if you have changed components or component values, then the modified circuit now belongs to you.
Post away.
But you're wrong about "darlington and cascode enhancements" they reduce distortion significantly. You won't convince me that the sound of non-linear Cbc is better than not having it.
Thank you Mr Self.
Here it is. Please bear in mind that although it might look sub-optimal in some ways from a purely objective analysis, this is what I ended up with after extended testing of various configurations and listening over long periods. There is a relay attenuator up front which expects to see 10k, hence the low input bias resistor. The output offset is high at 0.5V, but since the output is AC coupled and it sounded better without the feedback cap I left it at that, figuring the headroom loss was insignificant.
I never disputed the improvements in objective peformance that VAS enhancements bring, and who am I to argue with you two distinguished gentlemen anyway
All I said is that the VAS enhancements sounded worse (to me and in this particular circuit). Pointless arguing about that and I would not expect to convince you Mr Self!
I found the whole process very interesting, having expected that as the THD was improved by increasing gain in the VAS the sound quality would also improve, but this did not seem to be the case in practise. My methods were hardly scientific, but I was only doing it to suit my own ears after all...
The worsening of SQ mostly seemed to be an increase in midrange "forwardness" and "hardness" where eg strings took on a screechy quality. There are some nice audiophile adjectives for you
I did carry out some distortion measurements on the circuit as shown using a friend's Cyril Bateman rig from his capacitor distortion articles. Just looked up the measurements and it seems at 1kHz I was getting:
@ 1V input and 1V output:
2nd harmonic: -110dB, 3rd harmonic -113dB
@ 1V input and 3V output:
2nd harmonic: -106dB, 3rd harmonic -122dB
Not too shabby. In theory the VAS enhancements should have lowered these even further but the levels should already have been inaudible. I thought the overall SQ was audibly worse though.
Here it is. Please bear in mind that although it might look sub-optimal in some ways from a purely objective analysis, this is what I ended up with after extended testing of various configurations and listening over long periods. There is a relay attenuator up front which expects to see 10k, hence the low input bias resistor. The output offset is high at 0.5V, but since the output is AC coupled and it sounded better without the feedback cap I left it at that, figuring the headroom loss was insignificant.
I never disputed the improvements in objective peformance that VAS enhancements bring, and who am I to argue with you two distinguished gentlemen anyway
All I said is that the VAS enhancements sounded worse (to me and in this particular circuit). Pointless arguing about that and I would not expect to convince you Mr Self!I found the whole process very interesting, having expected that as the THD was improved by increasing gain in the VAS the sound quality would also improve, but this did not seem to be the case in practise. My methods were hardly scientific, but I was only doing it to suit my own ears after all...
The worsening of SQ mostly seemed to be an increase in midrange "forwardness" and "hardness" where eg strings took on a screechy quality. There are some nice audiophile adjectives for you
I did carry out some distortion measurements on the circuit as shown using a friend's Cyril Bateman rig from his capacitor distortion articles. Just looked up the measurements and it seems at 1kHz I was getting:
@ 1V input and 1V output:
2nd harmonic: -110dB, 3rd harmonic -113dB
@ 1V input and 3V output:
2nd harmonic: -106dB, 3rd harmonic -122dB
Not too shabby. In theory the VAS enhancements should have lowered these even further but the levels should already have been inaudible. I thought the overall SQ was audibly worse though.
Attachments
It's a pretty conventional Lin topology stage.
Do you think it sounds better than an audio targeted opamp like a 553x or a 4562?
It's clear tha you can probably get more drive out of it than an opamp, since there is no current limit, but at the levels you are talking about (1V out)I'd say this is hardly an issue.
Do you think it sounds better than an audio targeted opamp like a 553x or a 4562?
It's clear tha you can probably get more drive out of it than an opamp, since there is no current limit, but at the levels you are talking about (1V out)I'd say this is hardly an issue.
Thank you for that astute observation, the thought had never occurred, hence why I mentioned it a couple of posts ago...
In normal listening or A-B comparisons it sounds much better than the opamp preamps I've compared it to. I also did some single-blind testing of this cct against a single NE5534 configured with the same feedback arrangements and after careful listening to about 60 tracks found that I preferred the discrete circuit about 78% of the time.
