high current amplifier for waveform/pulse generator in the MHz range

Andy,

The OPA541 is unlikely to fulfill this requirement because its bandwidth and slew rate are not sufficient. You have not mentioned a key piece of information that is critical in understanding your requirements: What is the inductance of your load? This determines the voltage that must be applied to the load to create this current.

Regards, Bruce.

Hi Bruce,

thanks for your reply. My problem is rather unusual. The load is a ferrite E-core micromagnet which is used for exciting nanomagnetic samples at high frequencies. The inductance itself will likely vary during each cycle due to partial material magnetization saturation which is inevitable. The maximum inductance should be somewhere in the 500µH range but could be potentially higher. Unfortunately a prototype has yet to be manufactured, so I've had no chance to verify this experimentally. I'm basically searching for an amplifier that can provide a peak voltage output (for sinusoidal input signals) that is as high as possible at a frequency of approx. 1MHz.

Regards,

Andy

Andy,

At 1MHz, 500uH has a reactance of approximately 3k ohms. It would require 15000V to force 5A to flow into this load. This is roughly 1000-times the capability of the OPA541 or any similar linear op-amp-like device on the market. It's going to require some very special type of Mega-circuit to perform this function and this forum is not likely to have the expertise to offer much help.

Sorry we can't be of more help.

Regards, Bruce.

Hi Andy,

A review of our precision power operational amplifiers shows that they will fall short on the output current and/orgain bandwidth; especially the latter in relation to your original mention of a pulse generator. You may be able to employ one of TI's High-Speed operational amplifier as a driver for an external, complementary NPN/PNP output stage. However, it is probable that much consideration would be required with regard to compensation to assure stability with your intended load. The High-Speed Amplifiers E2E forum may be able to assist you if you decide to explore that approach.

The application may be better satisfied with conventional RF power amplifier designs, rather than an operational amplifier. By that I mean typical, heavy feedback, broadband RF stages where the drivers are transformer coupled class-A amplifiers, followed by a push-pull class-AB output stage. Many designs were covered in the applications section of the 1980s Motorola RF Device manual. A search on the web will bring up many of those older application notes. In particular, you may want to search for the application notes written by Helge O. Granberg. He developed a number of broadband power amplifier designs for RF communications that may be adaptable to your application. As an example search on AN-779; it provides information on a high-frequency 20W, 25dB gain amplifier. Note that these amplifiers were designed for terminated 50-Ohm loads, so your reactive load will require additional consideration.

Most of those application notes are getting a bit long on the tooth, but they provide the fundamental information necessary to design very capable RF power amplifiers.

Regards, Thomas

PA - Linear Applications Engineering

Hi Bruce,

I'm sorry, I was wrong about the maximum inductance I mentioned. The whole inductor setup is rather complicated, so I redid the simulation and instead of relying on the value yielded by the simulation tool, I calculated the inductance by hand and was able to peg it at 15µH, which is a fairly good estimate. But this is only the value for currents up to about 3.5A. For currents larger than that, the inductance will sharply decrease, as the ferrite material gets partially saturated. This means that the maximum impedance value for currents up to 3.5 amps is about 90 ohms. Above that value it will be considerably lower. Maybe I can reduce the 90 ohms with an adequate compensation circuit. What is the highest peak output voltage for a Texas instrument amplifier model that operates at a frequency of 1MHz?

Regards,

Andy

Andy,

3.5A and 90 ohms still requires 315V, well beyond our capability. Think in terms of V*A capability. Our most powerful amplifiers are in the range of 5A and 25V, or a 125 V-A. You need 1100 V-A . (I'm ignoring the frequency range issue and just looking at our most powerful amplifiers.) This assumes that you could transform the impedance of the load somehow. One way or another, you need make up the volts with amps, or vice versa, to get a product of around 1100 V-A.

It's not clear from your description but it sounds like you are trying to get short pulses of some type. I think Thomas had some good suggestions that might be along these lines. It may be possible to produce pulses of this magnitude with some type of pulse switching into a resonant network tuned to your frequency of interest. You must be thinking of how to get a V-A product in the 1000+ range and that means you are going to have switches (transistors or otherwise) that must handle lots of volts and amps. One way or another, this is going to be a circuit that should have you hiding behind a plastic shield. :)

Regards, Bruce.