John Curl's Blowtorch preamplifier part III

Status
Not open for further replies.
How are the bearings? Do the wheels wiggle? Are there any in the blade guides?

Can the saw cut a straight line in plywood?

Haven't tried the latter and we need to get at least one fresh blade as a dull one will invariably fail. The guides were known to need replacement/modernization. The frame/bearings are smooth and solid, no undue rattles or vibrations when we fired it up -- but we'll go through the whole thing and make sure it gets set up square.
 
www.hifisonix.com
Joined 2003
Paid Member
A bit retarded reaction, but just flying through the postings of this endless thread, I came across your suggested servo system, meant to remove DC from the output from the PA.
However to my opinion that is not what this topology is doing.
It should be seen as an Active Error Feedback or AEF, correcting distortion components caused by the PA, see image below.

When increasing the BW of the LPF in your schematic, it will do a very effective job of error correcting over the whole audio range, but DC at the input will come out as the same DC, not at all what a DC removing servo is supposed to do.

Hans

This is a bit like AFEC which I’ve used and is very effective at reducing overall distortion. 10x reduction, removal of DC offsets and a 30 dB improvement in PSRR

Here is a practical implementation Ovation High Fidelity - Model 1721

Here’s the write-up voverimg the technicalities Augmented Feedback Error Correction (AFEC)
 
Another great post by someone who knows what they're talking about
Talking about objectivity ?
Don't you realize it means somebody know what he is talking about when he think the same thing than you ?

And, about this thread, this is something I did decades ago. In a way closer to the reality: one gain 1X20db attenuator, one gain 10 stage, one gain 1 stage inverting, one attenuator, one gain 10 stage inverting, one gain 1 stage.
No need to explain why ?

Anyway, the result cannot be generalized. It depend greatly of the sources you listen-to and the equipment used to listen.
Let say the sources are the drivers and the listening system a racing circuit, this can quickly become a comparison between two formulas ones driven by ordinary drivers on a dirt track. See what I mean.(No surprise if one of us "don't understand " ;-)
 
Last edited:
A bit retarded reaction, but just flying through the postings of this endless thread, I came across your suggested servo system, meant to remove DC from the output from the PA.
However to my opinion that is not what this topology is doing.
It should be seen as an Active Error Feedback or AEF, correcting distortion components caused by the PA, see image below.

When increasing the BW of the LPF in your schematic, it will do a very effective job of error correcting over the whole audio range, but DC at the input will come out as the same DC, not at all what a DC removing servo is supposed to do.
Ha ha, I wonder if you had not understood my schematic ? I use this servo AND this optional error correction you suggest (that you can switch in or out). Not at all the same purpose.
The servo is a servo, as it has no effect on the audio bandwidth, but only on DC and ultra low frequencies.
Just lke any other. The difference is the audio singnal has been removed by the comparator in order to add ecciciency and minimize any sonic impact of the servo.
See the original:
Pizzicato, a 200W low distortion CFA amplifier
 
Last edited:
Maybe because they failed pathologically, Hallmark used a particularly bad 1000uF/10v cap to store the rectified AC (the pre-2005 ornaments were a real hack depending on the divider string of lamps right off of the mains). The cap leaked, failed, and a replacement worked fine.
What is you point of view, Scott ? That electrolytics used in the signal path are totally transparent ? And that all the people that noticed a difference were just fooled by some kind of rumor ?
 
Disabled Account
Joined 2012
Here is a nice Current-Mode Amplifier (CMA) example I am going to try and measure and listen.

comments? Needs some adjustment to schematic;


Current-mode-amplifier-circuit-diagram.png


THx-RNMarsh
 
What is you point of view, Scott ? That electrolytics used in the signal path are totally transparent ? And that all the people that noticed a difference were just fooled by some kind of rumor ?

I'm sorry these are PS caps that failed by chemical leakage what does that have to do with anything else? BTW I guess we are not getting a Xena and Gabrielle ornament in the foreseeable future.
 
Disabled Account
Joined 2012
I don't know what this straw man argument is regarding THD and polar caps is about. Walt and I showed in 1980 Audio article the distortion increase when voltage is across the cap. And how it goes down when lower voltage is across the cap. So, others more recently did same tests and find same results.

What is the big deal?

Of course there are many places where a volt drop is part of the necessary function and design. Filters.

We also showed the affect of zero vs DC bias on polar cap distortion.

many tests of polar caps do not appreciate the effect of the diode effect in those tests. Most models of electro do not show the poor quality diode of the polar cap but it is responsible for distortion if the signal applied to a non-biased polar capacitor. namely, 2H.

if your tests are below the diode threshold, you wont see much 2H but if a signal peak exceeds that diode volt, you get 2H. Depending on volt rating of the cap, the diode volt can easily be exceeded with zero vdc on cap. So a test must consider the test ac voltage level. Some test levels have been just under the diode turn-on voltage.

Here is a model you could use for sim.


electro leakage model.jpg



TH-RNMarsh


In 1980, most music sources were still from LP. Many amps/pre-amps at that time used the coupling cap to roll-off low freq artifacts (HPF) so rumble, flutter, wow et al would be attenuated.

Today with the new standard of DAC sources, Almost all CDPs/DACs analog outputs have signal strengths of 2- 2.3 volts. Peak levels can more easily cause distortion with unbiased polar/electrolytic coupling caps.
 
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
Another great post by someone who knows what they're talking about https://www.diyaudio.com/forums/ana...pamps-chained-measurements-4.html#post6010363
I enjoy your links, but they normally don't work for myself and those that have a different posts per page setting than you do, when you copy the address from the address bar, but they should if you copy the post# link in the top right of the post. Not meaning to be difficult, I've said it now I'll move on.
 
Richard,

I tracked down the article of the schematic you posted and as I suspected it is a video/RF amplifier and I am trying to understand its use in our application(audio). The author says that to maintain stability you have to limit the gain and I suspect the amp is only stable into a specific load and not a real world speaker.

Jam
 

Attachments

  • CurrentModeAmps.pdf
    253.4 KB · Views: 69
Last edited:
Disabled Account
Joined 2012
Richard,

I tracked down the article of the schematic you posted and as I suspected it is a video/RF amplifier and I am trying to understand its use in our application(audio). The author says that to maintain stability you have to limit the gain and I suspect the amp is only stable into a specific load and not a real world speaker.

Jam

Try it. Learn the characteristics of a CMA. Thats the reason for posting a CMA design. With 811 it is quick to DIY and easy to build/test.

Most are not stable at gain = 1 unless specifically made to be. You will get sever peaking at top of BW limit if you tried. Yes, the gain setting R's are critical to proper operation of CMA. Its isnt so willy-nilly values as in VFA.

It will need some tweeking as i said for my app. But, you could learn a lot about CMA circuits if you actually built a CMA.


-Richard
 
Last edited:
Disabled Account
Joined 2012
And AC can not to develop (so also distortion in audio band..) across capacitor if value is "sufficient" for given load and aplication as coupling capacitor. And some can not to accept this..That is all..

yes, you can go to a very large value. If cost, size dont matter. But as I said in foot note many use coupling cap to limit low freq response.

Several of them in series in a record/[playback system can lead to other issues, such as group-delay being excessive.

Going to all direct-coupled leads to entire system which over-all can sound more accurate. It really isnt a big issue of only one coupling cap but a typical system having many and the total affect also needs to be considered (and measured).

Typically, in VFA and LP system.... besides the recorder and all the electronics, mixing, microphone etc on the recording side.... a pre-pre, an RIAA pre, a line stage and a power amp. Each with an input, output and fb to grnd electrolytic coupling cap. How many is that total in just the play back of an LP. Or inside a CD player? At least 8-11 caps. each rolling off the low end.

Might be an audible benefit to eliminate them all.


IMO, of course.



-RNM
 
Last edited:
I have Richard, a variation of the circuit I posted.

Issues you usually have to add a buffer (which increases the complexity) and usually an additional gain and then add compensation to get the whole mess to work in harmony, no elegance here. I still don't see the need for DC to light bandwidth, could you explain that one to me.

Problem is, it might measure good but they sound flat and lifeless once you have it working (probably due to all that complexity).

Maybe Richard you should post a working circuit, that you think qualifies, that we could build and test.;)

Jam
 
Last edited:
I enjoy your links, but they normally don't work for myself and those that have a different posts per page setting than you do, when you copy the address from the address bar, but they should if you copy the post# link in the top right of the post. Not meaning to be difficult, I've said it now I'll move on.
Thank you. Does this work better? 12 opamps chained - measurements
 
Last edited:
Status
Not open for further replies.