NBTV without any moving parts. THE NIPTRIX

[Viewmaster]

Hi Albert

Thanks for the schematic post its nice to see how its done all nice to see something working and such but to learn i for one need and example !

.

Harry please don't be mislead by my circuits. They are just a 'pick and mix' bunch derived from bits of circuits from Jeremy, Grant Dixon, etc and the net
.
So be warned by Steve's warnings to me about the 4017.

I wouldn't wish for you to start anything based on my ignorant wanderings among the circuit circus.

It's just that I do like to try to do something on my own without bothering folk on this forum as little as possible.

As I said,I 'pick and mix', hold yer breath and see what boils up in the pot of ignorance

[Steve Anderson]

Circuit wise I am now commited to the 4017, for good or evil now, so we shall see.:

All understood Albert, let's hope my concerns are unfounded and all goes well.

Steve A.

[Harry Dalek]

Hi Albert

Thats fine thats what i do as well mix and match here and there see what works so thats no problem..... but i like to see how you go about it then i can go arrr thats how he did it !

I don't think any one has tried a matrix Led monitor on the forum ...i recall bigscreens one row of leds and a rotating 4 sided mirror that cuts down the number of leds but it is what it is .

I have used the 4017 a fair bit over the years one thing i think it can only count in one direction used it a bit for stepper motors only counting to 4 ..my ic driver i used for the stepper ULN2803 is made to drive leds as well it will drive 8 leds might be a thing to look into if you don't want the 4017 to drive the leds or want a bit more control.

[Viewmaster]

I don't think any one has tried a matrix Led monitor on the forum ....

As far as I am aware this is the first Niptrix using ultra bright white LEDs too.
So Harry and guys here's the 2 bags of lights and the switching board with trannies soldered in. More cross wires etc to be done and clocking board to be finished.
Then just the last little job of soldering 1500 odd LEDs into the antique frame to finish my
NIPTRIX frame up !

[Klaas Robers]

One thing is important: never stop the clock. So always continue scanning. Then you will not blow up an LED.

Look at the circuit of the EPROM NBTV picture generator. This circuit comes from Grant Dixon. However I added the genlock feature. The counters run on a continuous running cristal oscillator, but the counters can be reset by a very narrow line sync pulse or frame sync pulse. This will keep the scanning in synchronisation and at all if the video comes from a CD. If the video signal disappears the scanning continues.

Think of placing the LEDs in a matrix with a horizontal scanning and a vertical scanning. Only the LED on the moving crossing is lit, that is: is connected to the output. It is not simple to make this without investing huge amounts of components. Especially if you plan to overdrive your LEDs you need real currents and the output of a 4017 isn't powerfull enough.

I would think in the line direction to have 48 (or 64) emitterfollowers, e.g BC337 and for the horizontal direction 32 BC337 s to direct the outputcurrent to the selected column. But it is not an easy circuit.

[Viewmaster]

Klaas I think you will find in this thread I started on my NIPTRIX that is what I am doing.
It will be a matrix of 48 x 32 LEDs (NIPkow + maTRIX) addressed as per NBTV club standard. ie bottom right up in columns of 48 and right to left in 32 rows of LEDs, One switched on at a time and controlled by a single video modulator in the common earth return. (using your own single transistor
as in the handbook).

Re 4017 not being powerful enough....Not on it's own but driving NPN and PNP transistors (see my last post photo of all the trannies soldered in) it will be OK I think.
I have had an ultra bright LED running on 100ma at a few cycles per sec clocked by a 4017 into a 2N3904 ,for an hour without any problems.

Re clock not stopping....I have outlined how I would monitor the clocks and cut the video modulator off if I need to overrun my LEDs greatly.

But bear in mind, Klaas, that others who over run theirs did so partly because, or mainly because they used coloured LEDs whose light output is poor compared to my clear ultra bright white LEDs. I believe that this is the first time they have been used in a 48 x 32 matrix.

