Indeed, which is better? Depends on what you want, really. I love that the better I get my system and room the more I can hear the differences in recordings both good and bad. But I gotta admit, that's a pretty geeky thing.
Nothing wrong with a fellow who just wants to be transported to the concert hall or rock show to enjoy the music as it might sound there.
Nigel, about the Radford power amp:
I knew the engineer who designed the Radford power amp. He and I were 'colleagues' in 1976, and I even stayed over at his place in Copenhagen, on my way to my month in Finland, working with Matti Otala, in 1976. He happened to have a schematic of his power amp tacked on a wall of his apartment that I saw. IT was so complicated, that the OUTPUT STAGE was about as complicated as one of my entire power amps, at the time. He was very proud of the low distortion, but I personally 'knew' that it would not sound very good, as the design was too complex. I never listened to this amp, so your opinion confirms my prediction. (Of course, since it 'failed' in the marketplace tells me the same thing.)
I knew the engineer who designed the Radford power amp. He and I were 'colleagues' in 1976, and I even stayed over at his place in Copenhagen, on my way to my month in Finland, working with Matti Otala, in 1976. He happened to have a schematic of his power amp tacked on a wall of his apartment that I saw. IT was so complicated, that the OUTPUT STAGE was about as complicated as one of my entire power amps, at the time. He was very proud of the low distortion, but I personally 'knew' that it would not sound very good, as the design was too complex. I never listened to this amp, so your opinion confirms my prediction. (Of course, since it 'failed' in the marketplace tells me the same thing.)
Actually John, it failed on the market primarily because of its rather high price in comparison with what it offered at the time, and in comparison with its competition. Even then, circa 1975, "features" were becoming mandatory.
And of course, as you say, if it failed to stun with its sound, it had nothing going for it.
Quite a few British amps of the day didn't fare too well against the concentraed onslaught of the Japanese jugernauts, like Sansui, Kenwood, Pioneer, etc. Frankly, the Brits were much more about the sound than their competition, but they were rather expensive for those days of ever cheaper watts and lacked the "features". Essential things like high and low filters with two turnover frequencies each, tone control turnover frequencies and so forth, all critical items.
The German audio industry, even though it was much more powerful than the British one, also failed for exactly the same reasons, in addition to the fact that they kept their power outputs ridiculously low for the time - 30 Watts continuous and 40 Watts of "music power" and into 4 Ohms were as exciting as a wet rag.
And they weren't the only ones. John, do you perhaps remember a nominally US company called Craig? Well, they made probably the sweetest sounding audio item I had ever heard until that time, a receiver rated at something like 50 or 60 WRMS into 8 Ohms. I spent something like 3 months trying to arrange for transport and import of it, I was that hell bent, and by the time I did it, they were out of business - they could not compete with the Japanese, even if their own products were "Designed in the U.S.A., manufactured in Japan".
I never really got over that, I feel sad to this day.
That's echos a sort of regret on my part . I would like to have the amp back to play with . The speakers I would like to try with it would be some of the better KEF's like 107 which had an attempt at a resistive load . I doubt it would have changed much , I might be wrong ( the sound was like the amp had seen a ghost ) . The other thing to try at the same time is to make it into a buffer . Then ask it to mimic another amp ( a simple triode would do ) . Quad spoke of this as always working " gentlemen if so you prefer distortion " was what they said . I would also like to do the 50 Quad 303 end to end test . It was an advert of theirs in the 1970's in the shape of a question mark . See if it did sound as good as they said , akin to a tape deck in noise levels they said . Many say the dislike of the Quad 405 is the protection circuits it used . PA friends love the Quad for it's protection circuits . I sold a bunch of 405 's from Browns Restaurant in Oxford . Later I was told why they sold them . The load was under 1 ohm . My guess is the staff put the system together . I had no idea of this when I sold them . All the amps worked fine and none came back . Browns liked it loud sometimes so they were fully " tested " . If anyone wants to build a better mini PA parallel 405-2 's are perfect . Put a 4K7 pot in the plus inputs with the wiper to signal . Put a cheap speaker across the two speaker plus outputs . Using music balance the pot for minimum sound . First thing you will notice is that you can not get it perfect . That argues against some versions of null tests . Copy the resistor values . Run the output terminals in parallel . This was an official Quad use . Then get your nasty speakers out and hear how much better it sounds ( was Gale 401 when we did it , his friend had a Lecson which inspired the extra 405 ). The 405 inverts so for better drum and transients invert the speakers . Don't say the recording engineer sometimes gets it wrong ( people always do ) . Wrong is right in that case . It was a choice even if it was not planned .
