The definitive info on this is in the Holman paper I linked to.
There was a VERY careful & detailed investigation by one of pseudo-prophet Otala's pupils in AES Hamburg, 1981 looking for ultrasonic stuff from MC cartridges on clicks .. but it essentially showed that no further supa dupa slews were necessary. A well designed single OPA active RIAA stage would suffice.
As you say, the only requirement above the usual 20+dB overload level for a phono stage is INSTANT recovery from ticks & pops.
One 21st century requirement is that the initial preamp should be FLAT so as not to lengthen any ticks & pops. This allows MUCH easier & BETTER digital de-clicking (which should be of course in the TIME DOMAIN)
Wayne's thread on his website is probably definitive on this and has loadsa other useful PRACTICAL stuff that belongs in a book attempting to extend the SoTA on vinyl replay.
Last edited:
Yes. Even though mechanical tracking recovery time might be longer than good saturation recovery, it does no harm and it's good practice to keep electronics out of saturation if at all possible.
I scanned the Tomlinson Holman document thanks for the link Richard). In the summary he states a phono amp should have 100 mV overload capability at 1 kHz. However, at LF this would mean the corresponding overload capability would be 10 mV. The amplitude of subsonic warp, resonances etc can exceed this - he recommends some sort of HF filter so that this figure is at least 35 mV.
His reference gain, used through this paper appears to be 40 dB at 1 kHz which means the overload spec would be 20 dB ref a 5 mV input signal.
HF ticks and pops are not pre-emphasized (i.e EQ'd)' but during playback they benefit from the RIAA 20 dB/decade reduction in gain from 2120 Hz - differences between all active and other EQ methods notwithstanding.
The paper is based on extensive measurement of actual records so in my view this is the go to reference on RIAA overload.
His reference gain, used through this paper appears to be 40 dB at 1 kHz which means the overload spec would be 20 dB ref a 5 mV input signal.
HF ticks and pops are not pre-emphasized (i.e EQ'd)' but during playback they benefit from the RIAA 20 dB/decade reduction in gain from 2120 Hz - differences between all active and other EQ methods notwithstanding.
The paper is based on extensive measurement of actual records so in my view this is the go to reference on RIAA overload.
Last edited:
Self SSAD 2nd Edition tabulates measured data from the Holman paper and many other published sources (see pp. 205-216). In my view it is the go to reference on RIAA overload.
Holman was my bible for RIAA preamp requirements in Jurassic times when I did this stuff for a living.
Could you post the references that Self uses in addition to Holman? I'm sure Doug won't mind if you post the references though the table itself would be nice
_________________
Andrew, I presume you meant HP filterBonsai said:
Now here's something that is novel AND useful so suitable for inclusion in the new book.
How about incorporating a 12dB/8ve HP filter to replace the infamous IEC 20Hz roll-off in a simple single OPA RIAA stage?
I worked out a single implementation in Jurassic times and the "turn your page sideways algebra" was so traumatic that I went away and wrote my own Linear Circuit Analysis package to avoid ever having to do that again
(This was the days before sensibly priced and easy to use SPICE implementations ... or even 5534s ... were available)____________________
R2 3 4 & C1 2 can be any of the 4 RIAA networks per Lipshitz. Add the extra supersonic pole to taste.
R2 with R5 6 & C3 4 implement a HP filter which is near 2nd order Butterworth HP at 20.02Hz to replace the infamous IEC roll-off.
IMHO, instead of agonising over the accuracy of the IEC amendment, we should just consider it the simplest subsonic filter and welcome higher order filters with about the same 20.02Hz -3dB point.
"near 2nd order bla bla" cos at very LF, the non-inverting stage reverts to unity gain.
This would be difficult to implement accurately with switched gain so perhaps single examples at 40 & 30dB 1kHz gain would be good.
Besides, Holman & others show that there is a huge bump extending into the audible range from the cartridge/arm resonance and there is a good case for variable Q on this 20.02Hz filter
As Doug apparently shares my fear of mind bending algebra, perhaps Mark Johnson could work out these 2 simple cases when he has a spare 20min.
_____________________
BTW, the hype about overload margin needs to be put into perspective. It ALL hinges on where your volume control is set .. on the usual "RIAA preamp - volume control - some gain including your PA" system.
If we assume a 'normal' log volume control, this is about -20dB at half volume and usually -30dB at about 9 or 10 o'clock.
If your volume control is always set above half way, 20dB overload margin on the RIAA preamp is sufficient as the PA will always clip first.
If you never have your volume control set below 9 o'clock, 30dB overload is sufficient.
40dB overload margin is only needed if you always have your volume control below about 8 o'clock.
IME, 40dB 1kHz RIAA gain is usually good with +/- 15V OPAs ... except with cheapo high output cartridges where 30dB 1kHz gain (as in Doug's example) may be required.
Switchable 1kHz RIAA gain (with suitable attention to the required overload) may be good just to operate your volume control in a nice matched part of the track.
Attachments
Last edited:
It is very important to evaluate the need for a sub 20Hz filter. Most high grade turntables don't really need it. I use two pole 4Hz with an option for one pole. Do listen to this as the outcome is not where you would expect it to be. Voice for example. Doubtless this is me preferring something technically wrong. When I listen late at night to a warm friendly sound I don't really care too much why. It may not be wrong as often the bass of an LP is reduced to help the cut or get more playback time. The LF resonance and that EQ have to be factored in and is a matter of taste. In addition the LF pick up resonance might be 8Hz or 12Hz. That will be important. I tend to prefer nearer 8Hz.
My system needs EQ for the speakers. I would be very tempted to change the RIAA to see if both EQ's could be done inside the phono amp for simplicity. I would then EQ my line sources to suit. I can see how it would be difficult from the start. It might work. Knowing " real " overload margin problems would help. My speakers need HF boost. That would be the 75uS. The bass the 3180uS.
I spent the evening with Colleen watching VHS tapes. Fireball XL5 and Champagne ( was to be at New Year ). I built my system to work with high or low grade sources. The idea of EQ for each input would be very useful.
My system needs EQ for the speakers. I would be very tempted to change the RIAA to see if both EQ's could be done inside the phono amp for simplicity. I would then EQ my line sources to suit. I can see how it would be difficult from the start. It might work. Knowing " real " overload margin problems would help. My speakers need HF boost. That would be the 75uS. The bass the 3180uS.
I spent the evening with Colleen watching VHS tapes. Fireball XL5 and Champagne ( was to be at New Year ). I built my system to work with high or low grade sources. The idea of EQ for each input would be very useful.
You might like to look at Holman and his pics for LF response of typical arm/cartridge combinations. More in the B&K papers IIRC.
Of course, 6dB boost at LF may sound good to you.
Some high grade turtables may not need a 20Hz filter but most records do.
Of course Bill doesn't play warped records
Must get da flat ones from Ye Olde Unobtainium Shoppe.Radford made & sold the first sensible speakers with flat response down below 30Hz ... Dr. Bailey's Transmission Lines in the early 70s
Their amps from that period deliberately had fixed 40Hz 12dB/8ve built into their RIAA preamps
All I would say is be sure it's what you want. Use you ears. It seems many will not as some book or some person said this or that. I read that 3180uS wasn't a real curve, an RCA idea of what could work. It was a suggestion which became fact as time went on.The real curve was far more complex and would not be the same company to company. RIAA helped make a standard as the engineers listen via an RIAA amplifier to review playback at the pre-production stage. Every cutting engineer would treat exact EQ as a starting point, so should we.
My speakers do not overly mind if there are warps or defects. Yours might.
My speakers do not overly mind if there are warps or defects. Yours might.
Radford Monitor HiFi Stereo Speakers
Pair of same now relegated to being my dad's garage speakers.
Still good, damm good.
Dan.
Last edited:
Revisiting briefly the subject of bootstrapped floating supply RIAA phono preamps...
Some of you may find this citation interesting:
http://www.proaudiodesignforum.com/...S_v30n5_Feedforward_Floating_Power_Supply.pdf
Wide Bandwidth RIAA Phono Preamp With Feedforward Power Supply
Some of you may find this citation interesting:
http://www.proaudiodesignforum.com/...S_v30n5_Feedforward_Floating_Power_Supply.pdf
Wide Bandwidth RIAA Phono Preamp With Feedforward Power Supply
Nigel, the turntable is not the major issue - I agree most nowadays are very good when it comes to rumble issues. Its the LP's themselves that are either warped or off center causing high outputs at infrasonic frequencies - this is discussed in the Tomlinson paper.
I have a 30 year old Manhattan Transfer LP that is badly warped. If I play it without the filter the cones on my B&W 703's flap around horrendously. With the filter in, no issues.
-0.2 dB at 20 Hz; -18 dB at 8 Hz; -36 dB at 2 Hz - so virtually no effect at 20 Hz and I can't detect anything getting affected higher up in the bass region when its active.
My power amp is -3 dB at 2 Hz.
I have a 30 year old Manhattan Transfer LP that is badly warped. If I play it without the filter the cones on my B&W 703's flap around horrendously. With the filter in, no issues.
-0.2 dB at 20 Hz; -18 dB at 8 Hz; -36 dB at 2 Hz - so virtually no effect at 20 Hz and I can't detect anything getting affected higher up in the bass region when its active.
My power amp is -3 dB at 2 Hz.
When I designed my speakers I had a nightmare with well respected drive units hitting the end stops, SB Acoustics being the worst. Even BBC broadcasts could be a problem. I bought some PA drivers that on paper were much the same although fitted theory better. The idea was to build an optimum HP filter to suit. From the word go I had no need. Modern drive units could learn from PA units on this. This was a very unexspected bonus and allowed me a wide open RIAA. My Pro Audio friends just smile and say they could never use hi fi parts and suspect they even measure worse at domestic levels. My ears go along with theirs, many use their pro speakers at home. IB speakers should be good on this. Isobaric+IB types better still. True Acoustic suspension might have worked with the SB units.
Doing the maths of my pick up and arm. 8.5cu, mass of PU + arm 19 g. fo = 1/2Pi [(8.5 x 10^-6 x 19 ) ^0.5] = 12.5 Hz. I suspect that helps.
I come at this from a different direction. Most line sources are less important to me. Vinyl still my prime source. This is down to listening fatigue alone. I then make my line sources work by careful adjustment of level mostly. If you have the luxury of doing exactly what you want and hear how good a system like this can be there is no going back. The cheapest off the shelf speaker I know of to do this are Quad ESL 63 used in the higher dangerous grilles off state.
Doing the maths of my pick up and arm. 8.5cu, mass of PU + arm 19 g. fo = 1/2Pi [(8.5 x 10^-6 x 19 ) ^0.5] = 12.5 Hz. I suspect that helps.
I come at this from a different direction. Most line sources are less important to me. Vinyl still my prime source. This is down to listening fatigue alone. I then make my line sources work by careful adjustment of level mostly. If you have the luxury of doing exactly what you want and hear how good a system like this can be there is no going back. The cheapest off the shelf speaker I know of to do this are Quad ESL 63 used in the higher dangerous grilles off state.
Nigel, you've said this at least twice now.
Would you like to tell us what you hear when the EVIL 20.02Hz IEC (or shock horror
even higher slopes & cutoff) filters are engaged.I've done a small amount of (sighted) listening tests on this in Jurassic times and would be interested in your aural findings.
I made a preamp for the Audiophiles in Germany. It was compared against the Lyra
LYRA - Lyra_Connoisseur 4-2 SE Line
On the whole it was not bad. Then I was shown one of my designs with the bass filtering removed. Much closer to the Lyra. I was very surprised and agreed that high end speakers show it in a way that would never let me go back. I would imagine AR3X could be a cheap buy that could show this. Linn Sara also. Sara works very well with NAD3020 so not a silly idea ( external preamp for phono ).
I fully understand that people adding LP to a system that grew up with digital sources may not be able to do this. Speakers do seem more sensetive to warps than in my day. To be honest we wouldn't have been able to sell many of these speakers in the day. Port loading was thought a cheap design route with notable exceptions that still were thought suspect for head-bangers ( Spendor BC1 ). The use of simulation helped make them more respectable. Even then I have issues with that as often a small " error " sounds better. Celestion used a 500 uF capacitor in some speakers to both protect and offer a second order filter. I thought these worked very well. The press thought them to be yet another cheap design. This ignored Celestions engineering abilities.
What I would say is speakers that can cope might give you more than common ported designs into the bargain ( I hate the sudden death of the bass notes ). Try drinking straws in the port as did Celef. This raises the port frequency by about 10% and gives it better control. You might still need a gentle filter at 15 Hz.
The Quad ESL63 is not thought to be able to cope with real RIAA. They do !
LYRA - Lyra_Connoisseur 4-2 SE Line
On the whole it was not bad. Then I was shown one of my designs with the bass filtering removed. Much closer to the Lyra. I was very surprised and agreed that high end speakers show it in a way that would never let me go back. I would imagine AR3X could be a cheap buy that could show this. Linn Sara also. Sara works very well with NAD3020 so not a silly idea ( external preamp for phono ).
I fully understand that people adding LP to a system that grew up with digital sources may not be able to do this. Speakers do seem more sensetive to warps than in my day. To be honest we wouldn't have been able to sell many of these speakers in the day. Port loading was thought a cheap design route with notable exceptions that still were thought suspect for head-bangers ( Spendor BC1 ). The use of simulation helped make them more respectable. Even then I have issues with that as often a small " error " sounds better. Celestion used a 500 uF capacitor in some speakers to both protect and offer a second order filter. I thought these worked very well. The press thought them to be yet another cheap design. This ignored Celestions engineering abilities.
What I would say is speakers that can cope might give you more than common ported designs into the bargain ( I hate the sudden death of the bass notes ). Try drinking straws in the port as did Celef. This raises the port frequency by about 10% and gives it better control. You might still need a gentle filter at 15 Hz.
The Quad ESL63 is not thought to be able to cope with real RIAA. They do !
Last edited:
Walk in depth and better tonal colour. Mostly this needs baffle speakers or ones designed by lovers of ( German make Sonics for example ). This is a discussion I often had with designers. One needs to hear it to know it. Classical music in real life has these qualities. Andre Previn was well known for using the tricks of good hi fi to get a sound. Critics said he just doubled up the bass. I have to be a music snob now and say only classical or simple music can show this. The rest is taste. As Bob Stewrt said, we can only at best get one voice and one guitar in a domestic room and have hi fi. The rest is whatever compromise we like best. The Quads do help the room vanish.
One thing I seldom hear is Beethoven as it should be . Real life also. Disney got it right in the film Fantasia. It isn't stereo. It is a musical flower bed of grouped musicians that moves from bed to bed. It's nothing like what we get served on average. Real life Beethoven sounds woody like 78's ( low feedback designs ? ) with a hint of subtle >10 kHz treble. The same musicians playing Mahler later that night were modern with vast treble and no big woody qualities. It's all in how the music works and often is lost somewhere in the hi fi. Most often in the source material. I don't have answers for you as to how or why. Just to say learn what it is if you don't know. I constantly need to go to live music so that I don't unlearn. Much of real music is not that great. For example the soloist might be harder to hear than a recording.
One thing I seldom hear is Beethoven as it should be . Real life also. Disney got it right in the film Fantasia. It isn't stereo. It is a musical flower bed of grouped musicians that moves from bed to bed. It's nothing like what we get served on average. Real life Beethoven sounds woody like 78's ( low feedback designs ? ) with a hint of subtle >10 kHz treble. The same musicians playing Mahler later that night were modern with vast treble and no big woody qualities. It's all in how the music works and often is lost somewhere in the hi fi. Most often in the source material. I don't have answers for you as to how or why. Just to say learn what it is if you don't know. I constantly need to go to live music so that I don't unlearn. Much of real music is not that great. For example the soloist might be harder to hear than a recording.
Last edited:
Err.rrh! You compared your RIAA preamp against a LINE preamp
What incredible German engineering allowed it to act as a vinyl playback preamp?
I'm a musical snob too .. and don't agree with your comments on Herr Ludwig & Herr Gustav. But that belongs on another thread.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.