I need to control a CRT coil (vertical and horizontal, but I will use two circuits, one for each coil). The coil draws about 450mA at full deflection at 5 volts D.C. I will use three 27R/3W resistors to get a 9R resistor (in parallel) to get the required deflection.

My problem is, that the output needs to vary from 5 volts to -5 volts, the negative voltage for deflecting to the left, the positive for the right. How would I go about using some kind of transistor or powerful amplifier circuit to control this with a 0 volt to 5 volt signal, or even a -5 volt to 5 volt signal?

The load is inductive, so protective/fly back diodes will be required. This is where I come unstuck, because at -5 volts doesn't that mean the diodes will conduct...?

Any help appreciated, as I'm trying to build my own oscilloscope with this.

 

Can't imagine why you need 450 milliamps? Scope tubes are normally voltage driven (electrostatic drive) with a linear transfer function.

TV tubes are normally current driven and the drive circuitry has non linear corrections built in, specific to the tube involved, for the raster scan.

Either way to do from +5volts to -5 volts normally requires two transistors in a push-pull amplifier.

Provide more detail for more help.

 

Yeah, CRT's are electrostatically driven. The only reason you would need that current (probably more) is if you were driving a vidicon (electromagnetically drivin).

 

ELECTRONERD -

The vidicon isn't real likely - http://en.wikipedia.org/wiki/Video_camera_tube

CRT beams may be deflected by electrostatic plates (common in oscilloscope usage), or by deflection coils (electromagnetically), as in TV sets. Making a field by electromagnetic means can require a lot of current.

To tom66 -

The coil is being driven in both directions. You don't need the diodes, as the field never collapses, and how could you put them in circuit? - the current goes in both directions. A snubber diode would just short the coil current.

There is a reason that oscilloscope CRT's use electrostatic deflection, as the effect is quite linear and the circuitry is much simpler - even if you need a lot more voltage for the deflection. I don't know how far along you are, but look into the issue of deflection and see if electrostatic doesn't make more sense.

 

I'm unsure. All I want is to be able to deflect the coils, so as to make a primitive, possibly into the audio range, maybe a bit above, as a hobby oscilloscope. I can't afford the expensive ones, but I have little use for 1 MHz bandwidth.

The coils are in an old CRT portable TV, monochrome only. I use a portable TV for a few reasons: little dangerous electricity (12 volts, 1,000 volts out of the flyback), and because monochrome avoids colour fringing etc.

I can deflect almost full scale at 5v from a computer power supply, for a few seconds using a 10R 1/4W resistor, but it (the resistor) quickly gets way too hot so I can only do it for a few seconds at a time every 30 seconds.

I also plan to control the deflection using a PIC and a shift register... but that's another story. The plan is to eventually make a vector (X/Y, single brightness) monitor...

 

You want to be more careful than you think - that flyback voltage is going to be a minimum of 10 KV, and something like 2 - 3 KV off the focus divider.

You should learn a lot from the project. Among other things that uncorrected electromagnetic deflection looks horrible. I used to work with characteron tube that had two deflection coils. It was a data display. It took 4 -5 days to get the deflection corrections set up.

 

Nonetheless all the flyback is going to do is cause muscle spasms. Which could hurt. And break things... But anyways, I'm only messing with the horizontal and vertical circuits, I've cut the horizontal leads and I get a vertical line. I've wired these in and I can get a primitive oscilloscope right there. But I'm planning to build a horizontal time base, then a vertical deflection circuit.

This is just a hobby project. I do not really care about accuracy, or any non-linear effects. I just want to see waveforms. For example I could easily see a square wave or a sine wave without too much fuss. I can add extra 'correction' logic later.

If I am building a vector display, do I not need diodes? Because if I have the coil at one side (let's call it (1, 1) for top right) and the logic momentarily returns it to (0, 0) won't that mean I will get a back EMF?

 

Perhaps, but how fast does the voltage change? Does you circuit apply unipolar current? If so, put a snubber diode in - they are very inexpensive.

 

Hate to be stupid, but what's unipolar current? I'm guessing it means current flowing in one direction...

I will investigate snubber diodes. In my circuit, it's quite possible that the controller will quickly move between positions but in between them it may begin sending it back to the zero volt position (it doesn't seemlessly move between positions). Also, if the controller does demand movement to the zero position then it could cause these problems.

What about some back-to-back zeners of about 6 volts... this would suppress a surge/back EMF right? Most surges would be very large in voltage.