Passive 75 uS then 3180/318 uS active has raised the usual thought that it is unsuitable due to oveload margins. Years ago I did extensive testing of the reality of this and found it to be bogus. I even fed a Shure M44-7 9.5 mV into a stage intended for 500 uV ( 47 K, that's overloading from the word go ). Mostly it sounded nice! When it did show problems it was not unlike a tape deck in how it distorted and not like a digital system. I speculate that an all active design would be more fussy. Even the 250 uV input wasn't dreadful although the volume contol on the amp was as far down as it could go ( 3 % ? ). Alps high grade.
Why passive + active? I feel simplicity with OK specs gets you the best sound. As said before it makes the LP EQ at >2kHz very easy. Lets be very clear about this. Many pick ups are nothing like flat. This is most obvious around 7 kHz. A little bit of HF correction is harmless fun that could sound better. As my system allows in more sources I have had to make the turntable suit the rest. A small HF tweak is enough. When playing 78's although RIAA is wrong the HF tweak is still helpful. If asking a 78 tipped Shure M44 is my favourite. For fun I will share this. Shure M44-7 in SME 5 is one of the best LP sounds I ever heard ( best calculated dynamic range ? ). I don't usually like the 5 ( I have a Schroeder + Lyra Helikon and 401 and LP12 Ekos DL110 for all to use ). This was a Garrard 401 for Masterpiece in London who do 78's to CD. They also do LP's. They wanted a valve preamp with all the EQ's. Leak Varislope with a quiet HT and buffer amp used with very best EF 86's we could find ( Mullard ). The Leak needs a very high impedance buffer to work ( OPA2604 was at the time the better choice ). Ours had a gain of 3 if I remember, it had to run 10 metres of cable. The Leak is all active and only has two EF86 to do all the functions ( vast ), the EQ's were fiddled by Leak to get it to work due to lack of loop gain. Doubtless it's RIAA is not as accurate as we can build now. The first valve HT is not greatly different to op amp solutions with 55 V at the anode ( 135 V at the rail I think. via 330K ). 80V at V2 / 220K. This set up was part of a Cedar award winner for the Bing Crosby Radio Shows. Ironically Cedar wasn't used. These were the shows before Bing was advised to buy a share of Ampex ( Nazi Magnetophon, AEG I think ). Many do not know that Ampex made Bing rich rather than the obvious. The big discs Masterpiece transcribed were Bings first attempt to do three shows a day to suit the different USA time zones. The engineer Andy Pearce. The big format from the late 1920's talking films. 16 and 20 inch at 33 1/3. These ones were acetate, they can be shellac or vinyl. I built the record cleaning machine used. A Percy Wilson/Monks clone ( Audio and design Ltd Maidenhead , 1968, both they and Percy forgotten).
Douglas. You had great fun with a class A+C design. Would you risk a boosted op-amp ( rail switching ) to try to beat the Technics overload margin record ( I think you stated it in the past ). As said I don't belive it's required, even so fun to know. It would be interesting as anyone who ever looked at this said no way. I know special op amps exist to do this. It would be fun to do it the " wrong " way. Better still a Self higher voltage op amp to sell as a project. You would have customers I feel. It could be just a smaller version of your power amps. Motorola driver transistors as outputs? Make it class A if required. It could drive headphones or small speakers. It could be op amp No 2 in a chain. This allows simple RC filtering to op amp 1 on the same PSU as did valves ( +/- 12 V ? ). I've always told myself my next project is to do this. Must be 10 years I have been saying it.
