Should this work with sio2pc serial with 232 original design?

[Bikerbob]

Hi people, if you looked in the main A8 chat.. you would see I am getting my original Nick Kennedy designed 232 chip sio going.

 

I can run in DOS (boot from USB ) sio421 - the last dos based software from him.

 

But when I load RespeQt.. it says it canot load the Rs232 driver.. defaulting to USB.

 

James

[Joey Z]

Hi people, if you looked in the main A8 chat.. you would see I am getting my original Nick Kennedy designed 232 chip sio going.

 

I can run in DOS (boot from USB ) sio421 - the last dos based software from him.

 

But when I load RespeQt.. it says it canot load the Rs232 driver.. defaulting to USB.

 

James

Well, RespeQt doesn't have an error message like that, I'm pretty sure. RespeQt does not have separate RS232 and USB drivers. Are you sure you're running RespeQt? sounds like APE to me...

[+DrVenkman]

Well, RespeQt doesn't have an error message like that, I'm pretty sure. RespeQt does not have separate RS232 and USB drivers. Are you sure you're running RespeQt? sounds like APE to me...

 

 

Yeah, exactly. If you're running Windows, RespeQt should be configurable to any COM: port otherwise-accessible in Windows. If you're running Linux/Mac, RespeQt will use whatever entry your SIO2PC device has in the /dev hierarchy (for instance: /dev/tty.usbserial or something like that).

[Bikerbob]

OK sorry your are right.. that was APEtrial. But with AspeQT or RespeQt I get that emulation has started.. and nothing.

 

 

I edited this message.. Unplugging and plugging in a couple of times, rebooting the PC and the atari. I think maybe there might be an issue with the t40 laptop.. but not confirmed yet. Jezzz whats it got to complain about its about 1/2 the age of the atari its trying to feed.. lol

 

 

James

Edited April 3, 2016 by Bikerbob

[Roydea6]

 

Mine is set for DSR but there is an RI options also

Check Device Manager for the comport assigned. and put it into this Dialogue..

 

[Bikerbob]

OK, so not sure how to judge this.

 

AspeQt - working.. through 1050 and PR: connection.. last in chain running 57600baud (my dos ape boost)

 

RespeQt - not working.. not telling me anything.. just not working

 

APEtrial - not working.. giving me cannot load serial driver defaulting to USB (but when I run config wizard .. everything found good)

 

I can load one after the other.. same result each time.. only program that works is AspeQt. - No idea why.

 

James

[Joey Z]

OK, so not sure how to judge this.

 

AspeQt - working.. through 1050 and PR: connection.. last in chain running 57600baud (my dos ape boost)

 

RespeQt - not working.. not telling me anything.. just not working

 

APEtrial - not working.. giving me cannot load serial driver defaulting to USB (but when I run config wizard .. everything found good)

 

I can load one after the other.. same result each time.. only program that works is AspeQt. - No idea why.

 

James

Your settings in AspeQt match those in RespeQt, right?

[Bikerbob]

Absolutely, both AspeQt and RespeQt default to 19200 and RI handshaking. With AspeQt I can run Mydos Apewarp and run at 57600(?)

 

In SIO420 (DOS program) It also works NP.. of course DOS does not seem my Large GB partitions, but I guess I could run it storage wise from a USB stick.

 

AspeQt works.. so I am posting here, because this is supposed to be a fork of the same software, and maybe my design on the sio2pc is so old, no one really cares.. but I thought I would let you know.

 

James

[Joey Z]

Absolutely, both AspeQt and RespeQt default to 19200 and RI handshaking. With AspeQt I can run Mydos Apewarp and run at 57600(?)

 

In SIO420 (DOS program) It also works NP.. of course DOS does not seem my Large GB partitions, but I guess I could run it storage wise from a USB stick.

 

AspeQt works.. so I am posting here, because this is supposed to be a fork of the same software, and maybe my design on the sio2pc is so old, no one really cares.. but I thought I would let you know.

 

James

No, it should work. Something must have broke between when I forked it and now.

 

Please have a look here: https://github.com/jzatarski/RespeQt/releases

 

Try releases there starting with the earliest, and see if any work. If they do, let me know when they stop working. From there, we'll do our best to narrow down the problem.

[Bikerbob]

OK, I did that, I THINK.. not positive.. that release 1 works great and release 2 works good... I seem to get more read errors when using release 2 (in the status screen it will pause and then I get reading a sector x2)

 

So, not sure what was changed in release 3.. but that seems to be the one that broke it.

