Amplifier HELP / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Amplifier HELP

1 2 next

Move Thread

LAN_403

Paul I

Thu Nov 26 2009, 05:09PM

Registered Member #2411 Joined: Sat Oct 03 2009, 05:50PM
Location: South Texas, USA
Posts: 11

Maybe u guys can help me. I built the common source circuit shown in the attachment. In pspice it worked great, i got a 248v pk-pk sine wave with a 20v pk-pk input sine wave (60hz). But, when i built the actual circuit, my mosfets (irfp460) keep on dying (i believe when i power it on). To power it on, i plug in the ground wire and then i plug in the +150v and -150v at the same time (the best i can) to the bread board. Should i be powering it on differently? The caps are 220uf 160v and 50v 470uf. The 22meg resistors are used as a voltage divider so i can be able to see it on my oscilloscope. I'm also using a function generator as the input sine wave (should it be input before or after i power on the circuit?). The circuit has amplified once or twice, but died when i try to power it on again.

Thank You for your help

(irfp460 datasheet: http://www.datasheetcatalog.org/datasheet/stmicroelectronics/6179.pdf)

GeordieBoy

Thu Nov 26 2009, 06:13PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

You may be exceeding the gate-source oxide breakdown voltage during your power up sequence. That would result in the MOSFET failing with it's gate shorted to it's source terminal. This failure would most likely happen fairly un-eventfully, without the usual flash and bang that MOSFETs typically love to emit in high voltage circuits!

You could probably protect against this by putting a 20V zener diode between the source and gate pins of the MOSFET. It needs to be connected so that it prevents the gate from going more than 20V positive with respect to the source when the zener is reverse biased. It will also prevent the gate going negative with respect to the source because the zener will conduct in the forward direction.

You might also want to add a little resistance in series with your input DC blocking capacitor to limit the amount of current that can flow when you suddenly connect the source and this capacitor needs to charge.

Finally, I should point out that the IRFP460 is a switch-mode power MOSFET which has been designed to be optimised for switching operation. It is not really intended to be used in the linear mode. Such operation can result in HF-oscillations and also hot-spotting on the silicon die due to current hogging and negative Vgs(th) temperature coefficient.

-Richie,

Paul I

Thu Nov 26 2009, 06:38PM

Registered Member #2411 Joined: Sat Oct 03 2009, 05:50PM
Location: South Texas, USA
Posts: 11

Thank you for your reply,

Ill give your suggestions a try, though not today (eating turkey

). As for the mosfet, is there any you had in mind for this circuit? I was looking at the irf240, but BVDSS is 200v, i need one more than ~300v.

Hazmatt_(The Underdog)

Thu Nov 26 2009, 08:39PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

If you need high voltage you're probably going to have to use more then the one device, you will probably have to setup a push pull pair that drives another HV pair to spread the voltage difference.

If its audio range you ought to look at some of the audio amps.

If its beyond 100KHz then you're in a different situation, you'd need a push pull transformer coupled amp and go from there.

Aharrison23

Sat Nov 28 2009, 09:28PM

Registered Member #2484 Joined: Thu Nov 26 2009, 06:08AM
Location:
Posts: 2

Ok so being that I'm one of Pauls design partners on this project, I too am confused on why we were killing our MOSFETS without any indication. Either way I have decided to include a screen shot from Pspice of the other half of the circuit. Along with the voltage amplifier we are having problems with, we are using the output of that amplifier to drive a power output stage that is utilizing the push pull configuration. Everything works out nice on Pspice of course but it's just that voltage amplifier. Our requirement is that we are able to get a large voltage swing from the voltage amplifier. We would need a minimum amplitude of ~120 in order to serve the rest of our circuit (voltage multiplier). The push pull amplifier would just serve to provide current gain before a voltage multiplier. Any ideas on what else we are doing wrong or what else we could be doing better in order to achieve the large voltage swing with a variable sine wave input from 0~10V amplitude? It should be noted that we plan on using an op-amp with the feedback from the output in order to stabilize the gain of the amplifier. Also, if there is a way to achieve these criteria without using the large DC power supply that would help greatly on our finances! Comments and suggestions are greatly appreciated :)

Thank you
Amado

GeordieBoy

Sat Nov 28 2009, 11:46PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

MOSFET's can usually only safely tolerate about 20 volts from gate to source before the thin oxide layer is punctured. You don't seem to have any protection in your circuit to prevent the gate-source voltages of the devices from exceeding this level at power up. Think about what happens to the voltages in the circuit when you first apply power and the DC blocking capacitors have to charge up to their quiescent voltages.

