Communication Lost in a Smart UPS 1500 with USB

[Anonymous user]

We have an APC Smart-UPS SC 1500 for a Windows 2003 Server. We are using the USB/Serial converter cable because the server system only have a single serial port working for another application. The Powershute Business Edition connects and works good for a while, but every 2 or 3 days there is a "Communication Lost" alert in the console and we must to restart the APC services to restore the connection with the UPS.

If you know something that can help me to resolve and correct this, I'll appreciate any help. Thank You!

[Anonymous user]

Here's another one for you - communication is lost every couple of weeks (since install in Feb 08). The only thing that will get it working again is to unplug and then plug back in the USB cable (either on the server or on the UPS). The unit is remote to me, but thankfully there are a couple of savvy users I can get to do the job for me. Makes me look a bit unprofessional though!

Model of server/computer: Dell PE1900
Operating System: Windows Small Business Server 2003 R2
Powerchute Software version: Business Edition Basic (7.0.5.108)
Model/Serial number of UPS exhibiting the issue: Smart UPS 1500 (DLA1500I), AS0806230430
Please confirm you are ONLY using USB and NO serial communications on the UPS (Y/N): YES (USB only - the cable provided by APC)
Did you install the software with the same UPS connected? (versus using a different UPS to set up the software) (Y/N): YES
How did exactly you install this? (typical install, custom install, silent install, copied image from different machine, etc): Typical Install

The only other USB device that is plugged in to the server is an external USB hard drive (western digital).

Hopefully this will help. Please note that I've not tried the serial cable at all.
Thanks.

Message was edited by: jameshurrell

Message was edited by: jameshurrell

[BillP]

Hate to burst the bubble, but instead of 2 days it lasted about 4 and same issue again. Lost communication. Any solutions? Can't be rebooting a production server every 4 days.

[TheNotoriousKMP_apc]

What operating system is the server? This would be the first time the driver update didn't work after being applied to the prolific cable.

[BillP]

For all of you, who have remote systems, where USB cable unplugging is not so easy, try this (it works fine for me!):

1) Go to Device Manager and open Human Interface Devices
2) Right Click American Power Conversion USB UPS and select Disable
3) Wait a few seconds...
4) Right Click American Power Conversion USB UPS and select Enable

voila..

This reinstalls the driver for the device.

Good luck
Oliver

[BillP]

another thing to add to daily maintenance, awesome!

where are APC in all of this?

[Anonymous user]

Thanks for the tip Omerk.

I had a lost communication event this morning and i was all ready to try your tip. However, on remoting into the server I see that it had re-established communication all by itself. It has never done that before. Only 3 seconds between loss of communication and re-establishment. See attached Event Viewer screen shot.

I'd dearly love to get to the bottom of this annoyance. The only other event that occurs at exactly the same time as the loss of communication is a Success Audit log event in the Security log (538: anonymous logon). Doesn't give any further clues at to what's happening.

Message was edited by: jameshurrell

[BillP]

APC engineering is continuing to monitor this thread (as well as the cases we get via other methods of support) but we definitely need to know the operating systems being used that exhibit the problem and the exact model numbers and serial numbers ( located on the white stickers) on these UPSs so we can figure this issue out. we also need to know if there are any other USB devices you can tell us about or any other helpful details you can think of.

we appreciate those who have provided all of that information and will continue to work on it as long as the data keeps coming in to help. without it, it is difficult to pinpoint the issue unfortunately..

[BillP]

this will come across as somewhat rude but you guys have been "monitoring this thread" for something like 14 months and haven't come up with squat.

I've worked in retail and dealt directly with distribution in the past and when a manufacturer gets to this stage they start denying the problem and waiting for the units to all go out of warranty, never will i buy or recommend APC ever again, when such a basic part of the device is consistantly failing for so many users and this is the best you can come up with.

Hope someone got the boot over this one.

If there was a solution it should be here in this thread, a thread that's 14 months old now.

Message was edited by: power

[BillP]

I can sympathize with anyone experiencing a problem regarding USB and lost communication. I assure you that the problem is being investigated internally and the more reported cases that we receive have our utmost attention. I am not sure how else I can really word that.

Unfortunately, with a problem like this, there are several different factors that can contribute to the problem. ie. The UPS, Powerchute, the operating system, the USB cable, unsupported configurations, etc. We are still in the process of testing all factors in all combinations possible. Unfortunately, with a problem like this, there are also factors that can play a role that may cause our customers to see a problem that APC cannot initially reproduce. ie. Other server applications, server specific hardware, other USB devices, workload on the server in question, unsupported configurations, etc

In regards to your comment "when a manufacturer gets to this stage they start denying the problem and waiting for the units to all go out of warranty"..