 

James

[TheMontezuma]

OK, I did that, I THINK.. not positive.. that release 1 works great and release 2 works good... I seem to get more read errors when using release 2 (in the status screen it will pause and then I get reading a sector x2)

 

So, not sure what was changed in release 3.. but that seems to be the one that broke it.

 

James

 

This is really weird.

I just took my old laptop with a real RS232 port (Intel, single core, 1,5Ghz, Windows XP),

connected it with the SIO2PC cable (based on MAX232) to the ATARI and tested RespeQt-r2 and RespeQt-r3.

I had no problems at all.

 

The differences between r2 and r3 in regard to serial port handling could only improve it.

 

@Joe

have you compiled r3 with the same Qt libs as for r2?

 

@James

Tell us please more about your t40 laptop.

 

Which Windows do you have?

What is the CPU load when the SIO emulation is running?

[Bikerbob]

 

This is really weird.

I just took my old laptop with a real RS232 port (Intel, single core, 1,5Ghz, Windows XP),

connected it with the SIO2PC cable (based on MAX232) to the ATARI and tested RespeQt-r2 and RespeQt-r3.

I had no problems at all.

 

The differences between r2 and r3 in regard to serial port handling could only improve it.

 

@Joe

have you compiled r3 with the same Qt libs as for r2?

 

@James

Tell us please more about your t40 laptop.

 

Which Windows do you have?

What is the CPU load when the SIO emulation is running?

Well Montezuma our Laptops are same specs except I am running Vista on that machine.

 

I also have a Dell D610 I use for PP presentations it also has a serial port. I will bring it home tonight and try everything on there.

I know with the old 232 cables where it is in the SIO chain can effect it too. BUT with SIO420 or AspeQT it makes no difference where in the chain it is. ALSO with using the same ATR images the reads are better in general with AspeQt and SIO420 (almost never a x2 read required).

 

I am not using the laptops for anything else at the time I am running SIO2pc.. so I cant say for sure, but load should be almost nil.. Both the T40 and the D610 are clean installs of OS, both have no major background programs running. D610 is running XP though.. see if that makes a difference. I can also try the ubuntu version on the D610 as I have unbuntu Mate running on that as well.

 

James

[Joey Z]

 

This is really weird.

I just took my old laptop with a real RS232 port (Intel, single core, 1,5Ghz, Windows XP),

connected it with the SIO2PC cable (based on MAX232) to the ATARI and tested RespeQt-r2 and RespeQt-r3.

I had no problems at all.

 

The differences between r2 and r3 in regard to serial port handling could only improve it.

 

@Joe

have you compiled r3 with the same Qt libs as for r2?

 

@James

Tell us please more about your t40 laptop.

 

Which Windows do you have?

What is the CPU load when the SIO emulation is running?

r3 is compiled using the exact same libs and exact same process as r2. There should be no difference in that regard.

 

One possibility I have thought of, is that previously RespeQt and AspeQt were triggering on both edges of /CMD. Is it possible not every SIO2PC has the same /CMD polarity?

 

EDIT: I've got a really busy week ahead, and on top of that I'm sick. I may not be able to check in all that often in the next few days. But if I had the time, I'd start building intermediate windows versions of code that is from commits between r2 and r3, to try to narrow down the issue as far as I can.

Edited April 10, 2016 by Joey Z

[TheMontezuma]

One possibility I have thought of, is that previously RespeQt and AspeQt were triggering on both edges of /CMD. Is it possible not every SIO2PC has the same /CMD polarity?

 

Yes, inversed polarity of the command line in the SIO2PC cable would explain the issue.

However such (incorrect) cable would also never work under Linux, where falling / rising of the command line was properly handled from the very beginning (even in AspeQt).

Edited April 11, 2016 by TheMontezuma

[Bikerbob]

OK, tested on My Dell D610, no difference. So its not the hardware. I can get release 2 to work pretty much no issue. release 3.. no.

 

Hope this helps.. and otherwise while I am still using my sio2pc serial, I can run AspeQt and release 2

 

Is there documentation about the Printer emulation? is there something in this that also lets me telnet from the Atari?

 

 

thanks

 

James

[TheMontezuma]

Could you please try the handshake setting "none" with r3?

[Bikerbob]

YEP!! works with NO handshaking .. actually very fast.

 

Well I did not expect that.. but we are good to go.. call of the search team

 

James

[Joey Z]

YEP!! works with NO handshaking .. actually very fast.

 

Well I did not expect that.. but we are good to go.. call of the search team

 

James

This is not a solution, necessarily, merely another way to pinpoint the issue. I suspect that your SIO2PC wires the /CMD line directly to RI, rather than through an inverting RS232 driver.

