G'day all.

I need help from any of you technicians who are cluey with circuitry design for scanning and deflection.

Basically I have a few months ago completed my vidicon camera project http://www.troysvintagevideo.741.com/mycamera.html and the camera is working but there's problems, most notably with the vertical scanning. I am however starting to understand what's happening on the vertical scan. As I understand TR19's base is fed a 1v pulse every say 20ms and during that pulse period TR19 goes to saturation and shorts the current at the collector to ground and then during the pulse off period the 4.7uF capacitor charges up to about 14.5V, and then when the next pulse occurs the process is repeated which is suppose to create a sawtooth waveform. The problem is however the capacitor charges up exponentially giving a non-linear ramp which results in a picture that's vertically squashed at the top but stretches as the vertical scan progresses down the screen. I have experimented with increasing the capacitance which resulted in a better slope that looks more like a sawtooth but the peak voltage drops as capacitance is increased which is the nature of capacitors as they filter out AC ripple, this results in a better picture with more vertical width but more darker due to drop in amplitude.

I need to find a way to generate a sawtooth waveform from that 20ms pulse coming from the MC14528 chip with a linear ramp, so I was wondering if anyone here knows of what circuitry I could implement between TR19 and TR21 that would give me a linear charge up ramp to a peak voltage of about 14.5v in a period of 20ms? Any help would be MUCH appreciated.

Hmmm looking at your diagrams and waveforms, it seems that the 4.7u cap is being charged too quickly. TR20 is arranged as a constant current generator and this current should charge the 4.7u cap over the field period.

the dc voltage at tr20 emitter should be 5 volts LESS that the supply rail (15volts?) and the base voltage 5.6v less than the supply rail ( the zener voltage) These voltages should not change (much ) as the 5k pot is adjusted.

The actual current supplied to charge the cap is set by the 5k pot and hence the rate of voltage rise on the 4.7u cap.

If the circuit is correct check the zener voltage and the transistor is OK.

TR19 seems to be OK.

Brian S

G'day all.

I need help from any of you technicians who are cluey with circuitry design for scanning and deflection.

Basically I have a few months ago completed my vidicon camera project http://www.troysvintagevideo.741.com/mycamera.html and the camera is working but there's problems, most notably with the vertical scanning. I am however starting to understand what's happening on the vertical scan. As I understand TR19's base is fed a 1v pulse every say 20ms and during that pulse period TR19 goes to saturation and shorts the current at the collector to ground and then during the pulse off period the 4.7uF capacitor charges up to about 14.5V, and then when the next pulse occurs the process is repeated which is suppose to create a sawtooth waveform. The problem is however the capacitor charges up exponentially giving a non-linear ramp which results in a picture that's vertically squashed at the top but stretches as the vertical scan progresses down the screen. I have experimented with increasing the capacitance which resulted in a better slope that looks more like a sawtooth but the peak voltage drops as capacitance is increased which is the nature of capacitors as they filter out AC ripple, this results in a better picture with more vertical width but more darker due to drop in amplitude.

I need to find a way to generate a sawtooth waveform from that 20ms pulse coming from the MC14528 chip with a linear ramp, so I was wondering if anyone here knows of what circuitry I could implement between TR19 and TR21 that would give me a linear charge up ramp to a peak voltage of about 14.5v in a period of 20ms? Any help would be MUCH appreciated.


Something i never knew about before getting help on th nbtv forum was getting all the lines even you need feedback and a cap and resistor should do the trick.

Something i never knew about before getting help on th nbtv forum was getting all the lines even you need feedback and a cap and resistor should do the trick.

Not sure what you are getting at with this statement. You should not need to add anything to the circuit as shown.

As Brian said, TR20 is a current source, which will make a very linear sawtooth on the 4.7 uF. Check that the base of TR20 is at 14.4 volts and the emitter at 15 volts. Correct operation requires there is no other current into or out of the capacitor, hence the use of the FET source-follower for TR21. If TR21 is shorted, this could cause the waveform you see.

I did some calculations. The capacitor value can be calculated from the relationships:

Q=IT and Q=CV where Q is the charge in coulombs, I is current in Amperes, T is time in seconds, C is capacitance in Farads and V is Voltage.

Therefore CV=IT,

C=IT/V

In the circuit with the TR20 rheostat set to minimum (maximum charge current) and the maximum charge voltage to maintain a linear ramp at about 9.7 volts:

I= 5/1.2K = 4.2mA
One field duration is 1/50Hz = 20ms = 2x10-2 seconds

C= 4.2x10-3 x 20x10-2 /9.7
C= 8.7x10-6 = 8.7uF

So 8.7 uF is the minimum value of the capacitor for the 1.2K minimum TR20 emitter series resistance value.

