Television Magazine Testcard Generator

[See_Mos]

I started building a line separator using a sync separator IC and a PIC last year but never finished the project as I found the problem with my code and I too used NOP to pad out a couple of glitches.

I have now reproduced Malcolm’s test card and a few other extras like R, G, B, cross hatch, bars etc using PIC18F25K22. It’s a bit of overkill but I keep them handy. None of the peripherals are needed just one port for video components and one for the selector switch. I did have to run the device at 64MHz to get the gratings and run it on a crystal as the internal oscillator caused cogging verticals.

My colour encoder is an AD724 module that I bought on EBay last year. It has the advantage that no delay components are needed. At the time there were several suppliers but last week I could only find one and the price has more than doubled. The IC is readily available and I have done a board design in Easy-PC but never got around to having some made.

I haven’t been able to spend much time at the bench for several months so it is still on the strip board that I used for prototyping the project and only I only just got around to putting it in a case.

The thought has passed through my mind of doing a 405 line test card, it would be fairly easy compared to the colour one, but I don’t have the use for it and don’t have a B/W TV to test it on anyway.

[FRANK.C]

For those who don't want to programme a PIC a line separator can be found in the LM1881 data sheet.
I haven't built it but it looks straight forward and not too many components are needed.

It is possible to generate the circle using a PIC. An example is the 405 line test card that is in the Vol system A modulator. It uses a PIC18F24K22.
Timers are used to generate the circle. The background and centre of the circle are generated simultaneously.
The timers are loaded with values from a look up table during the line blanking period.
When the timer overflows a interrupt is triggered and the video out is switched from background to center of circle or visa versa.

Frank

[Chris55000]

Morning Symon!

Your Colour Castellations are in the right positions as they are for me and I would be happy to build them to the arrangement shown in your picture!

The Circle Pattern is generated by integrating the line and field active periods with respect to time twice to give parabolic voltages, and the arrangements shown using transistor Miller Integrators, fed into a transistor summing amplifier, shown in the PM5509 and the PM5515 S.M.'s are the simplest you can get and are amenable to a trial run on Breadboard or Veroboard!

This has the advantage, compared with a digitally generated circle, that you don't get jaggly patterns viewing the circle close–up!

Regarding the 130n special capacitors Philips used in the two field integration stages, four × 33n MKS (if you have a good LCR Meter it might be worth selecting them) 5 mm pitch capacitors in parallel are perfectly adequate as a substitute – this will give 132n theoretically, but if you find some on the low–side tolerance you could probably get them to 130n as near as makes no odds!

I will be designing a PIC microcontroller operated pattern selection board, operated by a rotate–and–push rotary encoder, with sixteen pattern select outputs, displaying the selected pattern on a 16 × 2 Character LCD Module, plus LED indicators to indicate each pattern on/off, the LED outputs would operate relays on the pattern select board, which has the outputs from each pattern circuit brought out to the relays on this.

The sixteen pattern selections I propose are :–

1) Complete Colour Test Card ;
2) Colour Castellations On/Off ;
3) Colour Bars ;
4) Red Raster On/Off ;
5) Green Raster On/Off ;
6) Blue Raster On/Off ;
7) White Raster On/Off ;
8) Grey Scale;
9) MultiBurst;
10) Pulse and Bar ;
11) Crosshatch ;
12) Vertical Lines;
13) Horizontal Lines;
14) Dots ;
15) Circle On/Off ;
16) All Black.

I propose three sound options:–

a) 440 Hz tone ;
b) 1 kHz tone ;
c) External Audio.

At the moment, I estimate eight to ten modules, including Trevor's PAL Encoder.

My prototype will employ an Astec UM1286 modulator, but I believe you can get five–pin ALPS ones, are these any improvement?

I will study the PM5508/5509 circuits to see if the special DEM/MATRIX patterns can be implemented without the need for awkward inductors, etc., altho' I will measure those in my PM5508 if there are any!

Now that the "Flowcode" Microcontroller Design Tool is free for hobbyist and non–commercial use, it makes designing a PIC (16F877 etc !) a much more attractive proposal than hitherto, having to try and piece separate bits of straggly code from innumerable different places all over the 'net did not appeal to me!

