In the 1970s i moved the preamp into the plinth under the tonearm to minimize the capacitive loading. I still claim that you load the cartridge for a Q of 0.6 and forget the frequency response. The formula for the Q is on page 411 of the old Radiotron Designers Handbook so I don't know why everybody gets the formula wrong. The approximate formula is Q=Rl*SQRT(C/L) where Rl is the load resistor, C the total load capacitance and L the cartridge inductance. The cartridge and load can be modeled as a two pole lowpass filter and a Q of 0.577 gives you a phase linear response. You don't use peaking (Q over 0.707) in an attempt to improve the frequency response.
There is a thread about this subject, without the common mistake of just looking at the electrical part and ignoring the mechanical part:
https://www.diyaudio.com/community/...ts-with-individual-transfer-functions.397815/
https://www.diyaudio.com/community/...ts-with-individual-transfer-functions.397815/
Considerable peaking is built into the mechanical side of things, and is easily visible with very low inductance cartridges including moving coils. One that can't be designed around is the vinyl compliance x stylus effective moving mass resonance that appears within or just above audio frequencies (for best quality moving parts and stylus shapes). Loading a conventional MM cartridge with a very small resistor (small enough to put the LR pole above audible) and adding make-up gain is one way to see the raw mechanical response.
National Semiconductor, in their 1970s manuals, popularized the idea of choosing a load resistance that combined with the cartridge's source inductance to give the 75uS RIAA pole. This was improved upon by SY in his Linear Audio article (and probably on his website) by choosing a compromise load resistance that split the 75uS pole between loading LR and a pole/zero in the preamp, to optimize for noise. These loadings also swamp the LC resonance and give a more raw frequency response.
All good fortune,
Chris
National Semiconductor, in their 1970s manuals, popularized the idea of choosing a load resistance that combined with the cartridge's source inductance to give the 75uS RIAA pole. This was improved upon by SY in his Linear Audio article (and probably on his website) by choosing a compromise load resistance that split the 75uS pole between loading LR and a pole/zero in the preamp, to optimize for noise. These loadings also swamp the LC resonance and give a more raw frequency response.
All good fortune,
Chris
Then move the electronics into the cartridge, there's plenty of room for a wire-bonded module! I don't have access to that technology though. I have seen a design where the preamp PCB was the headshell, that would also work!
Cartridges with a built-in electronic cascade were used in the 70s ( Toshiba, Kenwood , etc . ) . And it has not gained much traction .
There are such cartridges now ( DS Audio). The use of such cartridges required a special preamp . This was not a standard RIAA preamp .
Turntables of classical design are good for their versatility . You can use many different phonopreamps . And it does not require the alteration of the turntable itself . People of our age ( 50-60-70 years old ) usually have several turntables , a lot of cartridges and phonopreamps . The versatility of this equipment is very important .
Built-in pre-amps are back on low-to-moderately priced turntables. Some also have USB (!). They have a bypass switch. I suspect the switch gets used to eliminate the confusion of selecting "AUX" when one wants "PHONO".
Ed
Ed
Last edited:
Back in the 1980s REVOX had an expensive turntable with a built-in preamp and now they are advertising one again, REVOX is not low priced.
"Once again, Revox is reintroducing itself to the hi-fi world with the new Studio Master T700. This turntable features a built-in phono amplifier."
"Once again, Revox is reintroducing itself to the hi-fi world with the new Studio Master T700. This turntable features a built-in phono amplifier."
Fun project!
The NE5534 is an amazing opamp for its time and is still amazing value for the money, but it does show the pitfalls of the older semiconductor processes. For example, while the AVOL typically is 100 dB it is only guaranteed to be above 83 dB once part-to-part variation and variations in supply voltage and temperature are factored in. That's right in the spec table on page 5 of the TI data sheet. Compare that with the OPA1611/12, which is guaranteed to reach 110 dB (114 dB under light load) minimum and 130 dB for a typical part. Suddenly you have 30ish dB more loop gain available.
Splitting the gain into multiple stages as you've been doing is another approach to get more available loop gain. Do that with the OPA1611 and you'll get even better performance.
You may find that you'll make better use of the available loop gain in a multiple-feedback filter rather than the Sallen-Key. TI has a good app note on the topic.
I do get that the OPA1611 is umpteen times more expensive than the NE5534, but if you're only building one phono stage I think it's reasonable to spend the extra money.
Do note that the OPA1611 is not well suited for an MC phono stage as it has pretty high current noise. But it's great for an MM stage as far as I can tell.
Tom
The NE5534 is an amazing opamp for its time and is still amazing value for the money, but it does show the pitfalls of the older semiconductor processes. For example, while the AVOL typically is 100 dB it is only guaranteed to be above 83 dB once part-to-part variation and variations in supply voltage and temperature are factored in. That's right in the spec table on page 5 of the TI data sheet. Compare that with the OPA1611/12, which is guaranteed to reach 110 dB (114 dB under light load) minimum and 130 dB for a typical part. Suddenly you have 30ish dB more loop gain available.
Splitting the gain into multiple stages as you've been doing is another approach to get more available loop gain. Do that with the OPA1611 and you'll get even better performance.
You may find that you'll make better use of the available loop gain in a multiple-feedback filter rather than the Sallen-Key. TI has a good app note on the topic.
I do get that the OPA1611 is umpteen times more expensive than the NE5534, but if you're only building one phono stage I think it's reasonable to spend the extra money.
Do note that the OPA1611 is not well suited for an MC phono stage as it has pretty high current noise. But it's great for an MM stage as far as I can tell.
Tom
Isn't it the other way around? - MC stages due to the lower impedance and low output benefit from lower voltage noise. MM stages have high inductance and also higher output, these care more about current noise.
Yes, for a 500 mH MM cartridge, when you look at the total RIAA- and A-weighted noise, each pA/√Hz of white current noise has about as much impact as 12 nV/√Hz of white voltage noise.
Nick Sukhov would say it's even worse in practice, as the spectral distribution is different. The inductive source emphasizes the treble content of the current noise, and he dislikes the sound of that.
Nick Sukhov would say it's even worse in practice, as the spectral distribution is different. The inductive source emphasizes the treble content of the current noise, and he dislikes the sound of that.
check out the schematic of the Phono-RIAA MM stage from "The Pre-Amp" (first preamp model of Musical Fidelity) in post #5 (first PDF) under
https://www.diyaudio.com/community/threads/head-pre-for-denon-dl-103.155449/
The internal differential amplifier stage of the NE5534 wasn't connected - instead this is to find an external differential amplifier realized around the dual NPN transistor LM394 (predecessor of MAT02) connected on the pins 1+8 of NE5534.
I've never seen that before, but not in newer models either.
Maybe Tim de Paravicini's approach for reducing noise.
https://www.diyaudio.com/community/threads/head-pre-for-denon-dl-103.155449/
The internal differential amplifier stage of the NE5534 wasn't connected - instead this is to find an external differential amplifier realized around the dual NPN transistor LM394 (predecessor of MAT02) connected on the pins 1+8 of NE5534.
I've never seen that before, but not in newer models either.
Maybe Tim de Paravicini's approach for reducing noise.
That's actually another thing from National Semiconductor, possible because of the unique pinout of 5534s. Another variation has a dual NJFET input stage. Mid to late 1970s.
All good fortune,
Chris
All good fortune,
Chris
It is true. And it is not only for bipolars @input with their base noise current, but also for low input "standard" resistance 47 кОhm which generate noise current spectral density sqrt(4kT/R) . PS. I have also recently measured the frequency response of the Ortofon 2M Red cartridge with my JFE2140 phono preamplifier [ https://www.diyaudio.com/community/...e-85-dba-sn-ratio.387375/page-22#post-7414712 ] with 25 pF / 150 kOhm input and got an extremely flat frequency response up to 25 kHz without any humps due to mechanical resonance (but not only with better FreqResp but also with a big "passive cooling" noise advantage of 150 kOhm input vs "standard" 47 kOhm/250 pF). See here up to the end and with english subtitles:
Last edited:
0.01% distortion at 20hz is nothing compared to distortions occurring in the cutting of the LP.
I set mine for high frequency gain no higher then 10.
I set mine for high frequency gain no higher then 10.
I am getting a little bit frustrated with some ideas here. I am an engineer who does calculations.
If we want a 75us time constant, which is an approx 2 kHz rolloff, we might do that with 500mH and about 6K resistance.
Time constant = RC or L/R.
But is this a good idea? I don't think so. It pushes headshells and a lot of tonearms beyond their limits.
Most of the Phono Cartridge's efforts harnlessy dissipate into heat. I have said this once, and I'll say it again, High Inductance MM heads are broken and unfixable.
If we want a 75us time constant, which is an approx 2 kHz rolloff, we might do that with 500mH and about 6K resistance.
Time constant = RC or L/R.
But is this a good idea? I don't think so. It pushes headshells and a lot of tonearms beyond their limits.
Most of the Phono Cartridge's efforts harnlessy dissipate into heat. I have said this once, and I'll say it again, High Inductance MM heads are broken and unfixable.
Last edited:
I'm not sure what you are saying here.
To carry a 20kHz signal 80dB above the thermal noise floor at room temp. requires 33nW of power, which is about 5.7mV rms for a 1k resistance - which is fairly representative of a lot of MM cartridges.
However receiving that information does not require dissipating that power, a JFET (ignoring input capacitance) might pull 10pA at 5mV and thus absorb 50fW, about a million times less power than is in the signal. The source-drain circuit of the JFET will require a power level of 33nW or more to pass on the signal without inevitably adding noise.
For a 10 ohm MC cartridge 33nW is 570µV at 57µA, and this can be observed by the input transistors of an AD797 for example which has an input impedance of about 7.5k ohms differential, drawing 0.76µA of signal current open loop, and much less closed loop, again absorbing a tiny fraction of the signal power.
The power put out by the cartridge is not wasted, its essential to carry the signal, but it doesn't have to be dissipated in the amplifier to pass the signal on.
The Nature of our Audio systems is to convince ourselves that we are listening to the real thing,
As a person who attended many live Prom concerts in the vast Royal Albert Hall in London, I assure you this is is a difficult thing to do.
Our home room acoustics can never emulate a Big Hall. You feel the vastness and bass of the Albert Hall in your gut.
Above that a certain amount can be done, IMO.
I like this sort of loudspeaker:
If a cone bass is considered optimal, why should a cone tweeter be considered less? Naturally, I'd recommend a more line array approach, an MTTM for sure.
A question of dispersion.
But all depends on the input signal in the end. This is what you hear in an ideal world.
I do think high inductance MM cartridges are rubbish. You might have a better time with low inductance Moving Iron designs like Grado.
Amplification is a well-settled subject. Most Amps work. Ancient designs suffered from slew-limiting, but modern semiconductor devices are quite fast.
What I am saying is don't sweat amplifiers. Put your effort into transducers and damping your room. Hope it helps.
Best Regards, Steve in Portsmouth UK.
As a person who attended many live Prom concerts in the vast Royal Albert Hall in London, I assure you this is is a difficult thing to do.
Our home room acoustics can never emulate a Big Hall. You feel the vastness and bass of the Albert Hall in your gut.
Above that a certain amount can be done, IMO.
I like this sort of loudspeaker:
If a cone bass is considered optimal, why should a cone tweeter be considered less? Naturally, I'd recommend a more line array approach, an MTTM for sure.
A question of dispersion.
But all depends on the input signal in the end. This is what you hear in an ideal world.
I do think high inductance MM cartridges are rubbish. You might have a better time with low inductance Moving Iron designs like Grado.
Amplification is a well-settled subject. Most Amps work. Ancient designs suffered from slew-limiting, but modern semiconductor devices are quite fast.
What I am saying is don't sweat amplifiers. Put your effort into transducers and damping your room. Hope it helps.
Best Regards, Steve in Portsmouth UK.
Last edited:
Besides using unavoidably crappy parts this means passing another set of wires, presumably thicker than the signal tonearm wires to carry power. Mechanically this is not a great idea for a well designed tonearm, even the single set of wires is a serious technical challenge for tonearms with good bearings. And then there is the issue of damping the board, so it doesn't add all sorts of unpleasant sounds to the tonearm tube. Even more additional mass. It's not like a think tonearm mass is very important, but still...
But don't let some audiophile mumbo jumbo stop you. The world is awaiting the first USB turntable with a phono stage + ADC mounted inside a tonearm 😎
If you are not into extremely low distortion and want to keep the load on the cartridge as small as possible, you could mount a dual JFET in the headshell. You then have one JFET per stereo channel that you connect with its gate to the cartridge, drain to the centre conductor of the cable, source to ground (cable shield and the other side of the cartridge).
On the other end of the cable, you put a RIAA-equalized transimpedance amplifier per stereo channel, designed to have a DC input voltage high enough to supply the JFET. For minimum distortion at only a small noise penalty, the JFET should be biased a little bit into the triode region.
On the other end of the cable, you put a RIAA-equalized transimpedance amplifier per stereo channel, designed to have a DC input voltage high enough to supply the JFET. For minimum distortion at only a small noise penalty, the JFET should be biased a little bit into the triode region.
Some curious TT related matters.
https://www.vinylengine.com/turntable_forum/viewtopic.php?t=111297
https://archive.ph/x8hxz
https://calypso-tonearm.blogspot.com/
Use of remote headshell JFET's suggestion by @Preamp
https://www.diyaudio.com/community/threads/diy-tonearm-from-pcb-material.393271/post-7216422
https://www.vinylengine.com/turntable_forum/viewtopic.php?t=111297
https://archive.ph/x8hxz
https://calypso-tonearm.blogspot.com/
Use of remote headshell JFET's suggestion by @Preamp
https://www.diyaudio.com/community/threads/diy-tonearm-from-pcb-material.393271/post-7216422