The Radford valve amps were of sensible complexity . If I remember correctly the phase splitter was a pentode - triode long tail pair . The argument put forward was that it was easier to drive and retained some triode characteristics ( triodes are easy to drive so I don't buy that , more gain I would guess ? hi fi reviewers ! ) . As everything was warped in feedback the distortion was low . The valve was like the Dynaco and RCA amps , the 7199 . However not the same pin arrangement . 7199 is said to be an improved TV tube , the Radford one a normal TV tube . I liked that Radford idea as it is as nutty as some of mine . 90 % of my nutty ideas end very badly . 10 % don't .
Talking of nutty ideas . Years ago a friend tried to tell me about attaching a battery to the output of am amp to make it sound better . My friend was very good at building standard designs . He was a bit shaky technically . I fixed his " broken " oscilloscope by putting the knobs in better positions once . Recently I had an inkling of what he meant . Perhaps he had made a inwards offset of the speaker cone ? On a PA speaker that might give better distortion as the coil would be held in the linear range longer ? The simple physics of the way the coil moves in the magnetic gap causes curve distortion like a triode . If I was to offer this it would be gradual as it would alarm the customer to see the cone jump on start up . It would be under the control of a PIC or similar to change the offset when program levels increased . I could have mode A and B . Maybe C to have more curve distortion ( positive offset ) . Offsets up to 1 or even 2 volts might be OK ? My little Russian Rigonda TV had an SE amp . It had special speakers that coped with the full offset . When I hooked up my Wharfedale Lintons they worked very well doing the same . I was 14 and starting to learn . My little brother of 9 explained how SE amps works if I remember correctly and the offset . The tone of his voice said he didn't approve . I left my electronics books around and he had absorbed them ! He had absorbed books on plumbing also ( my dads ) . That helped me in later life .
About speakers getting rid of walls and showing the mixes . When phase response is of a high order and all the other things like polar response one can hear layers in the music . Undoubtedly not intended . Chicago and Earth wind and Fire come to mind . Simply superb recordings .
Another test I would like to try is the SSL test . David Mate had SSL move away form NE5534 ( don't know to what ) . He said some time I should connect 90 x 5534 end to end as they did . He said it was surprisingly OK seeing how awful the concept was . SSL was not at that time noted as being the best sounding ( what a surprise ) . They mostly sold to TV companies who could work very fast on a SSL as it was designed for easy use ( time is money ) . As I said before the belief in 5534 was so strong that engineers went around replacing early OPA series with 5534 . I guess because it sounded different it was said to be better ? The Kings new suit of clothes ?
BTW . I would like to try KEF 107 with my little valve amp . I might get to a damping factor of 5 if I try very hard . I suspect it will be not too bad . Also like to do frequency response with it and 303 as a control . Would be in the back garden if it ever stops raining . We have a hosepipe ban !!!!!!!.
Last edited:
I am almost sorry that I commented on the Radford amp. I formed an opinion just from looking at the schematic, and this was a bit 'hasty'.
I wasn't hasty ( and I did understand your humour , it was a bit long , sorry ) . I was inexperienced and judgmental then . Whilst I don't totally believe in measurements that amp should have sounded better . I would have loved to have optimized it and done the buffer test . All my other references to other Radford products and their ex staff make it hard to understand . Looking more carefully I see nothing too unusual . It must have been my dreadful taste in speakers then or something .
A Paul Kemble web page - the Radford HD250 and ZD22 amplifiers.
A Paul Kemble web page - the Radford HD250 and ZD22 amplifiers.
As one who started buying hi FI gear in the early 70's the must have amplifiers were anything from Sansui , well that is if you could not afford Mc Intosh 🙂
Apart from quad amplifers , which were only good for quad speakers ,(prevailing ignorance) never thought of buying British , hmmm there was an leak amplifier I do remember looking at but went for a Sony receiver instead ..
