Pin 5B: Is it TTL csync or not?

[Karbuncle]

I got sent several RGB cables to try out with the Jaguar while I develop OSSC optimal timing for it. One of the cables was RGB sync on luma, and I'm using that for now for safety concerns. Why safety concerns you may ask? Well because another RGB cable sent was wired for csync (pin 5B), but I found it didn't have any attenuation on the line. I contacted the maker of the cable and he swears he remembers measuring less than a volt under load for that output on his PAL console.

 

At any rate, I decided to start doing research and come to find out this line is actually "VSL" (vertical sync) and yet somehow also works as a csync line (no clue how that works). The 1995 Atari Jaguar technical reference document on page 8 actually claims the "VSL" pin is in fact csync TTL +5V.

 

I need to know resolutely: What is the load measurement of pin 5B for PAL and NTSC consoles? Is it truly a TTL csync output? If so, then RGB csync cables need to attenuate this line with a 470 Ohm resistor. Also, how is it the vertical sync pin doubles as a csync pin?

 

Much appreciated to those experts in this field for their help on this!

Edited May 27, 2018 by Karbuncle

[JagChris]

Thank you for your work with the Jaguar in this area.

[Zerosquare]

What is the load measurement of pin 5B for PAL and NTSC consoles? Is it truly a TTL csync output?

Yes, it is a logic level 2 mA output from Tom, without any buffering. So it's fine to drive a high-impedance sync input (a few kΩ or higher) directly, but not for a 75Ω input.

In the latter case, you should use a series 1200Ω resistor to drop the level to the standard 300 mVpp. But I wouldn't recommend it, since without a buffer it's not suited to drive a cable, and it doesn't offer anything over using luma as C-sync.

 

Also, how is it the vertical sync pin doubles as a csync pin?

It is software-selectable for either vertical sync or composite sync. In practice it's always set to composite sync, since the video encoder in the Jaguar requires it.

[Karbuncle]

Cool thanks for the info. I know looking at the schematic and inside the console there is an FB inductor (L13) on the csync line for AC decoupling, but of course this does nothing to attenuate the signal. One question though:

 

Does it really have to be 1.2K Ohms resistance? My understanding was that TTL csync typically just needs 470 Ohms resistance. Does the Jag need more because of the 2mA?

 

Also while sync on luma works great, there are some displays and devices that will only allow csync. I know a sync stripper can be added to the luma line in those cases, but if the Jaguar has csync readily available, then it would be easier to just make a SCART cable that uses it with proper attenuation.

Edited May 27, 2018 by Karbuncle

[Zerosquare]

Does it really have to be 1.2K Ohms resistance? My understanding was that TTL csync typically just needs 470 Ohms resistance. Does the Jag need more because of the 2mA?

The standard composite sync level for 75Ω inputs is 300 mVpp.

With a 5 V source, the voltage divider formula yields a series resistance of roughly 1200Ω. In practice it can be less since the source may not have zero impedance.

 

Note that even with a 1200Ω resistor, the current is roughly 4 mA, twice as much as the output is specified for. So I don't recommend it.

[Karbuncle]

Ste from HD Retrovision did some thorough testing both at load and internal, and also ran SPICE simulations for pin 5B. Here's what he had to say:

 

"Finally got this knocked out. Derived simulation model results and bench oscilloscope results attached for both the signals at the load and internal to the console to verify it still operates properly.

Series 680ohm resistor w/ 100uF to 220uF series capacitor. Actual capacitor is pretty loose as per the same reasons Bob talks about in that Saturn clip I linked earlier. 220uF probably makes more sense for cable manufacturers because they'll have that on hand. Therefore, I ran the attached simulation and bench tests using that. CSYNC levels internal to the console remain at the proper voltages as well.

No need to add logic buffers and such. This should be sufficient."

 

Below are screenshots of the scope and SPICE measurements:

 

Scope at load:

 

https://i.imgur.com/7J6u8eX.png

 

Scope internal:

 

https://i.imgur.com/6C6XYFi.png

SPICE at load:

https://i.imgur.com/xLalZbU.png

SPiCE internal:

https://i.imgur.com/WlTyYOJ.png

So there you have it, folks. Just add series 680 Ohms with series 220uf cap and csync is good to go!