There may (emphasis on the may!) be a means of programming a PIC to replace the ZNA134 and I have found code examples of this, but at the moment, I would suggest anyone who has a good working '134 sticks to it!

I have found an SAA1043/SAA1044 design in a B.A.T.C. book, and there is also Malcolm Burrell''s cheaper "Random Interlace" S.P.G. in a later issue as well!

I have begin drawing up all the individual circuit sections, but it's a very long job trying to unravel them all from Malcolm Burrell's and Philip's diagrams!

All diagrams and files will be on this Forum as I build and test the circuits!

Chris Williams

[Chris55000]

Addendum :–

The Sound Module is Rob Elliott's "Project 22" Lamp Stabilised Wien Bridge Oscillator, with just Range 3 in use, and two fixed resistors in the Wien Network are used to provide fixed 440 Hz & 1 kHz tones, selected by a small 5V DPDT relay

The "Pulse and Bar" pattern circuit I haven't been able to locate yet, I may have to go digging into the circuit diagrams for the Grundig VG1000 to see if I can filch them out of that!

Chris Williams

[See_Mos]

Thanks for showing me your pages Frank. Not being a TV enthusiast I had not looked at your projects, I only repair older industrial PAL monitors.

I doubt my old head could work out the detail needed to incorporate a circle in my PIC test card and you have done such a good job of the 405 line pattern that I would be wasting my time!

[Wendymott]

One pattern I found invaluable when at NEI, TV assembly in Bradford. To set white balance on a crt you need to do at High beam and lower beam currents. I selected Red raster with Chroma turned off. This gave a nice grey raster to set the lowlights. As the professional White balance meters were VERY expensive. I borrowed one. Set up a couple of sample 14" CTV's then used those to calibrate our "in house" Tri colour white balance meters. Worked a treat.

[Chris55000]

Addendum 2 :–

I have studied the circuits involved in generating the special DEM, DELAY and MATRIX patterns and as noted, this involves making subcarrier encoding circuits with access to the (R–Y) and (B–Y) inputs of the colour encoders, so that rules out the TEA1002/TEA2000 as these have no access to the (R–Y) and (B–Y) encoder inputs!

However it might be possible to utilize the TDA2501 for this purpose, which is still available, altho' a bit pricey at £15 each!

The PM5508 utilizes a set of four centre–tapped subcarrier transformers (originally supplied by Philips as a set of four) – whilst a former for these could be 3D–printed by measuring the originals, then unwinding them and counting the turns, etc., I wouldn't be able to document them from my unit as I don't have a vectorscope to check them and correctly align them afterwards!

The PM5508/PM5509 colour encoder circuits, are, as is usual for Philips's creations, bafflingly obscure to try and unravel!

Chris Williams

PS!

The Promax GV–798/898 series has a "DEM" pattern, only trouble is these cost thousands of pounds to buy secondhand and needless to say you won't get anything out of Promax as to how they work – I suspect they're likely to use horrible things like digital FPGA synthesis of the patterns!

[See_Mos]

The TBA520 + TBA540 uses R-Y and B-Y. see Malcolm's article page Television June 1979 page 413

[Chris55000]

I wonder if the LM1889N Nat. Semi. NTSC Encoder (see Colour Teletext Add–On Board Circuit, (PDF Page 18 of June 1979 "Television") is worth trying out?

The diagram is distinctly revolting to unravel, but it does have the "R, G and B" inputs marked on it, or has this i.c. not got the necessary means of altering the R–Y and (B–Y) gain/phase needed for the PM5508/PM5509's special patterns?

The great beauty of the "Television" Teletext Colour Board is that the circuit achieves a colour encoded result without the need for any awkward–to–find external inductors, (the article points out that altho' the chip was originally designed for the American 3.58 MHz subcarrier frequency, it seems to operate perfectly happily with a 4.43 MHz PAL one!) – that one might be worth a look and the chip isn't horrendously expensive!

I have to confess I still haven't precisely worked out how Philips achieved these "DEM/DELAY/MATRIX" patterns in the '5508 yet – I do have a '5508 but not a PAL vectorscope!

Chris Williams

[Philips210]

Looks like you're making some good progress Chris. It's good to see interest in developing the old 1979 project.

Ref the UHF modulator, I have only been experimenting with the Astec UM1286 which also includes an audio input. Trevor used a suitable modulator from an old VCR which is a good option.