I also have a feeling you should be using a matched pair of N and P-channel MOSFETs designed for linear operation in the output stage. However, most of my experience is in the switch-mode field so I won't say much more in this thread.

As for another way to achieve your design spec, it might help to explain what you are trying to do, what frequency, what the load actually is, and how much current it draws from the amplifier?

For example, if you just need a variable source of rough 60Hz power you can do that in the states using just a variac and step-up/isolation transformer.

-Richie,

Hazmatt_(The Underdog)

Sun Nov 29 2009, 03:34AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

You might want to try something a bit more conventional as a phase splitter before your power mos. This will in a way, control how much the gate swings, so you don't have that huge 100V range on the gate, It should me more like a couple of volts. And for this kind of operation you need linear mosfets for proper operation and I think those are TMOS, ~ $12 each I think.

Here's a rough design for you, you'll need a lot of tweaking on this cricuit, but it gives you the basics. It will need your op amp front end with feedback like you mentioned, I just set this up and tested that it could work. Also check out amplifier schematics online to get an idea of what's going on, there's a lot out there.

Anon01

Sun Nov 29 2009, 07:58AM

Registered Member #1589 Joined: Sun Jul 13 2008, 06:40PM
Location:
Posts: 70

I can supply a couple of mainstream class D car audio amplifier schematics, only a few larger (~2kW) ones. Shoot me a PM if you want.

The general concept is to LPF a PWM'd supply in order to achieve a sinusoidal output.

Steve Conner

Sun Nov 29 2009, 04:57PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Bah! It's probably suffering from parasitic oscillations amongst other things.

I'd suggest trying a 100 ohm stopper resistor on the gate. A stopper resistor is a low value resistor in series with the gate circuit, between the gate and everything else. Put it physically right at the gate pin: this is important.

I also agree with the need for a gate protection zener, given that you have a -150V rail there and you aren't bringing up the power gradually. Heck, try bringing up the power gradually.

Aharrison23

Mon Nov 30 2009, 12:03AM

Registered Member #2484 Joined: Thu Nov 26 2009, 06:08AM
Location:
Posts: 2

First off, I would like to thank you guys for the awesome suggestions. We will definitely try your advice tomorrow. I can't thank you enough for the other implementation suggestions. The only problem is that Paul and I are near the end of the first half of this project. I'm already in the process of writing up a final report for submission so....it's kinda too late for us to totally change up the circuit. We plan on keeping the amplifier circuit since this is what makes up a good chunk of the report.

Due to some of confusion about the purpose of this project, I will go ahead and give further detail on the expected outcome.

The objective was to design a circuit(s) that would be able to produce a maximum 2000V/200A for power pulsing. A couple of weeks into the project it turned into variable output.

Our approach to achieve this was, 1. Definitely use a voltage multiplier to reach the high voltage. 2. Amplify a small variable signal that would be sent to the voltage multiplier. This would allow us to control the final voltage through amplification and multiplication. We set a goal for the amplification. We decided that we would shoot for 200V amplitude from a maximum 10V amplitude from an op amp.

The multiplier is 16 stage voltage multiplier. We wanted 200V amplitude so there would be margin for error. For the first half of this project, we will not have a 200VDC power supply for the amplifier. The most we will be able to use is 150VDC.

We will use a capacitor bank that will have an equivalent voltage of 2000VDC and be around ~50uf for the pulsing.

The circuit included below is a working design but will not simulate once the multiplier is connected to the output stage. PSpice cannot handle the simulation. Instead we connected a small impedence of 8 ohms and a large impedence 100K in order to demonstrate the operation of the multiplier when the caps are not charged and once they are.

We have tested our physical multiplier with a variac to verify operation. We were able to go well above 2000VDC. The only thing we have left is to get our voltage amplification working since the power output stage shouldn't be too much of a concern as we will just add resistance in series of the multiplier to limit the current for transistor protection.

At this point, anything will help. This was our best shot at keeping it cost effective as we are on a tight budget

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob