NBSC Colour televisor construction diary.

[Rydepier]

Dominic, the circuit I am working from is dated 3 Jan 2011. There are some significant differences to that shown on your website. In the new version the vburst signal from the uv decoder is not connected and the vcxo is controlled from a point marked CTL in the phase detector. Have I missed something?

[dominicbeesley]

No, I just haven't updated the website and I've only got the plans on a different computer. I'll try and dig them out tomorrow. It should make more sense then.

In the meantime it would be good if you could get U, V and CTL traces with an SMPTE test card as the signal.

D

[Rydepier]

Will do, thanks Dominic. If the worst comes to worst I can email your circuits back to you if it would make it easier.

[dominicbeesley]

Right got em, they were on this machine anyway - just had to reinstall Kicad.

Right, one thing to check (which is not entirely clear from the new schematics) is that the U, V inputs to the phase detector sub-circuit on the subcarrier circuit should be connected to the "SC Clock U" and V outputs of the clock divider...not actually wrong on the circuit but a little confusing as U, V labels are used on other pages and it had me going a minute ago.

You should be able to do roughly what I said before but the labels are now different.

You should break the circuit at C1103 as marked on diagram and if everything is working as it should you should then be able to vary and reverse the strobing effects by adjusting RV1101. You're aiming for an almost stationary set of colours and for the colours to be the same all the way down (or along if you're in portrait mode) the colour bar. If the colours are not the same along the bar but stationary then your subcarrier-oscillator and video source burst signals are out by a multiple of 12.5Hz (see notes on side lock 5.1 of televisor website section). Keep adjusting one way or the other until you minimise the number of colour changes and speed of strobing.

If that doesn't work try adjusting L1002 on the oscillator board with the RV1101 near its mid-point to set the oscillator's mid-frequency. Then repeat, you may need to actually get L1002 adjusted on centre nearer one end of RV1101's travel than the other but try middle for now. Then repeat the previous paragraph, you may need to mess around with these quite a bit to get it all working reliably but once set up they can be left alone.

All this would be easier if you can measure the frequency exactly - does your scope have a frequency counter built in or do you have an accurate counter (i.e to 100 or so Hz accuracy at 6MHz) and to about 1Hz at 15kHz?

If you can't get the lock to work or it side-locks: are you playing back from a computer or a CD player. I have a number of CD players that play substantially slow or fast - they need to be very close to spot on for this to work. Some use crappy ceramic resonators to set their reference speed and are so bad that I'm told they drive people with "pitch perfect" hearing crazy!

A good way of checking them is to burn a CD with a test tone say 10Khz using something like Audacity and scope this and check the accuracy (again this needs to be a lot of decimal places to make sense). Really the 15Khz burst needs to be within a few Hz of 15Khz.

If you play back direcy from a PC, Laptop or tablet (even my phone will do it) it should usually be close enough.

Let me know how you get on...

D

[Rydepier]

I am going to have to make some small changes to the interboard wireing, nothing too drastic though. Can I just confirm the following:

1. The V Burst signal from U2 pin 4 on the UV Decoder board is no longer used.

2. That the only place the U, U-, V, V- go to are on the colour matrix board.

I recorded a 15 minute SMPTE test card last night and that will give me plenty of time to adjust the circuits.

[dominicbeesley]

That's right.

V-burst is still useful as a test point, if all is well it should be at "mid" voltage when colour is locked and will be flat during burst if not side-locked, it will roll around if unlocked.

U,V,U-,V- all just connected between decoder and colour matrix. sorry about the confusing labels.

D

[Rydepier]

I could not resist the temptation to try to get this working, so I checked the NBSC wiring loom and found that I had connected the U and V signals to the Hue control instead of the SC clock U and V signals. When I started the televisor up with an SMPTE test card playing there it was my first colour picture. The colours were crawling as Dominic suggested but the positive side to that is now I know the three colour channels are working.

Just as I was about to grab some oscilloscope traces I lost the video line sync signals and with it the frame sync. Nothing I tried would get them back and I suspect I have disturbed something as I have been pulling the boards around so it's back to troubleshooting the basic circuits. Still at least I know the colour circuits are basically working.

Chris

[Rydepier]

After several frustrating hours where I just could not track down the problem where the disk would stop locking after three or four minutes and the colour circuits would start misbehaving I finally noticed that my sync signals were only 7 volts.