My thoughts are that the devil is in the detail, however optimising this topology (or perhaps any) for THD and or noise alone looks nice on the spec sheet but doesn't necessarily equate to subjective improvement of SQ. Perhaps it comes down to the relative levels and distribution of the harmonics, though I guess if anyone really had the answers this would cease to be a topic of discussion and audio reviewers would have to go and get real jobs
On a related subject I've just finished repairing our Sony Trinitron 32" HD CRT TV (the last of the CRTs) and boy am I glad to have it back. Compared to our other recent model Panasonic 50" plasma with 3D processing and much higher resolution, the CRT is much more realistic - natural colours and tones, a picture that seems more 3 dimensional and much faster response time so that there is no delay or jerkiness with fast movement. The plasma has better measured specs in just about every category and yet everyone in our household prefers to watch the smaller CRT. Anyone else see any parallels with audio? And music is tied up with our emotions... I don't understand why many design engineers seem to be happy to just optimise THD when the consensus amoung experienced listeners is that it doesn't correlate well with SQ. Surely it must be possible to pin down the measrements that are actually important for SQ?
In normal listening or A-B comparisons it sounds much better than the opamp preamps I've compared it to. I also did some single-blind testing of this cct against a single NE5534 configured with the same feedback arrangements and after careful listening to about 60 tracks found that I preferred the discrete circuit about 78% of the time.
My thoughts are that the devil is in the detail, however optimising this topology (or perhaps any) for THD and or noise alone looks nice on the spec sheet but doesn't necessarily equate to subjective improvement of SQ. Perhaps it comes down to the relative levels and distribution of the harmonics, though I guess if anyone really had the answers this would cease to be a topic of discussion and audio reviewers would have to go and get real jobs
On a related subject I've just finished repairing our Sony Trinitron 32" HD CRT TV (the last of the CRTs) and boy am I glad to have it back. Compared to our other recent model Panasonic 50" plasma with 3D processing and much higher resolution, the CRT is much more realistic - natural colours and tones, a picture that seems more 3 dimensional and much faster response time so that there is no delay or jerkiness with fast movement. The plasma has better measured specs in just about every category and yet everyone in our household prefers to watch the smaller CRT. Anyone else see any parallels with audio? And music is tied up with our emotions... I don't understand why many design engineers seem to be happy to just optimise THD when the consensus amoung experienced listeners is that it doesn't correlate well with SQ. Surely it must be possible to pin down the measrements that are actually important for SQ?
It is fascinating how different perception is from specs. There are endless number of examples where less linear circuit gives subjectively better results. One example is old budget NAD 3020 amp. No mirrors, no CCS, yet sound that is subjectively better than more complex circuits. Great linearity is not guarantee for better perception of sound. But most engineers would lose their jobs (which are scarce these days) if they would use simple amateur circuits. They must compete for existence of their jobs. Listener is not priority, because specs look better before you buy something. But Doug (and Bob, and other great engineers...) have sensible approach professionwise, because it is much easier to please scopes and spectrum analyzers than countless personal tastes.
Well, that's what we were hoping you'd get around to. As you have appropriate DIY experience, preamps that fit both criteria, the axe to grind and the formal discipline to tackle the matter, then let's hear it.
It's ironic though, that the usual way now to instill "SQ" into an amplifier is to ensure it generates a specific harmonic distortion "profile". Cynical, innit?
High linearity circuits are one way to avoid unpleasant sound, not a guarantee that listener shall actually enjoy the sound. It is safe way for circuit designers. But since everybody has some taste, some coloration he/she prefer, the most successful circuits are those that contain desired coloration inherent in topology. In my experience, for most people, highly linear circuits with tone controls are less successful way in offering listener desired coloration. Therefore, simple circuits with idiosincracies in form of more harmonic distortion, are for many listeners better way of adding the missing sparkle, than treble boost with tone control.
I remember, few years ago, that studio users of Yamaha NS10 mini monitors, used to tape paper tissue (toilet paper) over tweeters, although it only makes response more wavy, due to reflections (not absorption).
I remember, few years ago, that studio users of Yamaha NS10 mini monitors, used to tape paper tissue (toilet paper) over tweeters, although it only makes response more wavy, due to reflections (not absorption).
My Finished DS Preamp
Hi all,
Thought I'd post a pic of my finished 2012 Doug Self preamp. It sounds wonderful and is clearly the most expensive piece of gear in my modest system, and my most challenging DIY project of the year. But it was worth it! Thanks again Doug for an excellent sounding design and implementation.
My thread on the build can be found here.
Rick
Hi all,
Thought I'd post a pic of my finished 2012 Doug Self preamp. It sounds wonderful and is clearly the most expensive piece of gear in my modest system, and my most challenging DIY project of the year. But it was worth it! Thanks again Doug for an excellent sounding design and implementation.
My thread on the build can be found here.
Rick
Attachments
Tone controls today are much different than ones from the past. Typical modern tone controls are super precise shelving circuits. Old school tone controls were often passive and quite imprecise; for example some controls shifted the turnover frequency as you adjusted them. And passive tone controls do introduce phase shift, but they do not introduce phase shift into the feedback loop.
With modern op amps we can build a much more precise passive tone stack than in the old days, and this might turn out to be a worthwhile exercise. A well designed loudness control can make a big difference too
, although purists loathe them. Some recievers had adjustable loudness controls via either a potentiometer or switches. It's way better than one size fits all which is sometimes useless in certain systems.Thanks Mooly. It was a fun, time consuming, and at times frustrating project, but I never got deterred. This was my first time using Front Panel Express for the front and back panels, but they turned out great. Laying out the panels was very trial-n-error, until I got it right. A nice, well-laid out, machined, and lettered FP adds a very professional touch and look. A little expensive, but not bad in the big picture.redjr... very nice job
I'm very happy with it and the sound continues to amaze me. It's everything I had hoped it would be.
Rick
Last edited:
Hi Rick,
That's a VERY nice job!
Is that a Front Panel Express front panel?
Where did the enclosure come from?
Cheers,
Bob
I was looking at that front panel too... enclosures and panel work are always major issues (for me at least ) and deter me from many a project.
I was just googling "Front Panel Express". I'd never heard of it before. For a "proper" project something like that is worth the expense for sure.
It is an excellent job.
) and deter me from many a project.I was just googling "Front Panel Express". I'd never heard of it before. For a "proper" project something like that is worth the expense for sure.
It is an excellent job.
Hi Bob,
Yes. FPE machined the front and back panels, but I sent them the blank ones that came with the enclosure. I did all the layout in FPE, plagiarizing from another member's (muskyhuntr) work where I could. It's a real trial-n-error process until you get it right. Unfortunately, my eBay supplier is not offering that exact same case anymore. Not sure if he'll have them in the future or not. As it turned out, the case worked, but was a little smaller than I would have liked. Layout was cramped inside, but I did make everything fit. I don't appear to be getting any induction, or ground loop hum, but all signal wiring is shielded. I've attached a picture of the final inside, sans the FP.
Rick
Attachments
Hi Mooly,I was looking at that front panel too... enclosures and panel work are always major issues (for me at least
I was just googling "Front Panel Express". I'd never heard of it before. For a "proper" project something like that is worth the expense for sure.
It is an excellent job.
I think getting just the right chassis, or enclosure for a project is critical to its success - for me anyway. This was the first 'advanced' project I've ever tackled, so it was good to know the FPE was available for the mechnical and final fit-n-finish part of the project. I don't have a drill press yet, but it's on my list of bench tools that is a must have for projects like this. FPE offers a nice service, and free software that gets the layout job done. All you have to do is upload your files, supply the material (not necessary), and in about 2 weeks you'll receive a nice package from UPS. The total machining and lettering cost me about $160USD.
Rick
Is that an assumption or have you simulated or measured it in this circuit? You may be right, but it also changes other things such as the amount of feedback.
While I'm not sure I buy into the whole feedback is evil camp, another thing I noticed with this design is that when I reduced the closed loop gain below that shown on the schematic, the sound got harder and the soundstage narrower. My conclusion at the time was the reason that the VAS enhancements made the sound worse was the extra global feedback, since the worsening of SQ manifested in similar ways. I don't see why this should be so and it was quite possibly the wrong assumption, but I'm yet to find an alternative explanation.
Well, that's what we were hoping you'd get around to. As you have appropriate DIY experience, preamps that fit both criteria, the axe to grind and the formal discipline to tackle the matter, then let's hear it.
It's ironic though, that the usual way now to instill "SQ" into an amplifier is to ensure it generates a specific harmonic distortion "profile". Cynical, innit?
Me? You're joking - to do this properly would be a major undertaking involving massive resources and (open-minded) experts from various disciplines. Even if I did have all those attributes and the time and equipment to do it, if I came back here with my own conclusions I'd be shot down in flames by all the hard line objectivists faster than you could say "ABX comparator"...
Perhaps so, but if a certain harmonic spread turns out to improve the fidelity or enjoyment of recorded music I don't see any cynicism, only a more sensible way of approaching design that considers such factors.