I'll send you a file, please see if it is readable.
Look at the result, people actually liked the article. Now if it makes it's way to resistor manufacturers...!
Last edited:
Simon,
😱 😱 😱 😱 😱 😱 😱 😱
Quel Dommage. You build a very noisy circuit methinks.
Below is one of mine.
Single shot 32K FFT, no averaging, output level is around 2V RMS. AP2.
With enough measurements averaged the noisefloor essentially disappears.
It is a compound stage using four triodes, uses both local feedback and distortion cancellation, the 2nd HD could be lower, but it does not sounds as good that way and higher harmonics are boosted that way, so it ended up the way it is.
The room these measurements where taken in BTW, is full of fluorescent lights, all sorts of switched mode supplies, computers and a bunch of really big isolation transformers (they make themselves known in the plot BTW) right behind the measurement place (got to move these one day)...
Ciao T
😱 😱 😱 😱 😱 😱 😱 😱
Quel Dommage. You build a very noisy circuit methinks.
Below is one of mine.
Single shot 32K FFT, no averaging, output level is around 2V RMS. AP2.
With enough measurements averaged the noisefloor essentially disappears.
It is a compound stage using four triodes, uses both local feedback and distortion cancellation, the 2nd HD could be lower, but it does not sounds as good that way and higher harmonics are boosted that way, so it ended up the way it is.
The room these measurements where taken in BTW, is full of fluorescent lights, all sorts of switched mode supplies, computers and a bunch of really big isolation transformers (they make themselves known in the plot BTW) right behind the measurement place (got to move these one day)...
Ciao T
Attachments
Just need to convert from PostScript to JPG I believe. Unless the newer software already exports to JPG or other bitmap graphic.
se
Simon,
😀
Put in some effort!
File
Export as (or some such)
Save as WMF GFX
Open in windows paint
Save as jpg
Open in Photoshop, shink to size (original is 3K pixels wide) , crop to taste and apply any "pen damping" needed... 😀
BTW, what I posted is not "pen-damped", in case anyone suspects...
Ciao T
😀
Put in some effort!
File
Export as (or some such)
Save as WMF GFX
Open in windows paint
Save as jpg
Open in Photoshop, shink to size (original is 3K pixels wide) , crop to taste and apply any "pen damping" needed... 😀
BTW, what I posted is not "pen-damped", in case anyone suspects...
Ciao T
Last edited:
Thanks, I do export as WMF and open in paint so I need to figure out where my color reversal is coming from.
Steve it is an AP2 no useful file outputs! WMF is no longer supported in the newer software for the AP2.
AP remains software primitive. The guys here complain because it can not be part of any test jig needing speed. They think the FFT is still written in BASIC 🙂.
Definitely off topic, but for those who also build speakers please note:
SawStop - The World's Leading Maker of Safe 10-inch Table Saws
This saw will absolutely prevent damage to fingers. It is now the number one selling table saw in the world. Why?
If you are a school or business and you don't have one of these and a student or employee hurts himself (even through absolute self-stupidity), you can be sued out of business.
If you are a hobbyist, even a small accident (eg, cutting the tip of the finger and ripping the nail off) will cost more in hospital bills than will the saw.
In the final analysis, it is far, far cheaper to buy this saw than to risk an accident.
In a way it is a shame, because it encourages lazy thinking and lazy practices. Perhaps they could wire it up to an electric fence generator that would give the operator a painful shock for making an error, rather than simply saving his finger or hand (at the cost of replacing the bladed and the stopping mechanism, a few hundred dollars -- maybe that is discouragement enough from making stupid errors!).
Most production houses use CNC, laser, or water jet cutting these days. Have they hit up Norm yet? 😀
You are kidding though, this thing triggers and it's 100% down, few businesses would tolerate this. I have yet to see one.
You are kidding though, this thing triggers and it's 100% down, few businesses would tolerate this. I have yet to see one.
Last edited:
SawStop:
"The blade carries a small electrical signal, which the safety system continually monitors.
When skin contacts the blade, the signal changes because the human body is conductive.
The change to the signal activates the safety system."
This product is Amazing! The hot dog demo is a must see.
(I love having all my fingers and limbs intact.)
"The blade carries a small electrical signal, which the safety system continually monitors.
When skin contacts the blade, the signal changes because the human body is conductive.
The change to the signal activates the safety system."
This product is Amazing! The hot dog demo is a must see.
(I love having all my fingers and limbs intact.)
Here it is: YouTube - SawStop Hot dog Video - Saw blade retracts within 5 milliseconds of accidental contact
This is Great
This is Great
Thank you for the kind words.
