exxos's Atari, Amiga & retro forum

@stephen_usher

Looking at the VBXL it seems to have its own on board crystals, which got me wondering "if I disconnect the 3.5Mhz crytsal, will it still work".

So I lifted a leg on the crystal, and lo and behold, I have a bright, shimmer free picture. If I touched the leg of the crystal on the pad, the shimmer comes back.

So, I don't know if the 3.5Mhz oscillator circuit is required, but I do now know it's the culprit. I have some new transistors, and will replace them one at a time to see if I can cure it, but very happy I now know 100% that it's that oscillator circuit.

I dare say it isn't needed since the VBXL now drives the video out, but it'd be nice to fix it at least.

I'm just running a quick 'burn in' test to make sure that the problem doesn't return, but so far so good.


:girldance:

Excellent news! :-)

stephen_usher wrote: 12 Apr 2022 16:11 Excellent news! :-)

Certainly is, and having left it for 30 odd minutes there is no sign of the shimmer back at all. I will try replacing the transistors to see if the circuit can be fixed.

The output from the 3.5Mhz oscillator circuit feeds the GTIA - this creates the TV picture, and since its output is no longer used as it is provided by the VBXL, it may well be safe to leave it disabled.

Just tested the 80 column demo of The Last Word, which suffered terribly from the shimmer, and it is really sharp and clear now :)

Still no damn VIC chip Meh!!!

:(

Ok, so it seems this is the problem.

The old video circuit has a 3.5ish Mhz oscillator circuit, which is now surplus to requirements with the VBXL installed.

The VBXL takes a 3.5ish Mhz clock from pin 3 of U24, a 74LS74.

When the old oscillator circuit is running, there are harmonics present on the signal from the 74LS74 which cause the shimmer. Removing the crystal from the old oscillator circuit prevents it running and stops the harmonics.

There is a however though.

The VBXL instructions are only for the 800XL and there are subtle differences I missed.

Firstly, Q8 and R110 are removed on the 800XL and this DISABLES the 3.5Mhz oscillator circuit. The 600XL on the other hand does the same with Q2 and R49. The 600XL has got a Q8 and it seems to be in the video circuit, but it's not right in this scenario, so I actually removed the wrong parts, and that left the old oscillator circuit running when it shouldn't have been, hence the harmonics.

So a combination of no 600XL specific instructions and a misread of the schematics led to the old oscillator left running.

I'll replace the current lashed up test setup as soon as I get time later in the week, but I am at least there now.

I just ran the full Numen demo and it worked 100% fine, so very happy with the results!!

At least you found the underlying issue. :dualthumbup:

stephen_usher wrote: 13 Apr 2022 09:21 At least you found the underlying issue. :dualthumbup:

Indeed, very happy to have got there in the end 😁

The 600XL living its best life:


Ultimate 1MB, VBXL and SIDE3 and stereo Pokeymax.

Still nothing for the Vic though 😕

:dualthumbup:

Well, it's a bank holiday weekend so I wouldn't expect any deliveries of VIC chips or anything else.

Talking about clocks... Have you checked the clock inputs on the VIC-20's VIC chip socket?

C35 is used to tune the oscillation, then it uses UB9 (7402) for the main clock,plus UB10 (74LS74) as a clock divider.

R7 and R32 control the bias on the composite video output.