The random interlace sync pulse generator TV Aug 1979 had an error in the article. I built that circuit but failed to get it to work. Perhaps I misunderstood the correction. Apparently there's no problem with the PCB itself, it's the circuit diagram. I will have a look at it again although Trevor's experience with it was not too good. I think the ZNA134 is preferable. There was also another random interlace spg, Video Sync Box in Elektor magazine, Feb 1984.
I haven't looked at the SAA1043 but intend to at some point. It was used in the Television Effects Unit, Maplin magazine, Feb 1991. Also Paul Stenning used this IC in his Video Test Card and Test Pattern Generator, ETI magazine, Dec 1991 to Jan 1992. That's an interesting design with two EPROMs one for the test card and the other for the separate patterns.

In Elektor magazine, Feb 1983 there was the Video/Audio Modulator that used the LM1886N/LM1889N pair. That looks a feasible option for additional colour patterns as the R-Y and B-Y colour difference signals are available.

A few other Television magazine articles by Malcolm Burrell are worth a look: Cruciform Generator, Sep 1978; Video Mixer, Feb 1981 and Making Patterns, May 1981.

Regards,
Symon

[jhalphen]

Hi to all,
Hi Chris55000, Philips210,

National Semiconductor LM1889 NTSC modulator application note enclosed with theory of operation & full component data for typical use.
Had it in my IC database.
Used it circa 1986 to built a SECAM/PAL to NTSC transcoder/modulator for use with the Panasonic CT-101 1.5" subminiature NTSC CRT colour TV.

Best Regards
jhalphen
Paris/France

[Philips210]

Originally Posted by jhalphen

Hi to all,
Hi Chris55000, Philips210,

National Semiconductor LM1889 NTSC modulator application note enclosed with theory of operation & full component data for typical use.
Had it in my IC database.
Used it circa 1986 to built a SECAM/PAL to NTSC transcoder/modulator for use with the Panasonic CT-101 1.5" subminiature NTSC CRT colour TV.

Best Regards
jhalphen
Paris/France

Thank you jhalpen for providing the data on the LM1889.

Regards,
Symon

[Chris55000]

I've now dug into the cheap Addlestone BH–22 Pulse and Bar Generator and it's quite an easily reproducible circuit requiring no awkward inductors or tuned circuits, or specialised i.c.'s, and circuit diagrams will follow when I've drawn them to my satisfaction!

Chris Williams

[Philips210]

Hello again.

Just a further update on the project.

I've been continuing my experiments with various PAL encoders including the original one based on the TBA520/TBA540 combination.
My main complaint with the original encoder is the very dark display due to the low amplitude luminance signal which is only around 400mV Pk-Pk. The composite video signal fed the Manor Supplies UHF modulator. The latter being quite happy with the low level video.
The test card generator didn't have any provision for a line video output.

I think improvements could be made in the way the luminance signal is generated on the logic board. The monochrome sections of the test card are resistively added in a summing network which seems to be a very low impedance path to ground. It's essentially an analogue signal, particularly in the case of the grey background to the crosshatch grid. On the encoder, the luminance is added to the composite sync and gated modulated chroma signals which again seem to have quite low resistance paths to ground. I reckon there's a better way to do this.

In August 1979 Television magazine, article on the Simple Sync Pulse Generator (alternative to the ZNA134E), there was also a recommended video buffer circuit for driving 75 Ohm loads. I don't think anyone could have tested this test card generator on a monitor or a TV with a video input capability, few sets with video input existed in those days. The thing is the video level has to be increased to an acceptable value, also to drive a normal UHF modulator hence the extra buffer/level shift circuit. Also, with the original encoder, the colour rendition is not that brilliant either which I think is due to inaccuracies in the scaling of the resistor values in the luminance and R-Y/B-Y matrices.

On the TEA1002 version of my encoders, I used the dc level shift and amplifier circuit to boost the luminance signal from the logic board. This circuit was kindly from Trevor (SeeMOS). I also employed this same circuit on the original encoder which has given it sufficient lift to be be usable and has greatly improved the display. It wlll require further work to obtain the correct signal levels. A further complication is the burst gating, the amplitude seems quite critical, if excessive there's colour drop out. If insufficient there's a green cast on the upper part of the display.