Why passive + active? I feel simplicity with OK specs gets you the best sound. As said before it makes the LP EQ at >2kHz very easy. Lets be very clear about this. Many pick ups are nothing like flat. This is most obvious around 7 kHz. A little bit of HF correction is harmless fun that could sound better. As my system allows in more sources I have had to make the turntable suit the rest. A small HF tweak is enough. When playing 78's although RIAA is wrong the HF tweak is still helpful. If asking a 78 tipped Shure M44 is my favourite. For fun I will share this. Shure M44-7 in SME 5 is one of the best LP sounds I ever heard ( best calculated dynamic range ? ). I don't usually like the 5 ( I have a Schroeder + Lyra Helikon and 401 and LP12 Ekos DL110 for all to use ). This was a Garrard 401 for Masterpiece in London who do 78's to CD. They also do LP's. They wanted a valve preamp with all the EQ's. Leak Varislope with a quiet HT and buffer amp used with very best EF 86's we could find ( Mullard ). The Leak needs a very high impedance buffer to work ( OPA2604 was at the time the better choice ). Ours had a gain of 3 if I remember, it had to run 10 metres of cable. The Leak is all active and only has two EF86 to do all the functions ( vast ), the EQ's were fiddled by Leak to get it to work due to lack of loop gain. Doubtless it's RIAA is not as accurate as we can build now. The first valve HT is not greatly different to op amp solutions with 55 V at the anode ( 135 V at the rail I think. via 330K ). 80V at V2 / 220K. This set up was part of a Cedar award winner for the Bing Crosby Radio Shows. Ironically Cedar wasn't used. These were the shows before Bing was advised to buy a share of Ampex ( Nazi Magnetophon, AEG I think ). Many do not know that Ampex made Bing rich rather than the obvious. The big discs Masterpiece transcribed were Bings first attempt to do three shows a day to suit the different USA time zones. The engineer Andy Pearce. The big format from the late 1920's talking films. 16 and 20 inch at 33 1/3. These ones were acetate, they can be shellac or vinyl. I built the record cleaning machine used. A Percy Wilson/Monks clone ( Audio and design Ltd Maidenhead , 1968, both they and Percy forgotten).
Douglas. You had great fun with a class A+C design. Would you risk a boosted op-amp ( rail switching ) to try to beat the Technics overload margin record ( I think you stated it in the past ). As said I don't belive it's required, even so fun to know. It would be interesting as anyone who ever looked at this said no way. I know special op amps exist to do this. It would be fun to do it the " wrong " way. Better still a Self higher voltage op amp to sell as a project. You would have customers I feel. It could be just a smaller version of your power amps. Motorola driver transistors as outputs? Make it class A if required. It could drive headphones or small speakers. It could be op amp No 2 in a chain. This allows simple RC filtering to op amp 1 on the same PSU as did valves ( +/- 12 V ? ). I've always told myself my next project is to do this. Must be 10 years I have been saying it.
"Passive 75 uS then 3180/318 uS active has raised the usual thought that it is unsuitable due to oveload margins. Years ago I did extensive testing of the reality of this and found it to be bogus. I even fed a Shure M44-7 9.5 mV into a stage intended for 500 uV ( 47 K, that's overloading from the word go ). Mostly it sounded nice! When it did show problems it was not unlike a tape deck in how it distorted and not like a digital system. I speculate that an all active design would be more fussy. Even the 250 uV input wasn't dreadful although the volume contol on the amp was as far down as it could go ( 3 % ? ). Alps high grade."
If you think a 9.5mV (I assume this is the 1 kHz or 5cm/s output) into an MC stage sounds 'mostly nice' then who am I to argue.
Active/passives require serious tradeoffs - so why do that in the first place? You get lowest noise and best dynamic range with all active at the expense of a bit more design effort. Lowest distortion in theory too because there is greater feedback at HF. The fact that people talk about 'boosted rails' as a solution to the dynamic range problem with active/passive speaks volumes. And we haven't begun to talk about the noise aspect.
If you think a 9.5mV (I assume this is the 1 kHz or 5cm/s output) into an MC stage sounds 'mostly nice' then who am I to argue.
Active/passives require serious tradeoffs - so why do that in the first place? You get lowest noise and best dynamic range with all active at the expense of a bit more design effort. Lowest distortion in theory too because there is greater feedback at HF. The fact that people talk about 'boosted rails' as a solution to the dynamic range problem with active/passive speaks volumes. And we haven't begun to talk about the noise aspect.
