Nut upsuser - Oct 2006 - Adding TrippLite SMART550 / Protocol 2001 Support

I have added preliminary support for the TrippLite protocol number 2001 
into tripplite_usb.c.  The attached file is supplied as a patch against 
today's SVN.

This patch adds support for TrippLite SMART550USB and some Omni models.

Tested are On Line, On Battery, Battery Good, and Battery Bad indication. 
Everything else seems to work, but this UPS is attached to a critical 
system, and I can only do limited testing.

As an aside, tripplite_usb.c seems to need nut ownership of 
/proc/bus/XXX/YYY in order to work, as opposed to /dev/hid/hiddevX.  It 
would be nice to get this "fixed".

------------------------------------------------------------
               Anthony Ciani (aciani1@uic.edu)
            Computational Condensed Matter Physics
    Department of Physics, University of Illinois, Chicago
               http://ciani.phy.uic.edu/~tony
------------------------------------------------------------
-------------- next part --------------
--- nut.20061011/drivers/tripplite_usb.c	2006-10-11 14:51:04.000000000 -0500
+++ nut.20061004/drivers/tripplite_usb.c	2006-10-04 20:41:57.000000000 -0500
@@ -243,7 +243,8 @@
 #include <math.h>
 #include <ctype.h>
 
-static enum tl_model_t { TRIPP_LITE_UNKNOWN = 0, TRIPP_LITE_OMNIVS,
TRIPP_LITE_SMARTPRO }
+static enum tl_model_t { TRIPP_LITE_UNKNOWN = 0, TRIPP_LITE_OMNIVS,
TRIPP_LITE_OMNIVS2,
+TRIPP_LITE_SMARTPRO}
 tl_model = TRIPP_LITE_UNKNOWN;
 
 /*!@brief If a character is not printable, return a dot. */
@@ -387,6 +388,9 @@
 		case 0x1001:
 			upslogx(3, "Using OMNIVS protocol (%x)", proto);
 			return TRIPP_LITE_OMNIVS;
+		case 0x2001:
+			upslogx(3, "Using OMNIVS2 protocol (%x)", proto);
+			return TRIPP_LITE_OMNIVS2;
 		case 0x3003:
 			upslogx(3, "Using SMARTPRO protocol (%x)", proto);
 			return TRIPP_LITE_SMARTPRO;
@@ -861,7 +865,7 @@
 		return;
 	}
 
-	if(tl_model != TRIPP_LITE_OMNIVS) {
+	if(tl_model != TRIPP_LITE_OMNIVS && tl_model != TRIPP_LITE_OMNIVS2) {
 		dstate_setinfo("ups.debug.S","%s", hexascdump(s_value+1,
7));
 	}
 
@@ -893,6 +897,44 @@
 		}
 	}
 
+
+	if(tl_model == TRIPP_LITE_OMNIVS2) {
+		switch(s_value[2]) {
+			case '0':
+				dstate_setinfo("battery.test.status", "Battery OK");
+				break;
+			case '1':
+				dstate_setinfo("battery.test.status", "Battery bad -
replace");
+				break;
+			case '2':
+				status_set("CAL");
+				break;
+			case '3':
+				status_set("OVER");
+				dstate_setinfo("battery.test.status", "Overcurrent?");
+				break;
+			case '4':
+				/* The following message is confusing, and may not be accurate: */
+				/* dstate_setinfo("battery.test.status", "Battery state
unknown"); */
+				break;
+			case '5':
+				status_set("OVER");
+				dstate_setinfo("battery.test.status", "Battery fail -
overcurrent?");
+				break;
+			default:
+				upslogx(LOG_ERR, "Unknown value for s[2]: 0x%02x", s_value[2]);
+				dstate_datastale();
+				break;
+		}
+
+		/* Online/on battery: */
+		if(s_value[4] & 1) {
+			status_set("OB");
+		} else {
+			status_set("OL");
+		}
+	}
+
 	if(tl_model == TRIPP_LITE_SMARTPRO) {
 		switch(s_value[2]) {
 			case '0':
@@ -949,7 +991,7 @@
 
 	status_commit();
 
-	if( tl_model == TRIPP_LITE_OMNIVS ) {
+	if( tl_model == TRIPP_LITE_OMNIVS || tl_model == TRIPP_LITE_OMNIVS2 ) {
 		ret = send_cmd(b_msg, sizeof(b_msg), b_value, sizeof(b_value));
 		if(ret <= 0) {
 			dstate_datastale();
@@ -1037,6 +1079,7 @@
 
 	switch(tl_model) {
 		case TRIPP_LITE_OMNIVS:
+		case TRIPP_LITE_OMNIVS2:
 			dstate_setinfo("output.voltage", "%.1f",
hex2d(l_value+1, 4)/2.0);
 			break;
 		case TRIPP_LITE_SMARTPRO:
@@ -1047,7 +1090,7 @@
 			break;
 	}
 
-	if(tl_model != TRIPP_LITE_OMNIVS) {
+	if(tl_model != TRIPP_LITE_OMNIVS && tl_model != TRIPP_LITE_OMNIVS2) {
 		debug_message("D", 2);
 		debug_message("V", 2);

On 10/11/06, Anthony J. Ciani <aciani1@uic.edu>
wrote:
> I have added preliminary support for the TrippLite protocol number 2001
> into tripplite_usb.c.  The attached file is supplied as a patch against
> today's SVN.
>
> This patch adds support for TrippLite SMART550USB and some Omni models.
>
> Tested are On Line, On Battery, Battery Good, and Battery Bad indication.
> Everything else seems to work, but this UPS is attached to a critical
> system, and I can only do limited testing.
Cool!
> As an aside, tripplite_usb.c seems to need nut ownership of
> /proc/bus/XXX/YYY in order to work, as opposed to /dev/hid/hiddevX.  It
> would be nice to get this "fixed".
The hiddev interface is not portable to other OSes, so we use libusb
instead. (See the nut-upsdev archives for more info on hiddev.) We are
slowly adding hotplug and udev rules for all of the NUT-supported
devices, and we should be able to change permissions for the TrippLite
UPSes with those rules.

-- 
- Charles Lepple

On 10/11/06, Anthony J. Ciani <aciani1@uic.edu>
wrote:
> I have added preliminary support for the TrippLite protocol number 2001
> into tripplite_usb.c.  The attached file is supplied as a patch against
> today's SVN.
>
> This patch adds support for TrippLite SMART550USB and some Omni models.
I merged this into SVN as rev 556. There are also a few updates to the
way that the protocol is reported, to help others figure out what
works with their UPS.
> Tested are On Line, On Battery, Battery Good, and Battery Bad indication.
> Everything else seems to work, but this UPS is attached to a critical
> system, and I can only do limited testing.
If you (or anyone else reading this) gets a chance to test the
low-battery status, that would wrap up the basic testing.
> As an aside, tripplite_usb.c seems to need nut ownership of
> /proc/bus/XXX/YYY in order to work, as opposed to /dev/hid/hiddevX.  It
> would be nice to get this "fixed".
I looked into this, and if you put the nut-usbups.rules file in
/etc/udev/rules.d (as tested on Ubuntu Dapper), it should change the
group of the node when you plug the UPS in. Other distributions will
likely have slightly different ways of accomplishing this.

-- 
- Charles Lepple