VLa81 noisy video signal

[AlfaRomeo]

Hi,
so, last year I solve to fix a TS1000, my first computer, broken 35 years ago by my girlfriend.
As the RF sintonizer doesn´t finded any video signal from the computer I made a video mod (the one with NPN transistor, 100 Ohm resistor and 100uF decoupler capacitor).
The computer didn´t give any video signal from the composite so, I bought a litle scope and, with my limited knowledge, I concluded that was the ULA that fried and bough a VLa81. Now the computer works but the video signal is very noisy

.
The 5V in the chips are 5.05V and I even changed the PSU but with no results. Then I changed the BC549 transistor in the mod by a 2N2222 but the result was the same noisy signal. So I desassembled the composite mod and made a mod with only 2 decoupler capacitors in parallel (100uF an 100nF in parallel) before the RCA but the result was the same.
The video signal from VLa output is this

I didn´t disconnected the old Clock signal to VLa81 but I think it doesn´t is the culprit.
Anyway, the Clock signal from the VLa81 output seems normal

Now I have out of ideas on what to make to find the origin of the noise in the video signaln and hope anyone can help me.

[zx-heinz]

Check the Ground Line. Do you have noise between the ground at the modulator and the ground at the 7805 regulator (heat sink)?

[BarryN]

Hi,
so, last year I solve to fix a TS1000, my first computer, broken 35 years ago by my girlfriend.
As the RF sintonizer doesn´t finded any video signal from the computer I made a video mod (the one with NPN transistor, 100 Ohm resistor and 100uF decoupler capacitor).
The computer didn´t give any video signal from the composite so, I bought a litle scope and, with my limited knowledge, I concluded that was the ULA that fried and bough a VLa81. Now the computer works but the video signal is very noisy NoisyScr.jpg.
The 5V in the chips are 5.05V and I even changed the PSU but with no results. Then I changed the BC549 transistor in the mod by a 2N2222 but the result was the same noisy signal. So I desassembled the composite mod and made a mod with only 2 decoupler capacitors in parallel (100uF an 100nF in parallel) before the RCA but the result was the same.
The video signal from VLa output is this VideoSignal_1.jpgVideoSignal_2.jpg
I didn´t disconnected the old Clock signal to VLa81 but I think it doesn´t is the culprit.
Anyway, the Clock signal from the VLa81 output seems normal ClockSignal.jpg
Now I have out of ideas on what to make to find the origin of the noise in the video signaln and hope anyone can help me.

I see what you mean. My RF output from my 210E ULA is better.

[AlfaRomeo]

Check the Ground Line. Do you have noise between the ground at the modulator and the ground at the 7805 regulator (heat sink)?

Thanks for the help, now I'm at work but will do that when I come home

[AlfaRomeo]

Indeed, the computer have noise between the modulator ground and the 7805 ground.
So, what is the best solution to reduce the noise?
I mesured the ground after the keyboard connection (with the keyboard membrane disconnected) and it already has same noise but not that much as the noise in the modulator ground, what could origin the noise?

[Bloodnok]

Hi,

Just in case some members don't know, I make the vLA81 (and a few other retro things). As with a lot of CPLD designs (due to their very fast switching speeds) there is going to be some impulse noise that will make it to the analogue domain with some resultant frequencies appearing in the video spectrum. It is difficult to eliminate it completely and needs to be done at source (following good design practices with ground planes etc. - some room for improvement no doubt). Trying to filter on the output is difficult as the impulse noise has harmonics in the video domain and attempts to minimise it just lead to blurry video. I have employed a 22ohm/1.5nF RC filter on the video output in later designs but this is mainly to reduce the level of signals around PAL subcarrier frequency (4.43MHz) as, if present, can sometimes prompt the decoder to think the video is colour and results in coloured fringing around vertical transitions.

For reference I've attached a capture of the typical noise levels that are present on the vLA81 video output. The test ZX81 has a simple 100uF capacitor in line with the vLA81 video output. The video is terminated in 75 ohms. The ZX81 7805 regulator has been replaced by a switching type and probably contibutes some noise as well.

Some monitors seem to accentuate the noise, some definitely show a pristine picture. Adjusting the brightness/contrast can improve the cosmetics.

The results that AlfRomeo are getting do appear poorer than usual and I'm working directly him to find out why. We'll keep you posted!

[1024MAK]

Is the phono output connector on the modulator clean? Especially the outer ground/0V connection.

I are the ground/0V connections where the modulator mounts on the PCB properly soldered? Sometimes these are poor, or dry jointed. Also check for PCB track damage.

With the computer disconnected from the power, test between the outer ground/0V connection of the phono output connector on the modulator and the heatsink tab of the 7805 voltage regulator using the 200 ohm resistance range on your multimeter.

What power supply unit (PSU) are you using? Is it a ‘modern’ switching (switch mode power supply unit - SMPSU) type?

Is there anything else connected to either the computer or the display, or anything nearby that could be the source of the noise/interference?

Mark

[1024MAK]

Hi

Forgive me, I don’t have a vLA81 to hand to look at. I presume you include plenty of supply rail decoupling capacitors on the module. Only as per typical Sinclair design, there is just the bare minimum of decoupling capacitors on the ZX81/TS1000 main board. This was not really a problem when they were being used with black and white CRT televisions connected via the RF / aerial output. But with high resolution LCD TVs connected via composite video, sometimes the picture quality can be improved if better decoupling is provided on the +5V rail.

For example, on 16K/48K/+ ZX Spectrum boards, often the ‘jail bars’ can be reduced or eleminated by fitting 470nF or 1uF in place of the existing decoupling capacitors for the +12V rail for the 4116 (or equivalent) ‘lower’ DRAM chips.

On ZX81 issue one boards, I sometimes add two 100nF multilayer ceramic capacitors between the 9V input socket and the voltage regulator. On the U.K. / European versions, the RF channel switch is not used, so there are some convenient PCB pads that can be used. See the photos here.

Mark

[Bloodnok]

I presume you include plenty of supply rail decoupling capacitors on the module

There's a 100nF in close proximity to each IO rail and I do remember spending a bit of time minimising the noise from both the LDO regulator and CPLD itself. The law of diminishing returns stepped in however and adding more decoupling didn't provide any significant reduction. I'm sure some further improvements could be made with layout, or adding an earth plane (the PCB is only 2 layer) but I'd be struggling to justify the additional work these days especially as there are quite a few vLA81s out there and (apart from a rare failure) are keeping their zeddies working fine, as far as I know (and I'm sure I'd be notified ).

[AlfaRomeo]

What power supply unit (PSU) are you using? Is it a ‘modern’ switching (switch mode power supply unit - SMPSU) type?

Is there anything else connected to either the computer or the display, or anything nearby that could be the source of the noise/interference?
Mark

Usually I use a PSU I adapted that only outputs 11,5v because the original TS1000 PSU outputs almost 14v but I already tried with the original PSU and the result is the same.

I already checked PCB for track damage or dry joints but didn´t found any. Nearby didn´t see anything that could be the source of the noise/interference and in same place I turn on my Spectrums and all work fine.

With the computer disconnected from the power, test between the outer ground/0V connection of the phono output connector on the modulator and the heatsink tab of the 7805 voltage regulator using the 200 ohm resistance range on your multimeter.
Mark

I will try it when come from work.