• State TI Thinks Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 287 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.

LM5168: Design help and transformer construction support

Tanmay Sutaria
Prodigy 155 points
Part Number: LM5168
Other Parts Discussed in Thread: LM5017, , LM5169

Hello,

I am designing a Flybuck converter with three outputs, spec as follows :
1. Input: 20 – 60Vdc
2. Outputs :  (a) Non-isolated(main) 5Vdc, 50mA with LDO to get 3.3V
                     (b) Isolated(1) 6.5Vdc, 150mA with LDO to get 5V
                     (c) Isolated(2) 6.5Vdc, 50mA with LDO to get 5V
(For ease of calculations, b & c can be considered as single output of 6.5V, 200mA)
3.    Operating Freq : 680KHz


We are considering the LM5017/5018 as I have used the same earlier for a lower power design or the LM5168/51699 as they are the newer/suggested devices. Exact part to be finalized based on current requirement and availability.

I have designed the transformer for LM5168 using the datasheet design example iand the calculator spreadsheet - attached.
The design arrived at is - Primary L of 58uH for Fsw of 680KHz.
Using larger Tr.primary L of 84uH would lead to a lower Iripple of 70mA(28%) and hence lower Ipk of 320mA - so LM5168 may be adequate, since 5169 is out of stock.
Other values - Cout 33uF, Ca 2200pF, Cb 1000pf, Ra 100k giving injected ripple of 27mV which should be adequate.
Hence required transformer specs would be L = 84uH, primary saturation current ~400mA, N1-N2/3 ratio of 1:1.3.

But there are multiple aspects which need clarification.
1. The 516x is suggested as newer replacement for the 501x and hence the design process would be same / similar.
However, when I tried to cross verify the design using different guides - both datasheets examples, AN2292(snva674c) and 501x EVK as examples there is wide variation in the results achieved, specifically the value of 'K' and delta-'I' calculated or to be assumed.
While Rt calculation for both families is different and clear, please advise which document is best suited to design for respective family.

2. Most ref. designs/examples - 5168 datasheet & EVK, 5017 example etc. arrive at/use 33uH primary inductance and are for Fsw of 750KHz.   
All BoMs specify 33uH transformer from Coilcraft, Wurth etc for the same. But the transformer datasheets specify L values at 100KHz - i.e. selected device is 33uH @ 100KHz.
Will not using the same device at 750KHz effectively reduce the L 'seen' by 516x/501x i.e. make the 33uH, 33/7.5 = 4.4uH ?
Is this scaling built into the design math? i.e. required L is scaled to 100KHz in the math?
We will be using a custom transformer - so what specs should I be giving to the manufacturer - 84uH @ 680KHz or 84uH @100KHz?

Using a basic overview (I may be totally wrong about this so please advise), impedance of L of 84uH @ 100KHz calculated at 680KHz comes to 358 ohms.
It would seem that for min i/p of 20V and o/p of 5V, such imedance value will not allow 250mA of primary load current. Is this assumption accurate?

3. Based on the above design is 5168 adequate or is 5169 required? When would it be availe in at least smaller qty - ~200nos.

Please advise at the earliest possible.

Thanks.

4454.LM5169_calc_revD.xlsm

  • 0
    TI__Guru 73570 points

    Hi Tanmay,

    Thank you for considering our products.  I will study this and get back to you by COB tomorrow.

    Regards,

    Youhao

  • 0 in reply to Youhao Xi
    TI__Guru 73570 points

    Hi Tanmay,

    Let me answer your 2nd question first:  The transformer selection is based on the ripple current, which is limiting the pk-pk inductor current ripples to be below a certain percentage of the full load dc current.  Since your load current is smaller than those examples, the recommended inductance is higher to achieve the similar ration of pk-pk current to the full load dc current.  

    Please do not be confused by the transformer datasheet.  It means the inductance was tested at 100kHz in their production but the inductance should still be very close to that nominal value at 680kHz which is the operating condition of your circuit.   For your 680kHz, you do need 84uH to meet the ripple current requirement as explained above. 

