I have now listened enough to say that this stage does not sound sharp ore spitty in any way. Quite the oposite. It is a smooth ride that reveils a lot of details, even at the low volume i am listening too. I get more and more the impression that many feedback stages that do not sound right have a problem with non ideal compensation. When i designed this stage i really made sure it works well in the time domain. I used a 10kHz squarewave for this purpose and i found interesting things. More to come.
The description of real music like smoothness, and complete mid bass presentation in low volume listening, rings familiar from my open loop. So you must have succeeded to come near to a non problematic loop. You made a single ended, high gain noise measuring instrumentation amplifier with EQ in essence. A difficult task.
Hi Ingemar ! The Topology goes further back then the HPS 5.1. First there was the Self MC pre-pre back in the 90th. Then there was the Colins circuit in AudioXpress that uses a Fet.
Around the same time there was Burkhard Vogels BUVO that had a RIAA wraped around the input stage, again with bipolar input. Mr. Vogel split the gain though. His input stage has 47dB of gain and then he adds an adjustable linear gain stage. I also published a circuit in Linear Audio Volume 0 that has that kind of gain cell but set up as a symmetrical INA. Looking even further back Mr.Self must have gotten his idea somewhere and i found his basic topology in "Halbleiter Schaltungstechnik, Tietze-Schenk". I have only the latest issue 13. The basic circuit is on page 497. How old that idea is i can not say but Vogel also refers to Tietze-Schenk so i asume it has been there for many years. There is no new thing under the sun ( King Salomon ). What i claim is novel with my circuit is that i do the amplification and EQ in one stage. That saves me 2 gain stages compared to the HPS5.1 and one gain stage compared to the BUVO.
i also think that i have not lost quality in the process. There is around 77dB of feedback availlable at 20kHz so distortion should be superbly low. The Fets i use are optimised for Audio and have a very high Early Voltage for example plus low 1/F frequency plus low GR noise. I have other details in the circuit that give excellent PSU rejection and i drive the Fets at a certain operation points that avoids heating and gate current. I have published cascoded Fet stages here aplenty since the beginning of my thread 19 month ago. I must admit that i learned a lot about noise from Synn08 and Scott Wurzer, although it was not always a smooth ride.
Around the same time there was Burkhard Vogels BUVO that had a RIAA wraped around the input stage, again with bipolar input. Mr. Vogel split the gain though. His input stage has 47dB of gain and then he adds an adjustable linear gain stage. I also published a circuit in Linear Audio Volume 0 that has that kind of gain cell but set up as a symmetrical INA. Looking even further back Mr.Self must have gotten his idea somewhere and i found his basic topology in "Halbleiter Schaltungstechnik, Tietze-Schenk". I have only the latest issue 13. The basic circuit is on page 497. How old that idea is i can not say but Vogel also refers to Tietze-Schenk so i asume it has been there for many years. There is no new thing under the sun ( King Salomon ). What i claim is novel with my circuit is that i do the amplification and EQ in one stage. That saves me 2 gain stages compared to the HPS5.1 and one gain stage compared to the BUVO.
i also think that i have not lost quality in the process. There is around 77dB of feedback availlable at 20kHz so distortion should be superbly low. The Fets i use are optimised for Audio and have a very high Early Voltage for example plus low 1/F frequency plus low GR noise. I have other details in the circuit that give excellent PSU rejection and i drive the Fets at a certain operation points that avoids heating and gate current. I have published cascoded Fet stages here aplenty since the beginning of my thread 19 month ago. I must admit that i learned a lot about noise from Synn08 and Scott Wurzer, although it was not always a smooth ride.
A builder of 5.1 experienced VLF transients. Syn08 explained him why:http://www.diyaudio.com/forums/analogue-source/167360-hps-5-1-a-13.html
It seemed endemic for such high VLF gain and an integrator servo by the explanation. Did you get any? Or you overcome that by utilizing less stages or due to another reason? Because you mentioned its destined commercial, no market would accept any transients for any time no matter how sort or start up. Dealers would perceive the phono as simply faulty. So its a crucial factor to check well it never occurs in yours, or add output delay relay, since it has been documented in an alike configuration phono. Just remembered that issue (non issue for DIY), so I thought I ring a bell. Sorry if you have covered that already.
It seemed endemic for such high VLF gain and an integrator servo by the explanation. Did you get any? Or you overcome that by utilizing less stages or due to another reason? Because you mentioned its destined commercial, no market would accept any transients for any time no matter how sort or start up. Dealers would perceive the phono as simply faulty. So its a crucial factor to check well it never occurs in yours, or add output delay relay, since it has been documented in an alike configuration phono. Just remembered that issue (non issue for DIY), so I thought I ring a bell. Sorry if you have covered that already.
Hi Salas ! I know problems like that and i call them motorboating. One stage i made had that problem. It was an earlier atempt of a Open Loop RIAA where i made a bias servo.
When i was cranking the system full i could see the woofer cones moving in and out at small amplitude with a low frequency.