Are you sure you have measured that with real records. I did and found it wasn't a real issue. Most things in audio that pass for fact and nothing of the sort due to the real situation. People will say at least get what you can right. I am not convinced.
Hiss is worth reducing. Do most have any idea how high LP distortion is? Very is the answer. Most do not own a turntable that gets close to reducing it. I have a JVC paralell tracking DD in my collection. It is so obvious when listening that it starts to solve a little of that problem, records sound new that aren't. If you hid it behind a curtain and said here is the lastest low distortion preamp I would forgive you for saying how wonderful. It has only a spherical stylus! The arm is super ridged ( looks like plastic and isn't ).
BTW. Dymanic range is mostly down to PU output. The Shure M44 given a special preamp designed for it even sounds that way. It's stereo separation is enough to be stereo. > 68 K suits it best.
Hiss is worth reducing. Do most have any idea how high LP distortion is? Very is the answer. Most do not own a turntable that gets close to reducing it. I have a JVC paralell tracking DD in my collection. It is so obvious when listening that it starts to solve a little of that problem, records sound new that aren't. If you hid it behind a curtain and said here is the lastest low distortion preamp I would forgive you for saying how wonderful. It has only a spherical stylus! The arm is super ridged ( looks like plastic and isn't ).
BTW. Dymanic range is mostly down to PU output. The Shure M44 given a special preamp designed for it even sounds that way. It's stereo separation is enough to be stereo. > 68 K suits it best.
The vinyl/pickup/tracking error certainly are the largest contributors to distortion in LPs, no doubt about that one. LP+cart output is well documented and yes there may be a lot of folklore around peak output levels. It will of course also depend on the music you are listening to. compressed modern pop recordings might allow you to get away with less, but I listen to a lot of classical stuff and some of it demands good dynamic range - i.e. as much as you can get off vinyl. I would not consider anything <20 dB o/load margin at 20 kHz as acceptable if you are serious about playback quality.
Last edited:
A thought pulled out of the air here. Overload margin at 20 kHz is often quoted. The Baxendale solution much like my own offers about 16 dB margin. The Self solution at a wild guess shoud be 20 dB better. It's all very debatable and requires some thought and measurement. What I would say is play with the gain of any passive type. I use 10 nF as I have a very big bag of them ( Polystyrene 1% ). That might be doing me a favour as I have to drive the network. I also have some 33 nF that I never got around to trying.
On an all active design I built I have a 2 uS passive output filter. This seems to resemble the sound of a passive-active design. If you listen carefully to an all active you might hear a longer tick on a scratch or even a tock when bad. The 2uS takes some of that away. I really do dislike this distortion. It makes the records sound more wornout. It is possible this happens further up the amplifier. If I am right about this the supposed advantage of an all active is not proved. Could it be the signal is too fast? All active in valves did not seem to do this.
On an all active design I built I have a 2 uS passive output filter. This seems to resemble the sound of a passive-active design. If you listen carefully to an all active you might hear a longer tick on a scratch or even a tock when bad. The 2uS takes some of that away. I really do dislike this distortion. It makes the records sound more wornout. It is possible this happens further up the amplifier. If I am right about this the supposed advantage of an all active is not proved. Could it be the signal is too fast? All active in valves did not seem to do this.
Here's an active RIAA stage that was introduced by Roberts in Popular Electronics March 1981. Also used by SSM for the SSM2015 application note and by me. (I think this may have been cited earlier in this thread.)
The P10 RIAA EQ is noteworthy due to the simplicity of its math. Perhaps Mr. Self will include it.
What's unique about the P10's RIAA EQ is the location of the feedback to the inverting input. (IC1D)
Ignore C4 and C5. If it is assumed that virtual ground is ground, the input pole is 75 µs passive, the 318 µs is applied as preemphasis and the 3180 µs is an active pole in the feedback loop. I've only seen this topology previously in these two places. The 75 µs pole may not be immediately obvious but the math for this RIAA EQ is ultra-simple. It also has the advantage by virtue of being inverting a 75 us response that falls off below unity gain.
