Could you explain, or refer me to a text, why this opamp does not need local decoupling?
"High Q (low ESL/ESR) capacitors such as stacked films or ceramics should not be connected directly across the regulator output!"
The error amp power is supplied via bootstrap from the regulator's output (this was a change from the 1995 articles incorporated in 2000) so it's in the control loop, the Z's of the error amp input and outputs, if using the Kelvin connection, differ at HF which is where you experience the oscillation.
At any rate, the oscillation problem seems to "mostly" relate to use of the AD797 and not the AD825 or LME49710. The AD797 allows you to attain output impedance in the low microOhms, however.
The error amp power is supplied via bootstrap from the regulator's output (this was a change from the 1995 articles incorporated in 2000) so it's in the control loop, the Z's of the error amp input and outputs, if using the Kelvin connection, differ at HF which is where you experience the oscillation.
At any rate, the oscillation problem seems to "mostly" relate to use of the AD797 and not the AD825 or LME49710. The AD797 allows you to attain output impedance in the low microOhms, however.
I see the link and post 766 referring to the output.
Where does it say that "this opamp does not benefit from local decoupling?"
Where does it say that "this opamp does not benefit from local decoupling?"
The supply to the op-amp is the same point as the output. So decoupling the op-amp therefore means placing a cap across the output. 🙂
Last edited:
I understand that.
Thanks IanAs.
That should be explicit in the guide.
That this circuit does not suit local decoupling that would connect the output directly to Power ground.
It seems it is not the opamp that does not suit local decoupling. It's this circuit implememntation where a local decoupling cap across the power pins of the opamp are effectively across the output without any isolating resistance in between.
My original question was
Thanks IanAs.
That should be explicit in the guide.
That this circuit does not suit local decoupling that would connect the output directly to Power ground.
It seems it is not the opamp that does not suit local decoupling. It's this circuit implememntation where a local decoupling cap across the power pins of the opamp are effectively across the output without any isolating resistance in between.
My original question was
and that was preceded with
A simple explanation as IanAs has given would have resolved this way back !
Last edited:
Question for Jan Didden
I have been following this thread from the beginning and it has been a real education. Many thanks to everyone who has contributed, especially jan.didden, jackinnj, and WaltJ, whose posts have been especially helpful.
In the v2.2 and v2.3 circuit boads, Jan added RC filters to the sense lines (R15/C11 and R16/C12), to prevent oscillation when the sense lines are overly long. If I understand the schematic correctly, this effectively defeats the remote sensing feature at very high frequencies.
Is there any advantage to omitting these parts (removing the capacitors and jumpering the resistors) when the sense lines are short? I am specifically wondering about cases where the regulators are powering high frequency circuits like D/A converters.
I have been following this thread from the beginning and it has been a real education. Many thanks to everyone who has contributed, especially jan.didden, jackinnj, and WaltJ, whose posts have been especially helpful.
In the v2.2 and v2.3 circuit boads, Jan added RC filters to the sense lines (R15/C11 and R16/C12), to prevent oscillation when the sense lines are overly long. If I understand the schematic correctly, this effectively defeats the remote sensing feature at very high frequencies.
Is there any advantage to omitting these parts (removing the capacitors and jumpering the resistors) when the sense lines are short? I am specifically wondering about cases where the regulators are powering high frequency circuits like D/A converters.
There's nothing wrong with leaving them out as long as it is stable. It's a precaution, but there are many variables when you start using longer remote sense lines that there's no guarantee it will remain stable. Also no guarantee it will oscillate 😉
OTOH it does not really defeat the remote sensing unless you are working with video. The RC time places the roll off far, far above audio. So that should be no worry. At those frequencies the regulator itself will have troubles posing as a low Zout anyway! And good quality bypass right at the load will fix that of course.
Jan
http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-1505.html#post4497078
http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-1467.html#post4465657
Does that mean we'd be a bit better off not be using the LM329 in this regulator? Which alternative?
I had a quick look at that LT3042 data sheet. Noise might good but what about 'dynamic response' as mentioned as being more important in audio?
Does that go for phono stage regulators too? Would we be better to use this Super Reg or those LT3042 (if that can be soldered on to something)?
Last edited:
The LM329 is heavily filtered so the only source for noise is coming from the opamp itself so the answer is no.
The LT3042 seems to be very nice except for max 20 V in, tricky to solder by hand and comes only in a positive version.
The LT3042 seems to be very nice except for max 20 V in, tricky to solder by hand and comes only in a positive version.
quote from Glassware audio:
"I used two positive regulators. How is that possible? How do you get a negative regulator out a positive one? You don’t; you make two positive regulators and stack the outputs to create a bipolar power supply."
New bipolar low-voltage power supply for solid-state projects
does this work for LT3042?
"I used two positive regulators. How is that possible? How do you get a negative regulator out a positive one? You don’t; you make two positive regulators and stack the outputs to create a bipolar power supply."
New bipolar low-voltage power supply for solid-state projects
does this work for LT3042?
Has anyone an idea how a LT1085 based power supply compares to performance of superreg power supply?
Ah good 🙂 Thanks.
I bought the parts to throw a few of these SR's together but then had to fix a pals amp which has taken longer that expected, compounded by my only oscilloscope releasing some of it's smoke.
Thanks for the link 🙂 That lead me to Google that exact link which found me the rest of the series.
Last edited:
I have built 5 SRs out of the PCBs of my design based on Didden's PCB. 3 of them are tested working perfectly.
1 of them has a problem with the positive rail. The LED is not turned on. The input is 18.3VDC from LM317 prereg. The output should be 15.3V. But I got 17V.
1 of them also has a problem with the positive rail. The LED is turned on. But the output voltage was measured at 7.5V only.
I cannot think of anything that were done wrong. They were built together so there should be no incorrect pin orientations of active devices and the sort. The only potential problem could be overheating of the SMD AD825 as I don't have the proper sharp soldering iron tip to do the job so I went through it multiple times using more solder and using desoldering wick to take out the solder between legs.
Where should I start looking at to fix them?
1 of them has a problem with the positive rail. The LED is not turned on. The input is 18.3VDC from LM317 prereg. The output should be 15.3V. But I got 17V.
1 of them also has a problem with the positive rail. The LED is turned on. But the output voltage was measured at 7.5V only.
I cannot think of anything that were done wrong. They were built together so there should be no incorrect pin orientations of active devices and the sort. The only potential problem could be overheating of the SMD AD825 as I don't have the proper sharp soldering iron tip to do the job so I went through it multiple times using more solder and using desoldering wick to take out the solder between legs.
Where should I start looking at to fix them?