Nige, while I would never dispute the need for as quiet and noise free as possible, everywhere and anytime, I do think a lot of this debate is, for one reason or another, relevant to line electronics, and especially power amps.
After all, while noise is an issue, it is only of the issues which we need to weigh against other issues. Noise we can deal with by careful selection of our input transistors. I find, for example, that using 2SC2240, or 2N5551, or BC546B (better yet C when I can find it, higher gain) usually takes care of noise issues well enough without compromising other issues.
After all, while noise is an issue, it is only of the issues which we need to weigh against other issues. Noise we can deal with by careful selection of our input transistors. I find, for example, that using 2SC2240, or 2N5551, or BC546B (better yet C when I can find it, higher gain) usually takes care of noise issues well enough without compromising other issues.
To a certain extent it's also about how much noise is in the given source component. When I was doing class D designs for a startup I got the amplifier noise down to extraordinarily low levels, of order 30uV rms in the audio band at the output for a ~100W into 4 ohm product. It turned out the best digital source we could find at the time was considerably noisier than the amplifier.
Which should still make you VERY happy, Brad, because you achieved the goal of the amp not being the bottleneck.
I was thinking of oddities which can happen. To use an extreme example, say you find that your input differential pair behaves ideally with a bias current of say 0.5 mA per tranny, but the VAS stage is Miller compensated with say 47 pF. This would reduce your slew rate to just over 10V/uS. Depending on your point of view, this may be enough, but personally I'd still like to see say 40 V/uS as the working value, not because I can mathematically prove we need that much, but because in my experience faster amps tend to sound more coherent than slower ones.
So, this may cause you to investigate why do you need 47 pF and not say 22 pF, and that leads to examining the whole circuit, the layout and the very PCB. Some changes may be required, but neverthess, it would still require you to get some more bias in, say 1 mA. This is double the "ideal" value, so now you have re-examine the noise figure.
That sort of thing.
I was thinking of oddities which can happen. To use an extreme example, say you find that your input differential pair behaves ideally with a bias current of say 0.5 mA per tranny, but the VAS stage is Miller compensated with say 47 pF. This would reduce your slew rate to just over 10V/uS. Depending on your point of view, this may be enough, but personally I'd still like to see say 40 V/uS as the working value, not because I can mathematically prove we need that much, but because in my experience faster amps tend to sound more coherent than slower ones.
So, this may cause you to investigate why do you need 47 pF and not say 22 pF, and that leads to examining the whole circuit, the layout and the very PCB. Some changes may be required, but neverthess, it would still require you to get some more bias in, say 1 mA. This is double the "ideal" value, so now you have re-examine the noise figure.
That sort of thing.
Last edited:
In valve design there would be little point as the op amp style valve design is not universally liked ( read disliked ) . In transistors it is and makes sense to do one fits all . The good old 2SA1085 seems almost a good universal input device . 2N4403 on steroids . It's high gain means it might be possible to run 3 deceives coupled to a MC pick up and not exceed a sensible current flowing through the PU. As I think Stan Curtice said most caps leak more that that ( Naim type circuits ) .
On that . Seeing as no voltage swing is required at TR1 . What is the ideal discrete op amp voltage for MC gain 1000 ( pre 75 uS and further 3180 / 318 gain 16 @ 1kHz ) . +/- 6V seems to offer 50 db overload margin , stage two if op amp is less . That is I need 15 mV from stage 1 max . I would be using a VAS CCS .
For years I have used 2SA970 PNP ( 872A 1085 and 716 ) . 120 V , very low noise and high gain . Soon I will be forced to use something different . I notice in the book ( D Self ) that the input transistors are in the nothing special category . 2N5401 comes to mind as a possible substitute if nothing special is OK . I would like to have a >100 V capability if possible as sometimes I build high voltage amps that resemble hi fi use . LM4702 will do that for that , prefer it is my own design . I have seen MPSA 92 used ( HH ) . Any thoughts ?
Noise figure . 8 dB ? OK ?
2N5401G - ON SEMICONDUCTOR - TRANSISTOR, PNP, 150V, 0.6MA, | Farnell United Kingdom
Posted this question in another thread ( D Self book ) .
Noise figure . 8 dB ? OK ?
2N5401G - ON SEMICONDUCTOR - TRANSISTOR, PNP, 150V, 0.6MA, | Farnell United Kingdom
Posted this question in another thread ( D Self book ) .
Years ago I was sent a LM3900N Norton amplifier quad package that keep appearing in it's little coffin in the loft . Noticing that TL074 is now cheaper than LM 324 I saw this one also and asked myself what if . Looking at applications some audio uses are stated . A mixer and sine wave oscillator of the SVF or similar type . In the thread about Douglas Self's new book current feedback was being lorded on the thread to the extent it got a bit heated . I was of the impression LM3900N was for making audio compressors of about 1% THD so never gave it a second look thinking the distortion was inherent . Is the Norton ( op ) amp worth a listen ?
http://www.mouser.com/ds/2/405/slos059-122258.pdf
The LM3900N quad "current-differencing amplifier" is only $0.34 each in qty 10.
The LM3900N quad "current-differencing amplifier" is only $0.34 each in qty 10.
I remember reading Jim Thompson saying how disappointed they were when LM324 didn't give more performance . I have grown to love LM 324 over the years . Little things it can do better than most . I was looking at the Linn Valhalla turntable oscillator + power amp . The LM324 is possibly the best that could be fitted regardless of cost . LM324 is an op amp which might be used in audio at a pinch .
LM 324 = Work horse , tough . Great where reverse voltages happen ( cars ).
TL074 = Seemingly able to do most things especially where high input inpedance would be useful .
MC 33079 . LM 324 with true hi fi potenial .
Would love ideas for reasonably priced quad packages . OPA 604 in quad , does it exist ?
I like quads most because I often build dead bug circuits . Apart from + and - all the pis are obvious which helps when testing .
If anyone has a useful LM 3900 circuit I would love to learn more .
Come to that LM 339 looks like is could work as an audio device . It has an open collector which can do 50 V . It is shouting VAS . Shame there is no comp pin . If anyone needs an LM 339 NOR just ask . I used the spare device to make a relaxation oscillator to substitute the NE555 I was going to use . This was fed to a 4017 . This was with two cheap chips making a sequencer and under voltage switch off ( NOR ) . Diodes used to lock the 4017 with the relays in bucket brigade . The final output locks the clock . To use all the LM 339 is how I like to do things . The relays via ULN 2804 . If you have never used CMOS as drivers the maximum current is at the stated output voltage . If you ask them to do more they drop the voltage . They often will be happy to do that . Alas no data given and no guarantees . This is useful to know as the HC series is limited to 6 V , the general 4000 series I usually assume to be 15 V ( < 18 ) . A 74HC4060 at 5 V will run a white LED without any series resistor with no harm to itself or 4060 . Add 10R to be sure . That's the lighthouse on the model railroad sorted . A watch crystal would give 2 second internals I think . Point is this is a general quality of CMOS ( Ron current limiter ) . Mike Balance ( real name , ex Rotel and Hi Fi World ) and I were going to put in an RIAA circuit using a CMOS Hex inverter in World . It was to be an April Fools joke the wasn't . Hex inverters can be converted into crude op amps with inverting gain ( usually 10 ) . You simply hook them up as an op amp and bias the input as if a single rail op amp . Never tried a +/- supply . I guess it should work . If I get a moment I should do it for fun . Why not 3180 + 318 + 75 uS in 3 sections . Crossover distortion noise ? The devices are very fast and might just cope . Shure M44 - 7 of 9.5 mV to be used .
Shure M44 -7
English . High output pick up cartridge .
Spanish . High output pick up cartridge .
Germany . Pick up cartridge with 9.5 mV output !!!!
LM 324 = Work horse , tough . Great where reverse voltages happen ( cars ).
TL074 = Seemingly able to do most things especially where high input inpedance would be useful .
MC 33079 . LM 324 with true hi fi potenial .
Would love ideas for reasonably priced quad packages . OPA 604 in quad , does it exist ?
I like quads most because I often build dead bug circuits . Apart from + and - all the pis are obvious which helps when testing .
If anyone has a useful LM 3900 circuit I would love to learn more .
Come to that LM 339 looks like is could work as an audio device . It has an open collector which can do 50 V . It is shouting VAS . Shame there is no comp pin . If anyone needs an LM 339 NOR just ask . I used the spare device to make a relaxation oscillator to substitute the NE555 I was going to use . This was fed to a 4017 . This was with two cheap chips making a sequencer and under voltage switch off ( NOR ) . Diodes used to lock the 4017 with the relays in bucket brigade . The final output locks the clock . To use all the LM 339 is how I like to do things . The relays via ULN 2804 . If you have never used CMOS as drivers the maximum current is at the stated output voltage . If you ask them to do more they drop the voltage . They often will be happy to do that . Alas no data given and no guarantees . This is useful to know as the HC series is limited to 6 V , the general 4000 series I usually assume to be 15 V ( < 18 ) . A 74HC4060 at 5 V will run a white LED without any series resistor with no harm to itself or 4060 . Add 10R to be sure . That's the lighthouse on the model railroad sorted . A watch crystal would give 2 second internals I think . Point is this is a general quality of CMOS ( Ron current limiter ) . Mike Balance ( real name , ex Rotel and Hi Fi World ) and I were going to put in an RIAA circuit using a CMOS Hex inverter in World . It was to be an April Fools joke the wasn't . Hex inverters can be converted into crude op amps with inverting gain ( usually 10 ) . You simply hook them up as an op amp and bias the input as if a single rail op amp . Never tried a +/- supply . I guess it should work . If I get a moment I should do it for fun . Why not 3180 + 318 + 75 uS in 3 sections . Crossover distortion noise ? The devices are very fast and might just cope . Shure M44 - 7 of 9.5 mV to be used .
