abbey road d enfer

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #20 on: June 17, 2019, 03:02:27 PM »
Im doing achoustic instrumentation in which the phase between mics is critical, so yes theres something wrong with LF phase shift in some applications.
There is nothing wrong with phase-shift due to the High-Pass response. This phase-shift never exceeds 90°, which ensures that signals are always in the same quadrant, so combine without negative interference.
You are confusing this with distance-related phase-shift, which creates a non minimum-phase system.
« Last Edit: June 18, 2019, 12:47:29 PM by abbey road d enfer »
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.


benb

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #21 on: June 17, 2019, 07:01:26 PM »
What's the cutuff frequency in a typical servo circuit? One Hz? A tenth of a Hz? It's well below the claimed audible 20Hz, regardless, and any phase shift in the audio band is insignificant. There are much more serious phase shift problems in any multi-driver speaker crossover.

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #22 on: August 07, 2019, 07:57:42 PM »
I think servos get a bad rep due to sloppy implementation myself.  And yes, there is still a cap in the signal path with a servo, but you're replacing a large cap for a better quality small cap. 
Some thoughts:
Ideally you start with a circuit with low offset, then there isn't such a huge swing to correct.  Myself, I usually make the time constant long on the integrator, and I like to follow the integrator with another filter a decade away which is good for ameliorating the audible effects of the thing in my opinion.   You need to juggle the loop gain of the servo vs the loop gain of the circuit.   Make sure the servo still has enough input voltage headroom to be able to correct etc.
I also prefer inverting integrators followed by a unity inverter if needed so I only use one cap.   IOW, I use a dual op-amp.  Make sure this also is a low offset amp of course 😆
Um, what else...  If it's a discrete circuit you're working on, sometimes wrapping the servo just round the output stage rather than the whole circuit is better.
Best thing is to tailor the servo to the circuit rather than cutting and pasting , otherwise, use a big cap and be done 😃

boji

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #23 on: August 07, 2019, 09:05:48 PM »
Loving this thread.  Might someone direct me to any reading material on the relationship of DC offset and tx performance or what makes a 'bad' servo bad? For example I'm confused by some mic pre kits that allow the user to jumper-bypass caps, as if to suggest it might be a 'tone' one is looking for in an output tx.

Thank you!

Edit: Hope you don't mind the interjected question, DF.
« Last Edit: August 07, 2019, 09:17:43 PM by boji »

abbey road d enfer

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #24 on: August 08, 2019, 01:59:25 AM »
Might someone direct me to any reading material on the relationship of DC offset and tx performance
Not specifically. You have to understand ferromagnetism and how permanent bias shifts the magnetizing curve. Then you understand what DC offset results and that it may not be such a big issue, within limits.

Quote
or what makes a 'bad' servo bad?
Like any piece of electronics, bad design. The two main issues with servos are the lack of control range and inadequate frequency response.

Quote
For example I'm confused by some mic pre kits that allow the user to jumper-bypass caps, as if to suggest it might be a 'tone' one is looking for in an output tx.
Coupling a xfmr via caps creates a resonant circuit, that ususally results in a hump at subsonic frequencies. It may actually overload the output stage if VLF noise is present. There are ways to damp this resonance, but actually it's not very often implemented.
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.

boji

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #25 on: August 08, 2019, 10:19:04 AM »
Thank you Abby! Edit: more reason to get familiar with scopes.
« Last Edit: August 08, 2019, 10:24:20 AM by boji »

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #26 on: August 08, 2019, 10:39:48 AM »
Not specifically. You have to understand ferromagnetism and how permanent bias shifts the magnetizing curve. Then you understand what DC offset results and that it may not be such a big issue, within limits.

Interestingly (to me at least),  the "silk mode" control on Rupert Neve's newer mic amps exploits this 'shifting within limits' by varying the offset applied to the primary of the output transformer.   

5v333

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #27 on: August 08, 2019, 12:01:26 PM »
What about using a gyrator on the output as a shunt to to ground with fc set to 0.1hz?

Nothing in series with the audio path or the feedback loop.

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #28 on: August 08, 2019, 12:44:11 PM »
What about using a gyrator on the output as a shunt to to ground with fc set to 0.1hz?

