Homebrew PCBs

>
> @Volkan Sahin:
> 
> Thanks for your response. If by some "IP problems" you mean problems 
> with Verilog/VHDL source for FPGA, I don't care too much about that.
> 
> I hope you haven't been bullied by Epson about protocol technicalities 
> and so you can clarify things there.
> 
> As things were presented, it looks as each piezo has two parts:
> 
> - one common to all piezos that is driven by a waveform on COM signal 
> that does actual pumping the ink either in or out
> 
> - per-nozzle signals that modulate common stream for each nozzle. 
> Nozzles are driven serially through C/M/Y/K channel, two bits per nozzle.
> 
> Start of the bits packet is signalled by "latch enable" and nozzles are 
> fired simoultaneously with NCHG.
> 
> Is this about right ?
> 
> As I understand your data, COM waveform has to be shifted in somehow, 
> and and you need 32 bits for that. Missing part is how and what exactly 
> do they mean.
> Looking at waveform, it seems that each waveform has less than 8 
> segments and if each one is described by 4 bits, probably 2 are for 
> length and 2 are for direction.
> 
> But looking at the scheme it seems signal is driven by the hardware 
> itself ( your board has push-pull for it), so it is unclear why and how 
> would one have to shift its digital representation to the chip and why.
> 
> Could you go over that part ( signal generation and their purpose ) more 
> thoroughly ?
>

 From: brane2 <brankob@...>
To: Homebrew_PCBs@yahoogroups.com 
Sent: Friday, July 19, 2013 11:08 AM
Subject: Re: [Homebrew_PCBs] Re: Driving EPSON printeahd drectly- volkan  sahin's email ?
 


  
@Volkan Sahin:

Thanks for your response. If by some "IP problems" you mean problems 
with Verilog/VHDL source for FPGA, I don't care too much about that.

I hope you haven't been bullied by Epson about protocol technicalities 
and so you can clarify things there.

As things were presented, it looks as each piezo has two parts:

- one common to all piezos that is driven by a waveform on COM signal 
that does actual pumping the ink either in or out

- per-nozzle signals that modulate common stream for each nozzle. 
Nozzles are driven serially through C/M/Y/K channel, two bits per nozzle.

Start of the bits packet is signalled by "latch enable" and nozzles are 
fired simoultaneously with NCHG.

Is this about right ?

As I understand your data, COM waveform has to be shifted in somehow, 
and and you need 32 bits for that. Missing part is how and what exactly 
do they mean.
Looking at waveform, it seems that each waveform has less than 8 
segments and if each one is described by 4 bits, probably 2 are for 
length and 2 are for direction.

But looking at the scheme it seems signal is driven by the hardware 
itself ( your board has push-pull for it), so it is unclear why and how 
would one have to shift its digital representation to the chip and why.

Could you go over that part ( signal generation and their purpose ) more 
thoroughly ?

 

[Non-text portions of this message have been removed]