Flat Phono Preamp Based on John's P10 and 2SK389

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
scott wurcer
Posts: 1
Joined: Thu Jul 16, 2015 2:45 pm

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Post by scott wurcer » Fri Sep 11, 2015 2:55 pm

ricardo wrote:
It's not that all FIRs are EVIL. It's cos most are Linear Phase which is truly EVIL. Its the easiest way to design FIRs and most DAW people haven't taken Digital Filters 101.

Analogue stuff is usually Minimum Phase ... Phase as God intended. Among the Holy properties of Minimum Phase is that if you EQ a Minimum Phase response with another Minimum Phase response, both the Amplitude AND the Phase response become flat.

Practically all electromechanical/acoustic devices are Minimum Phase. (I can quote only 2 examples in speakers which are not which only proves the rule). Analogue circuits are usually Minimum Phase too. You have to go to a LOT of extra complexity to do otherwise.

The same goes for the common and easy methods to do IIRs ... most classic Bi-quads etc are Minimum Phase cos they try to emulate analogue stuff.
My article in Linear Audio Vol. 10 is out. I cover IIR and min phase FIR using a trick that I have not seen elsewhere. The IIR coefficients are tablated there for <.005dB accurate response even at 48K sampling for RIAA and most of the historical equalizations. I only posted the FIR's for RIAA but supply the Python code to do any you might want. The min phase FIR at 96K is only ~5k taps but does .000002 dB and .000001 Degrees even down to 20Hz.

ricardo
Posts: 262
Joined: Sun Apr 10, 2011 7:24 am

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Post by ricardo » Fri Sep 11, 2015 6:14 pm

scott wurcer wrote:My article in Linear Audio Vol. 10 is out. I cover IIR and min phase FIR using a trick that I have not seen elsewhere. The IIR coefficients are tablated there for <.005dB accurate response even at 48K sampling for RIAA and most of the historical equalizations. I only posted the FIR's for RIAA but supply the Python code to do any you might want. The min phase FIR at 96K is only ~5k taps but does .000002 dB and .000001 Degrees even down to 20Hz.
Dis beach might have to get a real job so he can buy some 'real life' books like Linear Audio. :o

Guru Wurcer, is this where you do the naive Linear Phase FIR and then throw away half of it? I've glossed over loadsa trivial (??!!) housekeeping in an attempt to keep up my charade as a pseudo DSP guru :mrgreen:

More seriously, are there any free Media Players that allow an Arbitrary IIR with any number of feedback & feedforward coeffs to be loaded and used conveniently? I'll be happy with a total of 200 coeffs. which will match a 2k FIR nicely but will need 500 coeffs to match Guru Wurcer's 5k taps.

You can do this in Plogue Bidule but its hardly a convenient Media Player for da unwashed masses. :lol:
Last edited by ricardo on Sun Sep 13, 2015 1:13 pm, edited 1 time in total.

User avatar
JR.
Posts: 2418
Joined: Sat May 24, 2008 7:21 pm

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Post by JR. » Sun Sep 13, 2015 10:17 am

Time to upgrade your tunrtable
Image

If that (Japanese) table is too pricey you can get a cheaper (?) German one for a few $10k less.

Image

Looney tunes.... The cheaper one cost more than my house.

JR

User avatar
mediatechnology
Posts: 3275
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Post by mediatechnology » Tue Sep 15, 2015 8:12 am

Yikes.
I wonder how much the preamps cost.

I got the flat preamp boxed: viewtopic.php?f=7&t=753&start=1
Looking for a way to add the liquid nitrogen fittings to cool the front end. :lol:

User avatar
mediatechnology
Posts: 3275
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Cartridge Upgrade and Cart Internal Grounding

Post by mediatechnology » Wed Sep 16, 2015 7:04 pm

My article in Linear Audio Vol. 10 is out. I cover IIR and min phase FIR using a trick that I have not seen elsewhere. The IIR coefficients are tablated there for <.005dB accurate response even at 48K sampling for RIAA and most of the historical equalizations. I only posted the FIR's for RIAA but supply the Python code to do any you might want. The min phase FIR at 96K is only ~5k taps but does .000002 dB and .000001 Degrees even down to 20Hz.
Scott - Just now saw your post. Thank you for joining us!
I guess I'll have to get a real job too and subscribe to Linear Audio.

I'm pretty happy with how DC8's RIAA sounded compared to CoolEdit's RIAA but I have a DC8 learning curve.
I had strange things happen with it truncating files but that might be because I was using 32 bit float as my test source.

