[sdiy] SPI comms - how fast is reasonable?

[grant at musictechnologiesgroup.com]

SPI is faster but the chips I've been using recently have all been I2C
that's all.

One-way is certainly more straight-forward since the master controls all
the data flow in SPI.  Without knowing how you are dealing with the data
on the receiver end it's hard to say more. I assume the receiver has
hardware SPI.  Is it setup to be interrupt or polled mode? Is the data
acted on immediately or queued? How big is a packet? How big is the
hardware receive buffer? etc.  

GB


> -------- Original Message --------
> Subject: Re: [sdiy] SPI comms - how fast is reasonable?
> From: Tom Wiltshire <tom at electricdruid.net>
> Date: Fri, March 05, 2010 11:28 am
> To: grant at musictechnologiesgroup.com
> Cc: synth-diy at dropmix.xs4all.nl
> 
> 
> Grant,
> 
> No, the comms is one-way. I want a main CPU to send parameter data  
> and note on/offs to the synth voice(s). Assuming I can be reasonably  
> sure the data will get through ok (no reason why it shouldn't,  
> right?) then I don't need anything coming back. If not, I might need  
> a "data not received correctly" message going the other way, but that  
> is to be avoided if at all possible.
> Also, I hear what you're saying about the bottleneck being software  
> not the actual transmission speed. I'll do what I can in that regard.
> 
> What are your reasons for using I2C not SPI? Isn't SPI faster? Am I  
> making the wrong choice?
> 
> Thanks,
> Tom