Note: I see that you have the ideal ramp charging to 14.5 volts. That is not correct. It should only charge to a maximum of 9.7 volts since this is the point when TR20 saturates. Therefore something is wrong with the charging circuit such as TR20 being shorted?

I would suggest changing the capacitor to 8.2 uF or 10 uF to prevent the ramp becoming non-linear with the range of the rheostat adjustment.

Good luck: I trust you will find the problem.:yes:

Cheers,

Terry

I missed the point about the saturation voltage of TR20, so, yes the peak sawtooth voltage should be about 9.7. I don't think you need to change the 4.7 microfarad, though, because this should be in the range of the rheostat. Setting the rheostat to max 5k should get you a sawtooth that is only 3.4 v peak.

So, I agree a shorted TR20 is most likely the cause, or maybe TR21.

Don't mean to talk down to you, but are you sure the transistors are the correct polarity (PNP/NPN) and the b-e-c connections are not interchanged? In the past I made a mistake like that and got some very strange results.

Yes, I suppose that the capacitor change is not necessary. Perhaps the range of rheostat and limit resitor could be changed to prevent the ramp becoming non-linear? But then again, if the range is sufficient I guess there is no need to change anything there.

There is one otherthing that troubles me, however. That is the TR21 directly across the capacitor. Shorting the capacitor directly with TR21 is a bit vicious and might provide a bit of a strain on TR21.

I would put a resistor in series to limit the discharge current. with the ramp reaching about 10 volts, perhaps the peak discharge current could be limited to about 1 Ampere or so with a 10 ohm resistor.

I would be careful about inserting too high a value of resistance here. :scratch2: But the discharge time constant with 10 ohms is about 0.3ms compared with the ramp duration of 20ms. The capacitor must be sufficiently discharged during vertical retrace.

Yours conservatively,

Terry

Given the large range of the rheostat, I think the intention was to use a 'lytic with a minus 10% / plus 50% tolerance, and +/-10% for the resistors. So, you could design it closer, but not cheaper. ;>)

Agree that the TR19 is being pushed unnecessarily hard without a series resistor. (I think you meant TR19, not TR21?)

G'day all. I found the exact problem. TR21 is the culprit as when I removed it from the circuit I got a perfect sawtooth waveform at the 4.7uF cap and at a much lower voltage. I was using the MPF102 transistor which according to my tech friend is close to equivalent to the 2N3819 though I wonder if the slight differences with them would play a big factor in causing the capacitor to charge to quick?

Good news, it turns out that the MPF102 was defective, I bought a few more from my local electronics store and replaced the defective one and BINGO, I now have a perfect sawtooth waveform and the camera is now making good full screen pictures!!! :D I'm so glad this issue is FINALLY solved!!! :D I will post pictures shortly. The vidicon tube has very poor sensitivity (probably from wear n tear) so the best pictures I could make is pointing the camera to the computer screen. Tomorrow I will test it outdoors, I should get a much nicer picture with good lighting! :)

Here's a few pictures of my camera in operation and a pic of the sawtooth on my CRO. As seen the pictures are now full screen and the sawtooth looks like a proper sawtooth :D (although triggering on the scope is off)! As seen the pictures are very dim due to the vidicon tube's poor low light sensitivity but are otherwise looking good considering. There's still more refinements to be made to make this camera perform up to par with say a Sony AVC-3250 http://www.labguysworld.com/Sony_AVC-3250.htm but that will happen in due course! :)

Nasty DC coupling from Cap to FET
Since the FET is high Z, it wouldn't hurt to have something like a 47k resistor is series with the Gate ? I'm not surprised FET was damaged.

Glad that's fixed!
The third picture (showing the camera) looks like there could be stray light getting in (?). If so, you should try covering it.

G'day all. I found the exact problem. TR21 is the culprit as when I removed it from the circuit I got a perfect sawtooth waveform at the 4.7uF cap and at a much lower voltage. I was using the MPF102 transistor which according to my tech friend is close to equivalent to the 2N3819 though I wonder if the slight differences with them would play a big factor in causing the capacitor to charge to quick?

This extreme charging had to be due to a gate-drain short in the defective TR21. Normally the gate current should be zero, and either of these parts should work.