Nothing in series with the audio path or the feedback loop.

Not seen that, do you have an example?

I've seen (though not experimented much with) taking the output of the DC servo into a transistor as a voltage to current converter to tweak the quiescent of the top or bottom in an output stage.  It could be argued that this is not in the signal path.  No more so than the resistor that normally sets the current anyway. 

abbey road d enfer

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #29 on: August 08, 2019, 01:04:29 PM »
What about using a gyrator on the output as a shunt to to ground with fc set to 0.1hz?

Nothing in series with the audio path or the feedback loop.
It would be in parallels though; same possible adverse effects. More so, considering it would have to shunt a possibly hefty output stage.
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.


abbey road d enfer

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #30 on: August 08, 2019, 01:08:11 PM »
Not seen that, do you have an example?

I've seen (though not experimented much with) taking the output of the DC servo into a transistor as a voltage to current converter to tweak the quiescent of the top or bottom in an output stage.  It could be argued that this is not in the signal path. 
And that would be utterly wrong... Signal is voltage AND current. Both should be as little distorted as can be.
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #31 on: August 08, 2019, 01:43:52 PM »
And that would be utterly wrong... Signal is voltage AND current. Both should be as little distorted as can be.

Yes I'm aware what constitutes signal.   
So, tell me what's wrong with the attached any more so than a DC servo? 

Edit:  Thinking about this some more, bringing the transistor back to the other side of R24 might be more optimum.

Edit of edit:. A quick sim shows this version does work but requires some juggling, seems finecky...

« Last Edit: August 08, 2019, 02:52:55 PM by Winston O'Boogie »

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #32 on: August 08, 2019, 02:23:59 PM »
Like So:

Edit:. Another quick sim and juggling around of things and this works too, less finecky than above but...
« Last Edit: August 08, 2019, 02:54:35 PM by Winston O'Boogie »

abbey road d enfer

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #33 on: August 08, 2019, 03:39:03 PM »
Yes I'm aware what constitutes signal.   
So, tell me what's wrong with the attached any more so than a DC servo? 
My answer was really for 5v3333, who suggested brute force for hi-passing signals.
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #34 on: August 08, 2019, 03:53:42 PM »
My answer was really for 5v3333, who suggested brute force for hi-passing signals.

OK fair enough 😊
I am genuinely interested in your opinion as to whether you think shunting quiescent current would have less of an impact over say the last DC scheme I posted.  Although I suspect that, like me, you don't have too much of an issue with an appropriate cap.
Peace

abbey road d enfer

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #35 on: August 08, 2019, 04:51:42 PM »
I am genuinely interested in your opinion as to whether you think shunting quiescent current would have less of an impact over say the last DC scheme I posted. 
With regard to the 2nd implementation, I would be concerned as the compensation loop is modulated by the output signal; then I'm afraid shunting quiescent would also be subject to same issue. Practical current sources are not perfect.
It is just a first impression, though...
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #36 on: August 08, 2019, 05:17:51 PM »
With regard to the 2nd implementation, I would be concerned as the compensation loop is modulated by the output signal; then I'm afraid shunting quiescent would also be subject to same issue. Practical current sources are not perfect.
It is just a first impression, though...
Thanks.  Yep, I think you might be right.  I don't have enough interest right now to build it and test in circuit.  And I certainly don't trust that spice says it's ok, maybe that's my lack of expertise with simulating?
Anyway, cheers, appreciate the input.

5v333

Re: Do you know any other way besides transformers or caps to block DC?
« Reply #37 on: August 08, 2019, 08:01:41 PM »
I guess by putting synthetic inductance in parallell with an OT would also lower the ability of the amp to drive the lowend of the OT...



 

Related Topics

  Subject / Started by Replies Last post
2 Replies
3195 Views
Last post June 04, 2004, 12:49:18 PM
by Scenaria
8 Replies
2819 Views
Last post May 24, 2005, 05:18:00 AM
by clintrubber
14 Replies
13928 Views
Last post November 28, 2006, 12:49:02 PM
by Gareth Connor
6 Replies
335 Views
Last post December 05, 2019, 12:24:28 PM
by Rob Flinn