and low side is ~0.7W. Selecting a lower performance low side FET actually helps reduce the losses in the high side FET. The overall loss in the FETs is lower using the CSD18540Q5B for both. I selected the CSD18534Q5A for the low side FET and CSD18540Q5B for the high side FET. Overall FET losses are somewhat higher but it balances the losses in the FETs. Estimated loss in the low side FET is about 2.2W and the high side FET is about 2.6W. 

    Of course, these are all estimates and will have to be verified by testing. TI has a number of 60V FETs in this same package and there will be options.

    https://www.ti.com/lit/an/slpa009a/slpa009a.pdf

    Please review and let me know what you think. I can send you my spreadsheet if you are interested.

    Thanks,

    John

  • 0 in reply to John Wallace1
    Prodigy 60 points

    Thanks for the help John,

    After playing around with the spreadsheet you sent, I believe we're going to use the CSD18531Q5A and the CSD18540Q5B. We'll try using the CSD18534 and 18533 as well. The FETs will be placed pretty much right next to each other, so I'm a little bit worried about cross heating which will increase Rds in both FETs. 

    Thank you for your help, I'll say this resolved my issue for now and ask more questions as we get our prototypes in.

    -John