APC is certainly not denying that there is a problem and APC is not waiting for the units to go out of warranty.

APC is trying to determine the exact cause of the problem. The delay in resolving this problem is not a lack of investigation, but is a lack of re-produceable results within our testing facility. We have engaged some of our customers in an effort to determine the problem. Though we have received some good information, we have been unable to determine the exact cause of the problem.

I understand your frustration on this matter and do apologize that you feel this way. I hope to have better news on this problem for you very soon but I realize that still may not change your opinion..

[BillP]

A SOLUTION WORTH TO TRY FOR ALL OF YOU WITH USB PROBLEMS!

I had problems with a HP DL385G2 Server and a Smart UPS-1000 RM connected via USB sinc April 2008. Communication was lost always after 5-15 min.. Replugging USB helped, but only for some minutes. APC did support me well, however, we could not figure out where my problem is. In December I have set up a secong HP DL385G5 with another Smart UPS-1000 RM and I ran into the same USB problems. In January I had the chance to reinstall another old server system with Windows Server2003 and I connected on of my UPS via UPS to it and it was running fine!!! I investigated differences between the installations and I found out that my two HP servers have USB 2.0 ports, while my old server only had USB 1.1 ports. I remembered that there is a BIOS setting at the HP servers named "USB port capability" which is can be set to "USB 2.0" or "USB 1.1". I have changed this setting on my servers to "USB 1.1" and now communication is stable.

So all of you having USB problems please check, if your UPS is connected to an USB 2.0 port and try when it is changed to "USB 1.1". Please report your experience here, maybe we can help many frustated APC UPS users.

To check on which USB port your UPS is connected go to "Device manager" and change "View" to "Devices by connection". Find the APC UPS and check the USB Host controller type. If it is connected to a "Standard Enhanced PCI to USB Host Controller" it is USB 2.0 and maybe the cause of your problem. See the attached pictures for details.

Find some USB Host Controller background info here: http://en.wikipedia.org/wiki/Host_controller_interface
Here is also some info from Microsoft: http://www.microsoft.com/whdc/connect/USB/USB2support.mspx?pf=true#usb2support

I hope this info helps many of you.

[Anonymous user]

Hi Omerk,

This is indeed interesting. I will try this on my Dell PE1900. I can only do it temporarily as I have a USB external drive also connected for nightly backups and the file transfer speed would be drastically reduced at USB1.1 rate I guess.

Many thanks!

[Anonymous user]

Forgive me if my admin skills are lax in this area, but how can I disable a device that isn't showing?

[BillP]

i also have a USB hard drive connected for backups so forcing my USB to 1.1 would not be a solution for me either.

[BillP]

OK, good point... 😉

It might be that your underlying problem might be different, or that your system behaves different than mine and removes the device automatically when there is a USB problem with the UPS.

So, at least you should replug USB once and check in the device manager what type of USB host the UPS is connected to according to my last post. If it is at an enhanced host, you should think about trying to change USB capabilities to 1.1, if possible.

As Angela from APC posted, there are for sure many individual causes for the USB problems. For my systems the underlying cause was USB2.0 incompatibility. This could explain the strange behavior that even on cloned systems some will work and some not, since BIOS settings might still be different.

Some words about APC support: I was contacted by APC soon after posting and we are having since very good contact about finding the cause for my problem. We have exchanged even USB communication protocols and enabled APC Agent logging. So from my point of view APC is really considering this problem with high priority and they are working on a solution. But they need your input to track the problem, so if you have your UPS running please post your solution here.

[Anonymous user]

I really find this whole situation rather disappointing. It was pointed out in this thread in a post by Squmph that

"The opinion of the technician I spoke to was that this is an issue with the UPS's USB 1.1 controller chip. So hold your breath we might all be in for a big swap if you're in warranty..."

Omerk's post pretty much confirms that the problem is the chipset doesn't work correctly with USB 2.0.

Is it really that difficult to fix this problem - every other usb device I've ever used functions perfectly in systems that are on 24x7. I can only presume is that the rather large cost of recalling all these units to add a working chipset mean that APC is hoping that by dragging it out so long they can instead charge us to fix the problem as the units will be out of warranty.

[BillP]

i just want to say, and I will keep saying it, APC is absolutely NOT waiting for these units to go out of warranty so they can make people pay for replacements or repairs. thanks to omerk for sharing his experience.

