Author Topic: FSFR2100XS  (Read 2422 times)

0 Members and 1 Guest are viewing this topic.

Offline Marian_elfTopic starter

  • Contributor
  • Posts: 17
  • Country: ro
FSFR2100XS
« on: September 28, 2023, 08:51:54 am »
Hello.

I hope this is the right place for this topic, if not then i do apologize, but it did seamed the best place for it.
Anyway, my question is simple, does anyone know why Onsemi gave up on FSFR2100 series IC's and if there are other similar LLC controllers wit integrated mosfets?
LLC seams the obvious choice due to it's better performance and smaller size, and an IC with integrated mosfets saves pcb space and reduces the BOM costs.
So, why would they give up on this?
Is it because of e newer/better variant, or did they lost interest?

All the best.
"What's right is what's left if you do everything else wrong" - Robin Wiliams.
 

Offline Wolfram

  • Frequent Contributor
  • **
  • Posts: 382
  • Country: no
Re: FSFR2100XS
« Reply #1 on: September 28, 2023, 10:42:10 am »
Ah, that sucks. I've used the 2100 in a few products and it's such a clean solution. Probably the market didn't justify keeping such a product in production. Most commercial volume LLCs in the 200+ W class will require active PFC, which means you practically can't avoid having multiple power devices to cool anyways, and the benefit of replacing a pair of TO-220 or D2PAK MOSFETs with a single larger package incorporating the controller is limited. For smaller volume projects where RND is a major part of the cost, being able to do a multi-hundred watt LLC with a single chip in an isolated package and internal thermal sensing is very valuable, but probably these markets are not big enough to justify keeping a packaging line going for such a unique multi-die power package.

There's still the Power Integrations HiperLCS series of chips, but these look a bit more awkward to mount. I would consider just going for the FAN7621 series of controllers with external MOSFETs, or any of the alternatives from TI, onsemi, Infineon and the likes.
 

Offline Marian_elfTopic starter

  • Contributor
  • Posts: 17
  • Country: ro
Re: FSFR2100XS
« Reply #2 on: September 28, 2023, 11:55:24 am »
Thank you for your input.

I had thought about FAN7621 before this but on it's page on the Onsemi website, there is no info for the series, just the datasheet.
Digikey though says this one is obsolete as well, so...bad news.
I did asked Onsemi about this some time ago, but they didn't clarify much of anything, they just said that the closest replacement active part is NCP13992ACDR2G ( fact confirmed by Digikey as well ).
It is an active part but it seams much to involved for less than 300W ( comparing with 2100 ).
Way to complex for my taste anyway.
« Last Edit: September 28, 2023, 11:56:57 am by Marian_elf »
"What's right is what's left if you do everything else wrong" - Robin Wiliams.
 

Offline Faringdon

  • Super Contributor
  • ***
  • Posts: 1982
  • Country: gb
Re: FSFR2100XS
« Reply #3 on: September 30, 2023, 04:38:25 pm »
The 1000 series was good, but only if super well layed out, due to the onchip bootstrap hi side drive.....these things drive off 400v...and if bootstrap isnt perfectly layed out at that voltage, they just blow up.
Youre better off with discrete solution and gate drive transformers...or high side drive supply and signal referral up to it...via suitable digi isolator type chip with good v/ns
« Last Edit: September 30, 2023, 04:39:59 pm by Faringdon »
'Perfection' is the enemy of 'perfectly satisfactory'
 

Offline Marian_elfTopic starter

  • Contributor
  • Posts: 17
  • Country: ro
Re: FSFR2100XS
« Reply #4 on: October 01, 2023, 11:43:59 am »
I do have a number of 2100XS IC's from Onsemi and i have tested one on an improvised board and it functioned quite well, no magic smoke, good waveforms, and all. There can still be found some of them on Digikey, Mouser and Farnell i think.
I guess for a low series production they'll do just fine.

Gate drive transformers with all their discrete drivers and stuff, require too much pcb space to be financially viable at such a low power level.
But thanks anyway.
"What's right is what's left if you do everything else wrong" - Robin Wiliams.
 

Offline Faringdon

  • Super Contributor
  • ***
  • Posts: 1982
  • Country: gb
Re: FSFR2100XS
« Reply #5 on: October 01, 2023, 12:38:14 pm »
Thanks, but the bootstrap problem , can be one of progressive weakening of the part, due to the  negative spike voltage they get repeatedly...you are likely to possibly see premature failure...say after 2 years or so.
You wont get immediate magic smoke...it would indeed be better if we did see that...it is one of those incidious progressive weakening problems.

In here you can read of the CMTI problem in these bootstrappers...
https://www.infineon.com/dgdl/Infineon-Advantages_of_high-voltage_GD_ICs-ApplicationNotes-v02_00-EN.pdf?fileId=5546d4626e41e490016e5967d5894033

Read here page 16 of the demon bootstrapper problem, in this case with NCP5106...
https://www.onsemi.com/pdf/datasheet/ncp5106-d.pdf

...If after the above, you still havent turned your back on bootstrapping with 400V inputs...then read pg 15-16 of Dr Ray Ridley...
"Gate Driver design tips" Dr Ray Ridley
https://imserv.org/MFKForum/Datasheets/gate%20drive/gate%20drive%20design%20tips.pdf

...leave well alone.....embrace  the forgiven  Gate drive transformer.

China are the SMPS experts of the world these days......take apart any  Chinese PFC'd smps that you  like...you wont see bootstrap hi side drives on the post PFC smps stage...you will see gate drive transformers.
« Last Edit: October 01, 2023, 01:13:23 pm by Faringdon »
'Perfection' is the enemy of 'perfectly satisfactory'
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf