French ZX81 composite mod

[deadmeat73]

Hi All,

I have recently acquired a SECAM ZX81. It works, but my LCD TV does a terrible job with SECAM standard so I decided to do composite mod and I am a bit puzzled by it... the unit has the SECAM additional components, a zener in D9, a 1kohm resistor in R32 and a 2n2369 transistor with collector connected to modulator input.
I bypassed all that stuff that I'd like to keep installed for historical reasons and connected a 2N3904 in the usual way (100 ohm resistor to ground the emitter, base directly on Pin 16 of the ULA) - it kinda works but it seems a bit sketchy, I've tested it on a breadboard and sometimes I get picture, sometimes I don't but if I touch the components it appears but it's very noisy and washed out, I tought it was the cheap breadboard but even soldering components together does not fix... it seems that the TV have troubles synching the signal. So I scoped the ULA that is a 2c210 model so it should be the best possible but I really don't like the signal - on the internet I've found this that is for a 184E model

While my 210E reads like this, the levels seems off, it looks very noisy and it seems to not generate back-porch which it should?

I'm not an expert so I may be reading this completely wrong, but I'm wondering if the ULA is marginal, or if there's anything else in a SECAM model that can be causing problems, or if it is just normal production variances. And what's best to solve, different video mod? Investing in a video conditioning board (with back-porch??)? Or replace the ULA altogether with a vla81 (i would really like to keep the machine original but a NOS ULA costs way too much).

Thanks for any suggestion...

[1024MAK]

Have you removed D9 or lifted one end of it? The SECAM circuitry and the modulator need to be isolated from ULA pin 16.

My version of a simple composite video modification circuit is here.

Even with this, not all TVs are happy with the "not quite to spec." video signal. Hence, I normally recommend, where possible, to try another TV.

Mark

[deadmeat73]

Hi Mark,

Yes I did try to lift the D9 zener as a good measure I also removed entirely the 2n2369 and lifted R29, but the signal does not change. I also tried your circuit: didn't know it was yours I've just found it on the internet. It actually seem to give the best results, it seems to be the most reliable in syncing but the video is quite dark. However I am not so familiar with different parts (I have a little bit of theoretical knowledge but no practica), I have only a 2n2222a and 2n3904 npn available (this is the one I used), would any of these work in place of BC548?

Unfortunately I only have one LCD tv with composite input, but will try the commodore 1084 monitor and see what happens. I also have a fairly expensive upscaler that I use with my 64s may try that one as well.

Thanks a lot.

[1024MAK]

Did you get a picture via RF before you started work on the board?

A 2N2222 or a 2N3904 should work okay in my circuit.

Normally when someone says they have a dark picture, it's because their board has one of the first two ULA versions. But you say you have a 2C210E? More information on the different ULAs is on these forums, I suggest you read my summary here first (unless you already have ).

It is possible that your ULA is faulty.

Can you please post up a photo of what you are actually getting on your TV.

Also some more 'scope pictures. It's helpful if you also say what timebase/horizontal and what voltage/div / vertical settings you are using. As well as showing where the 0V level is on the display. Test on the composite output, not on the ULA.

Some upscalers are not at all happy with the video from a ZX81. So keep that in mind.

Mark

[deadmeat73]

Hi

I did get a very bad and noisy picture on the RF before trying to apply the mod. I didn't spend too much time with that because I thought my PAL TV wasn't tuning well with a very old SECAM output (I just used to test the computer was actually booting).

I can confirm that I have a 2c210e ula (datecode 8250).

For the picture, you mean RF? ATM it's difficult because I removed the transistor used for SECAM inversion and detached RF modulator power in order to do the composite mod (will take me a while to reattach). Which kind of info you plan to get from that?

For the scope picture the one posted was with 1v/div, zero where the 1 is displayed and 50us/div (so peak is about 4.4 and sync about 4v), the one on composite output you mean with your composite mod applied? Will do as soon as I get back home.

Thanks again.

[1024MAK]

For the picture, you mean RF?

No, I mean what you currently get via composite video. I'm not asking you to revert to RF. There is no point in doing that.

Which kind of info you plan to get from that?

A picture is worth a thousand words, or more like, it's easer for me to see rather than just going on a description alone.

... you mean with your composite mod applied? Will do as soon as I get back home.

Yes.

Mark

[deadmeat73]

Thank you Mark clear now, let me see what I can do...

So this picture is the TV output, the image is a bit blurry, greyish and there are very visible jailbars. Its not horrible but the TV sometimes loses sync or does not sync when powering on (if I try a couple of times it works). Note: I'm using a cheap switching bench PSU not the ZX81 old linear.

This picture is the oscilloscope reading at your circuit video output. As I did not trust my breadboard connections I've soldered the components on a striboard, the transistor is a 2N3904 the rest of the circuit is identical (10 ohm on collector, 100 ohm on emitter, 2.2k on base, 100uF in parallel with 100nF on the output). I did not install the 100uF between +5 and GND. The scale is in the picture.

This is instead at the circuit video input (i.e. pin 16 of the ULA that I remind you is a 2C210E) - I have a feeling that the ULA is marginal... but finger crossed. The scale is in the picture and v/div is different than the previous pic.

[1024MAK]

Ahh, so the first thing to try, is another PSU. Some cheap switching PSUs can produce noise that affects audio and/or analogue video systems.

The jail bars may indicate problems with a lack of decoupling on the +5V rail. I presume your board is using the "standard" 7805 voltage regulator?

But please try a different PSU first.

Mark

[deadmeat73]

So quite some time passed, but finally got some time to work a bit on this zx81 and like to share an update.

I've tried with the original power supply, linear regulated - works fine but no noticeable improvement vs my switching Chinese PSU... actually checked with oscilloscope the power output of the bench psu doesn't look bad at all.

Anyway, since I didn't want to mess again with stripboards and buy components I didn't have at home to test, I've decided to buy and try an smd pre-built zx8ccb as is is a cheap and low profile circuit (I wanted to install it inside the case but without removing the modulator) it looked a good fit. After installing got a mix of disappointment and excitement. My Sharp LCD TV from 2008-ish still have a lot of problems in synching no matter how I regulate the two trimmers but when it does sync the picture is amazing: crystal clear (as the circuit promises) … so I fetched from my basement the old commodore 1084SP monitor and tried.
Again strange behaviour... if I turn on the monitor first and then the ZX81 it synch immediately and have an outstanding picture even better than the LCD one … but if I turn the zx81 on and then the monitor, it doesn't sync, I need to turn off and on the zx81 - again no matter how I regulate the trimmers. Same if I switch the monitor from composite to RGB input and back to composite need to power cycle the zx81.

Not sure what else I could try. I'm pretty much convinced that the ULA installed in this zx81 (210 model) does not drive synch pulse down enough or something like that (btw I didn't want to, but at the end I've removed all components installed for the SECAM modification) … maybe it has always been like that, maybe it's marginal; not sure. I don't have another ULA or another ZX81 to test my TV sets (they work perfectly with ZX spectrums tough).

Anyway besides that sync thingy the ZX81 is fully functional: came with a 16 KB expansion (the owner made a cut out in the ZX81 box to store the computer with the expansion fitted, quite a nice work), and a small plastic keyboard to put on top of the membrane one. The additional keyboard is well used so I think this machine seen quite a lot of love in its days and now a bit more in 2024 as I'm trying a few programs.