Voltage Multiplier (5VDC to 6kVDC)- need help on Feedback network

Hi Guys,
Currently I'm working on Voltage Multiplier design which requires to convert 5VDC to 6kVDC.
One of the feature I need to add is 'output voltage regulation'. For that, I have adopted a simple resistor network feedback as we normal implement in the voltage regulators.

The resistor divided voltage is fed back to 555 timer to up/down the OSC output waveform amplitude. The method of feedback 555 timer to control the amplitude is still not finalized yet.

My top resistor has to be very big in value (5M ohm) as compared to bottom resistor (200k ohm). The issue is, this method is not practical to implement. For example, any small change in the value in the top resistor due to process change will give big output error.

Could you please give me some idea on how to implement this feedback network. Is there any other better circuit topology? with bear in mind that it has to be simple and low cost.

Please advice your opinion.

For accuracy you would have to do it differently but the cost would be much higher. There are two ways to do this if you use the simple circuit you propose:

1. use multiple resistors in series for the 5M. A 'normal' resistor is generally rated for around 250V before risk of internal arcing so it would be best to spread the voltage drop across maybe 25 smaller ones. Use a voltage comparator, comparing against voltage from a potentiometer across 5V and ground, take the output of the comparator to the reset input on the 555. When the voltage reaches the level you set with the potentiometer, it shuts down the 555 oscillator until it has dropped again.

2. use the resistor chain again but measure the current through it rather than a voltage tapped off it. Connect the bottom of the resistors to the base of a transistor, ground the emitter and use the collector as the 555 reset control signal. It's less accurate this way but uses fewer components.

You probably want the FET resistor in the gate not the source pin and you may need a Zener diode across the drain to source to clip excessive voltage peaks as well.

Brian.

My top resistor has to be very big in value (5M ohm) as compared to bottom resistor (200k ohm). The issue is, this method is not practical to implement. For example, any small change in the value in the top resistor due to process change will give big output error.

Click to expand...

I'm missing the reasoning behin this statement. It's a voltage divider, the accuracy depends mostly on the relative error of both resistors. At first sight, there's no essential difference between a 10:1 and a 2000:1 divider. Please reconsider.

In addition there are practical restrictions for the voltage rating and accuracy of high ohmic resistors, as addressed by betwixt.

Guys,
Thanks for the help.

betwixt,
I have added feedback circuits as per your suggestion, the prove of concept is works fine.