 

Basically, your SIO2PC has the opposite polarity on the /CMD line than is usually expected. It should be a relatively simple addition to put a handshaking polarity flip checkbox in the settings menu. I'll add it to the todo list.

 

Nevertheless, NONE handshaking should be fine as a quick fix for you, just keep in mind that mixing the SIO2PC and using NONE handshaking along with real atari drives in the chain could possibly cause data loss. It is VERY unlikely, but it is still a possibility.

Edited April 12, 2016 by Joey Z

[Bikerbob]

Hmmm.. ok.. well at the moment the point of me using sio2pc is to get ancient floppies moved to ATR.. so maybe release 2 is best for now.. or I will just run AspeQt until it is fixed.

 

I did post my original schematic by Nick Kenedy that I built. I am not a chip guy.. but should we not be able to see how polarity was from that?

 

James

[Joey Z]

Hmmm.. ok.. well at the moment the point of me using sio2pc is to get ancient floppies moved to ATR.. so maybe release 2 is best for now.. or I will just run AspeQt until it is fixed.

 

I did post my original schematic by Nick Kenedy that I built. I am not a chip guy.. but should we not be able to see how polarity was from that?

 

James

r2 also has a risk for data loss the fix for this in r3 is actually what caused your SIO2PC to no longer work. But I believe the conditions for issues with r2 when using hardware handshaking are even less likely than the potential issues with r3 when using NONE. Even so, the chances of data issues with r3 in NONE are absolutely miniscule. You just have to happen to have a piece of data that looks like a format command, basically, which means a chance of about 1 out of 1,099,511,627,776 that any random sequence of 5 bytes looks like a format command.

 

And I'll also add that this issue will NOT cause data loss on your real atari drives, only the emulated ATR's in RespeQt. And it will be obvious, RespeQt will log a format command that shouldn't be there.

 

EDIT: running some numbers, it seems that you have to transfer 128GB of random data, to real drives, before there is a .1% chance of having a random format command in there

Edited April 12, 2016 by Joey Z

[HiassofT]

You just have to happen to have a piece of data that looks like a format command, basically, which means a chance of about 1 out of 1,099,511,627,776 that any random sequence of 5 bytes looks like a format command.

It's very late here but I think you got the numbers wrong. Although there are 256^5 different 5-byte sequences only 3 bytes have to match: the drive number, the command byte and the checksum byte - DAUX1/2 are irrelevant here.

 

So out of the 256^5 5-byte sequences 65536 will trigger a format - a probability of 1:16millions.

 

That number should be halved as there are 2 format commands, so 1:8mio probability. And if you have more than one drive mounted the probability goes up again.

 

If the command match is implemented as a substring search the probability goes up even more. Consider a stream of 10 bytes you could either find a match at positions 1-5,2-6,3-7,4-8,5-9,6-10 - that's different to only checking for a match in disjunct positions 1-5 and 6-10. I'm too tired to do the math now so that's left as an excercise for the reader :-)

 

so long,

 

Hias

[TheMontezuma]

Not only the "format" command is critical. "Write" and "Write with verify" could hurt as well.

Actually if a device ID matches any of the emulated (active) devices and the checksum byte is OK, then the AspeQt / RespeQt would respond to it, possibly breaking the communication with the device on the SIO bus, to which the DATA frame (containing matching command frame) was sent.

[Bikerbob]

This software does not provide any R: handling does it?

 

I have been playing with APE.. and I have not had luck getting it to work with Bobterm so that I can telnet into BBSs.

 

Can this be done with RespeQt?

 

James

[TheMontezuma]

Hi James,

 

This software does not provide any R: handling does it?

 

No, it doesn't.

 

I have been playing with APE.. and I have not had luck getting it to work with Bobterm so that I can telnet into BBSs.

 

Try harder

And if you have extended memory I recommend you the "ICE-T" terminal software.

Edited April 21, 2016 by TheMontezuma

[Bikerbob]

Yes I did get the telnet to work, seems to depend on the system I am trying to get into.. It does not seem to play as well wtih some as others.

 

After using APE, AspeQt and RespeQt .. I have to say there is VALUE in the pay for APE - I do find the R: driver and the P: redirect to a PC printer of my choice.. etc.etc.. very nice. Not sure I think the price point is in line with the market.. but if I want to enjoy my 8bit that is something I will look at.

 

Thanks for the work on this software. .I would still test it if there is plans to fix the handshaking issue with the older 232 chip SIO2pc design.

 

James