Hello, can anyone help fill in the values of the resistors and capacitors in this circuit? I have seen the Vcc equals 50 volts in other schematics. I can do some basic calculations but I don't know much of how to find these values for this kind of device.
assuming I'm using either one of these transistors,
or
50V, that's quite a lot... Are you sure?
And uhm, I see values in the schematic... What's the problem?
50 V exceeds the maximum rating ( VCEO) for those transistors so is likely to kill them. Did you mean 5.0 V?
The values for R1 and C1 will depend on the pulse width you are using and the amplitude of the slope you require.
What is your application?
Russell.
And it doesn't generate a trapezoid, but rather exponential decay curve towards the supply voltage.
To generate a true trapezoid you would need a constant current source replacing the top resistor.
However I can see that using a large supply voltage would allow much more linear behaviour if the
signal is only a few volts - however the constant current source is a much better approach to linearization.
Well see most of these say 50 volts, like in this one, Series LR circuit. and this one, Figure 3-50A.Trapezoidal waveform generator, among other sites that have the same schematic image.
Well I figured out the values of those two, based on these two transistors, but I think I'm going to go for a higher voltage transistor because the application is the electromagnetic deflection of an electron beam. It needs to push the beam across every point of a sample. Usually for electrostatic fields a sawtooth generator is used but for electromagnetic ones a trapezoid generator is used.
Do you happen to have 50V to hand then?
Well I have two 36v power wall warts and can combine one of those with a 14.5v wall wart, after doing something to drop 0.5v off the 14.5v one. So yes,
You need high voltage for electrostatic deflection but for magnetic deflection it is the correct ampere.turns you need. The inductance of the deflection coils together with the deflection speed require will determine the voltage needed in that case. Do you have the deflection coils?
Russell.
There is that X Y problem again.
Maybe you should start with describing what it is you are trying to accomplish.
Why do you not use a combination of Arduino and a DAC (digital-analog-converter)?
You can program the whole shape of your output signal.
Above 100 KHz you might have less resolution when it comes to the signal shape.
(The Arduino runs at 16 MHz)
And do you really need 50 volts - nowadays that is quite exotic.
Well I figured out the values of those two, based on these two transistors, but I think I'm going to go for a higher voltage transistor because the application is the electromagnetic deflection of an electron beam. It needs to push the beam across every point of a sample. Usually for electrostatic fields a sawtooth generator is used but for electromagnetic ones a trapezoid generator is used.
OK, so rather than start with the circuit, let's start with what you need.
What voltage range do you need? A positive only sawtooth? Or one symmetrical about 0V?
I need a trapezoid shaped current to electromagnetically deflect a beam for a scanning electron microscope. I am somewhat familiar with my arduino and would greatly prefer to use it and its DAC due to its easiness and simplicity, how can I do this? I have seen a research paper that used another MCU for this purpose but it said to email the authors for the code, and I have not received a response from them.
For those who want to understand the background of this, sawtooth waves are used in electrostatic deflection, refer to the image 
see how it forces the beam to a path that will make it scan across a sample? However in electromagnetic deflection, a trapezoidal waveform is used, like stated in this Navy text
....and yes, I do have wire for the electromagnet coils, as well as ferrite core making supplies, and I also have what is needed to reach the vacuum, I am just trying to get the electronics like the deflection, my photomultiplier, my stigmator, etc. set up first.
Okay so this is that research paper I mentioned, and Fig. 3a is an example of how a DAC output could be made into a larger output, but this is a voltage signal. So I have found that first, people use the arduino "Due" DAC for this and then to get a current signal i was lead into this Due waveform page and then I found this page where people worked on getting a current output instead of voltage, and the common problem of the Due's 0.55v minimum output is fixed by the last poster in a schematic on his linked page, which I found on this other page. The last two linked pages also provide other recommended DAC IC names like the AD5420, setups with the XTR110, and even rail to rail conversion with the TLV2462. I have an UNO and a Teensy but not a Due, so I'm not sure yet which approach I'll take on this. After skimming through the Due page it seems like I can make a sketch from scratch of the shape I want to be made (trapezoid for electromagnetic response), save it as a waveform.h and then that route would work.
However in electromagnetic deflection, a trapezoidal waveform is used, like stated in this Navy text
Ha the Navy obviously have a different name for it than the rest of the world. In my book and many others this is a trapezoidal wave:-
On the grounds that mathematically it looks like a trapezium.
Still you can get the shape you want very simply from a Uno and the PWM output with a simple algorithm, there is no need to generate any lookup table for this.
You simply generate an interrupt from a timer. In the ISR ( interrupt service routine )you keep a track of the sample count. On the zero sample you output zero. On the next sample you output your offset, then for subsequent samples you increment the output and on the final sample you reset your sample count which produces a zero on the next ISR. Note with this the offset and maximum output will add up to 255.
If you want a greater dynamic range than this then you have to use an external D/A or a processor with one built in.
In order to scan the beam as in your picture you need a sawtooth current in the deflection coils. Now, depending on the scan frequency, if you apply rectangular voltage pulses to the coils the coil inductance will cause the current to rise in a linear fashion and can thus be made to give a sawtooth scan. That is the principle used in old analogue television line timebases with a few tweeks to improve linearity over a wide angle.
I guess the use of the "trapezoidal" waveform you describe is a compromise resulting from there not being sufficient inductance in the coils to perform the integration function accurately.
So, as I said before, you need to consider the speed of scan and the inductance of the coils before being able to determine the voltage waveform required.
Russell
Thank you all, I'm pretty sure i'll get a Due and what i've come to realize is that Fig. 3a in that research paper I linked is what i'll be using to carry the Due's signal to the electromagnetic deflection coils, at 12 volts. I mistakenly thought the coils needed more current.
Which Teensy? The 3.1 and 3.2 have a 12 bit DAC on board and are about as fast as a Due. Just one DAC, however.
I really don't see using the PWM on the Arduino filtered as suitable to generate this.
Cool project. You are basically building a TV camera. Have you thought about taking apart a small BW TV for the coils?
Is this a science fair project? One of the members of our makerspace is building a tunneling electron microscope, we were helping him troubleshoot it. The problem was caused by daisy-chained regulators in the power supply.
Why wouldn't the Due be okay for generating a signal that might fit into this circuit pictured in the top left corner?
I have a Teensy LC, Would a Teensy 3.2 do it adequately? It would generate a low output signal and then be amplified by the above top left corner circuit pictured before it gets sent to the deflection coils. It would be easier to use Teensy than other MCU's as its still Arduino IDE coded and the waveform can still be made with the waveform sketcher. Not sure how I'd do that on a non-Arduino MCU.
I'm using my Teensy LC for the visual data acquisition, trying to do exactly what Krasnow does in this video for this part of the project. Yea I do actually have I think three deflection coils from old televisions in my car, I was going to try to use them but decided I will make my own instead because they are too large and likely not UHV-compatible.
The Arduino Due has two 12 bit DACs, so it is probably better suited than the Teensy 3.2 which has only one DAC.
Is there a reason you have decided against using electric field deflection? Magnetic field deflection can get a bit tricky if you need to use a different frequency. It is simpler to keep a sawtooth voltage linear.
The Due's DAC has output range of 1/6 to 5/6 VCC (0.55-2.75V) for digital range 0-4095 (12-bit). Would need to level-shift to zero and amplify to get 0V-2.5V.