John tells me that his choice of 8200 pF||100pF was motivated by the fact he had a large quantity of 8200 pF in stock. If the Cs are changed to 10 nF the time constants and simplicity of the math become more obvious.
This is a version of Robert's P10 EQ I modified to provide jumper-selectable RIAA or line-level inverse RIAA. (For cutting.) The Secret Society of Lathe Trolls • View topic - RIAA/Inverse RIAA Circuit - Jumper Selectable
The 3180 µs is 10nF*318K, the 318 µs 10nF*31K8, the "74.99" µs pole (9K813||31K8)*10nF.
By rearranging the feedback connections a line-level Inverse RIAA response is possible. For inverse RIAA a 100 pF Cfb can be installed to limit ultrasonic response.
It's a super simple circuit. 10 nF 1% Wimas are available from Mouser at reasonable cost.
You will note that the P10 was also balanced input and realizes common mode rejection. (IC1C.)
RCA inputs are shown as they are rather than a fully symmetrical connection using shielded twisted pair. This was due to contact with Mr. Roberts from the patent attorneys of Logitek who somehow manged to patent a fully-balanced phono input at about the same time.
The P10 RIAA EQ is noteworthy due to the simplicity of its math. Perhaps Mr. Self will include it.
What's unique about the P10's RIAA EQ is the location of the feedback to the inverting input. (IC1D)
Ignore C4 and C5. If it is assumed that virtual ground is ground, the input pole is 75 µs passive, the 318 µs is applied as preemphasis and the 3180 µs is an active pole in the feedback loop. I've only seen this topology previously in these two places. The 75 µs pole may not be immediately obvious but the math for this RIAA EQ is ultra-simple. It also has the advantage by virtue of being inverting a 75 us response that falls off below unity gain.
John tells me that his choice of 8200 pF||100pF was motivated by the fact he had a large quantity of 8200 pF in stock. If the Cs are changed to 10 nF the time constants and simplicity of the math become more obvious.
This is a version of Robert's P10 EQ I modified to provide jumper-selectable RIAA or line-level inverse RIAA. (For cutting.) The Secret Society of Lathe Trolls • View topic - RIAA/Inverse RIAA Circuit - Jumper Selectable
The 3180 µs is 10nF*318K, the 318 µs 10nF*31K8, the "74.99" µs pole (9K813||31K8)*10nF.
By rearranging the feedback connections a line-level Inverse RIAA response is possible. For inverse RIAA a 100 pF Cfb can be installed to limit ultrasonic response.
It's a super simple circuit. 10 nF 1% Wimas are available from Mouser at reasonable cost.
You will note that the P10 was also balanced input and realizes common mode rejection. (IC1C.)
RCA inputs are shown as they are rather than a fully symmetrical connection using shielded twisted pair. This was due to contact with Mr. Roberts from the patent attorneys of Logitek who somehow manged to patent a fully-balanced phono input at about the same time.
Last edited:
May be I miss important thing and my apologies in this case, but the formula on top of your third yellow paper is very fast but incorrect.
Here is the result of my calculation:
Which, if correct, shows that there is another zero and the HF gain is 1.
Yes I cheerfully ignored that zero (Lipschitz calls its timeconstant "T6") and focused upon the pole-zero-pole (3180u, 318u, 75u) mandated by RIAA; Lipschitz calls these T3-T5. To cancel the T6 zero requires a passive RC postfilter which is not drawn on yellow sheet 1; nor is it drawn in Figure 8.3 of SSAD 2nd ed. It is however drawn in Figure 8.5.
It appears that you may have used some type of symbolic algebra software to generate your results, judging by the fancy fonts and mathematical formatting. That is exactly the type of modern, 21st century approach to circuit design which I advocated in post #173 above; good on ya. It's quite clear to me that YOU don't need Lipschitz!
Today you can hand it off to a symbolic algebra computer system like Maple and get The Right Answer immediately. The only "value add" from Prof. Lipschitz is that he and his grad students slogged through the algebra and checked it and double checked it (and forced the JAES reviewers to double-double check it), so that you don't have to. Now that you have Maple software, you don't need Lipschitz and you also don't need a gallon of coffee and two pads of notepaper.
The next step is to dump the full and complete Laplace transfer function into a nonlinear optimization program, and let it churn out the optimum set of component values which gives the best fit (minimum sum-of-squared-errors) to the idealized RIAA curve. Subject to whatever constraints the human designer has supplied, perhaps such as: (a) RIAA equalized gain at 1kHz = +33.500 dB; (b) Opamp AVol = 90 dB; (c) Opamp GBW = 15 MHz; (d) Minimum load on opamp output >= 200 ohms; etc.
I calculated it by hand and wrote the result on word because my writing is not easily readable .... but I verified with a friendly freeware called "sapwin".
Sure. As I didn't try Matlab, Python or Maple, I used excel's optimizer to calculate components values on post 149 schematics.
C'mon - stop hammering Lipshitz. Easy to talk down a solid piece of work done 40 years ago. It's like saying Newton added no value because he should have worked out relativity in the first place.
1. There were no symbolic math packages in 1979 available to average Joe. Things had to be hand calculated out.
2. A key point of his paper was to correct the misconception that you could just calculate the R&C values from t=RC in all-active designs
3. Your '747' effort missed out T6 - and you do need it - feed a 10 kHz square wave into an all active RIAA without the post filter via an Inverse Network and you will note some overshoot - The T6 post filter is designed to correct for that. And the requirement for that is something that's only going to pop out in a full and accurate circuit analysis.
4. How many people know the constraint T2.T3.T5=T1.T4.T6 without it being made explicit? A few guys up on their math, or who have investigated all-active RIAA in depth but not many others. Result: the current received wisdom that passive or active-passive is a better way to do EQ.
Peace
1. There were no symbolic math packages in 1979 available to average Joe. Things had to be hand calculated out.
2. A key point of his paper was to correct the misconception that you could just calculate the R&C values from t=RC in all-active designs
3. Your '747' effort missed out T6 - and you do need it - feed a 10 kHz square wave into an all active RIAA without the post filter via an Inverse Network and you will note some overshoot - The T6 post filter is designed to correct for that. And the requirement for that is something that's only going to pop out in a full and accurate circuit analysis.
4. How many people know the constraint T2.T3.T5=T1.T4.T6 without it being made explicit? A few guys up on their math, or who have investigated all-active RIAA in depth but not many others. Result: the current received wisdom that passive or active-passive is a better way to do EQ.
Peace
Why worship it today? We have Wolfram Alpha, Octave, Python, Perl, and Broyden-Fletcher-Goldfarb-Shanno, all of which are completely free. We have Excel and its free Solver. Why not wave a friendly goodbye to the ever-so-handy-back-when-Jimmy-Carter-was-President tables, and start using software? Software lets us find the optimum component values including the effect of the opamp's dominant pole. Software even lets us include the opamp's second pole. Software lets us include the opamp's finite open loop gain. Software lets us include the opamp's nonzero open loop output resistance.
One wonders whether the people who cling tightly to their dusty old tables, might perhaps be a little intimidated by software.
One wonders whether the people who cling tightly to their dusty old tables, might perhaps be a little intimidated by software.
Well, Doug. There's another chapter for your new book.
Please include the simple Lipshitz eqns for the 4 different networks to help *rse hu neber wen 2 skul.
For wannabe gurus hoping to emulate Mark Johnson, you can show the Wolfram Alpha, Octave, Python, Perl, and Broyden-Fletcher-Goldfarb-Shanno scripts and explain how they were obtained.
For the true idiots like me, a link to Mark's RIAA for Dummies programme.
Please include the simple Lipshitz eqns for the 4 different networks to help *rse hu neber wen 2 skul.
For wannabe gurus hoping to emulate Mark Johnson, you can show the Wolfram Alpha, Octave, Python, Perl, and Broyden-Fletcher-Goldfarb-Shanno scripts and explain how they were obtained.
For the true idiots like me, a link to Mark's RIAA for Dummies programme.
Last edited:
Thank you Guru rayma. I now remember this ... dredging up senile memories from 30+ yrs ago when I had more than 1 brain cell.
Mark, can you include this in your RIAA networks for dummies please?
Doug, can you reconfigure the Lipshitz eqns for the 4 networks to facilitate this method in your book please?
declicking with comparators, S/H, building mixed signal analog hardware seems very odd today given soundcards and computing horsepower
of course when you do go ADC/DSP/DAC then fixing up RIAA accuracy in light of your build's actual components becomes a relatively trivial additional soundcard/DSP calibration exercise
of course when you do go ADC/DSP/DAC then fixing up RIAA accuracy in light of your build's actual components becomes a relatively trivial additional soundcard/DSP calibration exercise
I have been looking for data on real LP HF content including surface damage and find nothing that suggests that a 16 dB overload margin at 20 kHz to be a cause for concern. What I did find was this!
Analysis of Vinyl Frequency Content
One idea that came to me in the bath is that LP playback like fishing is not exactly about the easy life and more about hobby. Thus if liking a passive-active circuit despite wise advice we should look at possible improvements. That is that the phono preamp could control volume and gain some headroom for careful listening. As I have hinted elsewhere the exact gain of an op amp might be a factor in liking a design. Ironically higher gain can be nicer. Remember that 0.1% THD in a vinyl system is nothing, not least if mostly second harmonic. Much of what is said is usually related to making many units for people you will never meet. You may have no interest in that and be happy to serve only your own taste.
For the first 23 years of my working life I repaired things. What I noticed was that most things that sounded very good were simple and often with good power supplies. Two makes of valve preamps of seemingly identical concept could be chalk and cheese. The one I really liked used a very short piece of TV coax preamp to power amp. I wonder how many were ruined by fitting " better " cable. Could it be it was a critical 2uS filter or whatever that I never guessed.
The idea that an all active circuit might be without any problems is against my instinct. I suspect if a passive-active verses all active was blind tested to all here the majority would think one was the other. The clue being suspected HF problems, sometimes sounding like a mildly dirty stylus or a hissy tick on a scratch.Tick-fizz if you like. Often good designs seem noise free untill you listen for clicks. This can make LP's you never should have bought sound OK. These preamps sound more open. Not lack of HF helping which is never does. Tock sounds worse that tick.
Would not real video op amps be best ( stage 1 )? Some even hint at audio use. They will drive 75R which is no bad thing. If used with care the low output is still useful. My preamp has a gain of 62 when 250 uV ( often fed 700 uV ) at op amp No1 ( 17 next ). This doesn't look to be impossible even with +/- 2.5 V rails. AD828 looks to be an interesting almost op amp device. It greatly stresses things that make me think it would be ideal for an all active circuit. High gain might avoid stability problems.
My brother insisted that only an inverting all active circuit would work. He said the gain formula says all you need to know. The non inverting gain never is less than one which must exagerate clicks. For an MC often inverting can work ( noise ). I have run a Lyra Helikon into 22R and was surprised how well it worked. 100R seems a winner if so for any sensible MC. The next stage could be 22uF (poly) + 1K8 as a sensible input that's also inverting ( non inverting total that's double inverting ). You should find the exact loading of an MC is not a big deal if the chain is strong. If you find otherwise perhaps the preamp is not as happy as your maths tells you. One other advatage of inverting is common mode problems. My instinct is it should be better. A filter I built recently seemed to prove this, could be I was just cancelling an identical error that I had caused by not using a ground plane. It was worth 6dB at 50 Hz.
Analysis of Vinyl Frequency Content
One idea that came to me in the bath is that LP playback like fishing is not exactly about the easy life and more about hobby. Thus if liking a passive-active circuit despite wise advice we should look at possible improvements. That is that the phono preamp could control volume and gain some headroom for careful listening. As I have hinted elsewhere the exact gain of an op amp might be a factor in liking a design. Ironically higher gain can be nicer. Remember that 0.1% THD in a vinyl system is nothing, not least if mostly second harmonic. Much of what is said is usually related to making many units for people you will never meet. You may have no interest in that and be happy to serve only your own taste.
For the first 23 years of my working life I repaired things. What I noticed was that most things that sounded very good were simple and often with good power supplies. Two makes of valve preamps of seemingly identical concept could be chalk and cheese. The one I really liked used a very short piece of TV coax preamp to power amp. I wonder how many were ruined by fitting " better " cable. Could it be it was a critical 2uS filter or whatever that I never guessed.
The idea that an all active circuit might be without any problems is against my instinct. I suspect if a passive-active verses all active was blind tested to all here the majority would think one was the other. The clue being suspected HF problems, sometimes sounding like a mildly dirty stylus or a hissy tick on a scratch.Tick-fizz if you like. Often good designs seem noise free untill you listen for clicks. This can make LP's you never should have bought sound OK. These preamps sound more open. Not lack of HF helping which is never does. Tock sounds worse that tick.
Would not real video op amps be best ( stage 1 )? Some even hint at audio use. They will drive 75R which is no bad thing. If used with care the low output is still useful. My preamp has a gain of 62 when 250 uV ( often fed 700 uV ) at op amp No1 ( 17 next ). This doesn't look to be impossible even with +/- 2.5 V rails. AD828 looks to be an interesting almost op amp device. It greatly stresses things that make me think it would be ideal for an all active circuit. High gain might avoid stability problems.
My brother insisted that only an inverting all active circuit would work. He said the gain formula says all you need to know. The non inverting gain never is less than one which must exagerate clicks. For an MC often inverting can work ( noise ). I have run a Lyra Helikon into 22R and was surprised how well it worked. 100R seems a winner if so for any sensible MC. The next stage could be 22uF (poly) + 1K8 as a sensible input that's also inverting ( non inverting total that's double inverting ). You should find the exact loading of an MC is not a big deal if the chain is strong. If you find otherwise perhaps the preamp is not as happy as your maths tells you. One other advatage of inverting is common mode problems. My instinct is it should be better. A filter I built recently seemed to prove this, could be I was just cancelling an identical error that I had caused by not using a ground plane. It was worth 6dB at 50 Hz.
I would not claim all-active is problem free. However, it is more problem free than passive or active passive topologies because you are not forced to make serious trade-offs between noise and overload margin. With a discrete JFET front end, +85 dB on an MM cart input and >30 dB O/load margin across the audio band is achievable (that's with the cart in situ - not shorted input). If that's not clean, I don't know what is!
I've never had a problem with ticks and pops being extended or HF not sounding right with all active. With all due respect Nigel, this latter point can only apply to your specific implementation - you cannot generalize like that about all-active RIAA. In any event, you would always have to use a suitable opamp for the main stage - I don't know what your exemplar used but 741/TL071 types will not cut it.
All active only uses a single amplifier stage - so it is in fact simpler than active-passive types where you have a gain stage and almost always a buffer. The only 'challenge' with all-active is in doing the network design properly and as you can see from the previous posts, that requires a bit of effort but it's not rocket science.
Wrt inverting RIAA, there is a noise penalty from the input resistor - no penalty with non inverting.
I've never had a problem with ticks and pops being extended or HF not sounding right with all active. With all due respect Nigel, this latter point can only apply to your specific implementation - you cannot generalize like that about all-active RIAA. In any event, you would always have to use a suitable opamp for the main stage - I don't know what your exemplar used but 741/TL071 types will not cut it.
All active only uses a single amplifier stage - so it is in fact simpler than active-passive types where you have a gain stage and almost always a buffer. The only 'challenge' with all-active is in doing the network design properly and as you can see from the previous posts, that requires a bit of effort but it's not rocket science.
Wrt inverting RIAA, there is a noise penalty from the input resistor - no penalty with non inverting.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.