Maybe not, but it’s always good to provide a lower impedance load than the 10MOhm intrinsic impedance of the input of the buffers. If the inputs are JFETs, they need a way to drain off charge.
Some measurements using a Cayin N3 DAP playing a 1kHz tone from a flac file. The Cayin has an OPA1688 as a source, so really can't be used to measure itself. But here is what this amp gets for 1Vpp and 2Vpp into 47ohms.
Attachments
It is good to be smart and have knowledge. It is even better to be kind and respectful to others.
I'm returning exactly as much respect as I was given earlier. Beside that respect I'm giving a useful information. How could this be a bad thing?
Last edited:
I've got doubts about the stability of the circuit in post 93, from what I can see from sims. But I've also some doubts about the positionning of the Tian probe in a multiple feedback situation.
Anyone could verify if this is correct ? If it is the case, I would reduce the 2K2 resistors before the buffers to 0, add a zobel (33R-100nF will do) between the input opamp and the buffers and once again increase the output zobel to 47r-100n.
Anyone could verify if this is correct ? If it is the case, I would reduce the 2K2 resistors before the buffers to 0, add a zobel (33R-100nF will do) between the input opamp and the buffers and once again increase the output zobel to 47r-100n.
Attachments
cutting all loops at once is the best for stability analysis
but the output amps are expected to be unity gain stable so its easier to trim loop stability by measuring only the outer loop
C1 should be local to the input amp only
Zobel load shouldn't be needed with relatively low C load on input op amp
but the output amps are expected to be unity gain stable so its easier to trim loop stability by measuring only the outer loop
C1 should be local to the input amp only
Zobel load shouldn't be needed with relatively low C load on input op amp
Thank you. A subsidiary questions: should we interpret the stability conditions differently depending on whether one or two loops are broken ?
I re-ran the sims with the original values (thx for catching the mistake on C1). The two sims (breaking one loop or the two) give similar final results. That's the first attachment.
Then I re-ran the sims, only fixing the output zobel. Now the two sims don't match anymore (2MHz and 25° difference). Second attachment.
And finally, I reduced the 2K2 resistor to 1R. The results might be somewhat closer but still... Third attachment.
I re-ran the sims with the original values (thx for catching the mistake on C1). The two sims (breaking one loop or the two) give similar final results. That's the first attachment.
Then I re-ran the sims, only fixing the output zobel. Now the two sims don't match anymore (2MHz and 25° difference). Second attachment.
And finally, I reduced the 2K2 resistor to 1R. The results might be somewhat closer but still... Third attachment.
Attachments
I just had a go with the OPA1688 myself. Really simple design i made yet it sounds amazing. Running Dual Supply and a pretty low component count as i am still new in the Audio design world. Impressed by the opamp. Though i got some harmonics i might need to work a bit with it is a good start!
What are you using to do the Frequency response that goes up to 10MHz?
What are you using to do the Frequency response that goes up to 10MHz?
the series 2.2k R will interact with the output op amp's input C (or rather the Z since the unity feedback and diff in C gives some positive feedback)
if there's an optimum I expect it to be in 10's of Ohms if not 0
measuring inside the output unity feedback will give different frequency intercept because there's more gain
I think it sufficient that both show adequate gain, phase margins, Dave Zan may be better tooled up on Tain/Middlebrook probe and some multiloop issues
if there's an optimum I expect it to be in 10's of Ohms if not 0
measuring inside the output unity feedback will give different frequency intercept because there's more gain
I think it sufficient that both show adequate gain, phase margins, Dave Zan may be better tooled up on Tain/Middlebrook probe and some multiloop issues
Last edited:
Shaq888,
Not justified at all. Pretend you are a professional when you write to others.
It is how the message is delivered. Acting poorly reflects on you, which will affect how much respect you can expect from others.
Try being silent if you can't figure out a reasonable response.
This site is mostly self correcting for one. You don't need to save everyone from something you deem is wrong or incorrect. That and there are ways of doing this that are not as confrontational as you tend to be.
Many times it is how you say something that can be more important than what you said. Ideas are more likely to be accepted when presented with some politeness and something to back them up.
-Chris
I am using Focusrite 2i4 at 44kHz - data presented up to 20kHz. Where did you AE 10MHz?
Glad you tried it - what schematic are you using? Higher 3rd order harmonics can be caused by metal thick film resistors on feedback path. This amp is naturally 2nd order dominant distortion.
Last edited:
You know, I decided to verify how long it will take to make a tracking supply for the input opamp and how good it will perform. It actually turned out quite simple and well performing, especially considering that I didn't optimize anything and pretty much used parts I could find on my desk. Took me around 15 min to solder and overall additional cost is ~ $0.5.
As earlier I used 3Vrms (or 8.5Vpp on output of the amplifier), which gives around 0.75Vrms (~2Vpp) of a common mode signal at the input.
I made 3 measurements for each test freq with balanced impedances (around 350Ohm) - in red, 1.3kOhm (yellow) and 10kOhm (green-ish, turned out almost yellow) signal source resistance.
Results for 1kHz
10kHz
mix of 19kHz+20kHz
I would say it performs quite reasonable in audio band. I can handle even 50k pot, but I would not use anything higher that 10k anyway.
But there is an inconvenience - it needs to be an independent circuit for each channel.
Last edited:
You know, if you read back to my very first post, you'll find out that I was trying to be helpful because that person looked puzzled with the simulation results. The responses i was getting back were far from being respectful, I'm not even saying about acceptance. I'm as much human as anyone else, I can stand it only as much as I did. Which comes to you first sentence - justified or not, people put their emotion in replies.
The same comes to lecturing others. Does any of you really believe that anything you said will have a positive affect?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.