Jump to content

various SRIO questions and NG hadware design


Recommended Posts

Posted

hey all

 

I'm just starting some "bus board" design for my moving fader and I just want to be sure about some details

 

I have some questions regarding SRIO chain, J8/9 and J1 J2

-Is it right that only RC1 is used for all Din and Dout? Why then rev5 Din/Dout fitted with J1/J2 have RC1 and RC2 linked ???

-If they are no use now for RC2, can I remove it from the line driver and use the channel for SI data?

-Do I understand right the only reason Din and Dout use Y cable is because, SI is not linked betwin J1 and J2 at the Dout pcb, and SO is not linked between J1 and J2 at the Din pcb,

Resulting for example, that I can just add a jumper-wire between SI J1 and SI J2 at a Dout card and link a Din at J2 in place of Y cable at J1

 

Due to design restriction I'm not able to respect J1/J2 spec for my bus connection, but I think i'm able to have J1 at start and J2 at end so the whole thing should be compatible in a global chain :smile:

 

Best

 

Zam

Posted

Hi,

 

with MBHP_CORE_LPC17 (and later MBHP_CORE_STM32F4) I started to separate the RC pins at the core side for more flexibility (e.g. so that the SPI interface could be used for other purposes).

So: on the rev5 DIN/DOUT board the bridge between the RC pins exists due to historical reasons, but it wouldn't be required anymore.

On the other hand, only a single RC line is required for both chains (to answer the actual question)

 

 

-If they are no use now for RC2, can I remove it from the line driver and use the channel for SI data?

 

No, because SI has a different direction (it is driven by the 74HC165 chip and not by the microcontroller)

 

 

-Do I understand right the only reason Din and Dout use Y cable is because, SI is not linked betwin J1 and J2 at the Dout pcb, and SO is not linked between J1 and J2 at the Din pcb,

 

no, this isn't correct (see answer above).

 

The main reason for the Y cable is, that this is the best way to get two independent serial register chains.

Mixed and/or unbalanced chains can result into electrical problems when multiple chips are connected together.

 

However, if you use line drivers between the "bus boards", potential electrical problems are "localized" and therefore less relevant.

But for common users who use MBHP_DIN and MBHP_DOUT modules the Y cable is "best practice"

 

Best Regards, Thorsten.

Posted (edited)

hello Thorsten

 

tks for clarification !!!

 

I still have a problem to see the difference between:

 

                         --> J1 Dinx1 J2--> J1 Dinx1 J2 --> etc...

J8/9 -->Ycable

                         --> J1 Doutx1  J2--> J1 Doutx1 J2 --> etc...

 

And:

 

J8/9-->J1 Dinx1 (with direct link J1SO-->J2SO) J2-->J1 Doutx1 (with direct link J1SI-->J2SI) J2--> etc...

 

 

 

Anyway i'm not sure you say there is some? maybe I don't ask the question in the right way, you just say this could run some data transfer qualities problems

 

Best

 

Zam

Edited by Zam
Posted

With such a configuration there is a certain danger for signal propagation delay issues with the result that the data value of one register won't be taken over by the following one with the rising clock edge (-> risk for a setup or hold violation).

It might work at your side, but it could also fail. This depends on the wire lengths, impedances, ambient conditions.

 

It's hard to calculate this (specially if cables and connectors are involved), or to give you a rule of thumb what is allowed or not.

But I remember that many years ago some MIDIbox SEQ users noticed problems, such as flickering LEDs or random button events, due to a wild mix of DIN and DOUT SRs similar to this circuit http://www.midibox.org/midibox_seq_photos/3.html but just a little bit different (maybe longer cables between the SRs, or whatever)

 

For the BLM16x16+X project I connected the DIN and DOUT registers Y like, and it works very robust: 

 

So - from my experience - I've to say that this seems to be a good way for stable shift register transfers.

 

You might find some literature about this topic via Google to get an independent opinion, and maybe more hints how to get your layout "first time right".

 

Best Regards, Thorsten.

Posted

Tks a lot for your input on this

Your right, I need for sure more literature around this, especially layout. I will.

 

The actual layout is close to 8 "DIOmatrix" (but with only one165 and one 595) daisy chained with 20cm wire between each (2x 10cm ribbon and 4 connector :wacko: ).

Maybe it's little ambitious, but for the moment it's my first option due to hardware restriction.

 

I have to check thing to be sure it's possible at my side, but second option is to run the long ribbon between diode/buttons and shift register in/out

 

Tks again, your so helpful. :sorcerer:

Best

Zam

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...