Author Topic: Need help with Op amp buffer  (Read 2123 times)

0 Members and 1 Guest are viewing this topic.

Offline hummusdudeTopic starter

  • Regular Contributor
  • *
  • Posts: 141
  • Country: us
Need help with Op amp buffer
« on: December 21, 2019, 05:31:56 am »
I'm trying to build a circuit that will buffer the output of an AD 9850 waveform generator module. The 9850 is very sensitive to even the slightest esd and frequently shuts off requiring a reboot of the waveform generator. I built the circuit in the attached schematic but it's not doing what I'd hoped. The rail splitter seems to work pretty well although the + and - values are not exactly even distance from virtual ground. R6 is 1M because I don't want to amplify the signal...just buffer it.

The datasheet for the LM741 says it will work without a negative ref voltage so I hooked it up with V- at 0V but the output would not go below 2V with 0V in. So I built the rail splitter hoping that a negative voltage for V- might do the trick. I imagine I'm getting some chuckles at this point. Obviously, I don't understand op amps well enough to figure this one out so hoping for some advice on how to buffer the AD 9850.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20363
  • Country: gb
  • 0999
Re: Need help with Op amp buffer
« Reply #1 on: December 21, 2019, 10:16:18 am »
The rail splitter looks fine.

The schematic isn't clear. Where exactly do you have the input and output leads connected to the circuit?

It could be a grounding problem, the output of the rail splitter must be the point from where all voltages are measured from. The 0V of both the input and output should be connected to the output of the rail splitter.

By the way, the LM741 doesn't have sufficient slew rate or bandwidth to make full use of the AD9850.
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: Need help with Op amp buffer
« Reply #2 on: December 21, 2019, 11:03:19 am »
The datasheet for the LM741 says it will work without a negative ref voltage so I hooked it up with V- at 0V but the output would not go below 2V with 0V in.

The ancient 741 does not have rail-rail input common mode range or output swing, typically it can only get within about 2v of the rails.
 

Offline avogadro

  • Contributor
  • Posts: 34
  • Country: hr
Re: Need help with Op amp buffer
« Reply #3 on: December 21, 2019, 12:29:04 pm »
It looks like you connected the diode pair to the GNDREF. Try removing that connection. I attached the schematic.
 
The following users thanked this post: unitedatoms

Offline unitedatoms

  • Frequent Contributor
  • **
  • !
  • Posts: 324
  • Country: us
Re: Need help with Op amp buffer
« Reply #4 on: December 21, 2019, 01:40:30 pm »
That's interesting splitter. Nice answer, @Avogadro
Interested in all design related projects no matter how simple, or complicated, slow going or fast, failures or successes
 

Offline hummusdudeTopic starter

  • Regular Contributor
  • *
  • Posts: 141
  • Country: us
Re: Need help with Op amp buffer
« Reply #5 on: December 21, 2019, 06:57:48 pm »
Quote
It looks like you connected the diode pair to the GNDREF. Try removing that connection. I attached the schematic.

Ah, thanks for pointing that out. The schematic was wrong...the diodes are not attached to GNDREF. I attached a corrected schematic. And that is the same schematic I'm working from!
« Last Edit: December 21, 2019, 07:12:50 pm by hummusdude »
 

Offline hummusdudeTopic starter

  • Regular Contributor
  • *
  • Posts: 141
  • Country: us
Re: Need help with Op amp buffer
« Reply #6 on: December 21, 2019, 07:09:41 pm »
Quote
It looks like you connected the diode pair to the GNDREF. Try removing that connection. I attached the schematic.

I do have the DMM reference point at GNDREF. Thanks for the observation. This is only my second forum post so still getting up to speed on how best to articulate things.

About the slew rate, if I interpret the datasheet for the LM741 correctly, 0.7 v/uS means max 0.7v swing at 1MHz. Since my generator is supposed to go to 40 MHz it would seem I need a buffer with a much higher rate. I checked the TL084 and it's slew rate is 13 v/uS. That would appear to be far to slow as well. It's true I'm working with ancient tech...benefits of being a student at community college. Is there a modern buffer that will handle 40Mhz 0-5v?
« Last Edit: December 21, 2019, 07:11:26 pm by hummusdude »
 

Offline hummusdudeTopic starter

  • Regular Contributor
  • *
  • Posts: 141
  • Country: us
Re: Need help with Op amp buffer
« Reply #7 on: December 21, 2019, 07:21:31 pm »
Quote
The ancient 741 does not have rail-rail input common mode range or output swing, typically it can only get within about 2v of the rails.

I wondered about that as I was taking the measurement. That's why I tried using the rail splitter...thinking that if I gave it a negative ref voltage than it might be able to output as low as 0 v. As it is it will output 0 v with about 5 v on the positive rail. But at 12V on the postive rail I'm getting -5.8 V on the non-inverting input. Can't help but notice that test point B is also -5.8 Volts.

The input signal will be 0 -5 v ac up to 40 MHz. Can you recommend a buffer that can work with that?
 

Offline avogadro

  • Contributor
  • Posts: 34
  • Country: hr
Re: Need help with Op amp buffer
« Reply #8 on: December 21, 2019, 08:14:50 pm »
Okay so I messed a bit with this circuit in ltspice and only way I was able to get a big unbalance is by flipping one or both diodes. Try to check the polarities of your diodes. Also what are you doing with the offset pins? Are they floating or are they grounded?
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: Need help with Op amp buffer
« Reply #9 on: December 21, 2019, 08:19:31 pm »
That's interesting splitter. Nice answer, @Avogadro

That is not nice splitter, that is absolutely un-necessary splitter. 99 out of 100 opamp circuits can be made without these splitters.

