Arjen,
When you say Input Riaa shorted, do you mean directly at the preamp's input or shorted at the head shell.
You should in fact shorten at the head shell, this can make a big difference in picking up EMI signals.
This will much better discriminate with the situation where without the short the Cart is driving the wires.
In your schematic the phono preamp's gnd is connected to 5, which should be the case when using the TT's powers supply.
But when using batteries, you can disconnect from 5 and make the whole phono preamp floating.
This galvanic separation could or could not give you an advantage.
But as Marcel suggested, it's also very informative to see what the spectrum does when moving the arm to another position.
Hans
When you say Input Riaa shorted, do you mean directly at the preamp's input or shorted at the head shell.
You should in fact shorten at the head shell, this can make a big difference in picking up EMI signals.
This will much better discriminate with the situation where without the short the Cart is driving the wires.
In your schematic the phono preamp's gnd is connected to 5, which should be the case when using the TT's powers supply.
But when using batteries, you can disconnect from 5 and make the whole phono preamp floating.
This galvanic separation could or could not give you an advantage.
But as Marcel suggested, it's also very informative to see what the spectrum does when moving the arm to another position.
Hans
Should it? The biggest spur with shorted cartridge is -119 dBV at the phono preamplifier output at 1 kHz. That's just over 1 µV at a nominal signal level of 150 mV.
When you measure over small enough resolution bandwidths, you always end up seeing spurious signals if there are any being generated anywhere. It helps to have poor measuring equipment that can't measure over very small bandwidths, then you don't see the spurs and also don't need to worry about them 😉
When you measure over small enough resolution bandwidths, you always end up seeing spurious signals if there are any being generated anywhere. It helps to have poor measuring equipment that can't measure over very small bandwidths, then you don't see the spurs and also don't need to worry about them 😉
Last edited:
Exactly 1kHz and multiples... or sometimes 8kHz and multiples ==> could be USB packet noise, an artifact of your soundcard+cabling, not the RIAA preamp. As usual, the test is to short the output of the RIAA pre. Since this sort of test is needed all the time I've made myself a few RCA cables with a short in the plug (and leaving the center pin unconnected), and that must go to the source side of things. The error scales with the shield resistance of the cable used.
This kind of error, similar to GND loops, is best eliminated with an USB isolator (and of course a floating soundcard supply, if applicable).
I also have some plugs that are shorted inside, so I can test this.
I couldn't agree more. In 'ye olde days' I was already satisfied if my build didn't pick-up a high power SW radio station. Nowadays we can look > 120 dB into the signal and 'see' things that I probably won't hear. I spent most of this Sunday trying to improve my build with some of the advice you guys gave, where the result was 'more of the same'. In the end, one of my two RIAA channels stopped working. which seems to come from one of the two muting relays.
So, where I could have been listening to music, I am now somewhat frustrated with the lack of results from my efforts. Sometimes 'good enough' should be good enough. That was one of the reasons I went for single supply, single stage RIAA design with a simple NE5532. I will call this a day and focus on putting the new transformer in my turntable (did I mention the TT is originally 120V~ brought back to a 230V~ environment - and wanting to get rid of the auto tranny?), try to fix the 'broken' channel, and start enjoying music.
Thanks to everyone who has contributed, especially to @MarcelvdG for his novel design.
This is a new incarnation of MarvelvdG's RIAA pre-amp, this time with 3rd order HPF. I've built it into my integrated amp (Denon PMA-600NE) replacing the original circuitry. I bought the amp just some 6 months ago. I suppose my warranty is now void. 🙂 The new set-up replaces the previous where I had put the RIAA amp in the turntable. The reason is that I wasn't satisfied with the 50 Hz hum and its switch-off behaviour, despite the muting circuit.
New schematic:
Schematic
Please note that there is around 10 dB attenuation (R4 to R5//Rin) to have a more agreeable volume control.
C4 is at 68 pF. My tone-arm wiring and connecting cable together have 115 pF making the total around 180 pF.
