CMOS timebase for oscilloscope

Hi,
DDS would be very nice, but then this would require some processing power and extra hardware, but it would only replace the VCO. My object is to try to keep it as simple as possible. But DDS may be a consideration for a future project. Thanks for the suggestion.

I am sure that there are better solutions to the DAC, but a simple counter chip and an R2R chip or HW resistors will probably do the job. But that is what I intend to find out. If my solution is not satisfactory then maybe I need to seek something more technical.

BR Harry, and thanks for the suggestions. It is always nice to receive some form of feedback, and even more "food for thought" :-)

hi,
if the oscillator needs to go up to only 100khz i still think that a simple DDS (direct digital synthesis ) signal generator is best. There are also microcontrollers like pic16f1709 with DAC pheripheral inside. And the generated signal is not dependent of the mcu clock (with respect to nyquist criteria) or temperature... only the way one transfers eprom data to the DAC
The function generator i built from the net uses r2r dac made separately but i think a factory made dac inside a chip is much better. If the resistors are not the same it can be seen on scope i have to admit that. And if only sawtooth shape is needed the whole memory is more then enough to map the waveform 'in high resolution'.

Hi Ivan,
I do remember that early colour tubes had colour-mask alignment problems, but I do not remember any of the details.
Thank you for your description, I did not realise that the alignment was so complex, which probably explains why early colour TV sets were so expensive.

I am, however, familiar with watching surreal TV pictures in the days we had Beta and VHS films rented, often viewed on NTSC TV sets. That was in Kuwait and Saudi. I remember NTSC being given as standing for "Never Twice the Same Color". Most videos we watched in NTSC were shades of yellow, green and purple. Very rarely saw blue.

-

All the older crystals I have are full-specification, unused NOS. They were bought by PYE Telecom, Cambridge, England, from http://www.quartslab.com/. The 16MHz crystals are also full-spec new crystals from a computer retailler chain in Sweden. They were selling off discontinued stocks at a bargain price, so I orderd all the packs from all their outlets.

The 70MHz crystal filters were used in later generation, synthesised PMR radios, also manufactured by PYE telecom. PYE Telecom have ceased to exist many decades ago, but I filled my pockets with much of the new, unused stuff they discarded, before I left. I still have a roll of 0.7mm Dia enamelled copper wire on a 20kg roll. The weight is 20.3kg, including the drum.

Today I was searching in the shed in Spain and found some 10MHz crystal oscillator modules. So all the modules and crystals are even numbers, and in quantity. Looks like there are some oportunities for a bit of inventive thought. For example, the 45.000MHz x3 = 135MHz, which is exactly 10MHz from the 2m band. 11.155MHz puts it in the FM section, repeater input. Or conversely, 45.000MHz x 3 from 145.600MHz is 11.6MHz, right in the middle of the max-gain of the TDA7000, and the TDA7000 will do NBFM quite easily.

The Meissner self-oscillating mixer is most stable at low frequencies, so a 16MHz with a Meissner 1.65MHz to 2MHz receiver, or a modified domestic radio, will cover 14MHz. Could be a really cheap and simple portable GC receiver using an old MW receiver with a converter.

I can see loads of possibilities for experimentation. I just wish I had more time.

BR Harry

Hi Harry,
do you remember first colour TV sets with with "delta" CRTs? There was special circuitry to get three independent electron beams into a small hole in the iron mask and then to their corresponding grain of luminophor - across the whole screen. It had about 40 variable coils and resistors, that influenced each other. When regulating them, the screen had to be observed in a microscope.
I remember a sports broadcast in 1975: commentary (for those who already had B/W TV sets):"The boys have yellow T-shirts." On the CRT: boys in magenta T-shirts!
Another TV set a bit later, damaged probably by a power spike. A cooking magazine. Sound:"We are going to prepare salad from red radish and green peas." On the CRT: blue radish and violet peas... Modern devices do not give us such surrealistic pictures!

I do not know Proteus. I have successfully tested Formica, its limited version used to be (I hope still is) free of charge.

The technology of loudspeakers was probably changed, when small and powerful neodymium magnets became widely used.

I have no idea about the crystals now. Did you mention, that crystals, available at sales for a few cents, usually have strange frequecies?

VBR from Ivan

Ivan wrote:Ivan
The linearity of your CRT may be affected also by its geometry. The electrons travel a longer way from the deflection plates to the edges of the screen than to its center. But I may be totally wrong at this point.
VBR from Ivan

Hi Ivan,
I am sure that you are quite correct. When it comes to linearity, there are all manner of things that can affect it. Distance from cathode to screen - never really thought about that. Thank you.

