Signal Inverter revisited

[victorf]

Last month I posted a request for help here:

[url]
http://www.zbasic.net/forum/viewtopic.php?t=252
[/url]

The problem was this. I am interfacing to a LCD display and send stuff to it using non-inverted serial. To me this means that data pulses are positive going, rising from 0V to 5V. This works well as I can send strings to be displayed on the screen and they do. This particular display has a keypad interface and I have attached a Grayhill 4 X 4 pad to it. The weird thing is that the key data (pressing 1 should return 49/&H30) is returned as inverted serial. To me this means that data pulses are negative going, dropping from 5V to 0V. Now it is not possible for the ZX-24 to handle this over the com link that is expecting non-inverted serial. I am using com3 to transmit and receive.

After discussing this on the above mentioned thread, I implemented the invert2 solution. I am using the ZX-24 as a signal generator to send a continuous stream of data (non-inverted serial) to the IN terminal of one of the XOR gates. If I don't place a jumper on this gate's input (see diagram), the OUT terminal is inverted and mimics the keypress signal. I then turn this OUT signal around and apply it to another gate's IN terminal hence mimicing a inverted serial signal and DO NOT jumper this gate's input, it's OUT shows an non-inverted signal of about 2V. This explains why the ZX-24 never detects the mock (or real) keypress. Why is this happening? What can I do to get this puppy working?

The truth table for the 74LS86 Quad XOR is:

A B Z
L L L
L H H
H L H
H H L

Jumper ON - NO inversion
Jumper OFF - inversion

Any enlightenment will be appreciated.

Vic

[dkinzer]

[zbasicandy]

[zbasicandy]

Further thoughts:

Not having a schematic is very hard but ... the section where they have the
RS232 connector J3 they had to invert the serial non inverted TTL using a RS232 converter IC or just an inverter TTL gate for 5 VDC RS232.

The "connector section" where you hooked up for the keypad keystrobe I/O is non-inverted per manual.

To convert this non-inverted keystrobe I/O pin J1 Pin 6 to RS232 you need a RS232 converter IC Note: You need on the other end also RS232.
OR
You could just "leave it alone" and just talk non inverted TTL.

My advice is to convert the keystrobe I/O to RS232 and use the PC terminal program to troubleshoot it first before using a microcontroller.

[zbasicandy]

Please dis-reguard!

[victorf]

First, let me say that I am no electronics person, having made my living doing
software for many years. I am sure that I do not speak the electronics language
the way I should as I do not know much. I defer to the members if this forum for
the expertize in electronics.

I want to thank all who have taken the time to respond to my queries. It is much
appreciated that you take the time to help!

Keep in mind that I have only a modest set of test equipment, a logic probe, an
old TEK analog 453 o'scope and a WAVETek multimeter. Thats it!

I am working on a project that is very important to me. I have been stalled for
the better part of a month because of a problem that I cannot solve. I need
help!

I have placed a ZIP file in the Files section that contains the following:

An expressPCB schemetic ZX24.sch, the schematic MUX.sch, the files MAIN.BAS,
COM.BAS and KBD.bas.

This project requires the ZX-24 to talk to one of 3 components via a serial
connection. To that end, I developed the board shown in MUX.sch. The purpose of
this board it to connect the com3 port from the ZX-24 to an attached component.
This is done under software control. Any one of the MUX ports can be connected
to the ZX-24 by passing to the routine SelectMuxPort() in MAIN.BAS the port
number. In the case of my problem this is port LCD (MUX J1) to which is attached
a TRM-425.

The Seetron TRM-425 is a 4 line by 20 character LCD display that has the
capability to store up to 95 full screens of data for display. You can call up
any one to the screens by sending a screen number to the TRM. The TRM has a non-
inverted serial interface on it's J1 port. It also has a RS232 compliant port on
J3. The TRM also supports an interface for a matrix keypad. I have a Grayhill 4
X 4 keypad attached to the J2 port.

Please note that in my OpenCom routine I have defined the COM3 port using the
flag &H08 in accordance with the DefineCom routine discussion in the SysLib
docs. I believe that this sets the Com3 port to non-inverting and 8 data bits.

I can reliably send a screen number to the TRM and display the screen on the
LCD. However, I cannot receive any keypresses from the TRM. I refer you to the
module KBD.BAS to see my code for doing this. The pertinent routine is
GetKeyInput(). I believe that I have implemented the keystobe handling properly
as I get the message "Got data ready strobe". Zero is always returned from the
routine Com3GetByte(data). I have tried adding a little delay in the routine
just in case them TRM is slow to send, but that didn't help. I have done the
experiment of attaching the TRM's RS232 port to a computer running a terminal
program and using the ZX-24 to manipulate the strobe line, I can see the
keppresses appear on the terminal screen. Press 1, see 1, press 2, see 2 etc. so
I know that the strobe handling works and the TRM is responding. However, when I
attempt to do this using the TRM's TTL serial port on J1.5, I get nothing. Since
only one byte is sent for each keypress, I am not able to see the results on my
o'scope. If I set the KeyStrobeTime value to zero and press a key as fast as I
can I can fleetingly see something. The TRM will return 49 ("1") if the 1 key is
pressed. In the call to Com3GetByte, value should be set to 49. NOT! Value is
always 0. Nothing seems to get to the recieve buffer rxq3.

This, then is the crux of my problem. What am I doing wrong? The TRM appears to
work properly but nothing seems to get to rxq3.

As I always say when asking for help:

Any enlightenment will be appreciated.

Vic

[DH*]

[victorf]

[DH*]

You don't necessarily have to have separate queues. I had 4 secondary serial ports using the BX-24. I just redefined Com3 as needed to assign it to each pair of pins. I could only send/receive with one device at a time but the only device that needed a near full-time connection would poll once per second if it had data so that was not a problem.

While I think having 4 full-time ports operating in the background is one of the biggest advantages of the ZX, it's not always needed.

The change I suggested (calling GetQueue with a timeout) should solve your current problem. Start with a long timeout, reduce it until you start missing keys and then increase it until you can reliably capture each keypress.

[DH*]

I spent a bit more time reading the manual for the TRM-425. The RS232 connection uses ±10V levels so, if you connect it to the ZX you'll need some current limiting resistors in the TX & RX lines.

Given this, I'm again of the opinion that your test with a PC and HyperTerminal proves nothing since it is testing different circuitry on the TRM. It's not testing the TTL level serial output.

It could be that the TRM is not outputting the TTL level serial signal or is outputting at a level below the threshold of the ZX pins.

Also, I wasn't reading your code as carefully as I should. Your routine should get the keypress if the signal is arriving.

[zbasicandy]

Victorf
Have you tally up the number of hours you have worked on this module?
From the looks of all the links on this forum it must be a lot!
You have to say to yourself, at some time, what other alternative solution do I have?
Goto www.elabinc.com and buy their keypad serial encoder chip for the outrageous price of $7.00. Part#EDE1144. Then bypass the Seetron- 425 "quality piece of equipment" and get on with your life. If you still can't make it work, I will personally buy the chip and a keypad and will produce the code for you FREE of charge.

I don't do this often, but on quality pieces of equipment I use a sledge hammer

[victorf]

[victorf]

dhouston wrote:

[DH*]

[dkinzer]

I believe that the (untested) code below is the ZX equivalent of the BX-24 code that Dave posted. The BX code exploits an implementation detail (the value at address 21) because there is no other way to achieve the desired effect. However, since the ZX implementation is different, that particular technique won't work. Fortunately, there is a System Library function that provides the necessary functionality.