Pictures:
Original situation
Modified to 3rd order @ 16 Hz HPF
Noise performance:
For this measurement the turntable + cartridge are connected to the RIAA amp. Rec-out of the integrated amp is going into a + 20 dB head-amp feeding into my soundcard. Since the output of the RIAA amp is attenuated by 10 dB, the total 'lift' is +10 dB. Since the noise floor of my sound card is too high, the noise at higher frequencies is drowning into the noise floor of the sound card, but the 50 Hz hum is clearly visible. The 50 Hz hum in the previous set-up was at -80 dB, now it is at around -93 dB, so that's an improvement. I could only get that in the previous set-up battery feeding the RIAA amp. - EDIT on a second look, I previously made measurements with the motor-platter running, and with the motor-platter off. I don't remember what I did here. Probably the motor-platter was off. I need to re-check this with the motor-platter running.
Cart + RIAA amp + 20 dB head-amp - motor platter running (or off)
Record playing silent passage:
All the noise drowns in the record groove noise.
Bode plot:
The bode plot (via an anti RIAA circuit) is not normalized (yet). I am surprised that I am seeing the response rise again, despite the implementation of an RC filter with a corner frequency of 67 kHz (R4 - C9). I might re-run this measurement to make sure there were no issues. Since the amp is a dog to take apart, I have no intention of fixing it, but perhaps someone can shed their light on why I am seeing this.
Bode plot
New schematic:
Schematic
Please note that there is around 10 dB attenuation (R4 to R5//Rin) to have a more agreeable volume control.
C4 is at 68 pF. My tone-arm wiring and connecting cable together have 115 pF making the total around 180 pF.
Pictures:
Original situation
Modified to 3rd order @ 16 Hz HPF
Noise performance:
For this measurement the turntable + cartridge are connected to the RIAA amp. Rec-out of the integrated amp is going into a + 20 dB head-amp feeding into my soundcard. Since the output of the RIAA amp is attenuated by 10 dB, the total 'lift' is +10 dB. Since the noise floor of my sound card is too high, the noise at higher frequencies is drowning into the noise floor of the sound card, but the 50 Hz hum is clearly visible. The 50 Hz hum in the previous set-up was at -80 dB, now it is at around -93 dB, so that's an improvement. I could only get that in the previous set-up battery feeding the RIAA amp. - EDIT on a second look, I previously made measurements with the motor-platter running, and with the motor-platter off. I don't remember what I did here. Probably the motor-platter was off. I need to re-check this with the motor-platter running.
Cart + RIAA amp + 20 dB head-amp - motor platter running (or off)
Record playing silent passage:
All the noise drowns in the record groove noise.
Bode plot:
The bode plot (via an anti RIAA circuit) is not normalized (yet). I am surprised that I am seeing the response rise again, despite the implementation of an RC filter with a corner frequency of 67 kHz (R4 - C9). I might re-run this measurement to make sure there were no issues. Since the amp is a dog to take apart, I have no intention of fixing it, but perhaps someone can shed their light on why I am seeing this.
Bode plot
I was thinking the same thing but hum and buzz was a problem. The SSM2017 low noise chip may work.
Coincidently I also went with the excellent NJM2068 in external RIAA.
Dead quiet almost as good as CD. Groove noise is much higher high frequency hiss is virtually eliminated.
It's an outstanding chip for MM.
I am using AT VM95C and ML.
Coincidently I also went with the excellent NJM2068 in external RIAA.
Dead quiet almost as good as CD. Groove noise is much higher high frequency hiss is virtually eliminated.
It's an outstanding chip for MM.
I am using AT VM95C and ML.
Last edited:
C501 has a negligible impedance at ultrasonic frequencies, but you have to account for the rest. That is, the parallel value of R4, R5 and Rin together with C9 + CCZN (where CZN stands for condensator zonder naam, capacitor without name) determine the low-pass filter time constant.
That explains the error, because I only accounted for R4 & C9. C(zn) is of an unknown value, but I take small. Once (if) I'm feeling like it, I might correct it, otherwise, it sound good to me as is.
@stocktrader200 Thanks, my cart is an Ortfon Red.
@stocktrader200 Thanks, my cart is an Ortfon Red.
The Ortofon had a great frequency response on the other site. I am using 100k as the load which improves the high end.
I have re-measured the noise floor with the motor-platter running, and it is identical to the measurement with the motor-platter off. So, 50 Hz hum is down to -93 dB coming from -80 dB, so that's an improvement.
This concludes my adventures with MarcelvdG's RIAA pre-amp design. Thanks to everyone for their input.
This concludes my adventures with MarcelvdG's RIAA pre-amp design. Thanks to everyone for their input.
It would be great to see an FFT plot of a loud music playing on a record to compare with the silent groove plot. Nice work. Thanks for sharing.
Your wish, etc.
Below is the comparison of some loud piece of music playing compared with the silent groove.
Loud music playing: White trace = 16 averages, Red trace = Peak hold
Silent groove
For the above I could only get 9 averages, because after that: The music started.
Below is the comparison of some loud piece of music playing compared with the silent groove.
Loud music playing: White trace = 16 averages, Red trace = Peak hold
Silent groove
For the above I could only get 9 averages, because after that: The music started.
Last edited:
Very cool. Thanks for measuring and posting that. Looking at the curves at 1 kHz, the noise is maybe - 93 dB and the signal is maybe - 33 dB for an SNR of about 60 dB. That's as good as any measured performance I've seen for record playback. You've got to be happy with your preamp noise at around -125 dB at 1 kHz being 32 dB below the record surface noise at -93 dB. So that's about as good as it gets.
I captured signal and noise graphs for a not so perfect Jethro Tull album.
A sort of loud section of track 1, from Storm Watch
Red: Record surface noise Black/Grey: Needle off record.
I was seeing around -125 preamp noise vs. -45 signal at 1 kHz. So my preamp noise was about 80 dB below the signal at 1 kHz and the surface noise was -90 db vs -45 db signal so maybe 45 dB SNR for the record playback of this old album. My preamp has a flat response, so that gives different curves here.
I captured signal and noise graphs for a not so perfect Jethro Tull album.
A sort of loud section of track 1, from Storm Watch
Red: Record surface noise Black/Grey: Needle off record.
I was seeing around -125 preamp noise vs. -45 signal at 1 kHz. So my preamp noise was about 80 dB below the signal at 1 kHz and the surface noise was -90 db vs -45 db signal so maybe 45 dB SNR for the record playback of this old album. My preamp has a flat response, so that gives different curves here.
Last edited:
Because of RIAA EQ the high frequency hiss is less then low frequency noise. The lower frequency noise is masked almost perfectly by the music. It's the > 5khz hiss that bothers me most. The type heard when playing cassette or weak FM signals
I wanted to share some more measurements comparing my MarcelvdG's RIAA amp which has a 3rd order HPF @ 16Hz, vs. an amp with a 1st order RIAA HPF @ 5Hz. In most commercial integrated amps, there is usually a 1st order roll-off in the RIAA amp only. The Denon amp that I've built the 3rd order HPF version in, originally had a 1st order HPF @ 3.3 Hz. Another amp that I own (an old Onkyo) has it's roll-off at 1.75 Hz. My 'bench test amp', an 80's Pioneer, has a 1st order roll-off @ 5 Hz. I thought it would be interesting to test this bench amp against MarcelvdG's design and see what the differences are.
You need to ignore the absolute values on the Y-axis, since both RIAA amps have different gains. I tried making them look similar by changing the Y-axis scaling.
I've captured two different sound tracks for each amp; a piece of silent groove, and a piece of music, trying to keep these pieces identical for both tests.
Below are the results:
Silent groove - 1st order HPF @ 5Hz (Pioneer bench amp)
You can see that the average level at 6 to 7 Hz is 67 dB over the average level at 1 kHz.
Silent groove - 3rd order HPF @ 16 Hz (Denon / MarcelvdG)
The average peak at 8 to 9 Hz is around 48 dB above the average level at 1 kHz. That's 20 dB better than the amp with a 1st order HPF @ 5 Hz.
Next, let's look at some piece of music:
Music - 1st order HPF @ 5Hz (Pioneer bench amp)
There is a peak at around 7 Hz, which is around 7 dB down in level compared to the highest averaged peak below 100 Hz.
Now let's look at what a 3 order HPF @ 16 Hz brings:
Music - 3rd order @ 16 Hz (Denon / MarcelvdG)
The peak at 7 Hz is still there but it's around 24 dB down in level compared to the highest averaged peak below 100 Hz. That's around a 17 dB drop in level compared to the situation with a 1st order HPF @ 5 Hz.
For the comparison, I have assumed that both amps are flat in the pass-band. For accuracy, I should have shown the RIAA transfer curve of each amp too. I did test the version with the 3rd order HPF @ 16 Hz (see post #187 above), but have not tested tested my bench amp. I might do that at a later moment, and post the result below.
You need to ignore the absolute values on the Y-axis, since both RIAA amps have different gains. I tried making them look similar by changing the Y-axis scaling.
I've captured two different sound tracks for each amp; a piece of silent groove, and a piece of music, trying to keep these pieces identical for both tests.
Below are the results:
Silent groove - 1st order HPF @ 5Hz (Pioneer bench amp)
You can see that the average level at 6 to 7 Hz is 67 dB over the average level at 1 kHz.
Silent groove - 3rd order HPF @ 16 Hz (Denon / MarcelvdG)
The average peak at 8 to 9 Hz is around 48 dB above the average level at 1 kHz. That's 20 dB better than the amp with a 1st order HPF @ 5 Hz.
Next, let's look at some piece of music:
Music - 1st order HPF @ 5Hz (Pioneer bench amp)
There is a peak at around 7 Hz, which is around 7 dB down in level compared to the highest averaged peak below 100 Hz.
Now let's look at what a 3 order HPF @ 16 Hz brings:
Music - 3rd order @ 16 Hz (Denon / MarcelvdG)
The peak at 7 Hz is still there but it's around 24 dB down in level compared to the highest averaged peak below 100 Hz. That's around a 17 dB drop in level compared to the situation with a 1st order HPF @ 5 Hz.
For the comparison, I have assumed that both amps are flat in the pass-band. For accuracy, I should have shown the RIAA transfer curve of each amp too. I did test the version with the 3rd order HPF @ 16 Hz (see post #187 above), but have not tested tested my bench amp. I might do that at a later moment, and post the result below.
Below are the bode plots of the two amps that I tested. I tried normalizing both to 0 dB @ 1 kHz, but the second one I could not get closer due to limitations of my test set-up.
The improvement in low frequency content that was observed in the above post can also be found by looking at the bode plots. There is around a 20 dB difference in gain at 7 Hz between these two amps.
What I was curious about was how much low frequency content can be present in a record. And there certainly is.
To me, these tests show that you're better off with a RIAA amp with some steep HP filtering.
1st order HPF @ 5Hz (Pioneer bench amp)
3rd order HPF @ 16 Hz (Denon / MarcelvdG)
The improvement in low frequency content that was observed in the above post can also be found by looking at the bode plots. There is around a 20 dB difference in gain at 7 Hz between these two amps.
What I was curious about was how much low frequency content can be present in a record. And there certainly is.
To me, these tests show that you're better off with a RIAA amp with some steep HP filtering.
1st order HPF @ 5Hz (Pioneer bench amp)
3rd order HPF @ 16 Hz (Denon / MarcelvdG)
Informal testing indicates my new preamp noise is about 20 ish dB below silent groove noise. Njm2068 adds virtually no high frequency hiss. It's also working with LPF at 20hz as a single chip. I used the OPA1656 RIAA circuit which is excellent. 100k loading is also revealing. 👌
What I should have said is that I was curious about how much low frequency content is present in the signal. In my case, the 7 - 9 Hz bump seems to be a cartridge / arm resonance, because when I tap the turntable, I can see that peak jump up in level. It is there, whatever the origin, and it's good to address it before it reaches my speakers.
PS - I also looked at a Philips amp I once owned and saw it's RIAA amp had a 19 Hz 1st order HPF filter. Philips, in my view, always put more practical thinking in their approach to spec's than their Japanese counterparts.
Last edited: