Video Polarity and Direct Injection
 

I have a set into which I'd like to directly inject video. My problem is the set only has one video amplifier tube so the video polarity at the grid of the video amp is positive. Injection after the video amp isn't doable as I don't have enough gain, and injection before the video amp isn't doable because the polarity is backwards. Can anyone tell me how flipping the polarity on composite video might be accomplished so that I may inject positive video into the grid of the video amp?

A phase inverter using a single triode shouldn't be too hard to slap together. No doubt the collective wisdom on here can ace a quick design.

A factoid worth bookmarking is that a tube's plate is opposite phase from the grid, while the cathode (if unbypassed) is in-phase with the grid.

Or with a single transistor:

Video Inverter near the bottom of the list.....

http://www.hobbyprojects.com/A/video.html

I built a crosshatcht generator using an EDN magazine "Design Idea" and tried the video out with a salvaged Astec modulator - no joy, so I rigged up a 2N2222A as an inverter, and voila....


BTW, it still works, and is more stable than my RCA portable unit...

http://oh3tr.ele.tut.fi/~ftp/video/videoinv.gif




Would a 2N4401 work in that circuit?

Sure - any low gain (hfe<500) NPN should work. If you get any ringing, try ferrite beads on the base and collector leads.

I built my 2N2222A inverter on a double sided PCB, with the bottom as a ground plane, and "Manhattan" style construction on the top. The circuit I used was in "Circuits Cookbook" from 1967....

The whole circuit takes advantage of the common emitter configuration for inversion - and - the common collector configuration gives you the non-inverted video on the emitter. Two in one!

https://en.wikipedia.org/wiki/Common_emitter
https://en.wikipedia.org/wiki/Common_collector

Probably need some gain
 

Can anyone verify that the circuit I posted should work? I can't get anything out of it.

You mean you've built it already? If it's not working, the transistor may not be biased correctly. First measure the emitter/base voltage, which should be about 0.7 V. If the voltage is too low or missing, you need to increase value of the base-to-ground resistor.

Or, if the emitter/base voltage is too high, the transistor will overconduct and saturate. So lower the voltage by increasing value of the base-to-12V resistor.

This is all assuming the transistor is NPN silicon type and supply is +12V.

BTW, the circuit shown will deliver only half the output level that it would if the non-inverting feature (which you don't need) were eliminated. The mod would be easy to do if the output level turns out to be inadequate. Here's a neat little tutorial on transistor biasing..
http://www.electronics-tutorials.ws/...r-biasing.html

How would I accomplish that?

By the way, I'm using a C1815. I got the emitter/base voltage to .7VDC. I have output now, but just a minuscule amount.

The problem is that the base is biased at about 1/2 of B+, which makes the emitter that (6 v) minus 0.7 volts = 5.3 v, and the collector drops to 12-5.3 = 6.7 v, so the transistor is almost saturated at one end of the circuit's output range. You need to reduce the lower base resistor or increase the top one to get about 3.7 volts at the base. Then the emitter will be at 3 volts and be able to swing almost from ground to 6 volts, while the collector will be at 9 volts and be able to swing almost from 6 to 12 volts.

You could try leaving the lower base resistor as 4.7k and increasing the upper one to 15k or 18k.

I always use a pot...
 

say 100K or so and adjust it till I get what I want with a scope on the output. Then measure the pot and substitute fixed resistors.

Since the non-inverting output feature is not needed, it could be eliminated. This would allow the output level to take advantage of the full supply voltage (rather than half as it is).

It's always useful to think of a transistor as a triode analog - with the emitter the cathode, the base the grid, and the collector the plate.

In the sketch shown, start out with R2 as 4.7K, and R3 as 1K. Then for R1, first try 100K and work down in value till you get 0.7 V on the base. Once a usable output is obtained, then play with values for R2 and R3 and see if the output level can be improved any, all the while tweaking R1 to maintain 0.7V on the base.

Edit.. Error in sketch, 1K resistor "R2" should be R3. Dumb, dumb.:o:screwy:

As a follow-on to the triode analog thingy, the analogy breaks down a bit since the base is biased positive to the emitter to set the operating point, while in a vacuum triode, the grid is biased negative to the cathode.

This circuit will not work as a decent video amplifier because there is no emitter resistor. The gain will be very large and very dependent on bias - in other words, without an emitter resistor, you have a switching circuit rather than a linear amplifier. EDIT: this has sometimes been used as a sync separator circuit - a very non-linear switch that changes state on sync pulses.

If you want to have both inverted and non-inverted outputs (although with gain = 1), go back to the circuit with equal emitter and collector resistors and make the upper base resistor about 15k or 18k as I suggested. If you want non-inverting with gain, use the feedback circuit suggested by kf4rca. If you need inverting with gain, you can used the single transistor circuit, but the emitter resistor needs to be smaller and the base bias voltage needs to be reduced so the collector voltage is at the middle of its possible range.

Second edit: I forgot to add that the gain for the inverted signal on the collector is close to the ratio of the collector resistor divided by the emitter resistor. The gain of the non-inverted signal on the emitter is always (very) slightly less than 1.

What is the 75ohm "dotted" connection to ground at the input?

That means that the circuit is designed for a 75 ohm input 1 volt video source (standard).

Thinking of a transistor as a tube* is a really conceptually poor design approach. A BJT is completely different in behavior/operation from a tube. A tube is a voltage amplifier and a transistor is a current amplifier. Assuming DC biases are correct and can be ignored and AC analysis of a transistor looks like a relatively low resistance resistor that measures input current and controls a current source based on the current through the resistor. A tube looks like a near infinite resistance the voltage across which controls a variable voltage source.

*Except for FET transistors which are functionally and behaviorally identical to triode tubes.

Ok.... so it's not an actual part, just a way of noting (in dummy terms) "video is supposed to go there" :thmbsp:

You probably want to actually install the 75 ohm resistor, so that your DVD player or other source is seeing the proper load impedance.

The exception would be if multiple inputs are "daisy chained" from a single video source. Then you install the 75 ohm terminator across the last input in the chain.

Transmission line theory 101 terminate your line with it's characteristic impedance to prevent reflections, ringing, reduced signal at terminals etc...Might be a good idea to brush up on (or look up the basics of) transmission line theory. If you are ever running a wired signal more than a few feet (especially if it is more than a tiny fraction of wavelength), or transmitting, T-line theory is crucial.

Question.
 

What is the polarity at the CRT? Are you looking at the schematic or a scope?
The CRT should be positive. Which means the driver tube input should be negative.
The output of the detector is usually negative, because it is transmitted that way. (Sync is peak power.)
Some of those detector circuits produce 5V of video.

There is no fixed polarity that MUST be sent to a CRT. The polarity of video you feed a CRT depends on whether it is cathode injected video or grid injected video....Both injection schemes were used back in the day*, and those 2 schemes want the opposite polarity of each other.

*In fact color sets used both at once to effectively mix the color difference and monochrome signals into RGB in the electron gun to save a few signal tubes in the design.

The output of the detector is positive on this set. The video is amplified once and then fed to the grid of the kinescope.

Progress! I've got output with the correct polarity using that circuit. I'll have to do some tweaking, but it's a victory for now. One question, is there a way to adjust gain in that circuit?

1) Please tell me you are not using the circuit with the emitter tied directly to ground instead of through a resistor.
2)Are you using the collector output (which is inverted)? In that case, the gain is equal to the ratio of the collector resistor divided by the emitter resistor. Without the emitter resistor, the gain is very non-linear and depends on the amount of current in the transistor, which varies the effective emitter impedance. So, to get the gain you want, divide the collector resistor by the gain needed and use that value for the emitter resistor.
3) a fine point to add: if the following circuit (connected to the collector) has a high impedance (like a tube grid), all is well, but if it has a lower impedance, closer to the collector resistor, it will reduce the gain because it reduces the total effective resistance at the collector.

Well to tell you the truth, I built a number of projects back in the day (late 50s to early 60s) using the common emitter configuration with no resistor in the emitter-to-ground leg. They worked fine with no nonlinearity problems etc. But these were all done with germanium transistors. Apparently that does not transfer well to later practice with silicon transistors. So I was 'waay behind the curve and gave crappy advice and should be duly reprimanded. :bash:

So by all means sub some resistances in the emitter leg per the more experienced advice given.

(In the early days of transistors the "triode" analogy was indeed used as a visualization aid for newcomers. For the purpose at hand, it disregarded the input-impedance disparity. Then when FETs finally came along, the input impedance truly mimicked a vacuum triode.)

A common deliberate use for the variable emitter impedance is in an IF stage with Automatic Gain Control (AGC). Even then, an emitter resistor is generally used, but the emitter also has an AC bypass capacitor directly to ground. The AGC feedback voltage is applied to the base of the transistor to vary the current in the transistor. This works fine as long as the IF signal is small compared to the DC bias, so the signal itself doesn't cause a variation in gain (which results in distortion). So, you may see AGC applied to the early IF stages and not to the last stage.

Because the video you want to amplify (1 volt) is large compared to the base-emitter drop (0.7 volts) it definitely could cause distortion; but when applied to the combination of the base-emitter junction and the emitter resistor, the emitter resistor keeps things linear.

Yeah, the emitter resistor all makes perfect sense now. The stuff I built decades ago was all small-signal devices, and low supply voltage (like 3 to 6 V). I was locked into the 'no emitter resistor' pavlovian habit. Moral of story: Dont be stoompid

I was using the circuit with the emitter tied directly to ground. I will experiment with an emitter resistor next.

By the way I have to say this is really fun to experiment with transistors. I've never done anything with them before...They're a lot smaller than tubes :yes:.

Pardon my waveforms
 

I had my eyeballs in upside down.
I've had good luck with video op amps like the MAX405. I got free samples but they may be hard to find these days. The only problem with video op amps is they require a bipolar supply and you might have trouble getting that in a tube set but solid state sets are usually no problem.

My issue currently is that I have too much gain with this circuit. How can I reduce the amount of gain at the output?

BTW: I am injecting at a tube grid, pin 1 on VT-6A below.

http://i1075.photobucket.com/albums/...pskqhweugb.png

You could use 2 resistors wired in a voltage divider configuration (or the adjustable version of it commonly known as a POTentiometer) at the circuit input or output.

Hot Damn I was hoping that would be the answer :banana:. What are the implications of adding resistance at the input vs output?

Looking at the circuit again...
At the input if you put it before the blocking cap you may have impedance mismatch with your coax, if you put it after you will change transistor bias (which could resort in distortion), if you ad another DC blocking cap you may create a RC filter network (and would have to make sure it does not attenuate in the video pass-band).

Feeding the output of the coupling/blocking cap into your pot/divider seems the most logical idea to me.

If you have a potentiometer of value somewhere between 2k and 2.5k, you could substitute it for the 2.2k collector resistor. If you have a 5k potentiometer, you could parallel it with a 4.7k resistor and substitute the parallel combination for the 2.2k collector resistor.

Connect the output capacitor to the potentiometer wiper instead of the transistor collector.

Here are two video input adapters
 

The left one is for a Heathkit GR2000 and the right one is for an RCA CTC74. Both are built on 2x2 circuit boards and attached to the IF module with silicone rubber. They output to the first video amplifier.
The Heathkit adapter uses a MAX405 video op amp and the RCA uses a compound pair of transistors. Both are configured for a 6X output. Both work equally well.
The choice for the video op amp was made because a bipolar supply was available in the Heathkit but this was not the case in the RCA.
The cable exiting the adapters is the video input cable.

OK, I remade the inverter following this schematic, with the exception of the 2.2 K resistor which is a potentiometer/parallel resistor of the same overall value. The output is greatly improved. One problem I am having is what looks to be vertical interference as you can see from this video. The vertical bars look like a horizontal problem in the television, but it's in the video itself. When I adjust the potentiometer you can see at some point the vertical bars change direction. Thoughts?

https://www.youtube.com/watch?v=G7Xc...&feature=share

Hard to guess without more details of the circuit and investigation with an oscilloscope. One possibility is that the point where you are injecting needs a DC-restored signal in order for sync separation to work properly. So, you are seeing AC hum (perhaps a separate problem) plus you are seeing it un-synced until you have enough amplitude. Then when it is synced, maybe it's OK, but maybe not. You didn't stop adjusting at a good point or make the video long enough to tell what's going on. You need to adjust for an intermediate level where it syncs and see if it remains in sync or loses it depending on the picture content. If it's unstable with video content, it probably means DC restoration is needed.

Please publish more of the TV's circuit and where you are connecting the amp, if you can.

I should have been more specific on what I was doing in the video. The beginning of the video is showing a blank screen because I have paused the source. The vertical lines are way easier to see when there's nothing on the screen. I am adjusting the potentiometer in one direction and at second 3 you can see the vertical bars stop and start moving in the other direction. Then at about second 6 I start moving the potentiometer in the other direction and you can see the vertical bars stop and move back in the original direction. Then at second 11 I played the source video again so you can see the problem isn't vertical hold in the set itself. The set lost vertical sync as you can see, but the bars are "in front" of that. I continue to adjust the potentiometer with the source material playing.

I'll shoot another video and give a play by play voice over of what I am doing.

Here's more of the TV circuit. I am injecting at pin 1 of VT-6A below and have removed the output from the detector that was going there.


http://i1075.photobucket.com/albums/...psysducdeo.png


Pin 3 of VT-6A goes to pin 6 of the seven pin plug P7.


http://i1075.photobucket.com/albums/...pschgot4io.png