About 4000+ soldered total joints are required, so far about 500 completed.
A nice job over Xmas to complete

[rommert]

Hi Albert.My display is roughly constructed in the way you discribe.
Deflection circuits using mostly 40xx circuits followed by transistor
buffers and the video circuit and sync separation from the design
from Klaas. The gamma correction calculated for maximum 160 mA.
(see my former post) The overload indication is blinking a led when
the input videolevel is exceeding the limits mentioned in the design
from Klaas. ( I use a potmeter on the input to accomodate for different
input levels )
I do not use a safety circuit untill now. The deflection circuits are
clocked by synchronised free running oscillators. If you monitor only
this oscillators, a disaster can still happen when a fault occurs in the
following counters etc.
For test purposes I made dummy loads with resistors for both de-
flection circuits. ( one with 32 resistors and one with 48 resistors )
As you can see on the photographs from Dr Zarkov I connected the
led matrix with flatcableconnectors, so the matrix can easely replaced
by the dummy loads.
My leds are 3mm diameter. When installed in diagonaly rows on
veroboard, the distance between each led becomes 2.54 x sqrt 2 mm.
This is abt. 3.5mm. Unfortunately the leds had a small rim on the
lower part which made them a fraction to wide. I cut this off with a
sharp stanley knife. Of course the veroboard has to be cut diagonaly
when the leds are mounted in this way.
I wonder if you also experienced color change when overdriving
the white leds.

[Viewmaster]

Hi Albert.My display is roughly constructed in the way you discribe.
Deflection circuits using mostly 40xx circuits followed by transistor
buffers and the video circuit and sync separation from the design
from Klaas. The gamma correction calculated for maximum 160 mA.
(see my former post) The overload indication is blinking a led when
the input videolevel is exceeding the limits mentioned in the design
from Klaas. ( I use a potmeter on the input to accomodate for different
input levels )
I do not use a safety circuit untill now. The deflection circuits are
clocked by synchronised free running oscillators. If you monitor only
this oscillators, a disaster can still happen when a fault occurs in the
following counters etc.

Yes Rommert, that's something else I hadn't thought of if the actual counters go on the blink.
Both Steve and Gary have suggested possible solutions but I don't believe
that either scheme would have checked the actual counting?

For test purposes I made dummy loads with resistors for both de-
flection circuits. ( one with 32 resistors and one with 48 resistors )
As you can see on the photographs from Dr Zarkov I connected the
led matrix with flatcableconnectors, so the matrix can easely replaced
by the dummy loads.
.

I too am going to use with same cable connectors as you. The dummy load idea is useful.

My leds are 3mm diameter. When installed in diagonaly rows on
veroboard, the distance between each led becomes 2.54 x sqrt 2 mm.
This is abt. 3.5mm. Unfortunately the leds had a small rim on the
lower part which made them a fraction to wide. I cut this off with a
sharp stanley knife. Of course the veroboard has to be cut diagonaly
when the leds are mounted in this way.
.

I have 5mm dia LEDs which will just fit into 2 hole spacing on veroboard.
I thought at first I could stagger them higher/lower to avoid the flange/rim problem, but like you I will have to remove that flange at 4 points.

I wonder if you also experienced color change when overdriving
the white leds.

They get so blindingly white that it is nor possible to tell. They are just white. I can see not tendency towards blue.

I would like to thank you again, Rommert, for drawing my attention to the light leakage between adjacent LEDs. Something I would not have thought of until too late.
With my LEDs an unlit one next to a lit one does show some illumination, so that's more work to prevent this from happening.

If it all goes well I hope to hang that antique frame (it's really eBay plastic!) up on the wall as a picture with a multi cable running down to the floor to the power supply/sync sep/clocking box.
But your own matrix really does look a very professional job which I cannot match, so I will hide all my workmanship behing closed doors

Thanks again for your suggestions and ideas.