No, the stage presented here does not have that problem but to stabilise the DC conditions
was not easy. A particular problem was the constant current source and i tried out at least 6 varieties from a simple Widlar current mirror to a double cascoded bandgap reference. The trick is to find a constant current source that has a high output impedance but is temperature stable in combination with the input stage. At an impedance of 50kOhm only 20nA produce an offset of 1V.
If this stage makes a nasty switch on transient i have to check. In commercial preamps i use a safety cap and a relay somewhere anyway but you are absolutely right. It is imposible to sell something on the High End market that is not functioning perfectly.
When i was cranking the system full i could see the woofer cones moving in and out at small amplitude with a low frequency.
No, the stage presented here does not have that problem but to stabilise the DC conditions
was not easy. A particular problem was the constant current source and i tried out at least 6 varieties from a simple Widlar current mirror to a double cascoded bandgap reference. The trick is to find a constant current source that has a high output impedance but is temperature stable in combination with the input stage. At an impedance of 50kOhm only 20nA produce an offset of 1V.
If this stage makes a nasty switch on transient i have to check. In commercial preamps i use a safety cap and a relay somewhere anyway but you are absolutely right. It is imposible to sell something on the High End market that is not functioning perfectly.
Those were the kind to look for with instruments during start up: http://www.diyaudio.com/forums/analogue-source/167360-hps-5-1-a-12.html#post2333029 The builder describes them as much alarming so you would have noticed if there was some already I believe. About the CCS on the particular application and type input cascode, yes it normally ''fights'' with it if not very special to it. That is why I prefer a relatively high resistor and high B+ clean PSU than CCS complexity and repeatability issues. Each PSU lends sonics of its own beyond noise susceptibility of stages no matter how good the PSRR in my limited experience, still the PSRR on noise has a highly measurable impact. I would go in lengths for a better PSU every time, even if it was only for feeding DC to top PSRR OP-AMPS, for its general impact beyond straight noise issues.
I think that builder could try reducing the 1kOhm resistor behind the servo but it is not my circuit so i hope the designer can help. The simple resistor solution is just fine but you also amplify the noise from the PSU that way. You also can not make the resistor value very high especially when you run the input stage on high bias. One way whould be to use a very high PSU voltage but i think this is not very elegant. In my circuit it whould not work. I really need the gain increase by presenting the input stage a high dynamic impedance. Anyway, Salas your PSU´s are excellent but i will not design more open loop phonos any time soon. Having gone thruw over hundred topologies including their varieties over the last 2 years i think i have covered the basics.
I wasn't proposing to adopt a load resistor since you already have arrived to a very well locking CCS for it. Just discussing other more traditional ways, that put the toll on the clean B+ of course, and increase care there. I mainly meant to say that I would make such an effort anyway for the PSU no matter the PSRR, because I feel it still makes its presence felt in different proportions in all circuits. But open loop and a few FETS VS hundreds of integrated transistors taking care of all problems in chips and active EQ loops are different beasts, maybe I am wrong, and would be much more impervious to PSUs. That 5.1 issue had been answered to the builder by the designer in the previous link. Those transient shapes I linked just to know the type they discussed, for reference when you will look for start up behaviour with instruments as you mentioned. I don't think they are there in yours, they would be very obvious, since you say you did not have any perceivable cracking noises. Did you have louder listening sessions by now?
Hi Salas ! I agree. Many successfull designers tell me that an amplifier is a modulated PSU and i got great results with your Simple Hypnotize, even in circuits with a good PSU rejection. As i mentioned here before i am not very experienced with PSU´s but i collected enough knowlage now to design my own. The CCS and the cascode bias i use are some kind of PSU anyway and i have already experimented with sofisticated cap multipliers. As a reference i will use battery supply and your newer version of the meaner shunts. We are in a process to do a PCB for that but it is not finished. If i need your help i will ask.
Louder session begins in 5 minutes. The system is warmed up already.
Louder session begins in 5 minutes. The system is warmed up already.
Hi Joachim, nice design, what exactly do you mean by "EQ" in this context?In your post #1940 schematic, how did you implement the current source load, it is very difficult to avoid current fighting with the JFET?In the same schematic, could you explain reason for the strange R8+C10 lead lag compensation connection to the +15V rail?In the same schematic, what is the role of the R7 + C4 circuit?What is the role of C2, C12, R16, R14? If that's the RIAA HF EQ, then the caps must be pretty large (I suppose the resistors are small, thenths of ohms, to keep the noise low).
Hi YWN ! I will answer later the day because if am in a listening session and have to pack and send some stuff out today.
Salas, i have now listened enough to give a qualified description of what i hear.
First and foremost the same musicians play the same songs so what i heard was not totally unfamiliar to me. The sound has a rock solid fundament in the deep bass and has even a measure of warmth that i like. The sound is more liquid than i whould call it airy but there is no loss of detail that i could find against my best open loop stages. The presentaion of that details is slightly different though. Everything is there, the coffing and grumbling the musicians make on the Decca Stravinsky played by the Academy. Pre echos, acoustic information, sounds the instruments make etc. It is just not presented in the foreground but rather in the backround. It is hard to describe but it feels like those details are presented on a deeper level like shopping in the Supermarket and looking one level down to the more affordable wines that are still as good as the ones that stare you in the eye. Tonal balance is accurate and by means of the slight warm rendering even beautifull. I never heard Paul Desmonds Alto more pretty on Take Five for example. This stage sounds more deep then present. A bonus in my system because i sit rather close to the speakers and my room is shallow in the depth dimension. Height and width are very similiar then in my other stages. Transparency and focus are excellent. As good as anything i build so far. Dymamics are dramatic, again ancored in the fundamental tone range. A good example is "Floundering" by Carly Simom where Sly and Robby do their thing extremely precise and tight. As always i whould like to make a car analogy. When my open loop stages are a ride in an open 1950th Mercedes SL through the winding roads of the Italian Alps then this stage is a ride in a Porsche Turbo at 300KmH, securely clamped in the seats, on the straight Autobahn from Hannover to Hamburg. It feels that safe and secure but still is exiting. So many landscapes flying by so quick.
Salas, i have now listened enough to give a qualified description of what i hear.
First and foremost the same musicians play the same songs so what i heard was not totally unfamiliar to me. The sound has a rock solid fundament in the deep bass and has even a measure of warmth that i like. The sound is more liquid than i whould call it airy but there is no loss of detail that i could find against my best open loop stages. The presentaion of that details is slightly different though. Everything is there, the coffing and grumbling the musicians make on the Decca Stravinsky played by the Academy. Pre echos, acoustic information, sounds the instruments make etc. It is just not presented in the foreground but rather in the backround. It is hard to describe but it feels like those details are presented on a deeper level like shopping in the Supermarket and looking one level down to the more affordable wines that are still as good as the ones that stare you in the eye. Tonal balance is accurate and by means of the slight warm rendering even beautifull. I never heard Paul Desmonds Alto more pretty on Take Five for example. This stage sounds more deep then present. A bonus in my system because i sit rather close to the speakers and my room is shallow in the depth dimension. Height and width are very similiar then in my other stages. Transparency and focus are excellent. As good as anything i build so far. Dymamics are dramatic, again ancored in the fundamental tone range. A good example is "Floundering" by Carly Simom where Sly and Robby do their thing extremely precise and tight. As always i whould like to make a car analogy. When my open loop stages are a ride in an open 1950th Mercedes SL through the winding roads of the Italian Alps then this stage is a ride in a Porsche Turbo at 300KmH, securely clamped in the seats, on the straight Autobahn from Hannover to Hamburg. It feels that safe and secure but still is exiting. So many landscapes flying by so quick.
Thanks, no need to rush.BTW, following Salas clues I did a quick lookup and apparently you pissed SYN08 (is he still a member here?) pretty badly: Hall Of Shame
I terminated the project. It is not my intention to hurt anybodys reputation or feelings.
I had the impression that i correctly mentioned syn08s many contributions but that was not clear enough i now find. I could argue on and on what the differences to my circuit are
but the similarities are too obvious. I made a mistake that i regret.
I also made a decission that from now on when i publish a schematic i do not hold back any details about how it it made to work. This is a DIY forum and the readers have a right to see that. I have mixed hobby and profession too much recently and i promiss to sepertae that better from now on.
Concerning constant current sources i am prepared to share what i know and i would like to start with the most simple implementation. More to follow untill the dust has hopefully settled.
I had the impression that i correctly mentioned syn08s many contributions but that was not clear enough i now find. I could argue on and on what the differences to my circuit are
but the similarities are too obvious. I made a mistake that i regret.
I also made a decission that from now on when i publish a schematic i do not hold back any details about how it it made to work. This is a DIY forum and the readers have a right to see that. I have mixed hobby and profession too much recently and i promiss to sepertae that better from now on.
Concerning constant current sources i am prepared to share what i know and i would like to start with the most simple implementation. More to follow untill the dust has hopefully settled.
I would like to ask the members if the content of my last posts about a Single Ended Fet NFB phonostage should be erased. If it contains anything that is not wellcome i would prefer that but i can live with the oposite decission to keep the content for everybody to see.
At the time of writing i was not aware about the consequences and i certainly had no offending intentions.
At the time of writing i was not aware about the consequences and i certainly had no offending intentions.
That certainly happened to my work with loudspeakers but after some time i realised that i should be proud of that.
I can understand that it is a hard pill to swallow especially when someone gives away good ideas for free and others make the money. Happend to me too.
I am no angel though.
I can understand that it is a hard pill to swallow especially when someone gives away good ideas for free and others make the money. Happend to me too.
I am no angel though.
Joachim,
Keep up the good work that you are doing, I read the/your post that you make with great pleasure and intrest, and I do learn a lot in the process. It is unavoidable that designers duplicate (at the least partly) prior work, as I see you did acknowledge that, no more action should be taken.
Regards,
Frans de Wit.
Keep up the good work that you are doing, I read the/your post that you make with great pleasure and intrest, and I do learn a lot in the process. It is unavoidable that designers duplicate (at the least partly) prior work, as I see you did acknowledge that, no more action should be taken.
Regards,
Frans de Wit.