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.

WEBENCH® Tools/LM5008A: Suggested design changes with Max Ambient Temperature

Part Number: LM5008A
Other Parts Discussed in Thread: STRIKE

Tool/software: WEBENCH® Design Tools

Here is the WEBENCH Power Designer design I am pondering:


Vout = 3.3V

Iout(max) = 250mA

Vin = 6v to 75v

This power supply will be used to power a small vehicle interface that will be mounted under the dashboard in a 12V or 24V vehicle.  It is conceivable that the interior temperature on a hot summer day could approach 80°C, but the vehicle interior will not be that hot all year around.  Input voltage spikes will be capped to 70V with a TVS diode, and a B2100-13-F Schottky diode will provide reverse polarity protection. (I will also use the same diode p/n for power supply D1.) Typical output current could be as low as 2mA or as high as 150mA.  There will be one inductive load -- a 3V signal relay -- that will share the 3.3V voltage rail with a PIC MCU and fingerprint sensor.  I may need to add 47uF to 100uF of added capacitance via aluminum electrolytic to ensure the 3.3V output voltage remains stable when the relay is energized.

Within WEBENCH Designer, when I set Max Ambient Temperature anywhere between 30°C and 75°C, the suggested schematic components and switching frequency do not change.  But when I go above 75°C, Ron & Rr & L1 increase in value, and it suggests 2pcs of Cout.  If I then set the Max Ambient to 85°C, those components increase again and the Frequency drops to only 50kHz.

QUESTION#1: Is it best to go with the suggested design shown at 75°C with its 128.781kHz switching frequency, or is it really best to go with the suggested design at 80°C which drops the frequency to 100kHz or 85°C with a 50kHz frequency?  What are the disadvantages to running at a lower frequency that I should consider?


I am considering use of a single Cout with improved specifications over the recommended Cout:

LMK316AB7226ML-TR: 22uF, 10V, X7R, Derated Capacitance (at 3.3V): 15.4uF

QUESTION#2: Is it not better to use a single temperature-stable X7R capacitor over the suggested 2pcs of GRM188R60J226MEA0D (derated capacitance = 7.9uF -- 15.8uF combined)?  Or is the 1m-ohm lower ESR of those two X5R capacitors in parallel more desirable than temperature stability?  I'm trying to strike the right balance between cost and performance here while minimizing output ripple.

Thank you.

  • Hello James,

    This part has a maximum recommended operating junction temperature of 125 C. It appears that Webench is trying to limit dissipation by reducing frequency. Once the program starts reducing frequency rapidly with increasing temperature, you have reached the edge of the window in which the part operates well.

    I recommend sticking with 75 C BOM since as frequency drops, the physical size of the converter increases dramatically and maintaining low inductor DCR will be hard.

    If you are sticking to X7R capacitors, a single capacitor is OK. The output capacitor may be slightly more inductive. Attention should be paid to layout to keep this inductance small.

    Regards, Robert

  • Thank you, Robert.

    You emphasized the importance of reducing inductance on the output capacitor. What then are your thoughts on my use of a 3V (50mA) relay on the output voltage rail?   The relay coil is an inductive load. I will of course have a diode across the coil to suppress kickback, but as mentioned in my opening post, I am thinking about using additional capacitance in the form of an aluminum electrolytic capacitor, possibly 47 µF or 100 µF, on the output voltage rail near the relay to prevent voltage dips when it switches. Is that a bad idea or are there other considerations in light of the fact I will have this inductive load?   The same 3.3V output rail will power my MCU and a fingerprint sensor.

    Thank you.

  • Robert, your reply to my follow-up question posted yesterday would be greatly appreciated. In the meantime, I have one additional question for you.

    Page 11 in the LM5008A Datasheet describes diode D1 as follows:

    The important parameters are reverse recovery time and forward voltage. The reverse recovery time determines how long the reverse current surge lasts each time the buck switch is turned on. The forward voltage drop is significant in the event the output is short-circuited as it is only this diode’s voltage which forces the inductor current to reduce during the forced off-time. For this reason, a higher voltage is better, although that affects efficiency. A good choice is a Schottky power diode, such as the DFLS1100. D1’s reverse voltage rating must be at least as great as the maximum VIN, and its current rating be greater than the maximum current limit threshold (610 mA).

    QUESTION: Does Vf really need to be as high as the DFLS1100, or could the B2100-13-F be used instead and still work fine in cases where the output voltage is shorted to GND?  I typically use the B2100 as a reverse polarity protection diode on Vin and it makes sense to use the same diode as D1 too.  Below are the two diodes compared:

  • Hello James,

    Your Schottky is OK. Forward drop is high enough and it has no kink in the forward characteristic indicating junction diode behavior in the guard rings of the diode.

    My comment concerning output capacitor inductance is aimed at the ~1 – 2 nF of series inductance associated with output capacitor. Since your relay will like have 1-10 mH of parallel inductance, it should not interfere with the operation of your converter. The time constant of the output inductor with the output capacitor will be much shorter that the time constant of the relay with the output capacitor so the output inductor’s time constant will dominate and is already addressed.

    An electrolytic cap next to the relay and a freewheeling diode are good ideas. You may also control the slew of the device turning off the relay to prevent excessive voltage. Most relay drivers have this feature built in.

    Regards, Robert

  • Thank you, Robert!