[Anonymous user]

I have just found this forum after suffering intermittent problems since November 2006!! I wonder if APC will honour my warranty?

I have searched a number of times but finally discovered this thread after I lost communication with the UPS at the weekend and suffered a power failure on Monday. The UPS failed to shutdown the server when I needed it most. I usually resolve this problem by disconnecting reconnecting the USB cable.

We are currently purchasing new server hardware and UPSs and unless this issue is resolved once and for all I will not be considering APC (American Powerless Corporation).

Model of server/computer: Dell PowerEdge 2900
Operating System: Windows Server 2003 Small Business Server SP2 v5.2
Powerchute Software version: 7.0.5 (Agent 7.0.5.108)
Model/Serial number of UPS exhibiting the issue: Smart-UPS 750/AS0626342187
Please confirm you are ONLY using USB and NO serial communications on the UPS (Y/N): Y
Did you install the software with the same UPS connected? (versus using a different UPS to set up the software) (Y/N): Y
How did exactly you install this? (typical install, custom install, silent install, copied image from different machine, etc): Typical Install

Message was edited by: jdjwright

Message was edited by: jdjwright

[BillP]

Hello

We have the same problem with Dell PowerEdge 300 with a APC 750 XL UPS. It just loses communcation and the only way I can re-establish communication is to unplub and replug the USB port. Has APC addressed this issue yet as it seems to have happened to alot of people.

Thanks

Bryan

[Anonymous user]

We have the same problem here. We have two Smart UPS 1500s and one experiences the problem and one does not. We get between 0 and 4 lost communication errors every day, for only ten seconds, which are emailed to myself. This only seems temporary and communication re-establishes itself but I get low battery warnings on my server and had to turn the power settings off to shut down the server when a low battery is experienced. I have change USB ports and it seems OK for a few days and triggers again. The problematic server is an Acer Altos G900 and is on the UPS by itself. The other one has more than one server on it and never experiences issues. Funnily enough the problematic server does not data log. Both are set up exactly the same.
I am going down the path of installing an interface expander and moving across to serial with this UPS and I will add another server to it so I can monitor all the servers and control shutdown all of them. This is another ridiculous feature that you can connect more than one server but can only control shutdown one unless you have an expander. This should be standard not just another money grabber for an obvious inferior product.

[BillP]

So is there any actual progress on USB only?

[TheNotoriousKMP_apc]

Hi,

If it's not the serial cable part # in my above post (940-0024) the UPS will turn off. It's a proprietary cable that sends the correct information on different pins than a straight-through.

[Anonymous user]

Hi,

We have the same symptoms ("PowerChute Business Edition has detected the following event: Communication Lost on Agent server01" every few days), but in our case connection is immediately and consistently reinstated on unplugging the USB cable at the UPS end and then plugging it back in again. (At first it was thought the cleaner had knocked the cable - now we know this is not the case.)

Smart-UPS 750 (SUA750I)
PowerChute BE 7.0.5 Build 108
Windows Small Business Server 2003 R2 Premium

Should we try the new Prolific PL-2303 Driver? The link leads to the installation kit, but no documentation. I am reluctant to install this on the server without further info. on exactly what it is. Does it replace a different driver or is it a new version of the driver already in use? What is the installation procedure, e.g. should the various PowerChute processes be stopped first?

JPL

[TheNotoriousKMP_apc]

Your SUA750I shouldn't be using a Serial To USB cable. It should be using either the 940-0024 serial cable that came with the UPS, or a standard USB-A to USB-B cable (that also came with the UPS).

In your case, you might want to brain-dead the UPS.

[Anonymous user]

Hi,

Apologies - I misread the earlier posts. I did not realise there was a USB-serial cable option.

We are using a USB-USB cable.

When we tried the serial-serial cable, it instantly crashed the UPS, which could not even be powered up again until the serial cable had been unplugged.

JPL

[Anonymous user]

The Part No. of the serial data cable we tried is "APC-940-00" (not 940-0024 as you say it should be), but was said to be suitable for the Smart-UPS 750 according to the retailer. The actual cable supplied with the UPS cannot be found.

The only reason a serial cable is being tried at all is in desperation because the USB cable keeps on losing its connection. It would be preferable by far to resolve the recurrent problem with the USB cable and just forget about the serial cable, which was only tried at all as a possible "quick fix".

(By contrast, the data connection has never skipped a beat in two years on my own system at home - Back-UPS ES 700VA / PC Personal Edition 2.0 / XP SP3.)