As mentioned earlier, I used an alternative reference oscillator quadrature coil. It's one from a Philips G11 decoder rather than the recommended GEC 2110 version. I've since found it's possible to avoid using the quadrature coil. In the Philips PM5501 TV pattern generator, which also uses the TBA520 in the encoder, a 90 degree phase shift network is used around the 4.43MHz crystal oscillator. I built this circuit and proved it works in the test card's encoder. Also, the TBA540 is no longer required.

I am hoping to add special colour patterns besides the colour bars and other colour sections of the test card. Access to the R-Y and B-Y colour difference signals is required to develop the circuits. Alternatively, an encoder that allows analogue RGB signals would be good.
On the TBA520/TBA540 encoder I had limited success varying the level of the RGB signals feeding the matrices. I am not sure but think this could be due to the burst gating. It's only applied to the red channel and not the blue and green channels. It's not a perfect circuit, even Malcolm Burrell mentions that.

I still think the TBA520 type encoder has got good potential but a more refined version of the circuit may be the answer. I've been looking at the BATC Handbook PAL Coder which also uses a TBA520 as the chroma modulator. It is considerably more complicated than the version in the test card generator. I intend building this circuit. One thing to note is the level of the RGB input signals. It specifies that these are nominally 0.7V Pk-Pk. In the case of the test card generator, the RGB signals are digital at TTL levels, which would need attenuating to feed the BATC encoder. I am not 100% sure but would a simple potential divider be the way to do this?

As an aside, I also built a TEA2000 encoder, the successor to the TEA1002 and I'm quite pleased with the results. I had a few teething problems with spurious patterning which I've largely eliminated by repositioning the crystal/trimmer and chroma bandpass filter components. The colour rendition looks very good. The internal matrices must be accurately weighted. It's a little easier to build compared to the TEA1002 as the burst gating and PAL switch square wave are internally generated within the TEA2000. It also requires the additional luminance buffer/level shift circuit due to the low level luminance signal from the logic board.

Please see a few attached pics of the latest developments:

Pic1 is the original TBA520/TBA540 encoder with the video buffer and level shift circuit. Extra socket pins have been added to allow quick component changing/evaluation.

Pic 2 is the reference oscillator circuit to replace the quadrature coil and TBA540.

Pic 3 is the TEA2000 encoder, again with many sockets and IC holders for further testing.

Pic 4 is the Test card display using the TEA2000 encoder.

It's an enjoyable project and there's no substitute to building and testing circuits. I've learnt a great deal with this project but still have large gaps in my knowledge which I hope to fill.


Re. the BATC PAL Coder, I'll try to post the circuit diagram soon.

All the best,
Symon

[Philips210]

Hello again.

As previously mentioned, I intend building the BATC PAL Coder to assess its performance against the TV magazine's one. It also uses a TBA520 as a balanced modulator to process the two colour difference signals resulting in a modulated chroma signal which is combined with video and composite sync signals finally producing a composite video output. Please see attached the overall block diagram and circuit diagrams.


I have been trying to figure out how to produce the special colour patterns which may be added to the test card or as a separate pattern. What I have in mind is along the lines of the 'brackets' in the Philips PM5544 pattern or the similar Test Card G. The four main parts of the brackets lie on the R-Y and B-Y axis. Represented as chroma vectors then:

The lower left ruby bracket is R-Y at 90° with B-Y=0
Upper left viridian bracket is R-Y at 270° with B-Y=0
The lower right light purple bracket is B-Y at 0° with R-Y=0 and
Upper right olive green bracket is B-Y at 180° with R-Y=0
The top French blue 'ears' to the brackets represent G-Y at 326° and the lower yellow/brown ears are G-Y at 146°.

Also attached is a pic of the PM5544 pattern and a chroma phase/amplitude diagram for the main vectors. Note the R-Y and B-Y axis depict the same colours as in the brackets of the PM5544. Not that this relevant at the moment but this particular pic of the PM5544 is displaying a non PAL signal as the 'colourless' sections next to the brackets are showing Hanover blinds.