Now that I have the preamp in a box I decided it was time to upgrade phono cartridges.
I did this on the cheap having two Stanton 681EE bodies and a 2/3 full three-pack of D6800EL styli given to me by a radio station many years ago.
I've basically been using the AT-96 for testing since it was also free.
The AT-120 from my youth needs a new stylus...

One thing I noticed about the AT-96 is that it has no grounding strap from either the left or right ground connection to the body.
All of my other carts do.
The ADC LME-2 head shell is plastic and galvanically isolated from any tonearm lead.
The AT-96 also has a plastic outer shell and plastic mounting ears.

The AT-96 tended to have - depending on arm position - electrostatic induced noise the left channel didn't have.
When the arm was hovering over the turntable all was well but on the tonearm rest it was about 3 dB noisier when the room's dimmer was on located about 3 feet from the cartridge.
Approaching the cartridge body or mounting screws the left channel remained silent but the right channel would have this slight noise increase.
I got the feeling that the AT-96 had it's internal shield or frame bonded to the right channel ground.
Playing a record both channels were quiet.

When I mounted the Stanton I purposely left the ground strap on.
Sure enough this cartridge had the same noise signature but in the opposite channel.
Touching the metal body produced bodacious amounts of hum in that channel which is what we should expect in a balanced and floating connection.
It unbalances the circuit.

The bonding clip to the cartridge metal outer body can be slid off the ground post to provide a true floating connection.
(The grounding clip can also be re-installed - the clip slips between the outer metal shell and the cart's plastic terminal block.)
When I removed the bonding strap both channels were equally quiet regardless of arm position.

For a true balanced connection the clip should be removed and the metal cartridge body allowed to float.
That way I get the best CMR in both channels.

The Stanton 681 sounds absolutely fantastic.
The AT-96 was OK but now that I'm done experimenting with the preamp it was worth the no-cost upgrade.

Image
Noise Signature of Balanced Input Phono Preamp Input Termination Stanton 681 on Turntable with RIAA EQ
0 dBFS is +8 dBu, Gain at 1 kHz approx 34 dB


Image
Noise Signature of Balanced Input Phono Preamp Input Termination Stanton 681 on Turntable Flat, No RIAA

User avatar
mediatechnology
Posts: 3275
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Comparing the NJM2068 to the OPA2134 in the Flat Preamp

Post by mediatechnology » Wed Sep 23, 2015 1:39 pm

I later found that the OPA2134 is actually better with a cart, compared to a resistor, due to its lower current noise.
In the midrange the NJM2068, NJM2114 or NE5532 out-perform the OPA2134 where it's higher voltage noise dominates.
Got some numbers or preferably some noise spectrum plots for this? Yes. I know it will vary with cartridge ;)
Got some numbers ricardo.

The following two captures are baseline measurements with a Stanton 681EE cart and the NJM2068DD installed in both channels.
The first is with RIAA EQ, the second flat.

Image
Phono Transfer System Balanced Input Preamp Stanton 681 RIAA EQ NJM2068DD In Both Channels

Image
Phono Transfer System Balanced Input Preamp Stanton 681 Flat NJM2068DD In Both Channels

These images are with the NJM2068 installed in the Left channel the OPA2134 installed in the right.

Image
Phono Transfer System Balanced Input Preamp Stanton 681 Flat EQ NJM2068DD Left Channel OPA2134 Right

Image
Phono Transfer System Balanced Input Preamp Stanton 681 RIAA EQ NJM2068DD Left Channel OPA2134 Right

The NJM2068 is slightly better in the midrange at 1/10th the cost of the OPA2134.
The OPA2134, considering that its a FET input with higher voltage noise nevertheless has outstanding performance.

The DC differential offset of both channels measured at the preamp output was <2mV so from a DC perspective they are also just about equal performers.

One more interesting factoid: The measurements above were made with the Phono Transfer System powered from a Meanwell P25A14E-R1B switcher.
The noise performance was identical to the linear bench supply.

emrr
Posts: 489
Joined: Sun Mar 02, 2008 11:27 pm
Location: NC, USA
Contact:

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Post by emrr » Wed Sep 23, 2015 2:00 pm

Those Meanwell supplies have not failed me yet.
Best,

Doug Williams
Electromagnetic Radiation Recorders

ricardo
Posts: 262
Joined: Sun Apr 10, 2011 7:24 am

Re: Comparing the NJM2068 to the OPA2134 in the Flat Preamp

Post by ricardo » Tue Sep 29, 2015 2:42 pm

I later found that the OPA2134 is actually better with a cart, compared to a resistor, due to its lower current noise.
In the midrange the NJM2068, NJM2114 or NE5532 out-perform the OPA2134 where it's higher voltage noise dominates.
Got some numbers or preferably some noise spectrum plots for this? Yes. I know it will vary with cartridge ;)
Got some numbers ricardo.

The following two captures are baseline measurements with a Stanton 681EE cart and the NJM2068DD installed in both channels.
The first is with RIAA EQ, the second flat.
.........
These images are with the NJM2068 installed in the Left channel the OPA2134 installed in the right.
.........
The NJM2068 is slightly better in the midrange at 1/10th the cost of the OPA2134.
The OPA2134, considering that its a FET input with higher voltage noise nevertheless has outstanding performance.
.........
The DC differential offset of both channels measured at the preamp output was <2mV so from a DC perspective they are also just about equal performers.
.........
Great stuff Wayne :)

Looks like NJM2068 on paper looks slightly better on paper (test gear?) than OPA2134 including the important 4kHz region. I sorta expected OPA2134 to do better in that region.

Were you trying to say the OPA2134 sounded quieter though it measured slightly worse? This is entirely possible and I've used this phenomena to cheat ... I mean present figures which more truly reflect the excellent user experience with my designs :D .

Do you have Inductance and Resistance for your 681EE? I make Stanton 680/1 (their previous top model) 910mH & 1k3

The one item we haven't added in this investigation is the effect of the optimum load capacitance. In the 1975 Boston Audio Society article I sent you as part of my MCamp.doc, you can clearly see the effect of this on V15 III which demands a very high capacitance.

I'm tempted to resurrect my spreadsheet based on that NI application note on MM noise :ugeek:

User avatar
mediatechnology
Posts: 3275
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Post by mediatechnology » Tue Sep 29, 2015 3:45 pm

Great stuff Wayne :)
Thank you!
Looks like NJM2068 on paper looks slightly better on paper (test gear?) than OPA2134 including the important 4kHz region. I sorta expected OPA2134 to do better in that region.
The NJM2068 was just a little bit better.
Were you trying to say the OPA2134 sounded quieter though it measured slightly worse? This is entirely possible and I've used this phenomena to cheat ... I mean present figures which more truly reflect the excellent user experience with my designs :D .
The OPA2134 originally measured worse than it actually was so it sounded almost as quiet as the NJM2068. But it turned out to measure really close so what I heard correlated to what I later measured. (Thanks to the benefit of shielding.)
I can't really A/B but I suspect the NJM2068 may still come in a tad quieter audibly based on its midrange advantage.
Do you have Inductance and Resistance for your 681EE? I make Stanton 680/1 (their previous top model) 910mH & 1k3
Stanton 681 927 mH 1.25K for another 681 found in the same box. (Am I lucky or what?)
The one item we haven't added in this investigation is the effect of the optimum load capacitance. In the 1975 Boston Audio Society article I sent you as part of my MCamp.doc, you can clearly see the effect of this on V15 III which demands a very high capacitance.
Well maybe we have....

I had a chance to check it with both 47 K and various capacitance.

This is 47K Ohm, 100pF added, approximately 100 pF cable, tonearm capacitance unknown. Total Cload approximately 200 pF.
Stanton 681 CBS STR-112 Group 1A.

Image
Stanton 681EE 200pF CLoad 47K Rload 1kHz 5cms Square Wave Lateral Modulation. Left channel is top trace.
I'm tempted to resurrect my spreadsheet based on that NI application note on MM noise :ugeek:
Let me scan what I have in the original Audio Handbook. It's completely different than the modern TI version and spread over two sections.

User avatar
JR.
Posts: 2418
Joined: Sat May 24, 2008 7:21 pm

Re: Comparing the NJM2068 to the OPA2134 in the Flat Preamp

Post by JR. » Tue Sep 29, 2015 3:46 pm

ricardo wrote:
Great stuff Wayne :)

Were you trying to say the OPA2134 sounded quieter though it measured slightly worse? This is entirely possible and I've used this phenomena to cheat ... I mean present figures which more truly reflect the excellent user experience with my designs :D .
:ugeek:
Back in the day I used a little battery powered gain stage for noise measurements. The 40 dB gain stage had IHF "A" (?) noise weighting curve built in. It ignored both low bass and very HF noise so made measurements look a bunch better than wide band, but it was also more representative of how human hearing sensitivity/response curves work. Back in the day I never used wide band measurements for tweaking noise.

JR

Post Reply