For a buffer for AD9850, there is zero to none need for a rail splitter.  The output signal out the DDS is already DC biased, the DAC is unipolar obviously.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15157
  • Country: de
Re: Need help with Op amp buffer
« Reply #10 on: December 21, 2019, 09:31:30 pm »
The DDS chips can give you 0-5 V amplitude directly from the chip. The output is differential with 2 current outputs that usually work with some 50-200 Ohms load to the filter circuit. Some cheap boards may only use one output. So the output amplitude is more like 0.5 V_pp.

So one will usually need more than just a buffer, but some amplification. In addition the output will usually loose half the amplitude to the 50 Ohms termination. So if one wants 5 V out at 50 Ohms the amplifier usually needs to go to 10 V with a 100 Ohms load. This may need a little more current than the normal OPs can easily drive.

To get 1 V_pp at 1 MHz it needs a slow rate of 3.14 V/µs as the derivative of sin(2*pi*f*t) gives the extra factor 2 Pi.

Many modules don't work well all the way to 40 MHz (e.g. limited by the filter, layout). So one may not need to have to amplifier all the way to 40 MHz.

Adjusting the amplitude with the DAC reference usually only works well for a small range.

The choice of the OPs depends one the supply (not all of the fast OPs like +-15 V) and the speed one is comfortable with in the layout.
One could consider something like AD811, AD8051 or some Ti THS... series chips.

 

Offline hummusdudeTopic starter

  • Regular Contributor
  • *
  • Posts: 141
  • Country: us
Re: Need help with Op amp buffer
« Reply #11 on: December 22, 2019, 02:06:49 am »
Avogadro...I double checked my circuit and the diodes appear to be in correctly. The offset pins are floating. My understanding is that the offset is useful for fine-tuning +- 15 mV and not required if the accuracy is close enough. I did try a 10k pot on the offset pins with the wiper going to V- and did not see any change.
 

Offline hummusdudeTopic starter

  • Regular Contributor
  • *
  • Posts: 141
  • Country: us
Re: Need help with Op amp buffer
« Reply #12 on: December 22, 2019, 02:21:32 am »
Kleinstein...thanks for the analysis. I should have been more specific about my needs. The 9850 board I'm using does give 0-5 v square wave out and 0-1 v sine wave out. I didn't mean to imply that I was going to amplify the sine wave to 5 v. It would be more useful to have the range, i know. But I'm trying to resist scope creep and keep this project constrained to the output of the 9850. Just looking to buffer the signal because the DDS keeps shutting down every time I change the lead connection and the output is so noisy my scope has a hard time triggering. I assume the EDS issue of the 9850 is CMOS related and the jitter might be a lack of shielding . That, and classmates have mentioned that they were able to reduce the jitter by buffering the output.

As for the slew rate issue, it sounds like I can do better than the LM 741 but 40MHz may be a bit ambitious for the scope of the project. I'm okay with that...it's my first project and it's already pretty good. If I can get it stabilized and working consistently it will help with job interviews. I'm attaching a couple of pics for fun. The LCD is driven by an Arduino Uno that recieves serial data from the Mega.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15157
  • Country: de
Re: Need help with Op amp buffer
« Reply #13 on: December 22, 2019, 10:00:18 am »
The rectangular output has even faster slopes. So the easy way would be to use separate buffers for the sine and rectangular signal. For the digital part something like 74HC04 / 74HCT14 or similar, using 4 gates in parallel could be a simple solution. 4x180 Ohm can combine the outputs and give it an approximately 50 Ohms termination and some protection.

For the sine wave there are plenty of OPs fast than the 741. It can get tricky on the prototype area to go all the way to 40 MHz as the fast OPs also need good decoupling and are often available in SMD only.  An important parameter is the supply that is available. If there is a +-5 V to +-15 V supply I would consider the good old NE5534 / 5532. It should be good to some 5 MHz. One should still have termination (e.g. 50 Ohms)  at the output, even if it is only to isolate the OP from the capacitance of the cable. With a 12 V supply one could use AC coupling to bring the sine signal to about the center and also have the output AC coupled. The supply splitter could be an option too to make it a +-6 V.

For the "noise" / interference, part of this could be due to the cabling. Chances are an electrolytic cap (some 100 µF) near the DDS board at the supply could help somewhat.

The offset pins at the OP are usually there to trim the OPs own offset, usually < +-5 mV. Using them to add more offset can cause problems.
 

Offline Kirill V.

  • Regular Contributor
  • *
  • Posts: 118
  • Country: ru
Re: Need help with Op amp buffer
« Reply #14 on: December 22, 2019, 10:02:44 pm »
Yes, using logical gates as buffers of a logical signal is very reasonable. But does a purely analog buffer not have any advantages here?
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: Need help with Op amp buffer
« Reply #15 on: December 22, 2019, 10:33:55 pm »
As for buffering the AD9850, why not using a modern high speed amp?
CMOS 5V  RRIO: OPA354, LMH6611,...
Cable driver bipolar CFA: AD811
cable driver: THS3xxx series

Using  a 741 or 553x will go nowhere.

There is a lot of these, and some of them are not even that expensive.
 

Offline hummusdudeTopic starter

  • Regular Contributor
  • *
  • Posts: 141
  • Country: us
Re: Need help with Op amp buffer
« Reply #16 on: December 23, 2019, 03:16:35 am »
Thanks for all the input and advice. I've decided to shelve this project for a couple of months until my next school break. Next term is a class on op amps and I'm going to try and tackle the buffering problem after I dig into the theory a bit. Hopefully I'll have some better questions then.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf