Construction Diary -- Part 1, The LED matrix/display

Original build of a televisor by a complete novice.

Moderators: Dave Moll, Andrew Davie, Steve Anderson

Postby Klaas Robers » Thu Apr 05, 2007 5:46 am

Andrew,
before you are going to control the spinning of the disc try it first. See that you can change the speed of the disc by hand and look though it to your flashing LED's. Piece of opaque paper or matt ground glass in between. You will be seeing pictures and you will look at the whole disc. This is the most breath taking experience that you will have. All next experiences with rock stable synchronised discs are peanuts compared to this first exercise.
User avatar
Klaas Robers
"Gomez!", "Oh Morticia."
 
Posts: 1557
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Postby Andrew Davie » Fri Apr 06, 2007 12:12 am

I think I still have something very wrong. Although my LED matrix does indeed pulse in a 'meaningful' fashion -- it is unlit when no track is played, unlit for track 2, bright for track 5, and pulses for track 20... I decided to view the waveform across the + and - of the LED power inputs. As can be seen from the first picture, there's a bizarre sub-component of the waveform -- the big partial sine wave.

So I decided to measure the waveform (track 20) at the collector with and without the LEDs present. That can be seen in the following two images. There appears to be a large amount of high frequency noise added when the LED matrix is connected (the first), and in particular the higher voltage 'colours' have been lost ...? The last image is the 'normal' waveform at the collector with no LED matrix connected.

The matrix is powered from the 17V (possibly 'ripply') DC supply, shared as a power supply with the sync board.

Can anyone diagnose these traces for me?
Attachments
tk20collector.jpg
tk20collector.jpg (63.42 KiB) Viewed 12105 times
tk20led.jpg
tk20led.jpg (55.2 KiB) Viewed 12105 times
tk20atmatrix.jpg
tk20atmatrix.jpg (67.89 KiB) Viewed 12105 times
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1584
Joined: Wed Jan 24, 2007 4:42 pm
Location: Queensland, Australia

Postby Klaas Robers » Fri Apr 06, 2007 1:22 am

Andrew, it is difficult to analyse the oscilloscope pictures, as I lack a lot of information.

1. I should know the sensitivity settings.

2. Is the ground of the oscilloscope under all circumstances connected to the ground of the circuit?

3. Is the ground of your 17 volts the same ground? I have seen some photographs some time ago without connected ground. You use best the same 17 volts that goes to the stabiliser on the LED-drive board. Forget about the other power board. The input supply voltage of the Driver board may be 25 volt as a maximum, so connect to the 15 volt outlet of the transformer.

4. Measure the 17 volt (or more) of the LED supply with the oscilloscopes ground connected to the ground of the PCBs. This should be an almost ripple free DC voltage.

5. A measurement of the collector voltage with nothing connected to it is a useless measurement. If you suspect the LEDs then in stead of the LEDs connect at least a resistor of 100 ohm from the collector to the +12V FROM THE DRIVER PCB!! and then connect the oscilloscope. You should see an inverted video signal "hanging below" the +12 volt line.

Try this too: With no oscilloscope connected play track 2-4 and gently turn up the brightness control. The LEDs should start to glow. The correct setting is a very dim glow or just no glow.

I got the feeling that the hum comes from the oscilloscope or from a wrongly connected second power supply. For the first the oscilloscopes ground should be connected to the ground of the PCB. Be aware that the ground of the oscilloscope is connected to its metal case, which is again connected to the safety ground of the power grid (230V). If the oscilloscope is ungrounded there is a big hum induced to the oscilloscopes ground from the power transformer and/or the mains filter circuit in the oscilloscope. This is a disadvantage of oscilloscopes, they are normally always grounded. So called "floating" measurements are virtually impossible.

For the second: please use only one power supply for the time being.
If you have two rectifiers connected to the same transformer coils there are very strange things to be expected.
User avatar
Klaas Robers
"Gomez!", "Oh Morticia."
 
Posts: 1557
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Postby Andrew Davie » Fri Apr 06, 2007 6:41 pm

It's probable I was just using the oscilloscope incorrectly again. I'd placed the black of the oscilloscope on the negative of the LED matrix, and the red on the positive. This time I did it with the red on the positive, and the black on a ground on the LED driver board.

I'm including the following images more as a nostalgic reminder for myself of how it all looked when I considered it to be "working". I am happy enough to move on to understanding how an electric motor works.
Attachments
track2LED_AC.jpg
track #2, AC-coupled display 0.5V/div, 0.5ms/div. LEDs are *just* showing some light.
track2LED_AC.jpg (67.12 KiB) Viewed 12090 times
track5LED_AC.jpg
track #5. same settings as track #2. I assume this is the sort of "inverse signal" mentioned.
track5LED_AC.jpg (67.01 KiB) Viewed 12090 times
workbench.jpg
The workbench area. I have highlighted the most important and frequently used component.
workbench.jpg (82.92 KiB) Viewed 12090 times
workingrig.jpg
Close-up of the rig in operation. Wires everwhere, but the LEDs glow.. when they should!
workingrig.jpg (73.05 KiB) Viewed 12090 times
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1584
Joined: Wed Jan 24, 2007 4:42 pm
Location: Queensland, Australia

Time for a clean up?

Postby Steve Anderson » Fri Apr 06, 2007 7:48 pm

Andrew,

Having scrolled through a few of the pages of this thread one thing I have noticed is the apparent suscepability to hum pick-up as shown in the bench photos.

Firstly sink a 2m metre earth rod into the gound nearest your 'workshop', (if your house doesn't already have one), then run a 10mm2 (10mm squared) minimum earth cable in as direct route as you can to your bench(es).

Get an Aluminium sheet cut to match your worksurface(s) less 10mm in each direction, (any thickness) then a safety glass sheet on top of that the same size at the workbench(es). Bond the earth cable to the Aluminium sheet(s) and ignore the usual building earth/ground, replace that with your new 'Technical' ground, still using the live and neutral as usual.

Although nowhere as good as a Faraday cage, it's a step in the right direction, it's the arrangement I have here. I also have a good mains filter to take out the worst of the spikes and 'bumps' in the incoming 220V, but with the buried power cables in most developed countries less required. All power here is overhead and prone to lightning strikes especially during the wet season which is due to start soon. PCs really don't like it! I have a line-conditioner and UPS for them.

...moving on...keep your OV/ground/earth leads to all connected devices as short as practible, between PSUs/PCBs/anything.

I also notice that I don't see a proper 'scope probe, just a coax cable with a couple of 'croc' clips on the end. Op-amps and most semiconductor circuits don't like the capacitance of the cable and the input capacitance of the scope. This can add up to over 200pF, many circuits will go unstable with this hanging on their output unless specifically designed for it.

The next thing is supply decoupling. Each op-amp ideally should have someting like a 10 to 100nf disc ceramic capacitor between its supply rails, two if using split supplies. In addition bung in an electrolytic cap of about 100uF from each supply to ground per board, don't forget electrolytics are polarized!

Attached is a pic of the extents you have to go with dealing with logic only running at 40MHz! A 10nF disc ceramic capacitor on the underside of the board to provide sufficient bypassing/decoupling.

The recent photos would appear that something is unstable, also build/purchase a couple of decent stabilised power supplies so the power source can be eliminated as a suspect.

Cheers, it's not easy, but it's worth it!

Steve A.
Attachments
cap_1.jpg
cap_1.jpg (85.44 KiB) Viewed 12005 times
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 4540
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: Time for a clean up?

Postby Andrew Davie » Fri Apr 06, 2007 11:51 pm

Steve Anderson wrote:I also notice that I don't see a proper 'scope probe, just a coax cable with a couple of 'croc' clips on the end. Op-amps and most semiconductor circuits don't like the capacitance of the cable and the input capacitance of the scope. This can add up to over 200pF, many circuits will go unstable with this hanging on their output unless specifically designed for it.


After reading the above about my alligator-clip probes, I had a dig through the various accessories that came with my oscilloscope. I found a rather odd single-tip probe with a 'box' on the end, so I hooked that up, and a wire to the ground connection on the front of the oscilloscope going to the 17V ground.

Interesting.

I see much 'cleaner' signals on the oscilloscope, but the values are not at all what I expected to see. The probe has 'done something' to the signal. A very clean looking signal, but obviously I'm going to have to learn how to read it properly. The 'box' at the end of the probe (just before the connection to the oscilloscope) is marked 10M ohm, 12pF.

It was fun having a 'new' look at the signals coursing through the circuits, just to see with a bit more clarity what was happening. All seemed as expected until I tried connecting the probe to the + of the LED matrix.

After a double-take, I realised what I was seeing here (see image) was the ripple on the rectified power supply! I was unable, for the life of me, to get this view using my old faithful alligator probe. Of course, as noted, I can't make sense of the actual readings of this yet. It LOOKS to me like the probe is reducing the voltage readings to 10% of their actual value, but increasing the sensitivity.

For example, when I measured the rectivied 17V supply, my reading of the value onscreen was 1.7V. The image below was with settings 50mV/div, 5ms/div. That would, then, uh.... give the ripple at roughly +/- 150mV (each division is 50mV, probe attenuates to 10% so actual value is 500mV... ripple is 60% of this (300mV)). Note: I had shifted the trace vertically upwards, so the absolute vertical position is meaningless.
Attachments
ripple17.jpg
ripple17.jpg (71.04 KiB) Viewed 12072 times
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1584
Joined: Wed Jan 24, 2007 4:42 pm
Location: Queensland, Australia

Postby Klaas Robers » Sat Apr 07, 2007 4:01 am

This a so called x 10 probe. It gives an attentuation of 10x. So when you read 1 V it is actually 10 volts.

Now you have to calibrate the probe with the 1V square wave. It should give a correct square wave on your screen. If not rotate the small screw in the connector-box until the square wave has no "sag". You will see that you can have "sag" as well as "rise". The horizontal line then should be 0,10 volt (because of the 10x attenuation). Use this probe in stead of the crocodile-clip coaxial cable.

But be aware that the oscilloscopes ground is still connected to the safety ground.
User avatar
Klaas Robers
"Gomez!", "Oh Morticia."
 
Posts: 1557
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Probe ground

Postby Klaas Robers » Sat Apr 07, 2007 4:08 am

Andrew, dig again in the junk that came with the oscilloscope.

You will find (a) short lead(s), say 10 cm, with at one end a small crocodile clip, on the other end some other type of clip. This clip can be clipped to the back-end of your probe and gives you a better ground than the wire that follows a different path to your oscilloscope. This is the professional way to use an oscilloscope.

We assumed that you knew this all. Stupid us....
User avatar
Klaas Robers
"Gomez!", "Oh Morticia."
 
Posts: 1557
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: Probe ground

Postby Andrew Davie » Sat Apr 07, 2007 11:01 am

Klaas Robers wrote:Andrew, dig again in the junk that came with the oscilloscope.

You will find (a) short lead(s), say 10 cm, with at one end a small crocodile clip, on the other end some other type of clip. This clip can be clipped to the back-end of your probe and gives you a better ground than the wire that follows a different path to your oscilloscope. This is the professional way to use an oscilloscope.

We assumed that you knew this all. Stupid us....


also ref: http://www.taswegian.com/NBTV/forum/vie ... 1&start=84

I found the short lead, with a sort of sping clip on one end and an alligator on the other. It just happens to fit snugly around the sheath near the tip of the probe, and that's where I have it. I did a measurement with the alligator clip connected to ground, and it does work nicely. Thank you.

As I said at the beginning, I am a complete electronics novice. I'd never touched the dial of an oscilloscope before starting this project. Everyone's help has been most useful. Thanks.
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1584
Joined: Wed Jan 24, 2007 4:42 pm
Location: Queensland, Australia

How to use a 'scope.

Postby Steve Anderson » Sat Apr 07, 2007 12:40 pm

Andrew,

Attached is a 'scope primer from Tektronix, it's aimed mainly at the newer digital scopes but still equally applicable to analoge ones like yours. There are many others on-line as well. Do a search on 'How to use an oscilloscope'.

Cheers,

Steve A.
Attachments
Scope Primer.pdf
(1.44 MiB) Downloaded 614 times
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 4540
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

'Scope probes.

Postby Steve Anderson » Sat Apr 07, 2007 1:14 pm

Andrew,

On page seven of this thread Gary posted some words about 'scope probes along with a picture. I guess both Klaas and I assumed you were using one.

Perhaps because it was posted on April 1st you thought it was an 'April fools' joke? (Just joking).

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 4540
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: 'Scope probes.

Postby Andrew Davie » Sat Apr 07, 2007 1:36 pm

Steve Anderson wrote:Andrew,

On page seven of this thread Gary posted some words about 'scope probes along with a picture. I guess both Klaas and I assumed you were using one.

Perhaps because it was posted on April 1st you thought it was an 'April fools' joke? (Just joking).

Steve A.


Sometimes I don't realise how relevant information is, until much later. I was aware of the posting, just not the importance.
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1584
Joined: Wed Jan 24, 2007 4:42 pm
Location: Queensland, Australia

Breadboards.

Postby Steve Anderson » Sat Apr 07, 2007 1:42 pm

Andrew,

Another suggestion is in addition the essential tools you already have, 'scope, multimeter etc, you get a 'breadboard' as they are known. The picture attached shows one I've had for some 15 years (hence the dust and crud on it!).

This allows you to build up a circuit without any soldering, adjust values and confirm the circuit works before committing it to a PCB or stripboard (Veroboard).

It speeds up development vastly and you don't burn yourself with a soldering iron as often!

They are available in various sizes, this one is one of the larger ones. They're not that expensive and repay their value on the first thing you develop.

The only downside is there is a little more capacitance between the strips than found on a PCB, but this is unlikely to be a problem with NBTV circuits or moderate speed logic circuits. I've used this one with logic running at around 100MHz.

Cheers,

Steve A.
Attachments
breadboard.jpg
breadboard.jpg (216.03 KiB) Viewed 12007 times
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 4540
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Klaas Robers » Sat Apr 07, 2007 11:06 pm

The disadvantage of these bread boards is the bad contacts. Now and then a circuit doesn't work because a contact is unreliable. But some moving of the wires might solve the problem.

I never use these things. When you are more experienced in soldering and desoldering also experimenting can be done in that way. At our QRL (my work at Philips Electronics) we have PCB's with only a grid of pads around plated through holes. The connections are made then by thin wires. I will add a picture
Attachments
aanstuurprint2.jpg
This is the components side of a driver PCB for 6 Nixies
aanstuurprint2.jpg (82.27 KiB) Viewed 12006 times
aanstuurprint1.jpg
This is the wire-side of the same PCB. The wires are "wire wrap" wires with insulation of Teflon
aanstuurprint1.jpg (63.69 KiB) Viewed 12006 times
User avatar
Klaas Robers
"Gomez!", "Oh Morticia."
 
Posts: 1557
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Breadboards & 'protoboards'.

Postby Steve Anderson » Sat Apr 07, 2007 11:38 pm

Gents,

Klaas is quite right in what he says, but the 'protoboard' can only be used generally once, akin to stripboard (Veroboard). I tend to develop on the breadboard just a small section at a time, transfer this to a stripboard and move onto the next section.

Using a breadboard for complete large and complex designs is a no-no. With that I agree. For small sections as shown I will admit that occasionally one does get an odd open circuit, but even with that, t's still a rapid way of checking that a simple part or section of a design performs as advertised.

For amateur construction Klaas's protoboard is fine for the 'final version', just simply can't get it here otherwise I might give it a go!

By the way Klaas, very neat!

Cheers,

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 4540
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

PreviousNext

Return to Andrew Davie's Televisor #1 (2007)

Who is online

Users browsing this forum: No registered users and 1 guest

cron