[Steve Anderson]

Both Steve and Gary have suggested possible solutions but I don't believe that either scheme would have checked the actual counting?

I'm not sure about Gary's version, but I think my suggestion does cover the counting stages as well as the clock source(s)....the monostables I suggested are triggered off the drive that feeds the bases of the row/column switching transistors. Not every row or column, just two (say) per row/column within the X/Y arrangement.

In the line direction the mono is triggered by the drive to pixel 1 transistor and 25, if either do not re-trigger to mono with (say) 1.5ms the modulation is shut off...ditto in the line direction but with a 50ms duration....somewhat over half a line/frame time respectively. A single dual mono and a couple of diodes would do this...

Albert, thanks for the PM, but I can't quite find what you're referring to, if I have missed something please can you point me in the correct direction?

Steve A.

[Viewmaster]

Both Steve and Gary have suggested possible solutions but I don't believe that either scheme would have checked the actual counting?

I'm not sure about Gary's version, but I think my suggestion does cover the counting stages as well as the clock source(s)....the monostables I suggested are triggered off the drive that feeds the bases of the row/column switching transistors. Not every row or column, just two (say) per row/column within the X/Y arrangement.

In the line direction the mono is triggered by the drive to pixel 1 transistor and 25, if either do not re-trigger to mono with (say) 1.5ms the modulation is shut off...ditto in the line direction but with a 50ms duration....somewhat over half a line/frame time respectively. A single dual mono and a couple of diodes would do this...
.

I find that rather difficult to follow but if I do have to overdrive the LEDs, will come back to you on that.

To Rommert....re change of colour of white LEDs.
I took 2 LEDs this morning and ran them at 150ma. This time they ran with a bluish tinge and after about 20-30 seconds they blew.

So now I only have 1,998 left

So it seems that 100ma is the top limit for my LEDs and even that is probably shortening their life. But in a typical NBTV picture full white is never on for long, except parts of a stationary test card.

Albert, thanks for the PM, but I can't quite find what you're referring to, if I have missed something please can you point me in the correct direction?
Steve A.

Is my face red? I have just PM'd you about it. I had wondered if I had upset you in some of the PMs in some way, by possibly leaving out a word in a sentence and so changing the meaning, maybe causing offence, but obviuosly I am mistaken. I'm getting too paranoid in my old age. !

OK you guys who are reading this....this isn't the start of a flame up so get back to your bloody soldering etc

[Viewmaster]

I would think in the line direction to have 48 (or 64) emitterfollowers, e.g BC337 and for the horizontal direction 32 BC337 s to direct the outputcurrent to the selected column. But it is not an easy circuit.

And here it is. Both column and row switching cmos and associated transistors on one piece of vero.
Rather cramped as you see....just gotta check for short circuits and track cutting now

[Panrock]

Sheer dedication!

I wish you Good Luck (and a good magnifying glass) !

Steve O

[Viewmaster]

Sheer dedication!
Steve O

I'm not sure who is ahead in the sheer dedication stakes, Steve, you or me.
Let's say we are knack and knack

I wish you Good Luck (and a good magnifying glass) !
Steve O

I believe it was you, Steve, who mentioned how repetitious some NBTV work is. Amen to that.

Have just finished drilling out 1,532 track cuts in 12" x 8" veroboard on alternative strips.
After a while I felt that I was in a kinda trance, staring at the drill and the shining copper strips.

[Viewmaster]

The comment on the interference between Leds is one useful bit of knowledge perhaps the led bricks is a better way to go...
Also perhaps instead of painting leds for the unwanted light between them place a thin black paper barrier along the rows.....

I have found that drinking straws are exactly a tight push fit over 5mm dia LED's. I have ordered a pack of black party straws (I've never been to such a party !!)

When cut into short bits will push onto the LEDs and so mask every one from its neighbour.
Just 1532 cuts required.

[Panrock]

Nice one !

Steve O