usb uirt learning problems

Paul Forgey · Post by **Paul Forgey** » Thu Nov 18, 2004 1:49 am

Apparently the girder plug-in has a 'burst mode' setting for this situation. I've got at least one device which only sends its ir codes in short bursts. Because of this, the learn progress never gets beyond 20% or so, and I can't learn the code.

Post by **ScottBot** » Thu Nov 18, 2004 9:05 am

Paul,

Try turning on the tracing for the UIRT hardware Interface. To generate a trace file:

Select the Hardware Interface from the Tree View.
Click on the 'Hardware Module Type' button at the top of the Interface Property list.
Click on the 'Enabled' check box to turn on tracing.
Try and learn the IR code again. This should generate data in the log file.
Turn the tracing off by clearing the 'Enabled' check box set previously.
Send the trace file (InterfaceTrace.*.csv) to me at '[email protected]'. The file can be found in the HouseBot\Logs directory.

Paul Forgey · Post by **Paul Forgey** » Fri Nov 19, 2004 2:21 am

I tried that but it didn't leave a log file.

Post by **ScottBot** » Fri Nov 19, 2004 3:32 pm

I'll see if there's anything I can do to implement the 'burst mode'.

Post by **ScottBot** » Sat Nov 20, 2004 12:09 am

Paul,

I've sent you a PM with info on a new version.

Paul Forgey · Post by **Paul Forgey** » Thu Dec 02, 2004 1:17 am

I'm continuing to have further trouble with the USB-UIRT in general. Both the test and released versions of the plug-in are behaving the same with these problems:

- Pronto codes just don't work. They do work in my pronto, so I know they are the right codes. They also _used_ to work, which is very puzzling. I've already contacted support at usbuirt, and we've tried swapping for another unit so we can eliminate the hardware. I've tried re-installing both HouseBot, the usbuirt drivers and the housebot plug-in. I know the released version of the plug-in's install program won't replace the test version of the plug-in dll; I am not still accidentally running the test version.

- Some codes which used to learn no longer do so.

I'm very confused and frustrated. Any ideas? I really don't want to go back to using my Pronto, but with most of my important codes not working, I can't really continue to use HouseBot either.

Post by **ScottBot** » Thu Dec 02, 2004 10:25 am

Paul,

So is this one problem or two? It sounds like you can't learn some codes and also some codes won't send.

For the codes that won't learn, are these the 'burst' codes, or are there others also? Here's a utilitythe Jon Rhees wrote for learning also. See if you get any different results by learning with it and pasting the pronto codes into the HouseBot plugin.

For the codes that won't send, do you know if anything is being transmitted? I use a 'flashing' emitter while testing so I can see if anything is actually being sent to the emitter. If you have one of these, it would be good to know if something is trying to be sent.

Paul Forgey · Post by **Paul Forgey** » Thu Dec 02, 2004 4:54 pm

ScottBot wrote:Paul,

So is this one problem or two? It sounds like you can't learn some codes and also some codes won't send.

Two problems. Some won't learn, and others won't send. Pronto codes directly entered don't work at all anymore.

For the codes that won't learn, are these the 'burst' codes, or are there others also? Here's a utilitythe Jon Rhees wrote for learning also. See if you get any different results by learning with it and pasting the pronto codes into the HouseBot plugin.

No, these are other non 'burst' codes which used to learn. I will try playing with this application and see if I can't make a more helpful diagnostic.

For the codes that won't send, do you know if anything is being transmitted? I use a 'flashing' emitter while testing so I can see if anything is actually being sent to the emitter. If you have one of these, it would be good to know if something is trying to be sent.

Yes, it is transmitting. I can see it both on the indicator LED and in the activity LED on my distribution block. I know the problem isn't my distribution block because the same codes which work and don't work behave the same when I use line of sight.

Paul Forgey · Post by **Paul Forgey** » Mon Dec 06, 2004 9:33 pm

The lrnhelper application sends and learns all my problem codes just fine. I've even compared the pronto data stored in the housebot database.

Post by **ScottBot** » Mon Dec 06, 2004 9:50 pm

Paul Forgey wrote:...I've even compared the pronto data stored in the housebot database.

And the codes match or don't match?

Paul Forgey · Post by **Paul Forgey** » Mon Dec 06, 2004 9:54 pm

ScottBot wrote:
Paul Forgey wrote:...I've even compared the pronto data stored in the housebot database.
And the codes match or don't match?

Sorry! Typed that too quickly. Yes, the codes do indeed match. What I see by directly viewing a copy of the .mdb file in Access matches what I see in HouseBot which matches the working codes in lrnhelper.

Is there anything in the hardware specific field which might help?

Post by **ScottBot** » Mon Dec 06, 2004 10:19 pm

Paul Forgey wrote:Is there anything in the hardware specific field which might help?

The hardware specific info for the USB UIRT goes something like

Code: Select all

"USB:R=0004:Zn=2;Rx=123456709012"

Where:

R=Repeat Count

Zn=Zone (0 = All Zones)

Rx=Receive IR Code (only present for IR codes that are received).

I'm at a loss to explain why the same code would work differently in the different software. Unless the lrnhelper app is doing something else that I'm not... I don't know. I'll ask Jon if there's anything special that his app is doing that I need to be aware of.

Paul Forgey · Post by **Paul Forgey** » Mon Dec 06, 2004 10:42 pm

Those field contents appear to be fine, except that most of the codes I used HouseBot to learn codes with don't have an Rx= value (the field ends with "Rx="). There's no correlation between working codes and non-working codes that have or don't have Rx= values. The vast majority of the codes don't have an Rx= value.

Something else that may or may not be helpful-- if I accidentally leave HouseBot running while running lrnhelper, lrnhelper seems to behave very similar to how HouseBot is failing. In this state, the same random codes I try that work in HouseBot will also work in lrnhelper, and same for the non-working codes. If I exit both applications and run only lrnhelper, lrnhelper then works properly.

Post by **ScottBot** » Tue Dec 07, 2004 12:56 pm

Paul,

Apparently the only difference between the way the lrnhelper app and the HouseBot plugin send IR Codes is the lrnhelper app has the repeat set to 10, whereas HouseBot allows the user to specify. Have you tried increasing the repeat count to higher values in HouseBot?

Also, do you see this when sending multiple codes quickly, or can you duplicate it with sending just a single code?

if I accidentally leave HouseBot running while running lrnhelper, lrnhelper seems to behave very similar to how HouseBot is failing. In this state, the same random codes I try that work in HouseBot will also work in lrnhelper, and same for the non-working codes. If I exit both applications and run only lrnhelper, lrnhelper then works properly.

This sounds like it should be a good clue, but I can't imagine how just having HouseBot running would effect the other application.

Paul Forgey · Post by **Paul Forgey** » Tue Dec 07, 2004 5:22 pm

Well, now lrnhelper is failing the same way. I think there's just an intermittence somewhere in the drivers or API. That seems to indicate the HouseBot plug-in isn't causing the problem after all. I'll persue this over at the usb-uirt site.

ScottBot wrote:Paul,

Apparently the only difference between the way the lrnhelper app and the HouseBot plugin send IR Codes is the lrnhelper app has the repeat set to 10, whereas HouseBot allows the user to specify. Have you tried increasing the repeat count to higher values in HouseBot?

Yes, to no effect.

Also, do you see this when sending multiple codes quickly, or can you duplicate it with sending just a single code?

Just a single code