How are we to reconcile these two statements?
Is John deaf or deluded?
Or is he actually right about ICs? Everyone here remembers a couple of years ago when John was asking what the best sounding IC was. Then he spent months asking for tips on how to make them sound even better -- current load on the output stage, power supplies, feedback networks.
Either somebody with the true secret of good sound was holding out on John, or John is an idiot, or the high end community is telepathically linked in a delusional state (Stereophile said that the resulting product was OK, but nothing near the best available).
Or maybe ICs are inferior to discrete for some other reason -- bad sounding silicon resistors (unlikely, as two of our highest rated and most popular products use the AD844 and all of the ARC preamps use logarithmic electronic pots for volume controls), low bias currents (maybe), or some completely unknown factor.
Or maybe feedback sounds bad.
(I reject Putzey's hypothesis, put forward in the recent "Linear Audio" that says as long as there is 30 dB of feedback across the entire audio band that the sound will be superior. There are dozens of holes in his article, starting with his pronouncement that a circuit will sound better when it has a constant amount of feedback across the audio band. How in the hell did he come up with that one? Show me the mathematical basis for that claim. Clearly such a claim could only come from a listening test. So what gives him the right to choose which listening tests to accept, and which listening tests to discard? Rubbish. If his claim is correct, why did the amplifier that followed his principles the closest -- the Halcro -- languish in the market for years, then enjoy a year or so of glorious glory, and then fade back into near-total obscurity?)
Both of our products made with open loop AD844s are Class A rated in Stereophile, while John's products made with a far more expensive and far more modern IC can only manage a Class B rating. Is it because his circuit uses feedback? You tell me.
I still claim that I could easily help Scott develop an "audiophile approved" IC that would own the audio market in the way that their analog video decoders used to own the DVD player market (before HDMI). But even if Sony, Panasonic, Pioneer, JVC, LG, Samsung, and every high-end company in the world used it, it might still be a money loser.
Before the crash of 2008, National hired Mark Brasfield (founder of MSB digital) to help them develop audiophile ICs. They achieved some degree of success, but I don't think it was financial -- Mark was one of the first to go when the layoffs started.
Hi,
I read it in the original now.
Well, he say's "if you do this it hurts". All I can answer is "then don't do that".
Loading down a VAS stage is about the worst way to broadband it, UNLESS we explicitly want to avoid all forms of looped feedback.
If we want to reduce the open loop gain, why not (as even D. Self suggests) first degenerate the VAS a little and secondly introduce a real element to the reactive local looped feedback around the VAS (annoyingly, D. Self suggests this too).
For good measure, we may even take this loop (just like the miller feedbacl loop) from the output of the power buffer stage.
Heck, we can even have first a inner loop from the collector of the transistor cascoding our VAS transistor and a second from the output of the power buffer.
And we may go to great length to isolate the nasty signal modulated impedance of the power stage from the VAS stage (say a Jziklai buffer with CCS load) and to minimise the drive impedance to the VAS (maybe using a fast opposite polarity transistor, so we also negate the Vbe in the process and provide inherent overcurrent protection to the VAS).
Of course, such a "Linearised and broadbanded" amplifier will not allow for a lot of global negative feedback. But then again, it may not have much need of it in the first place, it may even become optional...
Incidentally, I have not build SS Amp's that way (yet?), but I have several Tube amps like that (and adding positive feedback inside the loop to maximise available local loop feedback) under my belt...
Despite having tons of NFB, those tubeamp's sounded rather good. Then again, all feedback loops where very short, no more than two stages (I consider the OPT a "stage")...
Ciao T
I read it in the original now.
Well, he say's "if you do this it hurts". All I can answer is "then don't do that".
Loading down a VAS stage is about the worst way to broadband it, UNLESS we explicitly want to avoid all forms of looped feedback.
If we want to reduce the open loop gain, why not (as even D. Self suggests) first degenerate the VAS a little and secondly introduce a real element to the reactive local looped feedback around the VAS (annoyingly, D. Self suggests this too).
For good measure, we may even take this loop (just like the miller feedbacl loop) from the output of the power buffer stage.
Heck, we can even have first a inner loop from the collector of the transistor cascoding our VAS transistor and a second from the output of the power buffer.
And we may go to great length to isolate the nasty signal modulated impedance of the power stage from the VAS stage (say a Jziklai buffer with CCS load) and to minimise the drive impedance to the VAS (maybe using a fast opposite polarity transistor, so we also negate the Vbe in the process and provide inherent overcurrent protection to the VAS).
Of course, such a "Linearised and broadbanded" amplifier will not allow for a lot of global negative feedback. But then again, it may not have much need of it in the first place, it may even become optional...
Incidentally, I have not build SS Amp's that way (yet?), but I have several Tube amps like that (and adding positive feedback inside the loop to maximise available local loop feedback) under my belt...
Despite having tons of NFB, those tubeamp's sounded rather good. Then again, all feedback loops where very short, no more than two stages (I consider the OPT a "stage")...
Ciao T
More OT,
Charles,
The mandated workmans compensation insurance greatly reduces the right for an injured employee to sue for these kinds of issues in return for a guaranteed income when injured. But the UL is now requiring table saw protection, so as that phases in then it becomes negligent not to have such safety features.
Of course negligence is mostly covered by workmans comp. (I did recently have an issue when the Construction Manager's contract required me to provide coverage for them if they were negligent and injured one of my workers even though neither I or any of my other employees had any involvement. That actually was covered under the old standard form!)
The Delta Uni-Saw has been the standard table saw of small woodshops. In addition to taking off thumbs, it has the bad habit of throwing work pieces if they twist between the blade and fence. As real wood has grain that often is not straight, just ripping a "Figured" piece of wood will do this.
The SawStop is a refinement of technology that has been around for a while. The earlier versions used to discharge a DC current from a capacitor into the induction motor to get it to stop. The SawStop uses a slug that is faster but requires tool replacement. But it only slightly improves throwing wood issues.
As the next step up in safety is the European Sliding table with blade cover and riving knife. However these are much more expensive.
CNC machines that are useful for production cost way more! My experience is that for a small shop the saw is the better investment.
Charles,
The mandated workmans compensation insurance greatly reduces the right for an injured employee to sue for these kinds of issues in return for a guaranteed income when injured. But the UL is now requiring table saw protection, so as that phases in then it becomes negligent not to have such safety features.
Of course negligence is mostly covered by workmans comp. (I did recently have an issue when the Construction Manager's contract required me to provide coverage for them if they were negligent and injured one of my workers even though neither I or any of my other employees had any involvement. That actually was covered under the old standard form!)
The Delta Uni-Saw has been the standard table saw of small woodshops. In addition to taking off thumbs, it has the bad habit of throwing work pieces if they twist between the blade and fence. As real wood has grain that often is not straight, just ripping a "Figured" piece of wood will do this.
The SawStop is a refinement of technology that has been around for a while. The earlier versions used to discharge a DC current from a capacitor into the induction motor to get it to stop. The SawStop uses a slug that is faster but requires tool replacement. But it only slightly improves throwing wood issues.
As the next step up in safety is the European Sliding table with blade cover and riving knife. However these are much more expensive.
CNC machines that are useful for production cost way more! My experience is that for a small shop the saw is the better investment.
Nice sentiments but LG and Samsung in particular make this a fantacy senario. The BOM on a $39 DVD has no room for it. You probably would not be able to show a single high volume socket in an asian consumer product that uses any one of those National amps. Analog video decoders???? you must be mistaken, we have never had a high runner in mainstream DVD business. Those were for computer video out, originally developed for Apple. Digital delivery of video was inevitable, but maybe you can explain why the YPrPb on my TIVO looks better (or at least as good as) the HDMI.
EDIT- BTW the AD712 probably remains high in the audio sales list. The "op-amps for audio" babble on many data sheets is just that, I think half of it is from all the engineers that envision themselves as the ultimate DIY'er. 😀
Last edited:
Sy,
Though shallst Google or Wiki...
YPbPr - Wikipedia, the free encyclopedia
Ciao T
PS, I am using VGA feed to my projector. 😛
Why? DVI (which is essentially HDMI with a more solid and robust plug) does not work for the length of the cable...
You might say that in this case the difference between analogue and digital is "night and day", but you would not be able to tell it in a "blind" test... 😉
Though shallst Google or Wiki...
YPbPr - Wikipedia, the free encyclopedia
Ciao T
PS, I am using VGA feed to my projector. 😛
Why? DVI (which is essentially HDMI with a more solid and robust plug) does not work for the length of the cable...
You might say that in this case the difference between analogue and digital is "night and day", but you would not be able to tell it in a "blind" test... 😉
Probably not and possibly. There are other possibilities- he's not great with IC design, he has a prejudice because of the demands of fashion in his market, or that there are indeed audible deficiencies intrinsic to all IC opamps that he's ever tried, which no-one has to date been able to demonstrate by ear alone. I gather that you favor the last possibility.
- Status
- Not open for further replies.