Composite Mod

[rune]

Hi, can any of the hardware gurus help me with the Transistor/resistor composite mod. I bought a nice DK-Tronics keyboard with a Series 1 motherboard and 1984 ULA, Z80 etc soldered directly to the motherboard.

I could just about tune in through the modulator to TV and everything is working fine.

I thought I'd try the composite mod using a transistor and resistor. To my surprise it worked perfectly - a nice sharp K cursor! The transistor was soldered directly to the modulator 5V and video feed and output was through a phono socket.

Great, next thing is to remove the modulator and take the guts out to house the mod on a piece of Vero board. This is where it went wrong. I got a K cursor but also lots of static on screen. Once I got the good cursor but only for a second. I tried rewiring the transistor directly to the motherboard, changed earths etc. There was no difference.

Next I connected the board from the modulator to in parallel with the transistor by attaching crocodile leads between 5V, video and earth.

So, what does the modulator do that takes out the interference and can I replicate it using a few components?

Dave

[1024MAK]

Can you post up a picture of what you are seeing please.

The modulator will slightly load the output pin of the ULA (causing the voltage to drop slightly). Apart from that, it does nothing to the input signal. Although some of the UHF carrier may leak out.

Mark

[rune]

I'll do my best to replicate it tomorrow Mark.

I finally gave up and tried a couple of the CCB boards i got from Pokemon. One wouldn't hold the signal - I could tune it perfectly and turn the zeddy off/on and it was back to a black screen. Pressing on the R1 resistor could restore it but thats no use for a working machine. The other CCB board gives a 4 or 5 character wide cursor. Very similar to that discussed in another post.

I've ordered 3 new CCB boards - I just wanted to try using the transistor/resistor mod.

[PokeMon]

This is a calibration issue which may be caused by a bad ULA as well.
If you have a 4-5 char wide cursor than it means that the sync is not trimmed well and any black pixel would be detected as sync. This would than put a back porch in the signal which is about 5us or about 4-5 chars length.

[rune]

Hello Karl,

Im using two CCB units on the same motherboard = same 210 ULA. Im not complaining about the CCBs as I have used them successfully on other ZX81s.

But one does give a 5 char wide block. The K at the left then 4 black squares. We've talked about this before - I was using it on the earliest ULA then. This time its on a 210e. Also I'm following the same procedure on both CCBs One is fine but wont retain the setting after power cycling - the other gives that black block. I'll get the electronics guy at work to check/replace the variable resistors at some point.

What I was wanting to find out in my original post though was why the transistor/resistor composite mod works perfectly when connected in parallel with the modulator, but gets tons of interference and a broken K cursor without the modulator.

I'll try all three ways tomorrow and post pictures.

[PokeMon]

Please post a picture of the ZX8-CCB settings (trimmer position) as well.

[RetroTechie]

I've ordered 3 new CCB boards - I just wanted to try using the transistor/resistor mod.

Just for clarity: I assume you're referring to this mod?
simple video output for ZX81

Could be that the ULA output works like a current source, and needs (a bit more) load to produce correct voltages. In that case, adding a resistor (perhaps a few kOhm?) from ULA pin to ground may do the trick.

But perhaps more likely: maybe the modulator input was a 'light' load, DC level on the output makes the transistor work harder than necessary, and -to some degree- same goes for the ULA output. Have you tried adding a capacitor in series with the output? 100~220 uF / 10V with + to transistor "video out" pin, - to cinch connector center pin. This makes it easier for the transistor to maintain a DC level at its "video out" pin, and in turn makes it easier for the ULA output to work in its 'natural' voltage range. This could be taken further by increasing the value of that 100 Ohm resistor.

To my surprise it worked perfectly - a nice sharp K cursor! The transistor was soldered directly to the modulator 5V and video feed and output was through a phono socket.

Great, next thing is to remove the modulator and take the guts out to house the mod on a piece of Vero board. This is where it went wrong. I got a K cursor but also lots of static on screen.

Ooorrrr... something went wrong in the modding part there. Maybe damage from static discharge, or transistor was unsoldered/re-soldered and fried (or reversed) in the process? You could try with a new one. And this time clamp a temporary heatsink on each transistor's leg before soldering (example: 3rd photo on this page).

[1024MAK]

In this post is the circuit that I use.

I'm normally too lazy to remove the modulator. Also my LCD TV will display a composite video AND tune in to the UHF signal

I do have, and do like Pokemon's CCB boards

Mark

[rune]

Well, I have been out to the man cave and got some results. I have also mislaid my box with ULAs and ROMs so didn't get a chance to test the CCB boards.

I built the circuit using solderless breadboard, the results are exactly the same as the time as the soldered up version.

This is the circuit I built using a NPN transistor and 100 Ohm resistor:



and it produced a display with a lot of static:



So, I connected the guts from the modulator in parallel with the +5V and Video signal:



I got a nice crisp K cursor:



I then desoldered the RF wire going to the centre pin on the video socket, soldered the transistor onto the +5 and video leads on the outside of the modulator, brought the 100 Ohm resistor through from the case (earth, through the centre pin of the video socket and out to connect to the emitter on the transistor.



Putting the modulator back onto the board and connecting the +5V and video leads to thePCB, the sharp image still works and was left to soak test for a couple of hours.



All build back up and looking good:




I did try adding a 2.2K resistor between the +5V and transistor, but it did nothing. I couldn't build Marks circuit as I don't have those parts.

Im happy to accept it all works and in future adding the transistor and resistor is going to be much easier as I don't have to remove and gut the modulator. But I am curious why it needs to be present to get the good cursor.

[1024MAK]

I think this is a DC offset problem.

The input stage to the modulator loads the ULA output pin causing the DC voltage level to drop slightly. When disconnected, the output of the ULA is effectively unloaded (the transistor of the composite video circuit has a relatively high resistance input).

The simple composite "emitter follower" single stage transistor amp passes most of this DC voltage level to the output, and boosts the current at the same time (but does not boost the voltage).

Without a coupling capacitor (C1 & C2 in my circuit), this DC voltage level goes to the TV input stage. Some TVs get upset if the DC voltage level is too high (it upsets their internal amplifier circuits). Maybe this is what is happening here. Composite video is not suppose to have a DC voltage level as produced by this circuit.

When the modulator is connected, the voltage from the ULA is reduced enough to bring the transistor output DC voltage level within the TVs "happy" range.

This is just theory. But it's my best guess.

Happy computing

Mark