Author Topic: This is becoming quite a challenge, but I managed it.  (Read 5528 times)

0 Members and 1 Guest are viewing this topic.

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 8790
  • Country: fi
Re: This is becoming quite a challenge, but I managed it.
« Reply #25 on: November 22, 2018, 07:33:09 am »
Simple series resistors (say 47R) offer the benefit of providing some protection while being almost as cheap as no protection at all. 47R series + external schottkys to rails was used in Terasic's DE2 development board, for example, which we used extensively.

These small resistors protect from short circuits which IMO are the most probable way of killing IO on the "lab table". They also may protect from overcurrent damaging things through substrate diodes when you accidentally put a voltage on an IO pin of turned-off device - they also protect from small overvoltages such as 5V on 3.3V input, by limiting the IO cell diode current.

They won't properly protect from ESD, or severe overvoltages. But these resistors are almost free, and they prevent ringing of fast output signals by working as series termination, without slowing down the signal too much.
 
The following users thanked this post: helius

Offline technixTopic starter

  • Super Contributor
  • ***
  • Posts: 3508
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
Re: This is becoming quite a challenge, but I managed it.
« Reply #26 on: November 22, 2018, 08:24:30 am »
Simple series resistors (say 47R) offer the benefit of providing some protection while being almost as cheap as no protection at all. 47R series + external schottkys to rails was used in Terasic's DE2 development board, for example, which we used extensively.

These small resistors protect from short circuits which IMO are the most probable way of killing IO on the "lab table". They also may protect from overcurrent damaging things through substrate diodes when you accidentally put a voltage on an IO pin of turned-off device - they also protect from small overvoltages such as 5V on 3.3V input, by limiting the IO cell diode current.

They won't properly protect from ESD, or severe overvoltages. But these resistors are almost free, and they prevent ringing of fast output signals by working as series termination, without slowing down the signal too much.
I have 22R and ESD clamp diode packs, but those won’t fit here.

The boards arrived and preliminary testing suggests it works.
 

Offline technixTopic starter

  • Super Contributor
  • ***
  • Posts: 3508
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
Re: This is becoming quite a challenge, but I managed it.
« Reply #27 on: November 22, 2018, 11:31:33 am »
That board came back finally, and my first unit survived a tour in the city in my pocket.

I wonder what should the MCU provide to the FPGA? USB CDC is an easy one, ditto the 36MHz core clock. Now how should I make use of the SPI interface?
 

Offline technixTopic starter

  • Super Contributor
  • ***
  • Posts: 3508
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
Re: This is becoming quite a challenge, but I managed it.
« Reply #28 on: November 28, 2018, 04:39:33 am »
I will make it available on my Tindie store for you, after the boards arrived and assed tests. Just do expect it being a bit expensive though, FPGA's aren't exactly cheap.

I was just looking at the state of things this morning and was a bit shocked at just how cheap they've become. The lowest end Cyclone 10 part for example is under $8 each in single quantities and it has quite respectable specs. I've been using the same older parts for quite some time now and had not bothered to check on the current offerings in a while.

Here it is: https://www.tindie.com/products/maxtch/sushibits-duplex-mini-2/
 

Offline technixTopic starter

  • Super Contributor
  • ***
  • Posts: 3508
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf