Featured
Latest
A review of Frank C...
 
Share:
Notifications
Clear all

Forum 141

A review of Frank Cuffe's PIC-based 405-line pattern generator

10 Posts
6 Users
0 Likes
724 Views
Cathovisor
(@cathovisor)
Posts: 6391
Famed Member Registered
Topic starter
 

Over on the old forum many of us followed the excellent thread in which Frank designed and built a PIC-based pattern generator for 405 lines. As a refresher, it is here: http://www.forum.radios-tv.co.uk/viewtopic.php?f=15&t=12412

I promised Frank I'd do a review of it (eventually!) and so, here it is.

The premise.

For most of us, we make do with Test Card C from whatever source we have. TC C is fine and useful for general picture set up, but sometimes the "fussiness" of the pattern can be more of a hindrance than a help when fault-finding a TV set or setting one up, especially with an oscilloscope in tow that may be of less than wonderful triggering ability! It's quite telling that for normal line-up purposes in a studio environment, nothing more complicated than a sawtooth was used in the BBC in monochrome days - specials like Pulse and Bar were intended for more thorough circuit line-up, and PLUGE (Picture Line Up Generating Equipment) for adjusting picture monitors to defined levels of brightness (black cut-off) and contrast levels. To this end, Frank's generator addresses this by providing many useful single test patterns as well as a test card and line and field-rate synchronising outputs for an oscilloscope.

The unit.

The board can be seen in the link here; the only change I made to the prototype that Frank sent me was to change two electrolytics for Panasonic low-profile types so that when the unit is fitted into a case, it can be retained by the three fixing nuts attached to the BNC sockets without the caps fouling the panel. As you can see, Frank fitted his caps horizontally to avoid this problem.

Practical measurements.

I'm not going to touch upon the patterns provided and their functionality, as that was discussed earlier but rather, concentrate on the electrical measurements of the unit under test. What should be stressed is this: the unit is not meant to be broadcast-standard, but a close representation of it and similar in performance to a typical good quality type of pattern generator in use in a good TV workshop. Certainly, compared to contemporary 405-line instruments it far exceeds those for practicality, power consumption and stability and on that strength alone it is to be commended.

For reference, the following were the parameters which the unit was tested:

Oscilloscope: Teledyne LeCroy WaveAce 200MHz, terminated in 75 Ohms, bandwidth limiting applied (this removed a small amount of clock noise present).

Supply voltage: 10V DC from Farnell L30 power supply.

Test waveforms are compared with the specifications for the 405-line television waveform laid out in the document VTW-A1 published by the BBC as Technical Instruction P2, May 1979 revision.

So, first of all, amplitudes.

PICGen007.png

From this we can see that sync. amplitude is 294mV against a stated value of 300mV, which corresponds to an error of -0.18dB. A touch low, normally in the broadcast environment you'd expect ±0.1dB or better but as we've said - this isn't broadcast spec.

PICGen002.png

Vision amplitude on the peak white test pattern was measured at 680mV, which is -0.25dB low compared to a reference 700mV.

Now, this can easily have been caused by the values of summing resistors used to produce the video waveforms so it can be adjusted if more precision is required: however these values are not too far from an ideal 7/3 relationship.

Now, timing.

So far, the only elements that have given me a little cause for concern, and I may well be making an erroneous measurement are the pulse rise times, but:

PICGen003.png

This indicates a rise time between a nominal 10% and 90% of amplitude of 0.13µs, which is a little on the fast side (the spec. is 0.25 - 0.5µs between 10% and 90% amplitude). As a result, it could potentially give rise to out-of-band components in the vision signal which could interfere with sound.

The same goes for the sync. pulses:

PICGen1006.png

Again, a bit too fast for comfort. In broadcast systems the sync. and blanking pulses would normally be bandwidth limited by Gaussian filters.

As for pulse durations:

PICGen004-2.png

Sync width measured between half amplitude duration points is 9µs - smack in the middle of the 8-10µs spec.

PICGen005.png

Similarly here, the front porch measures at 1.69µs, comfortably within the bounds of the 1.5 - 2.0µs spec.

Field blanking measures within spec., the BBC spec. being 13 - 15½ lines plus 18.25µs.

PICGen2004.png

As Frank stated at the time, there are no half lines of video at the end of field 1.

 

Trigger pulse timing:

The very useful feature of separate line and field-timed trigger pulses was made available on the pattern generator, and below are two traces showing their relative timing. These are positive-going TTL level pulses indicated by the red trace on the 'scope.

Firstly, field rate trigger:

PICGen1004.png

...and in greater detail:

PICGen1003.png

For completeness, here's the line rate trigger pulse;

PICGen1005.png

Summary.

Given the minimal component count, Frank has achieved something truly amazing here. Although I have a slight reservation about the rise time of some of the waveforms, this should in no way detract from what is ultimately a marvel, and I can thoroughly recommend it to anyone looking for a useful 405-line pattern generator.

 
Posted : 16/10/2016 5:15 pm
crustytv
(@crustytv)
Posts: 11869
Vrat Founder Admin
 

Hi Mike,

Great write up and thanks for explaining your findings. It certainly helps ( well me it did ) to not only have a long page of technical text but backed up by visualisations, namely the oscillographs. It was great to follow Franks progress as he developed the PIC Pattern Gen project as very few if any do that these days so it was a rare treat. I wonder if he's ever thought of tackling a 405 standards convertor.

CrustyTV Television Shop: Take a virtual tour
Crusty's TV/VCR Collection: View my collection
Crustys Youtube Channel: My stuff
Crusty's 70s Lounge: Take a peek

 
Posted : 16/10/2016 7:51 pm
Terrykc
(@terrykc)
Posts: 4005
Member Rest in Peace
 

...sync. amplitude is 294mV against a stated value of 300mV, which corresponds to an error of -0.18dB. 

Vision amplitude on the peak white test pattern was measured at 680mV, which is -0.25dB low compared to a reference 700mV.

The real performance is, I believe, even better than it appears from these figures - it all depends how you look at it.

If we take the peak level of the composite signal it is 974mV (680 + 294).

30% of this is 292.2mV and 70% is 681.8mV, so the sync is actually 1.8mV low and the video is 1.8mV high. If the 10k summing resistor (R5) is carefully selected to produce a composite output of 1V peak, both levels would only be in error by ±1.85mV which is much closer to the reference than the -6mV sync level error and -20mV video level error in the figures above would suggest.

When all else fails, read the instructions

 
Posted : 19/10/2016 5:05 pm
Cathovisor
(@cathovisor)
Posts: 6391
Famed Member Registered
Topic starter
 

Except in the broadcast world we treat vision and syncs separately because often the syncs would be removed and replaced by another piece of equipment (e.g. vision mixer proc/stab amp), or would have originated as a non-composite signal (camera).

So, as I pointed out - there are errors that would not be tolerated in the broadcast chain (certainly not at studio level - a contribution would have just passed muster with a error of -0.18dB) but this is not a broadcast piece of hardware. The main thing is that there's a good relationship between sync and video so any hint of sync crushing/stretching would be shown in an observed waveform. It is an excellent generator and as I stated, these are minor matters.

There are some who feel that a piece of test equipment should be beyond reproach and their argument is valid - how can you tell if there's a fault if the equipment doesn't correspond exactly to broadcast spec? I seem to recall F. Livingston Hogg saying as much in the popular Television Engineers Pocket Book series.

 
Posted : 19/10/2016 8:25 pm
Forum 142
(@Anonymous)
Posts: 0
 

Hi Mike
Thanks for taking the time to do such a lengthy write up, the oscillographs are great. This is the first time I have seen a single field blanking period on its own, up until this I have only seen both together. The rise times aren't surprising as the filtering is minimal to put it mildly.

 

Chris said
  I wonder if he's ever thought of tackling a 405 standards convertor. 

Hi Chris

'Only in my dreams'. To be honest, one day I would love to build a converter but it's an enormous leap from a simple pattern generator to a converter and also a huge commitment. For now at least I will stick to simpler things.

Frank

 
Posted : 19/10/2016 8:30 pm
Cathovisor
(@cathovisor)
Posts: 6391
Famed Member Registered
Topic starter
 

Frank said
Hi Mike
Thanks for taking the time to do such a lengthy write up, the oscillographs are great. This is the first time I have seen a single field blanking period on its own, up until this I have only seen both together. The rise times aren't surprising as the filtering is minimal to put it mildly.

Lovely to see you here, Frank! - a very warm welcome from me thumb_gif

The single field interval is very easy to achieve thanks to you providing the line and field trigger outputs: you can do it by adjusting the scope's trigger hold-off control. That's the beauty of a digital 'scope - okay, we quietly ignore the aliasing you get winkbut for events that occur at large duty cycles, they can't be beat. A mate showed me a trick on my 'scope (he was torn between sticking with Tek or going for a LeCroy, he stuck with Tek) which is to pick out a single colour burst and examine its phase in the PAL eight-field sequence - in other words, to look at an event that lasts for 2.2µs in a period of 160ms!

 
Posted : 19/10/2016 8:44 pm
Forum 142
(@Anonymous)
Posts: 0
 

Cathovisor said
 in other words, to look at an event that lasts for 2.2µs in a period of 160ms!  

Hi Mike

That's some piece of equipment, it sounds like needle in haystack territory.

Frank

 
Posted : 20/10/2016 12:02 am
Alastair
(@alastair)
Posts: 294
Reputable Member Registered
 

Begs the question....

How accurate in comparison to this sig-gen are the usual source used these days, the Auroras/Arorai How are they for accuracy in amplitude, timing and rise-times?

This sig-gen Really Doesnt NEED to be even as accurate as it is--Its a great bit of work!

BBC et-al will never again transmit 405 lines again, so the only source of such video is by conversion, which probably wont be as accurate as this device anyway!

 
Posted : 28/10/2016 9:05 pm
Cathovisor
(@cathovisor)
Posts: 6391
Famed Member Registered
Topic starter
 

I neither own or use an Aurora so I can't comment on its accuracy regarding its signal processing.

The whole point of a test signal generator is that it does not produce a signal that leads the operator to think that a fault exists in the equipment under test, so they have to be beyond reproach. I'd have thought that pretty obvious.

 
Posted : 28/10/2016 9:20 pm
Red_to_Black
(@red_to_black)
Posts: 1748
Prominent Member Banned
 

A nice write up for an excellent  piece of kit, I was  following the original discussion on the Vrat forum.

Frank did an amazing job with a fairly minimal component count, I seriously bow to both contributors. cheers

 

Ps. Cathy I think you need to swap 'help' and 'hindrance'  around about here wink :

 but sometimes the "fussiness" of the pattern can be more of a help rather than a hindrance when fault-finding a TV set or setting one up, 

 
Posted : 16/11/2016 8:26 pm
Share: