TS1000 RF output

[bwinkel67]

I have one ZX81 (US version that I originally bought in '82) and two TS1000's (one I picked up at a yard sale 25 years ago and one I bought last year on eBay). The newer eBay one actually gives a pretty good picture with RF, even on an LCD TV. The other two just give dark grey on dark grey and I will experiment with moding the TS1000 I've had for 25 years since it has a broken keyboard.

Question, when I plug the TS1000 ram pack in, the picture gets a bit more distorted. Anything that can be done with that in RF mode (i.e. without composite mod)? I took the ram pack out of its case and the interference doesn't get worse. The case. being US, had a metallic finish for FCC purposes but the interference seems to be internal to the TS1000 more than the ram pack itself. I'd like to keep this one as RF so I'm just wandering if the picture can be cleaned up. Without a ram pack in just its 2K mode, the picture is almost perfect on both an LCD and CRT monitor -- almost no distortion which I'm guessing is rare for the ZX81/TS1000 line (I have a QL and its RF is actually really good so Sinclair was capable of making them good).

BTW, I timed the TS1000 against the EightyOne emulator running at speed 1x reatlime at 50fps (i.e. same speed as a UK ZX81 supposedly) and the EightyOne was about 75% faster which was surprising since I read it should only be 10% faster (time test was to fill screen with HELLO). The TS1000's run at the 60Hz rate for NTSC so I thought they should be faster than the 50Hz PAL versions...or am I wrong?

[Lardo Boffin]

If you have a SCART enabled TV you could try one of these:

http://www.fruitcake.plus.com/Sinclair/ ... bility.htm

[bwinkel67]

Thanks! I'm not looking for a hardware solution for this one, just trying to figure out if a) the interference that the ram pack adds is normal -- I think I recall that it is with my ZX81 back in the 80's -- and b) is there anything that can clean it up over RF (i.e. I'd like the RF output to be cleaned up not to find an alternative connection since I can just do a composite mod -- and will with the other TS1000) since it's such a crisp picture already so I might as well keep it stock :-/

Things I tried:

• plugged power brick (i.e. wall wart) on a different power circuit
• added extra power strip filter to power brick
• tried adding some shielding to ram pack in the form of aluminum foil lining the inside of case (this before I took it out of the case). It seems silly but was easy enough to try...is there some other material that would do a better job?

All these really for debugging purposes to see if anything simple would have any sort of impact. I haven't tried a different power brick since it's hard to find that type of connector but did read if you clean up the power that can clean up the signal. I'm guessing the ram pack may cause the TS1000 to draw more power and perhaps that is adding the extra interference.

I know that if you get a bad signal it's hard to clean up but since the non-ram pack signal is almost pristine, maybe I can keep it that way with the ram pack.

[Lardo Boffin]

I have noticed over the last few years of buying / refurbing and selling zeddies that plugging in a RAM pack has an effect on the picture, even with a composite mod board (such as the ZX8-CCB) in place.
With a comp mod in place it tends to be in the form of vertical light grey stripes going across the screen.
The same issue applies, but to a lesser extent with an internal 16K RAM modification.
I can only assume that on RF the impact is more severe.

I’m sure one of the technical experts will come along and explain why this happens!

As an aside there is an adjustable pot inside the RF modulator (on the UK version anyway) which can be rotated to ‘tune’ the picture to some extent. I have done this on a couple of zeddies and got a better picture on RF. It may be worth trying this with the RAM pack plugged in?

Just make sure you note where the pot was before changing it.

[Lardo Boffin]

Its hard to get a photo of this on my phone but this is with a 16K RAM pack:

The vertical lines you can see on the left are far more obvious in real life and go all the way across the screen. Just underneath the cursor is clear screen with virtually no visual artefacts.

The other patterns are the moire affect from the camera.

[Moggy]

I think Mark once answered this(apologies if misappropriating ) insofar as the rampack (Sinclair ones spring to mind) have a crude switching power supply built into them to derive the 12v rail the chips need and it's this switching noise that causes the vertical banding in the display.

I do know that the Sinclair packs seem to draw a fair whack of power more than others and if the power pack is weak or not up to spec the display can show a slow up/down weave effect or a faint black band travelling top to bottom.

[1024MAK]

Yes, as moggy correctly states, the Sinclair RAM pack has a simple (low cost, not simple in understanding) DC/DC converter circuit that produces a +12V supply and a -5V supply for the DRAM chips. This is in the form of of a transistor based free running oscillator, and it takes it’s power from the +9V line. Each time it switches (at high speed), there is a brief current ‘surge’ and this causes the interference. The interface gets into the computer via the +9V line and the 0V/GND line. That’s why the shielding (or lack of) makes no difference.

Solving this problem is not really practical if you want to keep things original.
It may be possible however, to reduce the visible symptoms, if you don’t mind soldering extra capacitors in.

The problem may be worse depending on which 9V PSU is being used. The symptoms may be slightly less if a higher current PSU is used compared to the stock 700mA PSU (in the U.K., the UK700, I presume the stock Timex USA PSU were similarly rated).

Most third party RAM packs cheated, they did not produce a “proper” +12V supply, instead they added extra smoothing and used the +9V supply (and hoped that the DRAM chips would work with the +9V rail that typically is at about 10.5V to 11V). They do however have a less noisy DC/DC converter to generate the much lower current-5V supply.

Mark

[1024MAK]

The vertical lines you can see on the left are far more obvious in real life and go all the way across the screen. Just underneath the cursor is clear screen with virtually no visual artefacts.

This symptom is due to a different interference problem. This time, it’s the actual switching signals of the Z80, RAM and ROM chips that are the cause. The ZX Spectrum computers suffer from similar vertical lines.

Back in the day, when most users used a small CRT TV, this was less of an issue, as the problem would have been slightly less noticeable on a small CRT, and, users would have not known any different.

I have not tried this, but the thing that could be tried, is to solder 100nF, 220nF or 470nF multilayer ceramic capacitors directly across the +5V/Vcc pin and the 0V/GND pin of each of the chips on the ZX81/TS1000 board. I suggest doing this on the underside. This should improve things. If you have the time, you can experiment to find the position(s) that works the best (as a capacitor on each of the chips is really overkill) and if a 100nF is good enough or if a higher value is needed.

Mark

[Moggy]

Excellent information as ever Mark.

I have noticed with Memotech ram packs that if you drop the input voltage too much below 9volts in an aid to cooling the regulator then they have a tendency to "drop out"

As a side issue I recently bought a cable to use my Atari STe with a modern TV and if I use it on my supa dupa million pixel RGB monitor those "jail bars" are horrendous but when used with my Mickey Mouse LCD TV then they are totally non existent.

[1024MAK]

I have one ZX81 (US version that I originally bought in '82) and two TS1000's (one I picked up at a yard sale 25 years ago and one I bought last year on eBay). The newer eBay one actually gives a pretty good picture with RF, even on an LCD TV. The other two just give dark grey on dark grey <snip>

Sinclair supplied three versions of the ULA chip that is used in ZX81 and TS1000 computers. Being that the machines were intended to be used with monochrome / black and white TVs, the first two versions did not generate a part of the video signal known as the ‘back porch’. Some (most) colour TVs use the signal level during the back porch to set a reference value that is used to determine the displayed intensity levels. So if there is no proper defined level (that is the signal does not include the back porch) the TV will adjust itself incorrectly.

The third (and final) ULA version (2C210E) does include a back porch generator.

See also this topic and this post

Mark