Shure M44 -7
English . High output pick up cartridge .
Spanish . High output pick up cartridge .
Germany . Pick up cartridge with 9.5 mV output !!!!
Last edited:
Once in a while the wrong thing works . This was an April fools joke . It would have appealed to some as it almost would have been tubes . We would have made certain it matched a good cassette deck ( noise , distortion ) with the Shure before publishing . I missed the deadline alas . I got a wonderful review of my Percy Wilson designed record cleaning machine . Demonstrated to Mike in my Golf using a cheap inverter from the cigar socket . Mike reviewed it and bought it !! Twas whilst showing him that the April fools idea was hatched . Me met in the car park where Inspector Morse was filmed in Bicester ( Bus Stop ) . The Half way point for us both . I have a great admiration for Rotel , Mike was a very good engineer . I know enough to fill a book . Best say they punch well above their weight . Like many companies the basic models were the better designs . Mike was the first person I knew of who produced a fully balanced amp ( NPN + PNP , LTP ) . This became the special high end Rotel I beleive ?
As you can see my job is loosely connected with similar engineering . It is impossible not to ask these question when part of a design already progressing . I must try a comparator as a power some time . That is bootstrapped collector load and a pair of Darligtons at 50 V . Might even throw in 3 diodes and a resistor for good measure .
As you can see my job is loosely connected with similar engineering . It is impossible not to ask these question when part of a design already progressing . I must try a comparator as a power some time . That is bootstrapped collector load and a pair of Darligtons at 50 V . Might even throw in 3 diodes and a resistor for good measure .
Just listening to the last night of the proms . Quad FM 3 easilly beats digital radio which is ironiic as FM has been digital linked since the early 1970's . From memory the standard was better years ago ( the organ more audible ) . I assume the transmission might be a universal system now and the FM is being fed via a high grade DAC via a DAB signal? I have no idea If Nicam is still used ? The FM 3 has more space . Such distortion as it has sounds like woodiness and goes unnoticed . The FM 3 has good distortion up to 50% modulation . In practice this is fine as the BBC seldom use full modulation . The joke often made BBC Radio 3 is 75% silent ( ??? ) . This is true . Often I switch on to nothing . 30 seconds later a note or two or perhaps the hashed tones of the undertaker giving the most esoteric of information .
DAB subtractive distortion , yuk . FM additive , not as bad . Analogy . TV which is over saturated is additive , under-saturated is subtractive .
Radio 3 is now playing Jeremy Dale Thomas , Winter Music ( ? if I heard correctly ) . I almost like it , somewhere between Hindemith and Stockhausen . Beautiful flute playing . Not the proms as you might guess . Albert Hall , not my cup of tea .
Norton amplifiers . This one has OK noise . Any applications apart from compressors ?
LM359 Datasheet pdf - Dual, High Speed, Programmable, Current Mode (Norton) Amplifiers - National Semiconductor
DAB subtractive distortion , yuk . FM additive , not as bad . Analogy . TV which is over saturated is additive , under-saturated is subtractive .
Radio 3 is now playing Jeremy Dale Thomas , Winter Music ( ? if I heard correctly ) . I almost like it , somewhere between Hindemith and Stockhausen . Beautiful flute playing . Not the proms as you might guess . Albert Hall , not my cup of tea .
Norton amplifiers . This one has OK noise . Any applications apart from compressors ?
LM359 Datasheet pdf - Dual, High Speed, Programmable, Current Mode (Norton) Amplifiers - National Semiconductor
Last edited:
Yes, it's interesting that normal, kitchen quality digital radio is pretty terrible to listen to - it gets something fundamentally wrong, which makes it very unpleasant to listen to on an ongoing basis.
Which is the quandary with 'digital' sound in general ... if you're going to do it, do it right!! - otherwise it's worse, subjectively, than pretty dreadful analogue ...
Which is the quandary with 'digital' sound in general ... if you're going to do it, do it right!! - otherwise it's worse, subjectively, than pretty dreadful analogue ...
My audiophile friend Paul R. certainly agrees with you on that. Better o.k. analogue than mediocre digital (he has some of the best of both).
Welcome to my world, gents.
I tried a few DAB tuners, dismissed the idea and spent time and money refurbishing my stable of analog tuners. My favorite, reVox B 760 (5 gang varactor), will wipe the floor with any DAB tuner I have heard to date, but also my No.2 tuner, Philips 22AH180 (6 gang varactor) and my No.3 tuner, Technics SG 570 (5 gang varactor).
Admittedly, all of the three have the usual specs (image, suprious, and IF rejection) above or well above the 100 dB mark, which makes them rather up market models even in their day, and all three are digitally synthesized (178 for the reVox, 1982 for the Philips and 1992 for the Technics).
I am one hard core radio man. When I was 6, in 1959, I used to get up at 6 AM on Sundays, go do my mother's shopping list, and be back in front of the radio by 7:30, when the Children's Theatre was on. Without fail, in hot summers and neck high snow winters.
In 1967-1970, the Brits sealed my fate. That was the day of many, many radio stations, all of them trying hard to woo in the listebers. The years of the DJ. When Ludovic Kennedy was emerging big time. A sweet memory I am very fond of, those 3 years.
In 1969, I visited the USA, specifically NYC, for the first time. The 3 weeks I spent there, with my portable "Belson" (Made in Hong Kong) transistor radio, introduced me to some really good NYC radio stations. Cool sounds. Distinctly different from UK, but great to listen to. That was when I heard Waylon Jennings for the first time, and today, I still listen to his complete works, which I own with considerable pride. And Cash. And Willy.
Today, at 60, 54 years or half a century later, when I get up in the morning, I make my bed, shower, make my mug of morning coffee, light up some tobacco product, and with the greatest of pleasures, turn my radio (today, tuner + preamp + power amp) on. Unless I have to go anywhere for longer than 2 hours, it will stay on until 11-12 PM non stop.
To be fair, I must admit I still have several FM stations which do give a damn about sound quality, so I get not-quite-but-almost CD quality of sound. Most others are down to PC/MP3 programming.
I tried a few DAB tuners, dismissed the idea and spent time and money refurbishing my stable of analog tuners. My favorite, reVox B 760 (5 gang varactor), will wipe the floor with any DAB tuner I have heard to date, but also my No.2 tuner, Philips 22AH180 (6 gang varactor) and my No.3 tuner, Technics SG 570 (5 gang varactor).
Admittedly, all of the three have the usual specs (image, suprious, and IF rejection) above or well above the 100 dB mark, which makes them rather up market models even in their day, and all three are digitally synthesized (178 for the reVox, 1982 for the Philips and 1992 for the Technics).
I am one hard core radio man. When I was 6, in 1959, I used to get up at 6 AM on Sundays, go do my mother's shopping list, and be back in front of the radio by 7:30, when the Children's Theatre was on. Without fail, in hot summers and neck high snow winters.
In 1967-1970, the Brits sealed my fate. That was the day of many, many radio stations, all of them trying hard to woo in the listebers. The years of the DJ. When Ludovic Kennedy was emerging big time. A sweet memory I am very fond of, those 3 years.
In 1969, I visited the USA, specifically NYC, for the first time. The 3 weeks I spent there, with my portable "Belson" (Made in Hong Kong) transistor radio, introduced me to some really good NYC radio stations. Cool sounds. Distinctly different from UK, but great to listen to. That was when I heard Waylon Jennings for the first time, and today, I still listen to his complete works, which I own with considerable pride. And Cash. And Willy.
Today, at 60, 54 years or half a century later, when I get up in the morning, I make my bed, shower, make my mug of morning coffee, light up some tobacco product, and with the greatest of pleasures, turn my radio (today, tuner + preamp + power amp) on. Unless I have to go anywhere for longer than 2 hours, it will stay on until 11-12 PM non stop.
To be fair, I must admit I still have several FM stations which do give a damn about sound quality, so I get not-quite-but-almost CD quality of sound. Most others are down to PC/MP3 programming.
Last edited:
Damn right, Brad, the man has a strong point. It takes just a short listening session to prove him right.
in one customer product we used other manufacturer of same opamp, but one got phaseinversion when input was bit higher, other had not.
in general not (always) good idea to play with opamps as (speed) comparators
The problem with OK ditital is like big TV . More detail required . I have found two qualities that good hi fi can show with what I will call MP3 for want of a better tag ( all OK quality digital ) . These are layers and tonal purity . Then one is left with an opaque but generally pleasant sound which suggests more than is there , that is allowing imagination to take over . Bad digital gives edgy and sterile qualities . The off switch is the greatest upgrade for that .
Hex inverter op amps ( 74 HCU 04 is it ? ) . Can amplify into many MHz . Could be useful . One guy said made a tube sound for a friend , I doubt it .
Norton amps and comparators also . One guy said if LM339 - is taken to ground with 47 pF ( nF ? ) it will work . Must have a play with that . I want to drive 320 V via a Cascode , it might work .
- Status
- Not open for further replies.