Hi All.

Has anyone encountered an issue using Ginny's serial class in Windows 10 when connecting to something like a scale?

In other versions of Windows (XP, Win7, Embedded, etc), when I query the scale for a weight, it comes back perfectly fine with 3 decimal points (eg: 2.456) kilograms.

Running the exact same code on a Windows 10 machine, it drops the hundredths and thousandths of a kilogram and rounds (eg: 2.5).

The exact same application, scale, rs232 cable and com port settings - just different machine and OS.

Any clues?
Thanks in advance, Craig Dawson.

Dear Craig:

On Friday, May 21, 2021 at 12:50:58 AM UTC-7, Craig wrote:
> Has anyone encountered an issue using Ginny's serial class
> in Windows 10 when connecting to something like a scale?
>
> In other versions of Windows (XP, Win7, Embedded, etc),
> when I query the scale for a weight, it comes back perfectly
> fine with 3 decimal points (eg: 2.456) kilograms.
>
> Running the exact same code on a Windows 10 machine,
> it drops the hundredths and thousandths of a kilogram
> and rounds (eg: 2.5).

Have you discussed this with the scale manufacturer? You can't have been the first person to experience this. Have you looked at the actual character buffer, to see what was received by your code?
> The exact same application, scale, rs232 cable and com
> port settings - just different machine and OS.

David A. Smith

On Friday, May 21, 2021 at 5:50:58 PM UTC+10, Craig wrote:
> Any clues?
> Thanks in advance, Craig Dawson.

https://www.coolgear.com/product/usb-to-serial-adapter-windows-10

Read some threads on chipset change and workaround USB to RS232..

On Friday, May 21, 2021 at 11:36:41 PM UTC+10, dlzc wrote:
> Dear Craig:
> On Friday, May 21, 2021 at 12:50:58 AM UTC-7, Craig wrote:
> > Has anyone encountered an issue using Ginny's serial class
> > in Windows 10 when connecting to something like a scale?
> >
> > In other versions of Windows (XP, Win7, Embedded, etc),
> > when I query the scale for a weight, it comes back perfectly
> > fine with 3 decimal points (eg: 2.456) kilograms.
> >
> > Running the exact same code on a Windows 10 machine,
> > it drops the hundredths and thousandths of a kilogram
> > and rounds (eg: 2.5).
> Have you discussed this with the scale manufacturer? You can't have been the first person to experience this. Have you looked at the actual character buffer, to see what was received by your code?
> > The exact same application, scale, rs232 cable and com
> > port settings - just different machine and OS.
> David A. Smith

Hi David,
Thanks for that suggestion. Much appreciated.
After speaking with the techs at CAS themselves, they have confirmed that the Serial to USB adaptors that use the Prolific Chipset often return corrupt data trains. These seem to be everywhere as they are very cheap to buy (so many customers have them of course). The more expensive ones that use the FTDi Chipset don't suffer the same problem with corruption of data. I have purchased one and am now just awaiting its arrival. Phil McGuinness also alluded to the USB chipset and he could be right on the money as well. Anyway, will test it when it arrives and confirm either way.
Craig.

On Saturday, May 22, 2021 at 12:23:22 AM UTC+10, Phil McGuinness wrote:
> On Friday, May 21, 2021 at 5:50:58 PM UTC+10, Craig wrote:
> > Any clues?
> > Thanks in advance, Craig Dawson.
> https://www.coolgear.com/product/usb-to-serial-adapter-windows-10
>
> Read some threads on chipset change and workaround USB to RS232..

Yes Phil - the FTDi chipset looks to be the solution. Thanks for the tip ol' mate. Will let you know how things go once it arrives.
Craig.

Dear Craig:

On Friday, June 18, 2021 at 3:59:33 AM UTC-7, Craig wrote:
....
> After speaking with the techs at CAS themselves, they have
> confirmed that the Serial to USB adaptors that use the Prolific
> Chipset often return corrupt data trains. These seem to be
> everywhere as they are very cheap to buy (so many customers
> have them of course). The more expensive ones that use the
> FTDi Chipset don't suffer the same problem with corruption of
> data. I have purchased one and am now just awaiting its arrival.
> Phil McGuinness also alluded to the USB chipset and he could
> be right on the money as well. Anyway, will test it when it arrives
> and confirm either way.

Thank you for keeping posterity fed with valuable information. You don't need to direct your response to me, the Group is fine. My last use of serial comms with scales was in 1990, and I did not use VO.

David A. Smith

On Friday, June 18, 2021 at 11:51:20 PM UTC+10, dlzc wrote:
> Dear Craig:
>
> On Friday, June 18, 2021 at 3:59:33 AM UTC-7, Craig wrote:
> ...
> > After speaking with the techs at CAS themselves, they have
> > confirmed that the Serial to USB adaptors that use the Prolific
> > Chipset often return corrupt data trains. These seem to be
> > everywhere as they are very cheap to buy (so many customers
> > have them of course). The more expensive ones that use the
> > FTDi Chipset don't suffer the same problem with corruption of
> > data. I have purchased one and am now just awaiting its arrival.
> > Phil McGuinness also alluded to the USB chipset and he could
> > be right on the money as well. Anyway, will test it when it arrives
> > and confirm either way.
> Thank you for keeping posterity fed with valuable information. You don't need to direct your response to me, the Group is fine. My last use of serial comms with scales was in 1990, and I did not use VO.
>
> David A. Smith

Yes David, no problem, just had to say thanks for you comment on the newsgroup- it steered me on the right path. Just grabbed a FTDi converter from a friend (only 2 doors down the road - and bugger me) and it worked no worries at all.
Goes to show - VO code is soooo good. Hardware - hmm, sometimes not so. (Prolific take note).

This is interesting as we use the class with barcode readers and everyone says FTDI converters, now I know why.

I thought it was reliability rather than corruption.