JPL

[TheNotoriousKMP_apc]

Let's go ahead and try and resolve that comm problem with the USB then.

Do the following:

Power down the load
Power down the UPS
Disconnect the comm cables
Unplug the UPS from the wall
Depress the OFF button for 4-6 seconds til you hear a click
Power everything back up
Connect the USB cable.

That should solve the problem.

[Anonymous user]

Thank you for your help.

I'll give that a try on my next visit and report back.

JPL

[Anonymous user]

The procedure was followed today - no problems.

Now we just have to wait and see if it has worked. Recent failures occurred on 31 July, 21 & 28 August, and 12, 15 & 16 September, and so no recurrence by the end of October probably indicates success.

JPL

[Anonymous user]

At 08:52 this morning, i.e. less than 24 hours after following the procedure, PowerChute Business Edition reported, "Communication Lost on Agent server01".

The fix did not work.

Any ideas what we could try next?

JPL

[Anonymous user]

So that others may avoid this product, here is the URL for the serial data cable that caused the UPS to crash:

http://www.novatech.co.uk/novatech/specpage.html?APC-940-00

The text ("APC offers the UPS link cable") and APC logo on the Novatech page imply that it is an APC product, but it appears not.

At the time we were not aware that the APC cable was to a proprietary spec.

JPL

[TheNotoriousKMP_apc]

Do you have another workstation with a barebones install of an OS that you could test this on? I'm starting to wonder if there's a software application running that's causing this to just timeout eventually. Usually after the brain-dead in this scenario everything works fine.

[Anonymous user]

No, ARD has no spare kit at all, and in any case would be unwilling to run the server without the UPS while any trial was conducted.

However, the thought is a good one. The server is an old Dell PowerEdge 2500SC (Pentium III 993 MHz CPU, 3 GB RAM) running only SBS 2003 R2 Premium plus anti-virus s/w and, of course, PowerChute. No scope for uninstalling anything there! About 70% of the USB communication failures have occurred outside office hours, with no obvious pattern.

I shall persuade them to buy another serial data cable, and this time insist it is purchased directly from APC.

[Anonymous user]

I am having the same problem as JPL.

Brand new Dell Poweredge 2900 III and a brand new Smart-UPS 1500 connected by USB cable.
Windows SBS 2003 R2 SP2

I've reinstalled Powerchute, to no avail
I've deleted various devices, to no avail

The connection will hold for 5-10 minutes, but then be "Lost". unplugging the usb cable and plugging back in will usually help restore it for a while...

I have noticed that the connection being lost coincides EXACTLY with the appearance of the windows battery tray icon...

I've also noticed that the "batteries" device in the hardware manager is currently called "HID UPS Battery", but I seem to recall it being called something like "APCUps Battery" at some point before.

HELP!

[Anonymous user]

I too am having the same issues on multiple client sites using Smart-UPS 1500's connected via the supplied USB cable to a onboard USB port on the servers, and PowerChute software. I have brain-deaded the various units, and set the power management on the USB Root Hubs on the Device Manager - with no solution - the USB disconnects after around a day. Unplugging the USB, and reconnecting it establishes connection again.

[BillP]

I had similar frequent loss of USB communication. Looked through the logs and found the UPS frequently switch between AC and input battery. Realized the input AC voltage is fluctuating near the threshold setting. Changed the threshold setting. USB problem resolved.

[Anonymous user]

Wow, what an unsatisfying answer.  So much for making use of modern communication methods!

On 4/22/2009 2:21 PM, Angela said:

in regards to this issue, let me post APC's official response to this. please dont shoot the messenger.

APC has found a small number of Smart-UPS customers experiencing intermittent communications loss when using the USB interface. To correct this issue customers are advised to use Serial Communications. If you are continuing to have communications loss, please contact APC Technical Support for assistance in configuring Serial Communications, or to request assistance in troubleshooting. Please ask the Technical Support Engineer to reference KBASE Document # 9651.+

i am going to lock this thread as far as asking for updates on the situation, etc but if you have an issue like this, please remember that you should be ruling out other things such as too many USB devices connected, faulty USB ports, etc before contacting APC support.

[Psychor]

I just posted about this issue with a brand new SMT1500C. I've owned many SUA and SMT units and I've never had this issue. This is ridiculous that this hasn't been addressed by APC yet.

https://community.se.com/t5/Datacenter-Forum/SMT1500C-and-loss-of-USB-connection/td-p/367347

[totoli]

I also suffered from the same problem.

All problems are resolved when changing to serial communication. (Com 1)