I measured the 12 volt supply to the electronics and sure enough it was only just over seven volts. I switched everything off waited five minutes then turned it back on, the supply voltage was now 12 volts. As I watched the voltmeter the supply voltage dropped to about 7 volts over three or four minutes. I switched the power off disconnected the colour boards and switched it back on, this time the supply voltage remained at 12 volts.

To cut a long story short I have a fan mounted above the PSU board and LED Driver board and I had disconnected this when I started work on the wireing for the colour boards. The additional power supply demands of the colour circuits raised the temperature of the 12 volt regulator (the only one NOT on a heatsink) and the output voltage dropped in response to the abuse.

I reconnected the fan and switched back on with everything reconnected and left the motor running, twenty minutes later the power supply voltage was still 12 volts. I guess I should have mounted that regulator on a heat sink!

The good news is that in the few minutes I had a signal The SMPTE test card displayed well, with the colour of the bars the same along the bar, but changing, or rolling. Now I have everything back to normal I will start testing in earnest again this evening.

Chris

[dominicbeesley]

I'd be surprised if the colour boards were taking more than 100mA - I bet its less than that. Sounds like something might be up, try a heatsink first.

Sounds like you're getting close and it will just be debugging the burst phase detector.

D

[Rydepier]

I didn't think the colour circuits were drawing very much current Dominic, I suspect the rest of the circuits were drawing close to 3 or 400 ma and with no free air around the regulator it heated up and just went into thermal runaway! The fan seems to have solved the problem and it was my cabinet design that caused the problem.

I just need to get the domestic work out of the way so I can get back to the testing.

[Rydepier]

Not so good. I downloaded an App for my Android phone that allows it to play wav files and it seemed to work quite well. The output is only 300mv peak to peak and it was a bit touchy to get the disk speed synchronised, but it displayed the colour bars well.

The colour was not stable and it was not possible to stabilise the colours using R1101. I switched back to the CD player which gives perfect frame lock, but just a swarm of crawlng colour with no defined bars of colour.

So it seems like I need to sort out the signal source and I am going into work next week to pick up the decent frequency counter and in the meantime I am going to tidy up the interboard wireing in the case and check the colour boards to ensure there are no wireing faults - sepecially the VCXO board.

I have not had any further problems with the power supply now the fan is running and one thing I want to look at is the current drawn by my main circuit boards. I have ordered some 2 amp 12v regulators to replace the current 1 amp ones and I am trying to squeeze in another heat sink.

I have posted a couple of oscilloscope traces, these are from the Colour Bar test card and were captured when I was using the Android phone as a signal source. The first shows the video frame pulse (blue) and the V signal (red). The U signal seemed to be almost identical. The second trace shows the signal at the centre pin of the Hue control and shows the Sub Carrier clock U and V signals being mixed in different portions.

Chris

Post edited 12 August 2012

[dominicbeesley]

Yes looks like the frequency is quite a way out. Did you try L1002 on the oscillator?

D

[Dave Moll]

Rydepier - in your last post you show two traces labelled "hue control almost fully clockwise". Should one of them read "almost fully anticlockwise"?

[Rydepier]

I am sorry about that, yes you are correct. I have edited the picture to show one of the traces reading "almost fully anticlockwise".

I finished rewireing the board interconnections last night and it looks a lot neater. I was just about to test my handiwork when I noticed your post Dave.

I will take a look at the VCXO board with the frequency counter tomorrow with luck.

Dominic, it looks like most of the colour circuits are working correctly and the problem lies with the control of the 6 mhz crystal oscillator. I have it in my mind to rebuilt that part 'dead bug' as perhaps the veroboard construction method was not the best way to go with this kind of circuit.

Chris

[dominicbeesley]

Certainly the parts around the crystal, transistor and varicap need to be built with reduced capcaitance.

I never got round to asking - what did you use for D1000? On the circuit it show a 1N4005 - which will work but it might be worth trying different diodes. I can't remember what my circuit uses but this is the "varicap" diode that tune the circuit.

If you get a frequency counter what you're aiming for is to be able to "pull" the frequency above and below the centre 6MHz by about 2Khz by applying either 0 or 12V to the CTL input.

Test that first with a frequency counter before pulling the circuit apart though...

Sounds like it is close! D