2002 2003 2004 <2005> 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 | Index | 2002 2003 2004 <2005> 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 |
<== Date ==> | <== Thread ==> |
---|
Subject: | Re: Failure probing for event generator |
From: | Korhonen Timo <[email protected]> |
To: | Dayle Kotturi <[email protected]> |
Cc: | [email protected], EPICS core-talk <[email protected]> |
Date: | Fri, 23 Sep 2005 13:50:10 +0200 |
Hi Dayle,although the problem was solved with Ge, I summarize it here as other people may have interest in this topic also.
The card was not recognized because it needs a special configuration sequence. The EVG/EVR series 200 cards have been implemented according to the CR/CSR specification in VME64/VME64x standard (chapter 10 in the VME64x ANSI standard.) That is, before the cards are visible in the normally used access windows (A32,A24,A16,...) they have to be configured. This configuration happens in a separate address space which is geographically addressed (those who know PCI are familiar with the concept.) There are no jumpers or address switches to be set but on the other hand if the configuration is not done the cards are not visible in the normal windows. This would give many nice possibilities but unfortunately the concept is poorly or not at all supported in vxWorks at least (one can of course customize the BSP to have a window to the configuration address space.)
For the EVG/EVR 200-series we implemented a 'hack' where in the eg/erConfigure the Tundra registers are saved, modified to give access to the config address space, the card is configured and finally the Tundra registers are restored to the original values. This is not the best method but at least it works. Except for the problem that Ge had: the mvme5100 BSP has a bug (or poor implementation) of sysBusToLocalAdrs
routine and the dev support needed another hack to circumvent this bug. Would the CR/CSR concept be worth a short talk and a discussion in the EPICS meeting? best regards, Timo Dayle Kotturi wrote:
Hi Ge, Three things I'd suggest (based on my experience with 100 series hardware): 1) make sure your EVG is properly "seated" in the crate. Even if LEDs are on, I've found that board can be incorrectly seated. This happened with 64x VME crates. 2) ensure the top rotary dial is set to '2' and the lower rotary dial is set to '0' 3) change your card number from '4' to '0' if you only have one EVG Step (3) may not be necessary to get it working. You need to get a '0' return code from EgConfigure before you can get any further. The same suggestions apply to EVR (rotaries at '1' and '8', change card to '0'). For EVR only, check that jumpers are set to '5' as is documented on the board itself. Dayle Kotturi LINAC Coherent Light Source Stanford Linear Accelerator Center On Tue, 20 Sep 2005, Lei Ge wrote:Hello, I'm testing the event generator EVG200, but my IOC failed to probing the event-generator board. Could somebody give me some suggestions? Thank you very much. --------------------------------------------------------- EgConfigure(4,0x2000) value = -1 = 0xffffffff = prsrv_cast_client + 0xff0f7b3f ErConfigure(6,0x1800,164,5) value = -1 = 0xffffffff = prsrv_cast_client + 0xff0f7b3f cd "/home21/home/leige/epicsApp/dbd" value = 0 = 0x0 dbLoadDatabase("baseApp.dbd") devApsEg: Failure probing for event generator... Card disabled devApsEr: Failure probing for event receiver... Card disabled value = 0 = 0x0 dbLoadDatabase("evgevr.dbd") value = 0 = 0x0 ---------------------------------------------------------------- Best wishes, Ge
-- Timo Korhonen PSI (Paul Scherrer Institut) CH-5232 Villigen PSI tel + 41- 56 3103262 fax + 41 - 56 310 3383 e-mail: [email protected]