New program version 622D

[Bob]

Version 622 does not work properly on some computers due to a communication problem, especially when a USB/RS232 adapter is used.

This new version 622D hopefully will take care of the problem.

www.w5big.com/prog_update.htm

-- 73 Bob

[Deleted]

My AIM4170 arrived yesterday. Following USB/232 difficulties, I downloaded AIM_622d as Bob recommended. This helped - but, computer blue screen crashes follow after a few scans and usually during attempts to calibrate the analyzer with a BNC/Jumper Adaptor. I have tried different baud rates (9600, 19200, 57600), different COM ports and two USB/232 interfaces but without clear conclusions concerning whether any particular configuration is more reliable. I have now ordered the Y-105 USB/RS232 interface recommended on your web-site. Will report what happens when this arrives.

Current setup: IBM ThinkPad laptop. Intel Celeron (M) 1.50 GHz processor running Windows XP (Home Edition). USB/RS232 interfaces are Belkin F5U103-XP. These interfaces work perfectly when connected to radio equipment (TS870 and PK-232MBX).

If USB/232 problems have been solved by others, please let me know.

John

[Bob]

john said:

My AIM4170 arrived yesterday. Following USB/232 difficulties, I downloaded AIM_622d as Bob recommended. This helped - but, computer blue screen crashes follow after a few scans

Hi John,

AIM version 622c had a subtle problem that allowed it to work fine with some versions of Windows but not with others. I changed one line and then it worked much better with XP, although not perfectly, as we've seen.

The USB drivers seem to be prone to this. A friend and I bought HP Pavilion DV9000 laptops at the same time with XP. Using the USB adapter sold by Array Solutions, these both run fine for 5 or 10 minutes and then crash with the blue screen of death (BSD). I tried a completely different USB adapter marketed by a big USB/RS232 chip maker and the results were no better. My understanding was that this type of total system crash would not occur with XP. That is obviously incorrect.

I've talked to the HP customer service people a couple of times but nobody has been able to fix it. They are friendly and they try to help but this seems to be a very serious software problem in the OS.

Maybe someone who is knowledgeable about the USB interface can help with this problem. I and a lot of others would appreciate it. If anyone is interested, I'll send them the code used to initialize the comm port for them to review.

-- 73/ Bob

[Deleted]

Hi Bob,

Your comments are encouraging! I'm glad to know that my problems are not unique. It is difficult to avoid the conclusion that since a variety of USB/232 interfaces are failing, the common thread is the AIM4170.

Kernel panics (aka BSD) are often the result of buffer overruns, or buffer management problems where a buffer address is corrupted (perhaps by overruns) and an illegal address is used by the driver with consequent calamity.

Can you confirm that you are using software handshaking (Xon/Xoff) for flow control? And that the USB interface is configured to use Xon/Xoff and that the UART in the AIM4170 will respond to Xon/Xoff?

On the assumption that the problem could be caused by buffer overruns, I changed the baud rate to 75. In an hour of working, I had one blue screen, and three cases where the scan stopped in mid-point; the red LED remained on, and after a minute or two the AIM reported that there was "No input from the IO port". This seemed to hang the USB port and a reboot was needed. I made several dozen scans with only these few problems and calibrated different connectors - which was usually impossible to do at high baud rates. Working at a very slow transfer speed has significantly improved reliability without causing any serious inconvenience - the scan rate is quite acceptable.

I hope this information will be useful.

I have programmed the USB interface (on my Mac) to gain access to a Velleman board loaded with an ADC, DAC, push buttons etc, but do not have experience with a PC. The Y-105 has Mac support and I'll investigate the Y-105 in detail when I get it.

When my Y-105 arrives, I plan to use a USB sniffer to log all traffic between the PC and the AIM4170 in the hope that I can spot where things go wrong. (The Belkin F5U103 is a pretty sophisticated interface and I hope that the traffic from the Y-105 will be easier to interpret.)

I am delighted with the AIM4170. The information I have already gained about a multi-band vertical and a small 3 element beam is startling....I have some work to do!

73/John

[Bob]

Hi John,

Thanks for the feedback on the USB.
The AIM4170 does not use handshaking at all.
There are only three wires in the cable that are used: send, receive, and ground. Any delay time between characters does not matter as long as the comm buffer does not overflow. During a scan there are 72 characters sent from the AIM hardware back to the PC after reading a single point. Then the hardware waits as long as necessary for the next command from the PC.
This works ok even with my 400 MHz computer running Win98.
I use a direct RS232 link on that one. Win98 doesn't handle USB directly.

The PC does a checksum on each data exchange so it can tell if something is wrong, although it may take a few seconds for things to time out.

This problem does have the earmarks of a buffer problem. My laptop does a random number of scans before it hangs up. Usually a hundred or more. With a direct RS232 connection to any computer, I haven't seen this problem. Somehow the USB driver is not exactly emulating the RS232 interface. Maybe you can get some data that will give me a clue about how to work around this problem.

73/ Bob

[Deleted]

Hi Bob,

Yes, I saw that no hardware handshaking is in use. A direct 232 to 232 link has no intermediate buffer so should work without handshaking in a regime where the PC sends a command for every point. However, the Belkin interface, for example, has only got a 64 byte internal buffer and has to translate between 232 protocols to USB protocols so may need to stop the data stream momentarily. The Belkin can handle software flow control.
I'm off to work tomorrow for 10 days (Long Island) so will continue with investigations after that. I'll do what I can to help. In the meantime, much progress is being made with measurements.

More later this month.

73/John

[g3nrw]

admin said:

This problem does have the earmarks of a buffer problem. My laptop does a random number of scans before it hangs up. Usually a hundred or more. With a direct RS232 connection to any computer, I haven't seen this problem. Somehow the USB driver is not exactly emulating the RS232 interface. Maybe you can get some data that will give me a clue about how to work around this problem.

Bob, any further news on this one? I have a laptop with USB but no RS232 port, and want to be sure that everything works with a USB/232 adapter before buying a 4170.

73
Ian, G3NRW

[zs6yi]

Hi Ian,
I'm using a laptop with the USB to RS232 adapter and the AIM 4170. The USB adapter and AIM was purchased from Jay at Array Solutions. Everything is working perfectly. I'm running Windows XP PRO.
So I’m sure you will have no problems.
73 Gary ZS6YI

[Bob]

g3nrw said:

admin said:

This problem does have the earmarks of a buffer problem. My laptop does a random number of scans before it hangs up. Usually a hundred or more. With a direct RS232 connection to any computer, I haven't seen this problem. Somehow the USB driver is not exactly emulating the RS232 interface. Maybe you can get some data that will give me a clue about how to work around this problem.

Bob, any further news on this one? I have a laptop with USB but no RS232 port, and want to be sure that everything works with a USB/232 adapter before buying a 4170.

73
Ian, G3NRW

Hi Ian,

I haven't figured out yet why some computers crash.
If anyone has any ideas about this, I would like to hear from them.

I've posted a file with samples of the code used for serial I/O. It's pretty straightforward, but I may have overlooked something. The file is at:
w5big.com/AIM_Comm.txt


73/ Bob