128k +2 Mysterious Videos issue.

A place to discuss the ZX80's and ZX81's younger brother - the Sinclair ZX Spectrum
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

1024MAK wrote: Wed Sep 23, 2020 3:20 pm If you enter the 128K or 48K editor, does this make any difference to the symptoms?

Also if possible can you use your RGB to SCART lead, at least until the fault/problem is found.

On that subject, have you tried it on another display?

Mark
48k mode makes no difference.

Reverted to the RGB Scart lead and it's still there, but is very much reduced. Instead of the very clear wave or pixels it's just a slow slight fuzziness that moves down the screen at the same speed as the pixels using VGA.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
User avatar
1024MAK
Posts: 3375
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 128k +2 Mysterious Videos issue.

Post by 1024MAK »

What test gear do you have?

Mark
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

Multimeter and PC Oscilloscope. Noticed with the QL-VGA that the waving corruption is even worse when you change the screen colour to something other than white. Black was ok, unless that's changed since the re-cap.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
User avatar
Tiger
Posts: 18
Joined: Tue Sep 01, 2020 10:37 am

Re: 128k +2 Mysterious Videos issue.

Post by Tiger »

Hallo Chris,
mit meinen Englischkenntnissen stoße ich nun leider an meine Grenzen - da Du ja in Deutschland lebst, verstehst Du es ja evtl., wenn ich in deutsch schreibe:
bei dem Fehlerbild tippe ich auf eine Schwingungsstörung die sich da am Videoausgang zeigt, also eine überlagerte Wechselspannung. Durch Kondensatoren und Elkos wird so etwas normalerweise herausgefiltert bzw. geglättet. Dem Schaltplan zufolge waren die Kondensatoren im Bereich des Video-IC TEA 2000 daher "verdächtig". Da sind aber auch einige Dioden verbaut (D25-D35), von denen eine defekte u.U. auch solch einen Fehler erscheinen lassen könnte. Vielleicht ein kleiner Tipp, der mir in solchen Fällen oft geholfen hat, die defekten Bauteile zu lokalisieren:
Versuch mal die Platine soweit zu lösen, dass Du die Platinenrückseite mit der Hand berühren kannst. Wenn Du dann während des Betriebes mit den Fingern über die Leiterbahnen im Bereich des TEA2000 und der umliegenden Bauteile streichst, sollte sich das Bild deutlich verbessern oder verschlechtern, sobald Du in den Bereich kommst, wo das defekte Bauteil liegt. Mit dem Finger würdest Du einen hochohmigen Widerstand bilden - keine Angst, es fließen nur geringe Ströme - und damit das evtl. defekte Bauteil überbrücken. Sobald Du solch eine Stelle gefunden hast, kannst Du Dir die umliegenden Bauteile einmal genauer ansehen. Ich hätte angenommen, dass durch den Austausch aller Elkos der Fehler nicht mehr vorhanden wäre. Ein defekter 22pF-Kondensator C36, C34 im Bereich des 35,469 MHz-Quarzes könnte aber auch dieses Schwingungsmuster (obwohl da eigentlich das ganze Bild betroffen sein müsste) erzeugen ...
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

I'll certainly try that. When the board is out it's easy to get a finger under it while it's on.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
User avatar
1024MAK
Posts: 3375
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 128k +2 Mysterious Videos issue.

Post by 1024MAK »

What is strange is that the interference (for want of a better word) occurs exactly three times in the 52us of the visible part of the width of the picture (the visible part of a PAL tv line is normally 52us). If you do the calculations, that makes it a 57.7kHz signal. But what is unusual is that it is all in step with the ‘interference’ on the lines below and above, hence the sine wave pattern. But above the menu, it’s just looking like random dots. Further, in the photo it very much looks like digital pixels. So I’m wondering if you have another machine with the same type of ULA. I would suggest swapping it to see if the symptoms change.

Using your PC oscilloscope, check the 5V supply rail to confirm that it is free of any unusual amounts of noise. If your PC oscilloscope has a high enough bandwidth, check the frequency of the main oscillator and that of the Z80 clock (pin 6 on the Z80).

Also you may want to run the program in this topic Test Card Screen and see what effect this has with respect to the interference. Please post up a photo of the results.

Mark
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

1024MAK wrote: Sat Sep 26, 2020 3:39 pm What is strange is that the interference (for want of a better word) occurs exactly three times in the 52us of the visible part of the width of the picture (the visible part of a PAL tv line is normally 52us). If you do the calculations, that makes it a 57.7kHz signal. But what is unusual is that it is all in step with the ‘interference’ on the lines below and above, hence the sine wave pattern. But above the menu, it’s just looking like random dots. Further, in the photo it very much looks like digital pixels. So I’m wondering if you have another machine with the same type of ULA. I would suggest swapping it to see if the symptoms change.

Using your PC oscilloscope, check the 5V supply rail to confirm that it is free of any unusual amounts of noise. If your PC oscilloscope has a high enough bandwidth, check the frequency of the main oscillator and that of the Z80 clock (pin 6 on the Z80).

Also you may want to run the program in this topic Test Card Screen and see what effect this has with respect to the interference. Please post up a photo of the results.

Mark
Will do Mark, thanks for that.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
User avatar
Tiger
Posts: 18
Joined: Tue Sep 01, 2020 10:37 am

Re: 128k +2 Mysterious Videos issue.

Post by Tiger »

Is it possible that it's a variation of the "snow effect"? http://www.zxdesign.info/dynamicRam.shtml :roll:

This effect occurs if the CPU produces uncontrolled /RAS-Signals. To prevent this you can pull /MREQ (Pin19) of the CPU high. I think a resistor <1Kohm or a wire from pin 28 (RFSHL) to +5V (pin 11) could help. :idea:
Normally RFSHL is connected straight with the expansion bus (25B), could it be that there is a problem with the bus connector?
User avatar
Tiger
Posts: 18
Joined: Tue Sep 01, 2020 10:37 am

Re: 128k +2 Mysterious Videos issue.

Post by Tiger »

Today I had nearly the same problem with "crawling dots scrolling over the screen". Tested an ZX Spectrum Issue 6a PCB. But an other Speccy, wich worked fine before also shows these dots. I used an adjustable bench top power supply and after I took an other power supply every thing works fine. Seems that the (two years fine working) bench top power supply suddenly produces interferences :roll: . So I think it is a good way to test your spectrum with an alternative power supply too. :idea:
User avatar
1024MAK
Posts: 3375
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 128k +2 Mysterious Videos issue.

Post by 1024MAK »

Certainly some switch mode power supply units (SMPSU) can produce interference on the video display.

Mark
Post Reply