But my screen is small and not flat. It will be interesting to drive it with a digitally created sweep and actually do some measurements, also do some physical measurements of the tube. I shall keep my fingers crossed, because linearity is something I really rely on (or have got me little heart set on ).

That must also be a huge problem for the old "fat" TV sets, whre the cathode to screen distance can vay be several cm. But I suppose spaghetti-westerns and police explosive care-chases are a bit like listening to rapp-music on HiFi, who will notice the distortion?

-

Anyway, I will probably, just for a laugh, post a couple of smaller projects I made to test Proteus the PCB CAD program. The first I made was a retro-AF amplifier, with blue and red transformers, etc. I managed to get the scale wrong. In addition to that I made the mistake of mentally viewing the board from the copper side. Lessons learned, so I may as well share the end product.

I am also working on a retro RF mixer + 455kHz IF amplifier, but with a custom plug-in oscillator coil to form a 6-pin version of Alexander Meissner's self-oscillating mixer.

In about the 70's small loudspeakers became a load of rubbish; flimsy and fragile things. Today there are loads of speaker elements that have a really good bass response, often used in those old retro-looking radios (with TDA7000 series) or the cheap Bluetooth external speakers. I find them in the rubbish bins all the time.

Maybe time to have a bit of nostalgic fun

-

BTW, I have just bought about 50x 16.000MHz crystals, and I have found in my junk 100s of 45.000MHz (15.000MHz fundamental) crystals. Also 100s of 11.155MHz crystals. I have also found some 70.000MHz crystal filters. I wonder if the crystal fundamentals themselves are 23.333MHz? These could be interesting frequencies. Any interesting ideas Ivan?

Very best regards from Harry - EA/SM0VPO (rig-less, - PS RadioWorld have promised to loot at the radio middle of December. Looks like I don't get it until after Christmas)

Admin wrote:I didn't know the 4046 oscillator was temperature dependant. I had one that was varying with frequency, but I had accidentally used a ceramic cap for the Ct. When I changed that for a poly cap the problem seemed to be cured.

I was quite surprised when I was looking at sweep linearity. With a small tube you look a lot closer, and you can see cycles at the sart and end of the trace have different widths, more-so than with a larger display.

Hi Harry,
I did many experiments with 4046 some 35 years ago (ughh ). The technology has advanced since then, so its stability may be better now. The capacitors I used were styroflex and resistors were metalized (no carbon ones).

The linearity of your CRT may be affected also by its geometry. The electrons travel a longer way from the deflection plates to the edges of the screen than to its center. But I may be totally wrong at this point.

VBR from Ivan

Hi Ivan,
Both sound fantastic ideas  
I like the idea of using the PLL to sync the timebase with the input. That could be really useful.

I know that the PSD-I in the 4046 can sync to frequency multiples, and I have used that instead of a small divider in one of my experiments. That sounds like a really simple idea that can help give a stable trace.

I didn't know the 4046 oscillator was temperature dependant. I had one that was varying with frequency, but I had accidentally used a ceramic cap for the Ct. When I changed that for a poly cap the problem seemed to be cured.

Anyway, at the moment my scope is sitting at home with masses of spaghetti-wires everywhere after trying different timebase generators, but now I have had the time to do some thinking I hope that I can produce a simple instrument that is really useful.

I was quite surprised when I was looking at sweep linearity. With a small tube you look a lot closer, and you can see cycles at the sart and end of the trace have different widths, more-so than with a larger display. I physically compared with a PC-scope. On the large 16" PC scope display you have to physically measure to see the differences.

But as you say, 7-bit should be adequate with capacitive smoothing. I have a 2-pole 6-way switch for the range selection, and I was thinking about using the second pole for 7-bit / 9-bit selection. But the thing to do is see how it performs - "suck it and see"

As regards the frequency counter, Ihave limited space, but I do have a single-chip counter. Perhaps I can add that directly to the input (after the amplifier) as a simple/cheap method.

Many thanks for your coments. Really appreciated.

Best regards from Harry - EA/SM0VPO

Hi Harry,
the circuit should work.
For a 60 mm long timebase, even 7 bit counter should perfom well: each step represents less than 0,5 mm on the screen. If the timebase is smoothed by a capacitor, it is adequate. The smoothing capacitor should be switched simultaneously with the range one.

I have one more rough idea: add a function closing the PLL and making use of one of the phase detectors in the 4046. This should enable automatic synchronization (not triggering) of the timebase with the input signal. If you attach a small frequency counter (a PIC with a LCD display?) say on the Q9 output of the 4040, you may get a direct frequency reading of simple waveforms.

My experience with the free running 4046 VCO is that its frequency drift is quite big. Probably voltage and temperature stabilization will be necessary. The instability does not matter in its original use, when the PLL is locked.

Very best regards from Ivan