How to best amplify a sinewave (0-25 MHz)? / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

How to best amplify a sinewave (0-25 MHz)?

1 2 next

Move Thread

LAN_403

Bjørn

Sat Oct 17 2009, 09:03AM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

The source is a AD9835 direct digital synthesis chip with a 4 mA current output that is connected to ground by a 300 Ohm resistor to make a voltage output at 1.36 V peak to peak.

What i need is 5 V peak to peak or 3.3 V peak to peak. The signal will be fed to a BUF634 for analog output and to a 74HCxx to make a square wave.

I would like it to work as nice as possible from 0.1 Hz to 25 MHz.

The voltages available are -5 V, 3.3 V, 5 V and 6.5 V,

The fastest operational amplifier I got is an OPA2353

It works pretty good up to 5 MHz (1.8x gain), Above that I get triangle instead of sine. Decoupling the feedback resistor gives sine but the gain falls to to 1 or lower.
The fastest transistor I got is a BFW92

(.PDF) The simple amplifiers I have tried work up to 10 MHz but never on a wide range of frequencies. In addition I got the usual stack of old CD and video players to pick parts from.

10 MHz and below are most important since those are going to be used the most, frequencies below 1 kHz I can do in software in the controller so they are not critical if that makes things a lot easier.

Steve Conner

Sat Oct 17 2009, 09:32AM

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

Hi Bjorn,

I once used the AD811 current-feedback video op-amp to amplify a pseudo-random bit sequence for driving a laser diode. I used two amps in parallel to drive 10V p-p into 50 ohms, and they were still making a nice eye pattern at 40MBit/s. So a 25MHz sine wave should be easy. No need for the BUF634, the AD811 has plenty of drive by itself, as it was designed to drive 50 and 75 ohm video cables.

Agilent's 20MHz signal generators use some THS... op-amp from TI as the output stage. It's like the AD811 but faster, and they use four in parallel.

Op-amps are getting faster by the day, my colleague needed a 44MHz IF amplifier with a gain of 16dB, and he found something that would do it with GBW to spare.

Proud Mary

Sat Oct 17 2009, 11:12AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Plain old L and C with a PLL, with a DLL? Bjorn?

Too old fashioned, I'dguess. ?

GeordieBoy

Sat Oct 17 2009, 01:06PM

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

I@d recommend the THS range of current feedback op-amps from TI too. I used THS3001 to amplify the output from an AD9834 DDS chip after the anti-imaging "Reconstruction" filter.

Just do a few calculations to see how much gain you need. Remember that the reconstruction filter needs to be terminated at the input side AND the output side. So that might mean 300R at the DDS side of the filter and another 300R at the amplifier side of the filter. In the pass band that makes the total load resistance for the DDS 150R instead of 300R. You need to take this into account when choosing the op-amp otherwise you will only get half the signal out that you thought you would get.

I@ve also used the THS3061 (or maybe it was something like THS3089) dual current-feedback op-amp in a balanced application. If the AD9835 gives balanced (double-ended) current outputs, you can use a balanced filter and keep everything balanced right up to the output. This is what Agilent (HP) have done in their signal generators. It gives lower THD when done this way. (For my application, the lowes possible 2nd harmonic distortion was very important, and the balanced approach works very well for cancelling even harmonics.)

As Steve said, several chips can be paralleled (through seperate terminating resistors) if you need to drive a 50R line with lots of current. And multiple devices can be cascaded if GBW product is too low to get enough gain over the required frequency range.

-Richie,

Dr. Slack

Sat Oct 17 2009, 01:29PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

I used THS3091 in my last commercial high voltage video output project. Three amplifiers, each driving a 150ohm series resistor, in parallel for 50ohm output, to get the 3rd harmonic distortion better than -70dBc with 5v output at 25MHz, but I've just had a look at your voltage rails, and I don't think it's so happy from +/- 5.

What will work well from +/- 5v, Maxim have shed loads of amplifiers suitable for video, we sprinkle the MAX4104 around like smarties in some of our test gear, and of course they are free if you don't want too many.

Bjørn

Sat Oct 17 2009, 02:02PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

Plain old L and C with a PLL, with a DLL? Bjorn?

Too old fashioned, I'dguess. ?

Too complicated, I am using a 70 MHz ARM 7 processor to control everything. Raw DDS output, processed DDS output, digital output, analog/digital input. The digital output can do any bit pattern of a certain length up to 35 MHz. The analog input can sample slow waveforms for playback or be used to control the frequency. There are also modulation inputs to select phase and frequency. To the left of the keypad there is a very nice rotary encoder made from an old stepper motor from a harddrive. Finally there is a port for uploading new firmware so I can just write some code and upload it each time I need a new feature.

Almost everything is from the scrap box, the oldest components are marked 1977 and it is working nicely, except the amplifier so it is restricted to low amplitudes at the moment.

Steve Conner

Sat Oct 17 2009, 02:16PM

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

Well, if it has to come from the scrap box, I suggest a series-shunt pair made with a couple of those RF BJTs you found.

Bjørn

Sat Oct 17 2009, 02:22PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

It does not have to come out of the junk box even if that is preferable. If it is going to sit on the desk for weeks while I wait for parts it might get buried in other projects not to be rediscovered for years.

ShawnLG

Sun Oct 18 2009, 02:46AM

Registered Member #286 Joined: Mon Mar 06 2006, 04:52AM
Location:
Posts: 399

In addition I got the usual stack of old CD and video players to pick parts from.

Look in your video playes, you may get lucky and find a video line driver OP amp. I have found one of thoes in a cable converter box and it has a bandwidth around 300Mhz. If bandwidth is an issue with a part you choose. Cascading several amps can increase bandwidth VS gain.

Dave47

Tue Oct 20 2009, 11:07PM

Registered Member #84 Joined: Thu Feb 09 2006, 01:06PM
Location: Dallas, TX
Posts: 47

GeordieBoy wrote ...

I used THS3001 to amplify the output from an AD9834 DDS chip after the anti-imaging "Reconstruction" filter.

I probably know more than I want to about the THS3001 :).

I usually recommend the THS6012 for projects here. It has lots of bandwidth and lots of output drive all in one package. You also get an extra one...

The THS6032 is pretty cool, too. I've had it unsolder its 25 ohm load right off the test board! Not suitable for this project, though.

David

1 2 next

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