I appreciate that. Currently the values of R1 and R2 are 10k and 31.5k respectively. For the simulation 4 PMOS's were needed to get enough current flowing through them. In practice I found that I got too much current with 4 PMOS's, so I got rid of 3 of them. What do you mean by "modes"?

Thanks,
Mike

 

I also want to know the load capacitance (Oscilloscope + cable). While I used 15pF and 10MegOhm. I will use your data on Monday. I am very interested in your experiments.

 

The scope I've been using most of the time is a Tektronix 2022B with an input impedance of 1M ohm +/- 2% and 11.5pF +/- 2pF. I'm not sure what brand the cable is as it doesn't have a name on it, but it looks like the one at this website: http://www.techni-tool.com/767TE1015
The data sheet for that mentions an input capacitance of 12pF. Thanks a lot for your help, I appreciate it.

Mike

 

I made a simulation. The result is a generation at a frequency of approximately 2.5 MHz. Remove managed to generate additional, serial RC circuit.
See on circuit.

 

That seems to suggest the circuit is unstable at that spot. So, how do I fix it?

 

Have you noticed additional resistance and capacitance? It is 56pF and 4.7kOhm.
To expand the bandwidth, you can try to reduce the amount of capacity.
Apparently simulation exaggerates the not stability of the scheme.

 

I propose a new, more perfect scheme. This system has a much smaller offset.
I have also improved the model. I added the resistance of transistor gates.
See

 

Oh sorry, I didn't get a notice of a reply. Thank you for your help.

Mike

 

I didn't have access to the campus labs for the past few weeks so I decided to rent the equipment rather than wait till the end of January. I've tested the circuit you gave me with the 56pF capacitor and 4.7k resistor and the output matched the input. After a few minutes I would see the output shown in the following photos. I noticed that, when I brought my hands near the circuit, the output would correct itself almost to a match to the input. I don't understand what's going on here. Sometimes I see the output match the input and other times I see the output as shown in the photos. I haven't touched anything in the circuit.

Thanks,
Mike

 

Try to increase the compensation capacitor to 100 pF. Today I do not have access to a computer. Tomorrow I will be able to more fully address this case.
You have to remember that watching you make changes to the scheme. Raise your hand you bring additional capacity and interference from external noise. Any method of reducing the impact on your circuit. Use a resistive divider. It resistors 1 megohm and 120 kilohms. The transmission coefficient is approximately 0.1. Connect the divider to the output of the amplifier and Increase the sensitivity of the oscilloscope to 10 times. 1 megohm resistor in parallel switched capacitor 3 pF. This will correct the frequency response of the divider.

 

Is it reasonable trying to refine the design while the hardware is tested in a protoboard?

 

I do not propose to change the circuit board. I mean that the divider is made assemblies of attachments.

 

Sailmike, here's a guess based on experience. I've noticed similar problems when prototyping other high impedance circuits on the same type of breadboard as you're using. The rows of contacts tend to have high capacitance, and that can create unexpected AC pathways that can lead to unexpected oscillations.
Try prototyping on a perfect board, and see if that changes your results.

 

What is a perfect board? It sounds like something that'll be permanent and I wouldn't be able to make any changes? Also, if Bordodynov's solution works and I later move the components to a circuit board, would it still work? It sounds like I'm compensating for the capacitance of the breadboard so I would be moving from higher capacitance to lower. Or am I overthinking this?

Thanks,
Mike

 

I'm having a problem with the power supply, it keeps switching automatically to constant current mode at 120mA of current. These MOSFET's can take a maximum of 80mA, so they may be destroyed. It does that after I set it for 5 volts and I see that constant voltage is shown on the front panel of the supply. It's a Hewlett Packard 6621A system DC power supply. I rented it along with a signal generator and oscilloscope because I don't have the use of the campus labs. This supply is all the rental place had.

 

Hi Mike. In the simulation, I used the option "cshunt = 1.2pF". This means that each node circuitry I added a capacitance 1.2pF. So I tried to take into account the parasitic capacitance of installation. I also took into account the capacity of the oscilloscope and cable. The scheme really working load capacity will be different. But you must remember that the connection introduces an additional capacity of the oscilloscope. This may dramatically change the process (call generation)I will highly recommend your fee shunt ceramic capacitor 0.1uF.

 

Turns out I had the DC power supply connect backwards! One of those duh moments. It easy to forget to check the very basic stuff first. Ok, the voltage divider didn't appear to work.

 

Turns out I had the DC power supply connect backwards! One of those duh moments. It easy to forget to check the very basic stuff first. Ok, the voltage divider didn't appear to work.

Ok, nw it's time to check each chip, transistor and each polarized cap to see if they are still working after the reverse voltage.

 

All the chips appear to be fine and there are no polarized caps.

 

All the chips appear to be fine and there are no polarized caps.

If none of the parts was damaged by applying a reverse voltage then you better run out and buy a Power Ball ticket. Obviously your stars are aligned.