PlayStation Development Network

Hi guys. I'm new here and also new on the world of psxdev

I want to wire a RS-232 jack on the place that originally have the serial I/O port of my scph 9001 becouse the PSXSERIAL USB cable is too expansive for me right now.

I know I'll need a TTL conversor but there any specific one? which is the relation of the pinouts if it would possible?

Which voltages are you aiming at? Real classic RS232 would be around +9V/-9V, and TTL would be 0V/5V, and the unmodded PSX serial should be around 0V/3.3V. I would assume that most modern "USB-RS232" variants are somewhere in that range, too (in that case you might get away without voltage conversion).

EDIT: When saying mordern "USB-RS232" variants, I meant those used with modern hardware (especially in microprocessor DIY scene and which do omit conversion to "classic" voltages). But of course, there are also "USB-RS232" adaptors with "classic" voltages (for people who want to connect ancient hardware to their PC).

nocash wrote:Which voltages are you aiming at? Real classic RS232 would be around +9V/-9V, and TTL would be 0V/5V, and the unmodded PSX serial should be around 0V/3.3V. I would assume that most modern "USB-RS232" variants are somewhere in that range, too (in that case you might get away without voltage conversion).

Isn't the 'real classic' RS232 at 13V, TTL is 0 to 3V3 and CMOS is 0 to 5V?

The actual spec for RS-232 is that "mark" (logic 1) is between -3V and -15V and that "space" (logic 0) is between +3V to +15V.

RS-232 devices should also be able to accept voltages between -25V and +25V without sustaining damage. They are not required to operate correctly outside -15V/+15V, though.

So +/-5V, +/-9V or +/-12V are all perfectly legal RS-232 signals, and any compliant device should accept them.

The signal levels on the PSX serial port are 3.3V CMOS - so they swing (and expect input voltages to swing) from just above ground to just below 3.3V. If you use an RS-232 level converter (or a USB-Serial chip) with 3.3V I/O, then it should work with no problems.

One thing to watch for is that the modem control signals on the PSX serial port are inverted from the normal convention, so if you want a fully wired cable (as needed, for example, for the Net Yaroze) then you will need to add inverters in those lines. Note that this is ONLY the handshake lines - the data lines have the expected sense and do not require inverters.

Also note that if you build a converter around certain chips (such as the FTDI serial chips), you can configure the polarity of these signals using the vendor software. In this case, you can just directly connect the lines.

Yeah, maybe TTL could stand for transistor logic in general, I've always seen it as pseudonym for 5V logic. CMOS is more about power consumption (in amperes) (eg. as in 27Cxxx cmos-eproms vs older 27xxx noncmos-eproms).

Ah, yes, the inverted signals are also very important to know about. Reminds me that I've some PSX SIO hardware & soldering notes/schematics here: http://problemkaputt.de/psx-spx.htm#pinoutssiopinouts
Note that the inverted signals are due to transistors on the mainboard (when actually replacing the SIO port by an RS232 port one could just remove the transistors to get rid of the inversion).

Thanks for the excellent responses

As a note, there are also resistors on the serial lines before the R3000A, so the PSX can take the 5V, but it's not the best idea. It is over-driving it and will probably cause long term damage, but I've done it before and it seems to be okay.

nocash wrote:Yeah, maybe TTL could stand for transistor logic in general, I've always seen it as pseudonym for 5V logic. CMOS is more about power consumption (in amperes) (eg. as in 27Cxxx cmos-eproms vs older 27xxx noncmos-eproms).

Ah, yes, the inverted signals are also very important to know about. Reminds me that I've some PSX SIO hardware & soldering notes/schematics here: http://problemkaputt.de/psx-spx.htm#pinoutssiopinouts
Note that the inverted signals are due to transistors on the mainboard (when actually replacing the SIO port by an RS232 port one could just remove the transistors to get rid of the inversion).

Frankly, "TTL levels" as a bit of a mess, as you might expect from such an old logic family. The chips run from 5.0V nominal, but the outputs don't pull up the way up to the rail. If you look at the datasheet for the TI SN7400 (TI were the original "TTL" vendor):

http://pdf1.alldatasheet.com/datasheet- ... 7400N.html

Then you can see that VOL(max) is 0.4V and VOH(min) is 2.4V - so the guaranteed swing is only 2V
VIL(max) is 0.8V and VIH(min) is 2.0V - so even with that smaller swing there is still an assured 400mV margin in both logic states.

The switching levels on CMOS are much more intuitive - below 1/3 supply is '0' and over 2/3 supply is '1'. You might also notice that if you run CMOS from 3.3V the actual logic levels are close to the ones used for 5V TTL. This was a deliberate choice, since it allowed the direct interconnection of TTL and CMOS circuits without level translators (although it's still good practice to use clamp diodes on the CMOS inputs).

thaks for the ansewrs
I thought I would not need an voltage conversor since I would use a TTL to usb and just shift the original serial to a RS232. Even this is quite advanced for me