    Regarding the transformer selection for LM5168 and LM5169, one criteria is that the Isat of the inductor should not be smaller than the IC's peak current limit, or the magnetics would become saturated during dynamic response.  From the datasheet, the LM5169's peak current limit is about 2x of the LM5168's, therefore, you will need to select a transformer to meet the Isat requirement.  By the way, your 400mA Isat appears to be on the low end and you'd better to raise it for LM5168 because the peak current limit can be 484max on the high end. 

    Regarding your question 3, both ICs should work but you do need to pay attention to the Isat differences. Please check our website frequently to get the updated stock info.

    Regarding the LM5017/8, it should be able to support your application, and you can choose it if you can find part.  The similar rule on Isat applies.

    Hope this clarifies.

    Thanks,

    Youhao

  • 0 in reply to Youhao Xi
    Prodigy 155 points

    Hi Youhao,

    Thanks for the detailed explanation. Clears up the transformer and part selection queries.

    Please do revert on the first point also - best design document / steps for 516x as we will need that part regularly going forward, specially the calculations / assumptions for values of 'K' and delta-'I'.

    Thanks.

  • 0 in reply to Tanmay Sutaria
    TI__Guru 73570 points

    Hi Tanmay,

    There is not absolute value that you must choose for "K".  Based on our experience, which is also aligned with the experience from the field, a general guideline for "K" is about 0.2 to 0.4, which would lead to a good compromise among the inductor size, Cout size, and overall efficiency. 

    Hope this clarifies.

    Best Regards,

    Youhao

  • 0 in reply to Youhao Xi
    Prodigy 155 points

    Hi Youhao,

    Thanks. Have completed the design based on you feedback and using the spreadsheet.

    Shared the details with the magnetics vendors but something seem amiss in the design as the primary turns they are able to arrive at is very large going into 1000s!

    It would be great if you can review my design and point out major error/changes required if any - shared privately.

    Also, is there a tool or guide to calculate the primary turns from the design parameters, If yes, please share a link.

    Thanks and regards,

    Tanmay Sutaria. 

  • 0 in reply to Tanmay Sutaria
    TI__Guru 73570 points

    Hi Tanmay,

    Could you share the magnetics vendor's design details?  1000 turns sound strange because you just want 84uH.  I am afraid something was miscommunicated.  

    The EVM magnetics has 33uH. With the same core, you just need 1.6x number of turns of the 33uH transformer to make it 84uH, if the air gap is not changed. This is physics:  on the same core, the inductance is proportional to the number of turns squared. The EVM magnetic should not have 600 nor ever close. Where does the 1000 turns come from? 

    Thanks,

    Youhao

  • 0 in reply to Youhao Xi
    Prodigy 155 points

    Hi Youhao,

    Have emailed the transformer design.

    I agree with your turns estimation, we are thinking along the same lines, but we cannot know the turns used in the EVM device or thc core they have used (except size). Also, our core is different - EE!3- so redesign/recalculation of transformer would be necessary. Will recheck with our vendor and revert.

    Thanks,

    Tanmay.

  • 0 in reply to Tanmay Sutaria
    TI__Guru 73570 points

    Hi Tanmay,

    Sorry I don't have a calculator for the transformer itself. You can just tell the magnetics vendor the Lmag, turns ratio and the operating voltages and switching frequencies, they can get a design quickly.   One idea if you want to wind your own transformer, please refer to the following article.

    Magnetic Design Handboob-slup132.pdf

    Best Regards,

    Youhao

  • 0 in reply to Youhao Xi
    Prodigy 155 points

    Hi Youhao,

    Thanks for the design guide, will be very useful for the future also.

    The turns were finally figured out ans samples made. Working quite close to design specs, just need a little fine tuning.

    Thanks for all you help.

    Best regards,

    Tanmay.

  • 0 in reply to Tanmay Sutaria
    TI__Guru 73570 points

    Hi Tanmay,

    Thank you for the updates and let me close this thread here and we can continue the discussion in PMS.

    Best Regards,

    Youhao