Quest for Spectrum +3 RGB cable

I recently acquired a Spectrum +3 from Zetr0. Thanks m8! :)

It came with an RGB lead, which I used to connect the Speccy into my LG Flatron monitor TV. It tunes in but doesn't refresh for some reason so it's kind of a still image.

I though't I'd make myself a proper RGB - Scart cable using the schematic found from this site: http://mts.speccy.cz/doc/128_rgb.pdf. It's on page 14.

It was fairly simple to do but it resulted with a black image as seen below. You can barely see the text. The problem is that the cable doesn't provide blanking signal that forces it to RGB in the LG and you can't select RGB mode manually for the input. On the other hand the +3 only provides composite sync and not composite video so the result is this "negative" image.

There are ready made cables available but most declare that you have to be able to manually switch RGB mode on from your TV set, so those don't come into question.

The +3 has two pins that provide +12V in its RGB jack. The other could be used for the switching pin of the Scart as is but for blanking the voltage needed to be dropped down. I did some research and found out that the simplest way to drop voltage down is to put a Zener diode in between. Based on this I drew a schematic shown below.

This time I decided to do the cable from scratch instead of using a ready made Scart cable. I bought the necessary parts, including a Zener diode with a breakdown voltage of 10, which in theory should drop the +12V to +2V, which is within RGB signal limits. It was also neater to collect all the six needed ground signals to one pin at the Scart connector.

After two hours of soldering I was ready to try my new cable. Well, it didn't quite work as I expected as the picture was the same, although with somewhat more contrast. I grabbed my multimeter and measured voltages from the blanking pin, resulting +3.125V, which is slightly over the limit. :Doh: I then found out that the Zener diode with a certain breakdown voltage value will exhibit a voltage drop of nearly that value. It seems that I have to try with a Zener diode of slightly higher breakdown voltage.

The quest continues...

@prototek

my friend, while a Zener is fine fo its limitation, you can now look at a resistor pull-down after it... I know its convoluted... I will get the resistor calcualtor out in a smidge to see which one you need.

I gave the Zener diode another try. This time with breakdown voltage of 11, which would yield +2,39V, when measured. That should be well within the RGB limits. Yet, I still get this negative screen as in picture. The screen also shifts about pixel's lenghts to left and right, which gives the impression that it was blinking.

What am I missing here? Blanking is within limits, scwitching gets +12V, composite sync is connected and RGB signals have the specified resistors.

I've been using 330 Ohm resistors as suggested in the document linked at the first posting of this thread. Now I found another instructions for +3 RGB cable that suggests 68 Ohm resistors. Which is the correct one to be used?

Hi
I had the same problem when making a RGB lead for my +2a.
As you have already identified the problem is the blanking voltage is not high enough, Paul Farrow in the article you have linked in post 1, explains why. The following is from his article.

"Since the Spectrum +2A/+3 only provides RGB outputs, it may be necessary on some TVs to connect the BLANKING pin of the SCART socket to 1V - 3V. However, this can not be obtained from one of the 12V lines of the RGB socket via a simple resistor voltage divider circuit. The following calculation illustrates this:

75*12 / (75+R) = 1 hence R=825 ohms

However, the 12V has a 1K inline resistor and the value of R calculated includes this. In effect, a negative resistance of 175 ohms is required! Connecting one of the 12V lines to the BLANKING pin directly yields 0.84V and this is below the specification of 1V - 3V"

As you can see the 1k inline resister in the Spectrum is causing the voltage to drop below 1v when connected to the SCART blanking pin.
My solution, I reduced the value of the 1k inline resistor.

John

OR!

you can provide +3v from the use of a simple CR2032 or C2025 =)

These with a bit of effort will fit with a coin-cell holder - all soldered up and heat shrunk (perhaps with a small on/off switch) into the head of a SCART connector..

As an alternative you could of couse use a non-fixed regulator, in a small circuit - take +5v from the motherboard and provide just about 3volts. - I can draw a schematic if you would like =).

this could be fit nice a securly in your +3/+2AB or even +2 and provide proper voltage for the SCART output... I would look at either using an unused pin on the RGB connector OR surplanting one thats is of no use.

or

You can provide 1.5v with an AAA battery holder attached to the cable; I did this with my C128.

Alternately you can tap into 5v off the motherboard and pass it through a 75 ohm resistor. This worked for my Tatung Einstein.

John

I had overlooked the bit about the Spectrum having the 1 kOhm resistor at the 12 V line. It does make sense now that I think of it. How can I still measure 2.39 Volts from the line that has been cut down with a zener diode? Is it because there's no load on the line?

I removed the Zener from the 12V signal for blanking. The screen was still black but there were a few bits in the screen that had a cyan color. Apparently the 0.84V that worked for Paul Farrow didn't work for me. I tried adding a CR2032 battery to the blanking of my first cable but that didn't do any difference. Have to try it to the other cable as the first one is missing the switching signal completely.

A little more detail about trying the blanking from the 12V line. It's like there was a crisp RGB picture on the bottom which was covered by a black ragged composite veil and the cyan color would peek through that. I tried disconnecting the switching but that didn't make any difference. I also changed the resistors on RGB lines to 68 but it only made the visible parts brighter, so I'm going to change them back. The CR2032 battery actually measured slightly over 3 volts so it didn't even show the crisp picture with colors.