Apart from quad amplifers , which were only good for quad speakers ,(prevailing ignorance) never thought of buying British , hmmm there was an leak amplifier I do remember looking at but went for a Sony receiver instead ..
Last edited:
It was really better...
I remember a test done by a french audio review by 1978 or so,
they measured the perfs of the most used op amps in audio gear,
the 5534/5532 was the best by a huge margin.
The datasheet say it all , even the so called modern op amps barely
reach its open loop gain above 1khz.
I have a tiny Wein-bridge signal generator which I have handy ( 2 caps , 1 x RA 53 thermistor , 680R bottom arm , now a dual op amp , 4 additional resistors , then rotary switch with parallel resistors for the other frequencies ) . It is about 0.05% distortion or better . Recently I needed a bit more output so converted it to a dual op amp ( 2K2 , 10 K inverting ) . The original was 5534 , the only half respectable dual I had to hand was a TLO72 . Up to 22 kHz no big difference . At 50 kHz unusable ( - 30 dB , perhaps because the circuit suited the original chip ) . The 5534 would be still be - 60 dB . I intend one day to build the Rosen circuit of Wireless World . I saw an update where it gets > -130 dB ( 134 ) . He was using OPA 604 I think of which I have a few . The Audio Precision is -143 dB I seem to remember . Even testing valve amps I think I need something that good . I looked at a Butterworth filter of 8 poles . That should get me - 30 dB at 2 nd harmonic . 3.6 MHz GBP required . That added to my little Wein-bridge might give > - 90 dB 10 kHz . Wein-bridge was a Hewlett Packard invention I think ? I was told because it was based on bridges of Vienna ( in print ) , a mans name I think is the real reason .
I am told a light bulb is as good as a RA 53 ( it is obsolete ) .
Mostly where I met 5534 was in not very good sounding equipment . The Hypex uses it and I am told it is preferred to other options tried . One of the LT range ( 0.9 nV and normal voltage rails , 40 MHz GBP , forget the number ) sounded very good to start with . As time went by it seemed a bit metallic . To be frank as doubtful as 5534 but in opposite ways . I think it had too much bandwidth ( I will try again as it was promising , a slight roll off next time ) ! I note LT still recommends a 1 mA CCS to -ve rail with it . Do people do that with op amps now ? I guess 5534 should be tried with it ? I have to say the only people who used 5534 that I never doubted were Rotel and Denon ( RA820 , PMA250 ) . Both budget priced and excellent . I noted the Japan version of PMA 250 had choke inputs to PU input . That's worth trying .
As I said before I would love a competition where the 5534 and 2N3055 ( 2955 ) are the key components ( BC550 / 560 / 2N5551/5401/ BD139/140 MPSA 42/92 ) . Transistor Chip Amp ( Spud - chip - potato ) . It is my belief that an improved Quad 303 and a bit of careful work in the pre amp could be made to sound better than much of the high end ( I have heard plenty ) . I would also give marks ( 20 % perhaps ) when a better op amp fitted . The one used would be the choice of the competition judges . It would be unknown at the start . The marks would be based on how well both did . I think speakers would be Quad 63's and Klipshe Heresy . A class A switch would be allowed . Although no price limit would be set it would be judged with a bias to inexpensive . For example multi cheap caps preferred to Auidiophile grade . 10 x 4700 uF being the upper limit ( 47 000 total ) . Not less than 6 caps ( to make PSU design difficult , none less than 2200 ) . I would use 8 myself split in two ( monobloc ) , 2 extra for drivers . 10 x 470 uF for other departments . all other caps below 100 uF . Rectifiers could be whatever you like ( I would hope to see EZ81 for divers / 5534 by someone ) . Chokes would be allowed if handmade and air cored . PCB dual or single layer . Transformer < 500 VA .
Last edited:
Wien (note spelling) conceived of his bridge in 1891. See
Wien bridge oscillator - Wikipedia, the free encyclopedia
The references are worthy of perusal, especially the late Jim Williams mods using Linear Technology parts.
