TLC271ID configured for unity gain, inverting, oscillates when output is loaded

There are nearly no op amps that you can hang large capacitive loads on the output and have it remain stable.  There are however application specific, external op amp  compensation techniques that can make the circuit stable.  I need some specific information from you in the answers to the following questions:

1) What values of C19, C21 and C26 do you want in a final design.

2) I need exact schematic of what connects to Output in your schematic.

3) I need exact information on the cable that connects your output to the amplifier and the load at the other end.

As you make any circuit stable you will lose closed loop bandwidth.  Based on the rest of your application requirements you may need a different op amp to achieve what you want.

At the link below you will find a 4 part PowerPoint tutorial on op amp stability that I think you will find useful. 

http://e2e.ti.com/support/amplifiers/precision_amplifiers/w/design_notes/2645.solving-op-amp-stability-issues.aspx

Hi Tim,

Thanks for the quick response. I will answer your questions in the same order as you asked them:

1) At the moment, C19 is not in place and the board has passed pre-compliance RF emissions/immunity testing, so it's not critical for this to be in place. C21 and C26 were needed to achieve pre-compliance - 10pF is what I used during testing, so I would like to keep them at this value.

2) The "output" on the schematic goes to a 2-pin MOLEX 70551-0001 connector.

3a) Again, we have little control over what cables our customers will use, but our main customer will either use two wire cable or possibly individual wires of around a 1 meter in length. Both options will use wires with multi-strand cores, with the cores around 1mm in total diameter. I have been using a two-wire cable that I cut-off of a pair of cheap portable headphones, with quite thin wires/cores. - there is a 2-pin Molex plug on one end (with a 3.3uF capacitor cut in to the signal wire to block the 2V DC offset) and a 3.5mm stereo jack on the other end; our main customer will either use a Molex plug at both ends or a screw terminal block at the amp end of the cable. I have been assuming that our customers will use an amp with AC coupling (see 3b below), but I'm starting to think that I should put a 1-3.3uF capacitor on the output of my board, to guarantee AC coupling...

3b) We do not have much control over what amplifiers our customers will connect to this output from our board. For testing, I have been using commercial off-the-shelf devices, such as a digital recorder (Edirol R-09HR), Soundcraft Notepad 102 mixer, and an active PC speaker from Creative. I believe they all have input impedances around 10k-20kOhm, but I do not have access to their schematics. Attached is a link showing the input stage of an amp that I was working on a couple of years ago. I have also seen one of our customers using an amp that had an ETOL P1200 line matching transformer on the input.

Yesterday, I tried changing the design to one with a closed loop gain of 10. I cut-in a potential divider on the input to C16: 82KOhms feeding C16, and 8.2KOhms from the input to C16 and 0V2; and increased R23 to 1MOhm. I still see oscillation - varies between ~260kHz and ~350kHz, dependent on whether the 3.5mm jack was plugged in to the amp or not.

I will now read through the PPT material you suggested and see what else I can try. Any suggestions/recommendations you can make will be very welcome.

Thanks and best regards,  Mohan.

Possible Input Stage.pdf

See attached PowerPoint for a detailed analysis.  This does not look like a loop gain stability issue.  That implies that there are other factors which I ask about in the attached presentation. You can ignore the Appendix for your specific current issue.  Since the TLC271 is an older part and since you mention a stability issue I needed to do  some op amp macro model adjustments to match closely with what the real silicon is based on all of the data in the datasheet.  Back when this macro model was developed it got the right Aol curve but not the right Zo (small signal AC open loop output impedance).  For your application loop gain analysis I used a corrected model. 

TLC271 Stability Analysis_Mohan.pptx

Hi Tim,

Thank-you for doing the analysis on my circuit - I am in some ways happy to see that the loop should be stable, but of course, this now means that I have to look for another cause for this "instability".

My design uses switch mode power supplies to generate the 5V and 3V3 rails. There is also a class-DG amp which shares the audio that goes to this pre-amp circuit - so lots of sources of potential noise. I have tried switching off my on-board regulators and fed the board from an analogue lab power supply, but it did not help. I have tried putting capacitors across the supply pin-7 and ground and also had no success. The only thing that appears to have helped to any extent is to put a 1kOhm resistor in the output of the op-amp, before the ferrite bead. This has killed the oscillation that I was seeing. Of course, this means the output is somewhat reduced, dependent on the load connected, but I think I'll have to live with this.

Unfortunately, I cannot share more information about my circuit on a public forum, but if we can switch to private email, I can give you more information about the circuitry around the op-amp, as well as some information about board layout, etc.

Thanks again and best regards, Mohan.

I am always concerned when oscillations seem to mysteriously disappear without the known root cause. Your 1k may be masking the real cause of the oscillation. You can email me direct at t-green1@ti.com so we can do some more detailed debugging.