Re: [stella] Paddle controller docs?

Subject: Re: [stella] Paddle controller docs?
From: "Charles R. Bond" <cbond@xxxxxxxxxxxxx>
Date: Wed, 18 May 2005 15:20:30 -0400
(Excuse the top posting)

Thanks for the tips from you and others. It still seems strange that there isn't a 
consolidated technical document dealing with the paddles. It's OK to 'search the archives' 
or 'examine the schematics' for specifics, but you'd think that Atari would have published a
programmer's resource with more info...

Anyway, I did find that (according the AtariAge schematic) the paddles use 1meg pots.
On the other hand, I found another source which said they were 500k linear taper pots.
Apparently the caps are .068microfarad and are on the board with the chips. These
values are consistent with delay times from less than one scanline to somewhat more than
one frame -- as has been stated.


C. Bond

-----Original Message-----
From: Eric Ball <ericball@xxxxxxxxxxxx>
Sent: May 17, 2005 7:49 PM
To: stella@xxxxxxxxxxxxxxxxxx
Subject: Re: [stella] Paddle controller docs?

I recently did some proof-of-concept code for paddles on the 7800 (which
uses a normal TIA to interface with the paddles), so it's fairly fresh in my

From: "C. Bond"

> 1) Is there a schematic of the electrical circuit for the
> paddles, including component values?

Check the schematics on AtariAge (under 2600 archives).

> 2) What is the approximate minimum and maximum delay times
> from the start of charge to the activation of INPT0? (or
> INPT1, INPT2, etc.)

minimum is instant, maximum is more than 1/60th of a second (probably more
than 1/50th), but I don't think it's much more.

> 3) What is the best way to interrogate the paddles?

discarge the cap during VSYNC, clamp it near the start of your kernel, count
scan lines in the kernel

kernel code kinda like this
 STA WSYNC  ; synchronize to start of line
DLIP0 LDA DLI_A  ; restore A register

Although this is from a 7800 where cycle counting isn't necessary since you
can't update registers while the screen is being drawn.  2600 code should
implement standard equal path tricks.

Note: you don't have to sample INPTx every line, that just gives you the
greatest resolution.  A 2LK could sample every other line and get 120
positions instead of 240.

Archives (includes files) at
Unsub & more at

Archives (includes files) at
Unsub & more at

Current Thread