The crowning achivement of Hofer's Audio Precision oscillators is particularly in making something that manages both very low distortion at the same time quite fast settling when changing frequencies. It is the subject of patent(s) which I've only heard about, or I'd give an explicit reference.
Other than satisfying your curiosity, what would be the point?
The NE chip was a quantum leap forward, no argument, but that was in 1978. This is 34 years later, Nige, as this time we have quaite a few op amps which will beat the NE hands down.
As just one of many examples, do tray Analog Devices' 829. A bit finnicky, but once you get the hang of it, it's miles ahead of the NE. It will NOT serve as a direct replacemens in already made cisruits every time, but with just a bit of design, it will work like you wouldn't believe.
Then the 2N3055/2955. Yes, they are even exceptionally robust, you need a hammer to damage them, but they are s-l-o-w by any criteria you care to use. If you want that type of sound, which I also happen to like, I suggest you look at Motorola/ON Semi's MJ 21195/21196. They are at least rated at 250W each, and will at least hit 4 MHz wasy, and in most case, quite a bit above that. The plastic pack version, sooooo much more convenient for mounting, bear the same numbers, but have an "L" added to the model type, so they're MJL... .
If you are really hell bent for a power amplifier which uses an op amp for its input stage, you should have told me, I can unload on you a fair number of such attempts, but bear in mind, I haven't tried any.
There is no doubt such an approach can achieve results very much in the true High End sound quality - I should know, my own Karan KA-i180 uses Burr-Brown 2604 op amps throughout with, as far as I have been able to determine over the last 8 years, no ill side effects.
Nevertheless, I still prefer all discrete, thank you. I thought you did too, you've posted a few times how you'd love to do an all discrete op amp ?!?
Nigel, you never cease to amaze me. I had a famous audio colleague who wanted a contest of making audio amps and preamps with just Radio Shack components. I politely declined. Why limit yourself? Why not TRY to make the best sounding power amp at the lowest price possible? Then you can use cheap, but IMPROVED devices, and not something designed over 50 years ago.
I have a recollection of the 2N3055, from more than 50 years ago. While in college, and forced to take (at the time) primary classes, like History, Chemistry, Math, etc. I took a night Adult course at night in Electronics. It was virtually ALL germanium, and it appeared that nothing better was going to be made, which made future improvements look incremental, at best. Then at the very end of the course, the instructor, said that the 2N3055 was invented, and maybe just maybe, quality audio was possible with these devices. Yet, I would not bother to race with a 34 Ford engine, even though it was a V8, and popular for many years. Would you?
Why not sponsor a contest to make the best audio amp with the active parts priced under $5.00? Can you do that?
I have a recollection of the 2N3055, from more than 50 years ago. While in college, and forced to take (at the time) primary classes, like History, Chemistry, Math, etc. I took a night Adult course at night in Electronics. It was virtually ALL germanium, and it appeared that nothing better was going to be made, which made future improvements look incremental, at best. Then at the very end of the course, the instructor, said that the 2N3055 was invented, and maybe just maybe, quality audio was possible with these devices. Yet, I would not bother to race with a 34 Ford engine, even though it was a V8, and popular for many years. Would you?
Why not sponsor a contest to make the best audio amp with the active parts priced under $5.00? Can you do that?
When it comes to the 5534, we have wrung it out so many times, it is silly! It WAS a breakthrough at the time, BUT it was NOT perfect. I know because I did comparison tests with it and a discrete design that I normally used.
However, we went further. We REPLACED the input devices with a jfet pair, by using the offset pins, and we made one of the BEST IC's available. It was so good, that Dave Wilson bought them from us to put into his WAMM Equalizer, for his $100,000 speaker (at the time). That is progress, rather than thinking that the 5534 is 'good enough'.
However, we went further. We REPLACED the input devices with a jfet pair, by using the offset pins, and we made one of the BEST IC's available. It was so good, that Dave Wilson bought them from us to put into his WAMM Equalizer, for his $100,000 speaker (at the time). That is progress, rather than thinking that the 5534 is 'good enough'.
But alas, now those specialized, well designed audio parts are history and we are left with high volume parts needed in cell phones, towers or iSomthings. Op amps are sure not the end all, but neither is full discrete for all applications. If I were to whip up a reasonable active DIY type crossover, am I really going to do significantly better than a modern op-amp when I am playing back a CD? Ins't the speaker still the biggest problem?
