Author Topic: Faster 74HC595  (Read 1634 times)

0 Members and 1 Guest are viewing this topic.

Offline OwOTopic starter

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Faster 74HC595
« on: February 14, 2020, 12:48:55 pm »
I need to drive a slow (30MHz) parallel interface with ~26 pins some distance away from a FPGA, and want to avoid using 26 IOs. The idea is to use shift registers to deserialize the data at the receiver side. The lines will go over fpc cable and be driven differentially, and will be terminated on the FPGA side. So far I've found the 74LVC595 which goes up to 120MHz, but this is still a bit slow for my liking (I would prefer at least 8x serialization factor or 240MHz). Is there a faster 595 or other shift register IC? I would like to keep chips cost below $2 for the whole solution.
Email: OwOwOwOwO123@outlook.com
 

Offline Wolfram

  • Frequent Contributor
  • **
  • Posts: 414
  • Country: no
Re: Faster 74HC595
« Reply #1 on: February 14, 2020, 01:53:22 pm »
A GMSL or FPD-link II receiver could be an option, if you have enough FPGA resources available to implement the transmitter.
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 7375
  • Country: de
Re: Faster 74HC595
« Reply #2 on: February 14, 2020, 02:16:22 pm »
ECL might be a tad inconvenient, but gets you to 2.8 GHz shift clock:  ::)
https://www.onsemi.com/pub/Collateral/MC10EP142-D.PDF

Edit: Ah, sorry. Just realized that they are no longer available from ON, but only from Rochester Electronics, at boutique prices...
« Last Edit: February 14, 2020, 02:19:34 pm by ebastler »
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Faster 74HC595
« Reply #3 on: February 14, 2020, 02:17:13 pm »
You can take a look at Maxim's MAX9206/9208: https://www.maximintegrated.com/en/products/interface/high-speed-signaling/MAX9206.html

I'm sure other vendors have similar ICs as well if you're not comfortable with Maxim.
 

Offline OwOTopic starter

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Faster 74HC595
« Reply #4 on: February 14, 2020, 03:03:29 pm »
ECL input isn't a problem; actually it is desirable because it's easier to terminate correctly. (LVCMOS can only be source-terminated for the receiver to see correct voltage levels). However I do need 3.3V CMOS outputs to drive the parallel interface.

MAX9206 looks like it would do the job if I can find it for a good price (3 are needed to get 30 bits).

I see a lot of LVDS receiver chips for really cheap ($0.5), but need to find datasheets... Not sure if these can be used as dumb deserializers (they are usually for driving LCDs).
Email: OwOwOwOwO123@outlook.com
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Faster 74HC595
« Reply #5 on: February 14, 2020, 03:10:32 pm »
MAX9206 looks like it would do the job if I can find it for a good price (3 are needed to get 30 bits).

Yep, those are a bit pricy. Look for "deserializer" with the usual vendors and see what you get.

I see a lot of LVDS receiver chips for really cheap ($0.5), but need to find datasheets... Not sure if these can be used as dumb deserializers (they are usually for driving LCDs).

LVDS receiver chips in general are just the differential front-end and do not include any deserializer, hence the low cost, so don't keep your hopes high too much on those.
 

Online Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: Faster 74HC595
« Reply #6 on: February 14, 2020, 06:56:35 pm »
Look for LVDS serializer/deserializers. Try TI, they took over a lot of the National parts.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17428
  • Country: us
  • DavidH
Re: Faster 74HC595
« Reply #7 on: February 16, 2020, 10:31:48 am »
Why not deserialize in two stages?  Use a very fast deserializer, perhaps made up of ECL flip-flops since they will be easier to get, to then drive multiple 74LVC595s?
 

Offline chris_leyson

  • Super Contributor
  • ***
  • Posts: 1549
  • Country: wales
Re: Faster 74HC595
« Reply #8 on: February 16, 2020, 10:57:28 am »
You could try something like an SN65LVDS94 serdes receiver, it will give you 4:28 channel expansion and a parallel data rate of up to 68MHz. You would only need four IOs on the FPGA and a few differential LVDS drivers. EDIT: These, or similar, are often used for SDI video LVDS interfaces.
« Last Edit: February 16, 2020, 11:01:25 am by chris_leyson »
 

Offline OwOTopic starter

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Faster 74HC595
« Reply #9 on: February 16, 2020, 01:10:29 pm »
Thanks for the suggestions, I just found a whole bunch of LVDS receiver chips that are pin compatible with the SN65LVDS94 (MS90C386, GM8284C, etc). They all have a deserialization factor of 7 and have the same bit layout, which seems to originate from the DS90CR286.
I'll probably use this one which is cheap enough and goes to 175MHz (1.2Gb/s): https://lcsc.com/product-detail/LVDS-ICs_Hangzhou-Ruimeng-Tech-MS90C386B_C231931.html
Email: OwOwOwOwO123@outlook.com
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf