Another mistake

I do basically that in my latest chip design, except the second OpAmp is rail-to-rail input and can output 100mA rail-to-rail... the first OpAmp is integrating and also rail-to-rail output

It's probably singing... did you 'scope it? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.

Instead of directly coupling them in the (+) input, try a resistor between the output of the first OP-amp and input to the power amp.

I don't know the BW you're looking for? But my suspicions tell me if you use your first configuration and maybe something like a 2k R at least to couple them, you may see different results. also, I noticed that Op-amp has compensation inputs, have you terminated them properly? Just asking because I don't see it here ?

Btw, aren't those kind of old units? No matter, if they work I guess its fine.

Jamie

You probably exceeded the CM range of the power amp. A suitable voltage divider in between might make it work, assuming the power stage is much faster than the gain stage.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

er

=A0 =A0 ...Jim Thompson

=A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 | =A0 =A0 et =A0 =A0 =A0|

=A0|

=A0 =A0 =A0 |

ide quoted text -

I've always got a 'scope on circuits. It looks like I'm over voltaging the input of the second opamp. I've heard about phase reversal when you over drive. Now maybe I know about it.

Tomorrow, I'll try some diodes in the power supply leads to the first opamp. (Is there abtter way to drop a little voltage?)

George H.

k

wer

ss

k

wer

ss

Opps, (I hit the send button without typing)

Thanks Jamie I did try a few k ohms in the second circuit. (hmm maybe I need more?) The first circuit works fine. (Well except for the cross over distortion) I'm looking at a frequency of 20 kHz.

I was trying to do all this limiting with diodes and resistors, but the opa134 kept over powering whatever I hung on the output.

Do you have a better power opamp that does DC?

I was on the TI web site and noticed they now have power opamps that will do

100V. for ~$10. Anyone tried them?

George H.

Hey, George -

I got a pointer to this one a few weeks ago right here. It is a TCA0372DP2G. It is 1A output but only goes to 40V. I had not seen your requirement of 100V. HTH.

Cheers, John S

I've found issues with some circuits with 0 gain on the - input like you're doing. since that is a power amp, you basically are putting a very low Z on the (-). Normally when you do this, other items in the circuit after does not have a low Z. Not all amps behave the same like this how ever, in stead of a hard return to the - input with the power amp, try an R and still use the original configuration..

What I think may work is a high Z in the feed back. With charge that is in the (-) input, you should be able to get the cross over to switch differently and not have the distortion visible that you're now seeing..

It is worth a try any way..

Jamie

Opps, (I hit the send button without typing)

Thanks Jamie I did try a few k ohms in the second circuit. (hmm maybe I need more?) The first circuit works fine. (Well except for the cross over distortion) I'm looking at a frequency of 20 kHz.

I was trying to do all this limiting with diodes and resistors, but the opa134 kept over powering whatever I hung on the output.

Do you have a better power opamp that does DC?

I was on the TI web site and noticed they now have power opamps that will do

100V. for ~$10. Anyone tried them?

George H.

============================================= George,

these links at TI may be relevant to the issue on hand:

I was looking specifically for yet another app note on composite amplifiers, but could noy find it, maybe I'll be able to get a copy from an old disk...

Thanks, Jure Z.

I?ve got an opamp driving a unity gain power opamp.

|\ OPA134 Vin---|+\ |\OPA544 | >--+----|+\ +-|-/ | | >--+--Vout | |/ | +-|-/ | +--10R--+ | |/ | | +-------+ R | GND

With 10 ohms hanging on the output there is a bit of cross-over distortion. Hey, I thought, I?ll just feed back from the output and maybe the first opamp can ?goose? the second to reduce it.

So I changed it to this,

|\ Vin---|+\ |\ | >-------|+\ +-|-/ | >--+--Vout | |/ +-|-/ | +--10R--+--+ |/ | | +-------+ R | GND

Didn?t work at all??? The output of the first opamp was near the negative rail and the output of the power opamp was about 1/2 of the positive rail. Mucking about (adding some C from the first opamp back to it?s inverting input) I got the thing to work. But when I re-power or hit the rail with the output the thing stops working again.

Do I need to limit the voltage swing going into the second opamp?

Thanks, George H.

********************* Yes. You're exceeding the Vcm of the 544 which is typically 4V inside each rail. Min is 6V inside each rail. Art

Oh, the 100 Volts would be for a different Piezo circuit.

40 V would be plenty for the present circuit. (+/-15 V rails.)

I'll give it a look see.

Hmm, If I'm getting an amp...It sure feels like I want (at least) a TO-220 type of package. I was thinking of bolting it to the front panel... A 8"x12" piece of aluminum. (I'll need some sort of Sil pad.)

George H.

to

er

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

rs,

..

Wow, Thanks Jure. That's a bit more than I wanted to do.

But I see I can mix up the gain. I wanted to keep the power opamp at minimum gain. That looked like the lowest distortion.

George H.

On a sunny day (Wed, 28 Dec 2011 13:27:55 -0800 (PST)) it happened George Herold wrote in :

Sometimes 'crossover' can be eliminated by a bit of feed forward, in this setup R1 drives the output for very small signals. It should not oscillate, as it is driven from a very low impedance. The size of R1 depends on the size of the cross-over region and the power required at the output (say load). I have used someting like this is power RF amps (150kHz) ... many many years ago.

----R1--- |\ OPA134| | Vin---|+\ | |\OPA544| | >--+----|+\ | +-|-/ | | >--+---Vout | |/ | +-|-/ | +--10R--+ | |/ | | +-------+ R | GND

r

Sounds like you invented an analog latch.

to

back

re-power

=A0 =A0 ...Jim Thompson

=A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 | =A0 =A0 et =A0 =A0 =A0|

=A0|

=A0 =A0 =A0 |

=A0|

|

food.- Hide quoted text -

If you do consider a resistor between the diodes to hold them well into the conduction curve.

?-)

to

ack

power

=A0 =A0 =A0 ...Jim Thompson

=A0 =A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 | =A0 =A0 et =A0 =A0 =A0|

=A0 =A0|

=A0 =A0 =A0 =A0 |

=A0|

|

- Hide quoted text -

Thanks, It was the common mode voltage as Phil H, suggested. I threw away some signal and then added it bakc and got it to work. But no improvement in the cross over distortion. :^(

George H.

quoted text -

Did you try what I suggested in the feed back loop ? And that is to use a R in the feed back instead of a hard loop back?

This helps to reduce Ringing on fast transitions also, I find that when some amp devices get to the cross over, there isn't enough shoot through current and the (-) input is left with a low Z loop line connected to a diode junction that can act as a rectifier for stray RF, this will certainly give you a little noise.

By using a resistor in the feed back for the - input instead, you can take advantage of the charge that is sitting at the (-) to give you a slight gain increase to help you move that crossover a bit faster.

Also, increasing the R at - input which will lower the Q of any induction the feed back loop may have on it.

It's just an observation. I've had to use R's in the feed back of a unity stage amp instead of a hard loop for these reasons.

You may want to observe your output at crossover for what looks like a ringing. This could indicate the outputs are not getting on soon enough to quench the loaded C and prevent ring.

Forcing the (-) input hard can cause the charge internally to actually work against you in some cases. A lot depends on chip design.

Jamie

Hi

I've no idea if this would work at all, but what about reducing the gain of the 1st opamp before feeding the 2nd?

|\ Vin---|+\ |\ | >--+-----|+\ +-|-/ | | >--+--Vout | |/ | +-|-/ | +-100R--+ | |/ | | | | +-10R-------+-------+ | R | GND

--
bye
!(!1|1)

ars ago.

Cute.

The brutish way is to load the OPA544 output, forcing it to run class A. Cleans up LM324s nicely.

-- Cheers, James Arthur