PS. I did not know Wilson used active eq. Not surprised, but the one time I heard them it was not mentioned.
PPS: Power outrages have prevented the next test cabinets construction verifying my wife's golden hearing is sensitive to the tweeter quality, and not the amp. Soon. Very soon. I got good feedback on the last test, but need to do the pair, fix some cabinet issues and a tad more final voicing in the midrange.
PS. I did not know Wilson used active eq. Not surprised, but the one time I heard them it was not mentioned.
PPS: Power outrages have prevented the next test cabinets construction verifying my wife's golden hearing is sensitive to the tweeter quality, and not the amp. Soon. Very soon. I got good feedback on the last test, but need to do the pair, fix some cabinet issues and a tad more final voicing in the midrange.
One student of Stanford University decided to build a networked computer long time ago from parts that can be bought in a store... Another student compiled Berkeley Unix for it. It become a foundation of Sun Microsystems...
Out of my own interest I cranked up my Google Fu and found a few interesting things. These go back to Tektronix and are state variable rather than Wien bridge, but the interesting part of this one is the "leveller" or amplitude detection circuitry. It sums several phases through diodes to give a smooth-looking peak envelope and effectively "sampling" the waveform several times per cycle:
Low distortion oscillator
4,560,958 "State variable oscillator having improved rejection of leveler-induced distortion" This takes the signal at the variable gain stage and combines it with an inverted output signal to partially cancel the distortion:
Patent US4560958 - State variable oscillator having improved rejection of leveler-induced ... - Google Patents
With "suitable modifications" these could conceivably be applied to Wien bridge oscillators.
An interview with Hofer: "He has received 13 patents,..."
How new products challenge audio test - 2010-06-01 04:00:00 | Test & Measurement World
How new products challenge audio test (continued) - 2010-06-01 04:00:00 | Test & Measurement World
Thanks for that research benb! One of the questions Hofer used to ask job applicants was to describe the overload issues with state variable filters, according to Cal Perkins.
@John Curl
John, you really need to go out of California more. 😛
Did you know that some of the most liked British brands, such as say Arcam, continued to build products with 2N3055/2955 all the way into the 90ies?
If you were to somehow make all 2N3055s disappear, you would deal a very possibly crippling blow to the German electronics industry today.
If you think I'm overdoing it, pick up the last say three issues of Europe's largest, multiningual popular electronics magazine "Elektor", and be amazed at how many circuits are based on 2N3055. Some whacky, but some really good ones as well.
This should surprise no-one. It's been around longer than most electrical engineers today, it has gone through so many variations that by now, it's one of the few semiconductors we really know all about. It's dirt cheap, and you can probably buy them in the depths of the Amazon jungle, the crocs are selling it as a sideline.
Word for word, ditto for the NE op amps.
The thing is, so many people learnt their electronics, in colleges or at home, using them simply because they are so readily available, they are cheap and rather versatile. And, let's be honest about it, they are reliable. Sturdy old workhorses.
I have to admit that over the years, I have heard a few low power amps based on the 2N3055/2955 which did sound very, very decent, while not my idea of audio perfection, no obvious pitfalls.
Also, do not forget the realities of this world, even if they are (happily) the past now. Not too long ago, this was a very much divided world, and not everybody could follow up on the latest technology, and even if they did, they could not lay their hands on it. Even if in 1966 my dad quite legally bought a German Ford car in the old Yugoslavia from the importer, even if getting a passport was a 7 day routine here, even so we couldn't buy everything we knew existed. A part of the reason was the US ban on exports of electronics to what they saw as "the other guys".
I bought my first IBM PC in 1984 - but to get it, I had to travel to London, buy it there and ship it back home because its export was banned by the US government.
Motorola had a representative office in Belgrade it seems like forever, but the stock they sold was VERY limited, mostly to their takes on European transistors, also sold by everybody else's representative offices. MJE 150xx power transistors? You gotta be kidding! MJL 3281/1302? They were introduced 10, maybe 12 years ago, and I managed to get my first batch just a few days ago.
So, under these conditions, do you still wonder why the 2N3055 is so very much alive and well? Even the local semiconductor manufacturer made them under licence. Generations were weaned on it. I haven't actually counted, but the local shops are still full of projects based on them, must be at least 20 versions of this or that.
Have you ever wondered why I shy away from Japanese transistors, which are total routine for you? It's because my choice of them is severely limited even today, where was I supposed to buy them, try them and make my choice?
Quite simply, people tend to buy and use what they are well acquainted with and what is available to them. This is, of course, a moving target. For example, for the last 5-7 years, we have seen a steady supply of Toshiba's 2SC5200/2SA1943 power transistors locally, and not unexpectedly, most new projects tend to use these and similar devices from other sources.
I don't know how you work, but I like to try out my first sample of whatever using the cheapest available parts, such as capacitors. I use Italian made A/V caps, the standard fare, which is pure and simple junk, but it's cheap, so if it blows, who cares? My reasoning is that if it works reasonably well with such parts, it will work a whole lot better with decent components. And most of the time, it is indeed so.
John, you really need to go out of California more. 😛
Did you know that some of the most liked British brands, such as say Arcam, continued to build products with 2N3055/2955 all the way into the 90ies?
If you were to somehow make all 2N3055s disappear, you would deal a very possibly crippling blow to the German electronics industry today.
If you think I'm overdoing it, pick up the last say three issues of Europe's largest, multiningual popular electronics magazine "Elektor", and be amazed at how many circuits are based on 2N3055. Some whacky, but some really good ones as well.
This should surprise no-one. It's been around longer than most electrical engineers today, it has gone through so many variations that by now, it's one of the few semiconductors we really know all about. It's dirt cheap, and you can probably buy them in the depths of the Amazon jungle, the crocs are selling it as a sideline.
Word for word, ditto for the NE op amps.
The thing is, so many people learnt their electronics, in colleges or at home, using them simply because they are so readily available, they are cheap and rather versatile. And, let's be honest about it, they are reliable. Sturdy old workhorses.
I have to admit that over the years, I have heard a few low power amps based on the 2N3055/2955 which did sound very, very decent, while not my idea of audio perfection, no obvious pitfalls.
Also, do not forget the realities of this world, even if they are (happily) the past now. Not too long ago, this was a very much divided world, and not everybody could follow up on the latest technology, and even if they did, they could not lay their hands on it. Even if in 1966 my dad quite legally bought a German Ford car in the old Yugoslavia from the importer, even if getting a passport was a 7 day routine here, even so we couldn't buy everything we knew existed. A part of the reason was the US ban on exports of electronics to what they saw as "the other guys".
I bought my first IBM PC in 1984 - but to get it, I had to travel to London, buy it there and ship it back home because its export was banned by the US government.
Motorola had a representative office in Belgrade it seems like forever, but the stock they sold was VERY limited, mostly to their takes on European transistors, also sold by everybody else's representative offices. MJE 150xx power transistors? You gotta be kidding! MJL 3281/1302? They were introduced 10, maybe 12 years ago, and I managed to get my first batch just a few days ago.
So, under these conditions, do you still wonder why the 2N3055 is so very much alive and well? Even the local semiconductor manufacturer made them under licence. Generations were weaned on it. I haven't actually counted, but the local shops are still full of projects based on them, must be at least 20 versions of this or that.
Have you ever wondered why I shy away from Japanese transistors, which are total routine for you? It's because my choice of them is severely limited even today, where was I supposed to buy them, try them and make my choice?
Quite simply, people tend to buy and use what they are well acquainted with and what is available to them. This is, of course, a moving target. For example, for the last 5-7 years, we have seen a steady supply of Toshiba's 2SC5200/2SA1943 power transistors locally, and not unexpectedly, most new projects tend to use these and similar devices from other sources.
I don't know how you work, but I like to try out my first sample of whatever using the cheapest available parts, such as capacitors. I use Italian made A/V caps, the standard fare, which is pure and simple junk, but it's cheap, so if it blows, who cares? My reasoning is that if it works reasonably well with such parts, it will work a whole lot better with decent components. And most of the time, it is indeed so.
- Status
- Not open for further replies.