Noise on TV output

inaki

I have moved this question from my former thread, where I was asking about PAL mode.

Now I get the picture but I get as well·a lot of noise over the picture.
By noise I mean shadow lines moving on top of the picture.
In the TV Terminal demo the noise makes the text to tremble slightly.
The noise is strong enough to make the TV mode pretty annoying.

I have tested the video·demos with two different Propeller configurations, one my own, and the other the Propstick.
I have changed the video cable and used two set of resistors. No changes, the noise is still there.

Do you get a clear, solid·picture ?
I saw in the demos provided by Beau that the picture seems absolutly clear.

I am using PAL mode (bit 0 on _mode set to 1). Anyone else using PAL ?




▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Mike Cook

Sorry I can comment on the PAL mode, but I'm using a NTSC Sony HR broadcast monitor and the video output from the Propeller is rock solid.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Mike

Gadgetman

I'll try as soon as the Propellers and Propstick arrives.
(Assuming I have the right resistors, and doesn't screw up completely)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...

Pezi_Pink!

The PAL specification can be VERY picky / fiddly when coding the raster for a given tv set.

Although the Propeller PAL mode may be coded very closely to the "standard" (although there are several sub specs for PAL as well) it may well be that your TV just doesn't quite like the signal.· Do you have another set you can test it on?· The top few lines after the overscan can get "bent" and colour can phase in and out slightly with only minor timing variations between the code and tv set.

When I got my XGS, although it was tested on several PAL sets in the USA, I coudln't even get the PAL demos to show any colour without substantial modification.· I'm not sure how the Propeller handles the colour burst and the different frequencies for PAL/NSTC·so I can't comment on that.

Although I am English, wherever possible I will use NTSC as it is simply much more forgivng.· Not exactly a great solution, but if you can get a multi NTSC/PAL set or a NTSC set it may well save a lot of headaches in the future.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beware the pink squirrel!

inaki

Pezi,
I tried the TV tests against two different monitors. The noise is the same on both.
I think I can get an NTSC monitor. I will try to see what happens.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Pezi_Pink!

Hey inaki,

It would seem very strange that the problem is *identical* on both sets.

That would seem to indicate the signal generation software/hardware rather than the tv? I'm still wating for my Prop, I'll let you know how it goes on my sets once it arrives.

Is it possible you have some device nearby such as a desk fan that is causing the interference?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beware the pink squirrel!

inaki

Yes it might be some device is causing the interference.

However, if I connect other video input on the same table to the same TV I get a very clear picture. If there is an interference generator in the nearby should not affect *any* video source ?

By the way, I have tried also·an USB TV capture device (Avermedia USB 2.0) to get the picture but I get nothing. It·supports NTSC and PAL but it does not work. I suppose the video signal feature may be a problem for some TV devices.



















▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

LoopyByteloose

It might be the resistor values.

I see 1.1K, 560, and 270 ohm.
Is this really an aproximation of a ratio of one; one-half; and one quarter?

That would ideally be 1.1K, 550, and 275 ohm.
If so, finding a way to tighten the values might help.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

······································································ Warm regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

cgracey

Both NTSC and PAL signal generation should be perfect. They are to spec, and there were no compromises. The NTSC is 3.579MHz color and the PAL is 4.433MHz color.

It's important that you clock the Propeller with at least 40MHz internally. Remember not to go over 10MHz on the external crystal, as the PLL always multiplies it by 16 and 160MHz is the upper design limit. If you're seeing noise, it might be that there is insufficient capacitance between VDD and VSS, but this is probably not the case.

There's a slight chance that the video generator on that particular COG could be broken since our tester is still evolving to trap more failures, and the video generators haven't been·tested yet. By the end of the month, this will be tested for. One way to find out if that COGs video is broken is to start up another COG at the very start of your application, so that when the TV driver is subsequently launched (perhaps within tv_terminal), it will go into the next COG, instead. This can be done with the following code:



VAR

· xcogstack long[noparse][[/noparse]8]

PUB yourapplicationsfirstpub

· cognew(xcog, @xcogstack)·· 'first like of code in first PUB

· tv.start (or term.start(pins))



PRI xcog

· repeat·· 'loop forever



If it turns out that the video generator is bad, we'll get you another chip right away.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.

Phil Pilgrim (PhiPi)

Hi Inaki,

Are you using a noisy switching power supply or a wall transformer with inadequate filtering? (Some "DC" wall transformers -- primarily those designed for battery charging -- don't have filter caps.) Either one could be allowing noise or "hum" on the 3.3V rail, which would result in the behavior you're seeing. If you have a scope, take at look at your 3.3V line, and make sure it's rock solid and noise free.

-Phil

Paul Baker

Also physically remove the propeller from the TV, if I remember correctly, locating the propeller near the TV creates interference, use the length of RCA cable to prevent this.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
1+1=10

Phil Pilgrim (PhiPi)

Also, Inaki, you didn't mention whether you had a direct baseband connection to your TVs or were going through a modulator. If the latter, Paul's comment is especially poignant, since the TVs' RF section is not being bypassed and can pick up all manner of interference. Also, if there's a switch on your modulator (or TVs) that says "75-ohms", "terminate", or somesuch, make sure the switch is in that position. Switches like this allow daisy chaining your baseband connection, wherein only the set at the end of the chain gets terminated. But without any termination, your video is likely to misbehave badly. If there's no switch, and if in doubt, you can check the input with an ohmmeter. You should get a reading of 75 ohms.

-Phil

LoopyByteloose

Inaki,
It just occurred to me that some RCA connectors are not very secure. The ground lead is a solder tab that is held by a nut. If you have that kind, just try tightening the nut.

This all could be a loose ground wire.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

······································································ Warm regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

Coley

OK guys, I need some help here please.

I have a DLP-PROP and can't seem to get a stable NTSC or PAL image on anything from widescreen TV to professional grade composite monitor (NTSC & PAL).

The images are there and quite clear but there is no colour and the whole screen is jittery.

I do get some colour glitches, say two a second but it is a very short flash.

I have tried the example from Chip suggesting starting two cogs and I am no better off so I guess the video generators are fine.

None of the TV or VGA outputs signals are great, and I have double and triple checked my resistors - everything is as it should be.

The ONLY thing that I can think of now is the crystal speed of 6MHz which is generated by the on-board FTDI chip.

I've attached a modified program that I have been using, bear with me because this is my first day with the Propeller...

Regards,

Coley

Mike Green

A 6MHz crystal should be fine as long as you specify the crystal frequency in your program (which you've done). Can you post your schematic? Is your power supply well filtered?

Coley

Mike,

Thanks for the speedy reply, my schematic is as per the prop demo board

P12 --1.1k ----|
P13 --560R----|----Composite Video
P14 --270R----|

And of course a ground connection.

The supply is straight off the DLP-PROP which is supplied via the USB connector.

I've tried this on my home machine and my laptop with the same results.

All demo's run just fine but the video signal is screwed.......

Picture is more stable if I strip out the chroma but still slightly jittery.

Any ideas?

Regards,

Coley

Mike Green

Sorry, I don't see anything obvious. I haven't tried two TV outputs at the same time, but it ought to work (assuming you've got a 2nd set of resistors on pins 8-10).

Coley

Yes I have a set of resistors on pins 8-10 and the give exactly the same result as pins 12-14.
That's what made me think that it was a problem with the crystal....

Mike Green

There may be some problem with the DLP-PROP clock. You might e-mail FTDI support for help

cgracey

If I'm reading the DLP-PROP schematic correctly, there is no crystal. The FT232R chip outputs ~6MHz from some internal clock generator. It is unlikely that this signal is close enough to the ~6.000MHz MHz required for resynthesis into ~3.579545MHz or the PAL equivalent. Monolithic clock sources are just not that exact. For the purposes of the FTDI FT232R, that 6MHz only needs to be accurate enough to serve as a short-term lock·for the USB signal, after internal multiplication to probably 48MHz. To get colorburst frequencies to register with TV's, they need to be quite exact. As far as that jitter goes, try looking at the FT232R's·6MHz output on a scope and see if you see jitter there. That would explain why you see jitter after the Propeller's PLL, too (unless it's just a colorburst-error-induced jitter).

Here's the DLP-PROP datasheet:

http://www.ftdichip.com/Documents/DataSheets/DLP/dlp-prop-dsv10.pdf
·


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.

Coley

Chip, you are right to say the DLP-PROP has no crystal, I already knew that.
I should have been more precise when I said that it was a problem with the crystal.....
I will scope the output from the FTDI chip, it may be that it is not fit for the purpose of generating stable video signals.
I have ordered a starter kit from Farnell in the UK so I will wait until that comes and compare the two platforms.
I have to say though from first impressions and despite my 'problems' that you have done a fantastic job with the Propeller, it is phenomenal!!!

Thanks go to you and Mike for your help....

Regards,

Coley

Coley

Ok got the starter kit today and guess what it works just fine.
So it would seem as though the clock output from the FTDI chip was just not stable enough to generate a good video signal.
I would say the DLP-Prop board is good just not quite good enough for time critical applications.
It will be winging it's way back to FTDI in the morning.

Regards,

Coley

Phil Pilgrim (PhiPi)

What's really odd about this whole episode is that FTDI made such a fuss in their FT232BM specs (an older device) about how accurate the external 6MHz clock source had to be!

-Phil

Coley

Well I checked it out with the scope and it was averaging around 6.314 MHz but the waveform was all over the place.
Double checked against the 5MHz crystal on the Prop Demo board and it was rock steady as expected.
It would seem that when first started up the FTDI gives out as near to 6MHz as I can tell but once attached as a USB device everything goes awry....
It could just be faulty I guess, I will give them the benefit of doubt until i get a reply from tech support.

Coley.

bambino

When I ohm the input on my B/W portable TV set I get 0 ohms. There is no switch like the one mentioned in this thread. Is that a possible reason for my inability to get a picture? I have tried it with 4 different platforms and no luck.

I did get the graphics demo to work with one of the platforms, but TV text, and Terminal are out!

Mike Green

Your TV may have a transformer input in which case it may measure 0 ohms with a meter. I'm guessing that you're talking about the antenna input to the TV (broadcast). If you want to use broadcast mode, you have to modify the programs (the parameter table). The comments at the end of TV.spin describe the parameter table and the table itself is in tv_text.spin. With a 5MHz crystal, it works best on channel 3 with a carrier frequency of 60MHz. If you can replace your crystal with a 6MHz one, you'll get a better signal on the other low VHF channels since the channel spacing is 6MHz.

bambino

Yes, The input physical antennia connection measures 0 ohm's.
The only way I have gotten it to work is in broadcast mode with TV graphics. The TV terminal & Text I have never got right. It seems ,with the Graphics in broadcast mode, the further away from the prop I take it the better the reception.

Mike Green

I did get tv_text to work with a little Sony portable B&W TV with an analog tuner, but I could only get a good signal on channel 3 at a distance of 30cm - 1m with a short (10cm) antenna plugged into the demo board's video output jack.

bambino

Well, there's hope then.
At any rate, I got it at a Flea market for $5 dollars to make an alarm clock out of it with the Propeller embedded in the battery compartment. The Radio in it works fine so it isn't a complete loss. My first attempt at it used a 6Mhz crystal so if that was better, the other flaws in my system wouldn't let me see it.