Creating the luminance and colour difference signals can be achieved by combining the red, green and blue signals in resistor matrices that obey the following equations:

Y(Luminance) = 0.299R + 0.587G + 0.114B

Rearranging the luminance equation then;

R-Y = 0.701R - 0.587G - 0.114B and
B-Y = 0.886B - 0.299R - 0.587G

So by accurately apportioning the resistor values the correct level luminance and colour difference signals can be produced.

Something I'm trying to understand is how do you work out the proportions of the RGB signals for a particular phase angle? For example at 90° where R-Y is at a maximum positive value the ruby colour contains mostly a red component with small green and blue content.

I think to find the proportions of RGB in any phasor then a trigonometric operation is required. The R-Y axis is represented by the sine function, sin 90° = 1 and sin 270° = -1
As the colour difference signals are in quadrature, the B-Y axis is then a cosine function ie cos 0° = 1 and cos 180° = -1.

In the chroma phase diagram, the normal colour bar colours are again represented by their unique phase angle and level. So I assume the trigonometric expressions could be derived from this information. I don't really know where to start and hope that someone can assist me on this. Thanks.

Regards,
Symon

[Chris55000]

It's not too difficult to do, all you need to do is resolve the R, G and B signals from your basic colour bar generator circuit into (R-Y) and (B-Y) components - FIG. 1 - you don't need the (G-Y) component to produce these colours, then invert the (R-Y) and (B-Y) components as shown in FIG. 2 - if you get the ratios correct your PAL encoder should resolve these into the correct colours for the main L.H. and R.H. parts of the brackets!

For the bracket's "ears", it is only necessary to matrix the resulting (R-Y), -(R-Y), (B-Y) and -(B-Y) components as shown in FIG. 3 in the appropriate ratios!

If you want to provide "colourless bars", you then feed (R-Y) into the PAL encoder such that it is summed without any line-by-line switching, and add line-by-line switching to the (B-Y) signal - a study of the PM5544 circuits with reference to the matrix gating for blocks H5 to H7 inclusive and V3 to V29 for the left-hand "colourless information", then you need to locate what matrix output is gated for H35 to H37 inclusive and V3 to V29 inclusive for the right-hand "colourless information" - if my explanation isnt too clear I'll try and draw it out as soon as time permits!

Once you've located the matrix circuits that provide the "brackets", "brackets ears" and the "colourless information", you can filch them out of the PM5544 book as a starting-point - I think you need to refer to the pages for "Unit 7".

Chris Williams

[Chris55000]

If the Forum Software has downsized them too much, here are PDFs of the drawings I made!

Chris Williams

[Philips210]

Hello Chris.

Thank you for providing that useful information, I'll have to carefully study it
I haven't yet looked at the PM5544 technical data but must do so soon. It's good to see how theyve done things.

Regards,
Symon

[See_Mos]

The buffer part of the second circuit that Simon posted looks interesting. I might give it a try to see if it overcomes the black level drift problem that I had with the circuit that I used.

[Philips210]

Hello again,

Just a short update on progress with the Test card generator.

I've changed the size and position of the yellow and white castellations on the right hand side of the card, making the yellow ones smaller whilst increasing the size of the white ones. This being closer to the arrangement in Test Card F.

Also added to the card is a little green square located in the black part of the letter box for chroma/luma fit checks. This was obtained by taking the line and field drive waveforms of the buffered outputs from the ZNA134E. These drive waveforms are applied to four 74121 monostables and ANDed together with mixed blanking to produce a square. This is the applied to the green channel via an OR gate with the other green channel patterns. Delay and width for the line and field waveforms is adjustable by means of four presets to centralise the square within the black letter box.
I will look at using two 74LS221 dual monostables to reduce the IC count and the power consumption.

I am also looking at replacing most of the 74xx ICs in the original 1979 project with either 74LSxx or 74HCTxx versions. In any case, I think the 74121s could be replaced with 74LS221s. One 74LS221 to replace two 74121s. In a similar vein, one 74LS175 to replace two 7474s and one 74LS393 to replace two 7493s. This will reduce the overall IC count and save power. Replacing the rest of the original standard TTL with LS versions will further cut the power consumption. Going for LS will be slower than standard TTL but I will give it a try to see if it's successful.

I've also been building a PAL coder using the LM1889 and LM1886 matrix ICs but haven't yet completed the board just yet.

Please see attached pic of the latest card.

Regards,
Symon