Interesting, I've been contemplating a similar front-end for some time - have you built it ?
Guessing that C8 & C14 is cable capacitance ?
Seems like quite a large cap to hang on pin 5 of the AD844 ?
My biggest worry is what kind of voltage one will get between the two inverting inputs since we're running without feedback.
...and how the output buffer will behave without feedback.
The rest of the circuit I've yet to wrap my head around, got any links for those 'enhanced RIAA ideas' ?
Very interesting circuit 🙂
This RIAA is built "live" and works normally - ie. good.
C8 and C14 form, together with the cable capacity, the upper limit frequency of the circuit.
The capacitor together with the resistor to pin 5 (TZ ) forms the time constant for high frequencies (incl. Neumann correction!).
And from there we do the desired amplification of this first stage.
The voltage between the two inverting inputs is what comes out of the MC cartridge.
Now I use Rondo Bronze. It used to be Benz Micro Glider.
You can make sure everything works without feedback by using some simulation program.
C8 and C14 form, together with the cable capacity, the upper limit frequency of the circuit.
The capacitor together with the resistor to pin 5 (TZ ) forms the time constant for high frequencies (incl. Neumann correction!).
And from there we do the desired amplification of this first stage.
The voltage between the two inverting inputs is what comes out of the MC cartridge.
Now I use Rondo Bronze. It used to be Benz Micro Glider.
You can make sure everything works without feedback by using some simulation program.
Hi Step, thanks for contribution.
I'm also confused about function of first stage. How do you calculate gain of U1 and U2?
If for instance you connected 0.3mV - 5Ohm cartridge, what voltage do you get between points 39 and 40?
I'm also confused about function of first stage. How do you calculate gain of U1 and U2?
If for instance you connected 0.3mV - 5Ohm cartridge, what voltage do you get between points 39 and 40?
Looking at the AD844 datasheet, the input impedance of the inverting input is 50 ohms typical. So the cartridge is looking into a pseudo short circuit - so the current noise dominates.
Again from the datasheet, the current from pin 5 is identical to the input current. So putting a resistor R say from that node to ground defines the gain as R/(Rc +50) single ended where Rc is the cartridge resistance.
The RC network attached to pin 5 seems to deal with the RIAA 75uS time constant (74.8us with the component values)
Again from the datasheet, the current from pin 5 is identical to the input current. So putting a resistor R say from that node to ground defines the gain as R/(Rc +50) single ended where Rc is the cartridge resistance.
The RC network attached to pin 5 seems to deal with the RIAA 75uS time constant (74.8us with the component values)
Apart from the 75us thing, this is similar to Borbely's Proteus design (Linear Audio V6, Sept 2013). This is discrete, and has a current input. But an internal high impedance node (taken from joined transistor collectors) has a resistance Rfb to ground. The gain is then Rfb/Rpu with Rpu being the cartridge resistance.
Hi Drbulj,
The gain of U1 and U2 depends on the resistance of the cartridge and the resistor at point 5(TZ ).
In your specific example (0.3mV - 5Ohm) - we will get 5.8mV in points 39 and 40 (for each separately). Or respectively between the two points 11.16mV.
Yes - the cartridge "sees" 50 ohms.
Of course the gain depends on the resistance of the cartridge and the resistor at point 5(TZ).
A few comments:
1)There is no official pole for von Neumann, but Stereophile assumes 3.18usec is the most logical, where you used 5.5usec.
2) Because your 0.3mV 5R cart is in series with 2x50R, you will see 21mV between both AD844 outputs, because
0.3m/(2*50R +5R)=2.8uA and 2.8uA*(3K4+270)= 10,5mV per AD844
3) In that setting, this Cart will give you a weighted S/N of 37dBA.
Hans
1)There is no official pole for von Neumann, but Stereophile assumes 3.18usec is the most logical, where you used 5.5usec.
2) Because your 0.3mV 5R cart is in series with 2x50R, you will see 21mV between both AD844 outputs, because
0.3m/(2*50R +5R)=2.8uA and 2.8uA*(3K4+270)= 10,5mV per AD844
3) In that setting, this Cart will give you a weighted S/N of 37dBA.
Hans
The configuration is perhaps more suited to high R moving coil cartridges - for example the DL103 with 40R.
Incidentally I just measured the inductance of my spare DL103 on my AC bridge at 1kHz. It is 100uH (actually 105uH on my 0.1% accuracy bridge).
Craig
Incidentally I just measured the inductance of my spare DL103 on my AC bridge at 1kHz. It is 100uH (actually 105uH on my 0.1% accuracy bridge).
Craig
I don’t think so.
140R produces almost twice as much noise as 40R, a total waste of S/N.
Not to have this additional noise penalty of the 2*50R, the Cart itself should have a much higher resistance for this circuit, it’s more likely suited for an MM.
Hans
Can't argue with that Hans. The current noise might be low, but the Johnson thermal noise in the 50 ohm input R absolutely dominates.
For MM the design requirements are totally different of course
For MM the design requirements are totally different of course
Hi,
for the 2nd, 3rd stages I wouldn´t use AD797 or LT1028 which are both optimized for low impedances (and preferably balanced impedances in the LT1028 case), but rather modern lownoise JFET-Input OPAmps which might be superior for filter applications.
For U5 I´d suggest to have a look at balanced line receivers like INA134/137, or equivalents from THATCorp.
jauu
Calvin
for the 2nd, 3rd stages I wouldn´t use AD797 or LT1028 which are both optimized for low impedances (and preferably balanced impedances in the LT1028 case), but rather modern lownoise JFET-Input OPAmps which might be superior for filter applications.
For U5 I´d suggest to have a look at balanced line receivers like INA134/137, or equivalents from THATCorp.
jauu
Calvin
A moving coil in a magnetic field is definitely a voltage source. Induced voltage is proportional to rate of change of flux linkage. A current source's output voltage would rise without limit if unloaded...
Yes - it is. But there is also an "original" and serious play on words - which is true!
What Carlos Candeias said there is in the logic of my words:
https://www.stereophile.com/content/bmc-phono-mcci-phono-preamplifier
Last edited:
Hi,
@StepResponse: there are numerous JFET OPAmps around, or correctly JFET-Input OPAs.
TI lists many quite renowned 'Audio-OPAs' like OPA134/2134/4134, OPA604/2604, OPA627/637, OPA827, and more actual types .... OPA1641/1642, OPA1656 iIrc.
Do they sound better than bipolar OPAs? Don't know about the tastes of anyone else, but I like them and they are easy and rather non-complicated to implement.
Their input noise is low enough to cater for MMs and the tiny input bias allows for direct coupling of the source without affecting it.
This and the extremly low current noise figures are a strong pro in filter applications and low-offset/high-gain applications.
jauu
Calvin
@StepResponse: there are numerous JFET OPAmps around, or correctly JFET-Input OPAs.
TI lists many quite renowned 'Audio-OPAs' like OPA134/2134/4134, OPA604/2604, OPA627/637, OPA827, and more actual types .... OPA1641/1642, OPA1656 iIrc.
Do they sound better than bipolar OPAs? Don't know about the tastes of anyone else, but I like them and they are easy and rather non-complicated to implement.
Their input noise is low enough to cater for MMs and the tiny input bias allows for direct coupling of the source without affecting it.
This and the extremly low current noise figures are a strong pro in filter applications and low-offset/high-gain applications.
jauu
Calvin
In the case of a MM input, the best and least expensive show in town is the good old NE5534. That has an excellent combination of voltage and current noise that suits the MM cartridge impedance of mid hundreds of ohms and a fair chunk of a Henry inductance.
The only opamp that betters it is the much more recent OPA210, with a dual OPA2210. It has the same current noise as the 5534, around half the voltage noise, and lower frequency 1/f corner. It is also 3x more expensive, and in practical terms only finds you less than 1dB improvement in SNR with real world cartridge models at the input.
So bang for buck the NE5534 is still the best show in town for MM.
For MC, Bob Cordell's VinylTrak takes some beating. Discrete with FET input. Single ended, but here he suggests a balanced option https://www.cordellaudio.com/preamplifiers/vinyltrak.shtml
Not current input of course - VinylTrak is a voltage amplifier.
I've mentioned before that Borbely's Proteus is a discrete current mode phono stage with passive EQ. Without feedback it has an input R of around 3 ohms - so with my Zu-audio DL103 at 40 ohms that is pretty much OK. There is a feedback option that takes the input R to very low indeed, and suits lower resistance MC cartridges.
The only opamp that betters it is the much more recent OPA210, with a dual OPA2210. It has the same current noise as the 5534, around half the voltage noise, and lower frequency 1/f corner. It is also 3x more expensive, and in practical terms only finds you less than 1dB improvement in SNR with real world cartridge models at the input.
So bang for buck the NE5534 is still the best show in town for MM.
For MC, Bob Cordell's VinylTrak takes some beating. Discrete with FET input. Single ended, but here he suggests a balanced option https://www.cordellaudio.com/preamplifiers/vinyltrak.shtml
Not current input of course - VinylTrak is a voltage amplifier.
I've mentioned before that Borbely's Proteus is a discrete current mode phono stage with passive EQ. Without feedback it has an input R of around 3 ohms - so with my Zu-audio DL103 at 40 ohms that is pretty much OK. There is a feedback option that takes the input R to very low indeed, and suits lower resistance MC cartridges.
Hi StepResponse,
These days I'm on holidays and not so much at computer, but looking interesting conversation on phone sometimes. I see much more experienced members than me commented already.
Personally I'm not familiar or friend of video CFB opamps for such use, neither I know much about them. But if you see benefits that I don't see, might be AD811 with Rin of only 14 Ohm could be better choice here. That one is proven as excellent I/V converter for DAC, but there it is dealing with completely different sort of signal.
Further, some comments :
These days I'm on holidays and not so much at computer, but looking interesting conversation on phone sometimes. I see much more experienced members than me commented already.
Personally I'm not familiar or friend of video CFB opamps for such use, neither I know much about them. But if you see benefits that I don't see, might be AD811 with Rin of only 14 Ohm could be better choice here. That one is proven as excellent I/V converter for DAC, but there it is dealing with completely different sort of signal.
Further, some comments :
- Why 3 stages, amplification and RIAA can be comfortably fitted in 2
- Why R42-43 and C8-14? capacitors should be replaced with higher value resistors to loosely reference inputs to gdn, IMO. MC doesn't need capacitive load.
- If R9 and R6 are connected to each other and not ground (better use just one 500 Ohm) only much better CMMR will be achieved with no downsides
- Same probably applies for R11 and R19, but I am not familiar with special pin5 of ad844, so cant say for sure
- Also not sure what is the purpose of C2 C3 R7