• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 201 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LM2675: LM2675

Jason Pearce
Prodigy 20 points
Part Number: LM2675
Other Parts Discussed in Thread: LM43601

Hi

I have a question on the LM2675 in negative regulator mode, which the datasheet says is possible with this regulator.

We have two circuits using the adjustable version of this regulator (LM2675-ADJ) , configured as -5V and -12V. 

However, app note:

https://www.ti.com/lit/an/snva022e/snva022e.pdf

discusses generating negative voltages with the LM267X series.

It suggests that a capacitor Cc must be used to stabilise a right hand zero, and the value must be <=100uF and Cc must be less than CIN. By what factor should Cc be less than CIN? 25%, 50%?

Figure 6 in the above app note provides a 12V to -5V design where CIN is 45uF total, and Cc = 22uF so just less than 0.5*Cin in this example. In the example the capacitor appears to be ceramic as there is no polarity indicator on Cc, (although no detail is given and Cc is stipulated a low ESR although no suggestion of ESR value is given). 

The loop response appears to be predesigned, and (fairly) wide ranging in terms of components. The datasheet provides tables of suggested LOUT and COUT values. But in the datasheet there are no details on crossover frequency relative to either Fsw or the components used. The app note says test output waveforms for stability which isn't helpful after the device has been designed in: some advance indication of response and stability is far better.

A negative regulator is not supported in web-bench as it looks.

Are there problems with the LM2675-ADJ than using the fixed -12 and -5V parts to generate the negative voltages?

Thanks in advance for input to this question.

regards

  • 0
    TI__Genius 16885 points

    Hi Jason,

    Thank you for your questions. 

    In regard to Cc, Cc is recommended to be less than 100 uF to yield the best results. Unfortunately, there are no more guidelines we can provide that will guarantee stability past testing the system yourself. There are some suggestions as to improving stability in section 12. I would suggest beginning with the circuit given in the app note and going from there. I would also recommend keeping Cc at a value such that it is always under CIN's de-rated value.

    As for the ESR, it seems that you want to keep this as low as possible. The only part that has a minimum ESR is the output capacitor and this is detailed in section 7 of the app note.

    You can use the adjustable version of LM2675 as well, but you need to connect the feedback resistors from GND to VOUT. See section 2 of the app note for more insight.

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Hi Joshua

    thanks for your response, but there's nothing here that the app note doesn't suggest, with respect. I'm not very keen on building then rectifying problems when the information is better known at the design stage. This isn't my design anyway.

    Regarding feedback resistors, yes this is done. Our circuit is very similar to the app note save for absolutes values, but Cc is much large which can be reduced. That's the first thing I noticed.

    One confusing thing is why this device is not supported in WebBench? This suggests an issue, unless TI are pushing newer, better regulators for designs. I appreciate this is a an old Nat Semis regulator. I put in my design parameters and the LM2675 does not even come up in 28 suggested circuits.

    Thank you.

    Jason

  • 0 in reply to Jason Pearce
    TI__Genius 16885 points

    Hi Jason, 

    I do not have any knowledge as to why this device is not supported in Webench. Webench is supported for this device but the inverting buck-boost topology is not on Webench for every device, but this does not suggest that there is something wrong with creating an inverting buck-boost circuit for these devices. I was able to find an inverting buck-boost for the LM43601 on Webench, if you want to stick with Webench and the LM43601 is able to meet your design criteria.

    I did see that the LM2675 does have a PSPICE model and this can be used to simulate various waveforms in PSPICE for TI. This should be able to help you validate your design before building it. 

    Thank you,
    Joshua Austria

  • +1
    Prodigy 20 points

    Hi Joshua

    thanks for your response.

    The trouble with Webbench is it requires resource to maintain it.

    Thanks for your suggestion for the LM43601: however I'm currently stuck with the LM2675 as the board has been spun. 

    I have PSPICE TI and am building a model but getting the AC information out of it is proving problematic. I'll close this thread down now. Thanks for your help.

    Kind regards

    Jason