From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mout.gmx.net ([212.227.15.15]:55269 "EHLO mout.gmx.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755168AbaAVLxi (ORCPT ); Wed, 22 Jan 2014 06:53:38 -0500 Received: from minime.bse ([77.20.120.199]) by mail.gmx.com (mrgmx102) with ESMTPSA (Nemesis) id 0MbJTE-1VnJal3xQa-00In1z for ; Wed, 22 Jan 2014 12:53:36 +0100 Date: Wed, 22 Jan 2014 12:53:34 +0100 From: Daniel =?iso-8859-1?Q?Gl=F6ckner?= To: Robert Longbottom Cc: "linux-media@vger.kernel.org" Subject: Re: Conexant PCI-8604PW 4 channel BNC Video capture card (bttv) Message-ID: <20140122115334.GA14710@minime.bse> References: <52DD977E.3000907@googlemail.com> <1c25db0a-f11f-4bc0-b544-692140799b2a@email.android.com> <7D00B0B1-8873-4CB2-903F-8B98749C75FF@googlemail.com> <20140121101950.GA13818@minime.bse> <52DECF44.1070609@googlemail.com> <52DEDFCB.6010802@googlemail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <52DEDFCB.6010802@googlemail.com> Sender: linux-media-owner@vger.kernel.org List-ID: On Tue, Jan 21, 2014 at 08:59:55PM +0000, Robert Longbottom wrote: > On 21/01/2014 19:49, Robert Longbottom wrote: > >Here are some high-res pictures of both sides of the card. Scanned at > >600dpi (300dpi the tracks were very close). Good idea to scan it by the > >way, I like that, much better result than with a digital camera. > > > >http://www.flickr.com/photos/astrofraggle/12073752546/sizes/l/ > >http://www.flickr.com/photos/astrofraggle/12073651306/sizes/l/ ok: - The Atmel chip is an AT24C02 EEPROM. Does one of the 878As have a PCI subsystem ID? - The 74HCT04 is used to drive the clock from the oscillator to the 878As. - The 74HCT245 is a bus driver for four pins of the connector CN3. - The unlabled chip is probably a CPLD/FGPA. It filters the PCI REQ# lines from the 878As and has access to the GNT# and INT# lines, as well as to the GPIOs you mentioned. The bypass caps have a layout that fits to the Lattice ispMACH 4A. - There is no mux or gate between the BNC connectors and the 878As. The BNCs are on MUX0. MUX1 is connected to the two unpopulated 2x5 Headers. So the UNKNOWN/GENERIC card entry should have the BNC connectors on its first V4L input. Have you tried passing pll=35,35,35,35 as module parameter? Daniel