Author Topic: Inductor value choice for PI3526 Buck DCDC module?  (Read 291 times)

0 Members and 1 Guest are viewing this topic.

Offline zenerbjt

  • Regular Contributor
  • *
  • Posts: 68
  • Country: gb
Inductor value choice for PI3526 Buck DCDC module?
« on: September 06, 2020, 04:04:00 pm »
Dear Engineers,
The PI3526 Buck DCDC module looks like an Ace from Vicor.
We wish to use it with an external inductor we design ourselves instead of the recommended one by Eaton.
Please answer some questions about the choice of inductor value  for  the PI3526?

PI3526  Buck DCDC module datasheet
http://www.vicorpower.com/documents/datasheets/ds_PI352x-00.pdf

No Need to read the below if you already know info on inductor choice for PI3526,  and may kindly share it here.
_____-------_____------______------_______-------______-------______

Our Spec:
Vin = 14S (42V-59V), Vout = 12V, 11A.
 
Inductor choice for PI3526 Buck:

With “nominal” Vin=48V,Vout=12V,Iout=18A,Fsw=450kHz……and the datasheet recommended 480nH inductor, the Buck would be just inside DCM from 34-60Vin.
As such, I am guessing that the PI3526 Buck is meant to be operated in DCM, or rather, very close to the CCM/DCM boundary? Fig 79 of datasheet shows the PI3526 switching frequency being increased as load  reduces…..thus keeping the Buck near the CCM/DCM border and preventing too high peak currents.
This seems to be saying that the internal diodes in the PI3526 are not Schottky type diodes?.... since if they were Schottky, then there wouldn’t be much need to operate in DCM or close to it.

Please could you confirm that the internal diodes in the PI3526 are not Schottky type? We ask because if they are Schottky type, then we will pick an inductor which is higher value than 480nH, and have the Buck operating  well into  CCM, so as to reduce the peak inductor current, and reduce FET RMS currents. However, as you know, operating in CCM can result in more “ground bounce” and general noise, which tends to need slightly more LEB time, and a bit stronger current sense filtering….if the PI3526 module wont be able to tolerate this extra noise, then it would be a poor choice for us to pick a higher value inductor and operate in deeper CCM.

Also, as you know, the dynamics change between CCM and DCM operation. The inductor value is also a feedback loop parameter. So we do not want to push it toward instability by selecting the wrong value inductor. Its not possible to adjust the  feedback compensation components in PI3526, as they are all inside the module, "locked away".

Also, are we correct in assuming that if we select an inductor  of say 3uH for operation at Vin=48V, Vout =12V 11A,  then the switching frequency would be 450kHz (the lowest available)? (it would be in CCM at 450kHz with I(peak) = 14A and I(pedestal) = 7.5A. If the internal electronics in the PI3526  “waits” for the discharge of the inductor before turning the FET back ON, then the frequency would be even lower than 450kHz. Please could you  advise on whether the PI3526 has such an “inductor discharge detector”?
Please could you also advise what is the lowest possible switching frequency that the PI3526 would operate at? ie, would it ever go below the lowest F(sw) shown on Fig 79, page 27?

So, we don’t want to be “tied in” to a  particular  Offtheshelf inductor, due to Lead times and availability. We would need to design our own inductor.

So to summarise, please may we request what is the freedom of choice in inductor value selection for PI3526?..... bearing in mind that our  operating conditions are Vin=42-59V,Vout=12v, 0-11A?


 

Offline tom66

  • Super Contributor
  • ***
  • Posts: 4434
  • Country: gb
  • Electron Fiddler, FPGA Hacker, Embedded Systems EE
Re: Inductor value choice for PI3526 Buck DCDC module?
« Reply #1 on: September 06, 2020, 04:48:44 pm »
When you start asking questions like:

"Please could you confirm that the internal diodes in the PI3526 are not Schottky type?"

that is where you ask a FAE/sales/tech at Vicor.

P.S. I would think that the chances that for an 11A output, that anything other than a Schottky or sync FET is used, is close to zero.
 
The following users thanked this post: zenerbjt

Offline zenerbjt

  • Regular Contributor
  • *
  • Posts: 68
  • Country: gb
Re: Inductor value choice for PI3526 Buck DCDC module?
« Reply #2 on: September 06, 2020, 07:19:44 pm »
Thanks, yes, it definetely uses sync FETs...but as you know, the diode parallel to the FET conducts just before the FET is turned ON, and again  just after the sync fet is turned off, so if its not Schottky, then there will be  significant reverse recovery if done in CCM.

I think FAE's from any company only get very interested when you have big orders ready to go.
 

Offline butech

  • Contributor
  • Posts: 11
  • Country: de
Re: Inductor value choice for PI3526 Buck DCDC module?
« Reply #3 on: September 09, 2020, 07:51:39 am »
I think you will not have much luck, if you choose a much larger inductance value. The regulator (as I understand it) is designed to operate with ZVS (zero voltage switching), without any additional external networks (as the datasheet indicates). Therefore you will need such a large ripple in the inductor current, that the sync fet current becomes positive (from drain to source) prior to its turn-off instant. This will allow the (negative) inductor current to charge/discharge the main switching node and therefore ZVS will be achieved. If you try to force the converter into CCM than you will defeat its initial design objective, as ZVS of the main switch will not be possible. Therefore I think you have to use an inductor with a low inductance value, with low AC losses (large current swing).
 
The following users thanked this post: zenerbjt


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf