Phono Preamps

A resource for technical documentation. Datasheets, application notes, instruction manuals, books and links to resources are found in the Document Library.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Phono Preamps

Post by mediatechnology »

John -

That's a nice circuit. I started liking it from R26 and R27 onward. It's not often you see series resistors to limit ripple current but the capacitors sure appreciate it.

Wayne
User avatar
JR.
Posts: 3697
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Phono Preamps

Post by JR. »

I'd love to claim some grand design scheme, but i probably had more voltage than I needed, and I didn't mind spreading out the peak current over more dwell. It didn't need to be 2 resistors either but it probably made the layout work better or something like that.

Several things about that design are way over engineered or just symmetrical with no real benefit (like output driver).

I was most proud of input stage that was open loop (sort of)... closed loop for DC servo, but no NF for HF transients. The raw input gets rolled off by RIAA 75 uSec pole..

You could hit that input with lightning and it wouldn't overload, until about 2-3V pk-pk. :lol:

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Phono Preamps

Post by mediatechnology »

I want to experiment with the technique of transcribing LPs with flat preamplification and performing hiss, click and pop removal in DSP prior to application of DSP-based RIAA de-emphasis. People have reported good results doing it this way since the surface defects which cause clicks, ticks and pops are HF events. The reasoning is that it's easier for DSP applications to remove with less artifacts. I've used Virtos before on de-emphasized recordings and I'd like to see how it performs with RIAA equalization after Virtos.

This may require recording at reduced level, say -20 dB, to prevent overload. I have already obtained the RIAA de-emphasis filter files for Cool Edit 2000. What do you suppose would be a reasonable preamp gain if there is no RIAA de-emphasis using a typical MM cartridge?
User avatar
JR.
Posts: 3697
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Phono Preamps

Post by JR. »

mediatechnology wrote:I want to experiment with the technique of transcribing LPs with flat preamplification and performing hiss, click and pop removal in DSP prior to application of DSP-based RIAA de-emphasis. People have reported good results doing it this way since the surface defects which cause clicks, ticks and pops are HF events. The reasoning is that it's easier for DSP applications to remove with less artifacts. I've used Virtos before on de-emphasized recordings and I'd like to see how it performs with RIAA equalization after Virtos.
I had several customers use my old P-10 this way. As I understood it, de-clicking before the RIAA playback EQ meant you were dealing with shorter events. Running the brief click transient through the LPF of RIAA, spreads out the energy in time, making it harder to remove, and involves punching a larger hole in the sound when the smeared out click is muted. Further, depending on the click mute strategy, there may be HF edges from muting the click, that the following LPF of the RIAA EQ will smooth out.
This may require recording at reduced level, say -20 dB, to prevent overload. I have already obtained the RIAA de-emphasis filter files for Cool Edit 2000. What do you suppose would be a reasonable preamp gain if there is no RIAA de-emphasis using a typical MM cartridge?
You may want to confirm results with your favorite cart, but I wasn't the only designer to use flat gain before an inverting EQ stage, while my topology for the EQ was unique AFAIK.

Looking at the Jung preamp you posted he took a more conservative approach than I with a gain of only 6x in the flat stage. I used 3x that gain from a raw clipping POV, while my dual outputs doubled that gain again. In hindsight that differential amp probably should have run at -6dB to normalize clipping across the entire front end. FWIW I don't recall one customer complaint about front end overload.

IIRC nominal cartridge output voltages were mV level at 1kHz, so applying inverse RIAA eq would make peak input voltages around tens of mV at 20 kHz. It might be instructive to look at actual cart output in response to clicks. but I expect a falling response at HF would keep that manageable.

Note: I was using noisier opamps for my EQ stage than Jung for my EQ stage (roughly 3uV vs .5 uV input noise), so he could get away with less gain in his front end than me for similar noise performance.

So short answer I got away with 30+ dB gain... in a revisit I would probably use quieter opamps in the EQ section and take out at least the 6dB in the diff amp. While I would still keep the EQ values high to stay with all film caps in the audio path.

Note: the electrolytics in my front end gain stage are inside the NF loop so their (bad) behavior is minimized. (I know Wayne knows this, but this comment is for other casual readers). Not to throw stones at the Jung design, after the fact bench work I did on my '78 DJ preamp discovered a weak link in the electrolytic cap in the gain leg with a 22uf working into a 360 ohm gain resistor. The ESL of this cap contributed tens of degrees of phase shift at 20kHz. While I am not arguing the audibility of phase shift at 20khz, I prefer my RIAA to follow the phase and amplitude response as defined by the standard EQ time constants. So while I haven't ever bothered to measure one I would be apprehensive about Jung's 100uF working into 3k or so.. but I have been over engineering phono preamps since the late '70s, :roll:

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Phono Preamps

Post by mediatechnology »

I had several customers use my old P-10 this way. As I understood it, de-clicking before the RIAA playback EQ meant you were dealing with shorter events. Running the brief click transient through the LPF of RIAA, spreads out the energy in time, making it harder to remove, and involves punching a larger hole in the sound when the smeared out click is muted. Further, depending on the click mute strategy, there may be HF edges from muting the click, that the following LPF of the RIAA EQ will smooth out.
Thanks John. I built a P10 front end. Very cool, simple, yet elegant.

Using flat processing I did find that I could be more agressive in the use of Virtos, the DirectX plugin. One of the things it allows you to do is listen to the residual and dial in the maximum click duration it will process along with the threshold. What you can do is set the controls to just the point where you start hearing musical events in the residual. What I found was that I could have it process much longer clicks, at lower thresholds than before without artifacts. I was suprised I could set the duration so long. IIRC I used up to 1.5 mS or so. But it didn't "hole punch" the audio like it would have being processed post-EQ.

The gain that I'm using is 30 dB with an AT120. By the time it makes it through the soundcard to the EMU mixer app and Cool Edit I have no idea what the overall gain is.

But, with 30 dB of preamp gain and typical glass mixer gains (it says its 0dB) I'm seeing Cool Edit peaks of around -3dB in the "flat" file. So, 30 dB preamp gain or so is pretty close. Of course, after Cool Edit RIAA de-emphasis normalization is required.

The preamp overload point with +/-15V supplies (and a 1240 not a 1246) is about -8 dBu so there seems to be plenty of headroom.

I used 2SK389s as the front end FETS, a TL072 and a THAT1240 for the differential-SE stage. I played around with a bunch of different op amps and didn't see any significant improvement.

The P10 front end performs quite well. Most of the THD we're seeing is from the generator.

Image
User avatar
JR.
Posts: 3697
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Phono Preamps

Post by JR. »

Another observation about phono preamp design... While you are looking wide band, The RIAA EQ will roll off and understate THD measurements in preamp measurements. Back in the day I rolled my own two-tone (19-20kHz) IMD analyzer that i found very revealing for phono preamp design, since the 19 and 20KHz was heavy lifting for the circuitry, and the 1kHz IMD product was boosted 20 db by the RIAA...

The difference between my hot rodded IMD test measurements and simple THD in the context of phono preamps was huge.... another historical footnote.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Phono Preamps

Post by mediatechnology »

Hmm...Did those two-transistor preamps in cheaper gear crap out under that challenge?

It's brutal listening to the un-EQ'd playback due to the HF energy. That's the same stuff the front-end hears.

I think I'll add a RIAA EQ stage for a monitoring side-chain. It would be handy to have the RIAA EQ switchable for monitoring during transfer and auditioning click removal while playing back the flat file.
User avatar
JR.
Posts: 3697
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Phono Preamps

Post by JR. »

mediatechnology wrote:Hmm...Did those two-transistor preamps in cheaper gear crap out under that challenge?
I tried not to spend time measuring or even looking that closely at other people's designs**, but my earlier LM387 based designs revealed some blemishes with the two tone testing, while looking wonderful with simple thd+n
It's brutal listening to the un-EQ'd playback due to the HF energy. That's the same stuff the front-end hears.
yup... 40 dB rise from 20 Hz to 20kHz... hard to listen to that critically.
I think I'll add a RIAA EQ stage for a monitoring side-chain. It would be handy to have the RIAA EQ switchable for monitoring during transfer and auditioning click removal while playing back the flat file.
I am partial to my designs, and for some extra subtle back story, the 3180 usec pole in the feedback of the tl074 almost mimics the compensation pole in the open loop gain for flat loop gain margin.. ohhhhhhh

Of course this is just more esoterica.. but a lot of thought went into those old designs,,, the odd looking extra resistors around the 318 usec zero was in response to a corrective tweak suggested by S P Lipshitz (yes "the" SP Lipshitz). I met him at a AES show and he was a nice guy, despite me being pretty much illiterate about his higher math.

JR

** I did have one failed joint venture with big name guy who shall remain nameless for now... we parted our ways over design differences. it's a shame because his name would have sold a lot of kits, but even back then I had standards, and strong opinions.
Cancel the "cancel culture", do not support mob hatred.
ricardo
Posts: 262
Joined: Sun Apr 10, 2011 7:24 am

Re: Phono Preamps

Post by ricardo »

Gotta admit I'm amused at these supa-dupa Phono preamps. In da old days, I spent a lot of time on discrete OPAs and also how to stick LN transistors on LM301 etc.

When NE5532/4 appeared, I swept them all into the Don't Recycle bin.

NE5532 has almost perfect noise & BW performance for Moving Magnet RIAA.

The Dynamic Duo (Lipshitz & Vanderkooy) have a JAES article on dreaming up accurate RIAA.

3 important articles tell you what signals to expect. Shure show the maximum recorded velocities on real vinyl (JAES). Tim Holman (? the THX man) translated this to what you need in a real preamp (Audio magazine?). One of the false prophet, Matti Otala's, students did a brilliant AES preprint in Hamburg 1981 which essentially proved that the zillion V/us slew rates he was touting as essential were NEVER encountered in real life vinyl reproduction.

Wat dis mean?

Straightforward NE5532 with one of the Dynamic Duo's RIAA networks, direct coupled to cartridge, with sensible operating levels is as good as it gets for MM RIAA preamps. I defy anyone to do a better preamp for use with real life MM cartridges. You'd be hard put to measure any improvement, let alone hear any ... unless you are using a Golden Pinnae circuit with crap RIAA networks and poor overload.

For Moving Coil cartridges, I like

http://f1.grp.yahoofs.com/v1/gHldTuaW_G ... README.doc

I took the Leach circuit and by throwing out extraneous bits and changing others, transformed its performance from mediocre to the lowest noise such device in the known universe. For real life MC cartridges, this hugely complex circuit introduces negligible distortion.
User avatar
JR.
Posts: 3697
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Phono Preamps

Post by JR. »

ricardo wrote:Gotta admit I'm amused at these supa-dupa Phono preamps. In da old days, I spent a lot of time on discrete OPAs and also how to stick LN transistors on LM301 etc.
remember the feedforward compensation on the 301?
When NE5532/4 appeared, I swept them all into the Don't Recycle bin.
You and many others... between the 553x and 07x bifet you can cover most audio speed bases.
NE5532 has almost perfect noise & BW performance for Moving Magnet RIAA.
Perfect is noiseless with infinite gain, etc.
The Dynamic Duo (Lipshitz & Vanderkooy) have a JAES article on dreaming up accurate RIAA.
It is remarkable how much effort designers put into getting RIAA accurate to 0.1 dB and then didn't bother to dial in cartridge termination capacitance. I sold a little dip switch PCB with 4 caps that allowed you to add caps in parallel to cart termination with 25 pF resolution.

RIAA was disappointingly silent about how to handle below 20 Hz. The IEC proposed pole at 7950 uSec made a lot of sense (IMO).

All who designed their EQ to follow 75 uSec pole forever (me included) were ignoring that cutting lathes run out of boost before too long, but like the RIAA i don't care as much about response above 20 kHz as long as you don't generate spurious audible artifacts down in band so IMO real poles reject RF better.
3 important articles tell you what signals to expect. Shure show the maximum recorded velocities on real vinyl (JAES).
Yup, but RF happens so one needs to be able to roll that off harmlessly too.
Tim Holman (? the THX man) translated this to what you need in a real preamp (Audio magazine?).
Tomlinson Holman? the APT guy, etc... He did one crazy ass design where the input LTP had a JFET on one side and a bipolar transistor on the other side... Interesting concept, but i never got comfortable enough with it to consider it in anything I ever did.

The THX deal was a bit of a racket... You want your amps approved for THX theater use... you need to do it their way exactly... They don't walk on water and know more than manufacturers about how to make power amps... but they made a lot of gravy with that franchise.
One of the false prophet, Matti Otala's, students did a brilliant AES preprint in Hamburg 1981 which essentially proved that the zillion V/us slew rates he was touting as essential were NEVER encountered in real life vinyl reproduction.
The sundry new distortions imagineered by Otala was a running (slewing) joke with real designers. Unfortunately he captured the imagination of audio phools around the world, and even had some silly distortion tests find their way into real test equipment.
Wat dis mean?

Straightforward NE5532 with one of the Dynamic Duo's RIAA networks, direct coupled to cartridge, with sensible operating levels is as good as it gets for MM RIAA preamps. I defy anyone to do a better preamp for use with real life MM cartridges. You'd be hard put to measure any improvement, let alone hear any ... unless you are using a Golden Pinnae circuit with crap RIAA networks and poor overload.

For Moving Coil cartridges, I like

http://f1.grp.yahoofs.com/v1/gHldTuaW_G ... README.doc
Your link didn't open. for me.
I took the Leach circuit and by throwing out extraneous bits and changing others, transformed its performance from mediocre to the lowest noise such device in the known universe. For real life MC cartridges, this hugely complex circuit introduces negligible distortion.
I remember the oddball battery powered Leach head amp, and it was interesting but didn't float my boat as practical for anything.

I also sold a MC version of the P-10 using 2sb737s (.4 nVrt Hz) bipolar input devices and got kind words from Peter Aczel (Audio Critic magazine). But I got the sense he was looking for something more (like an advertisement) _ before saying anything good in print... He traded me a pair of loudspeakers he was pimping for one of my preamps... I knew the guy who designed the speakers too, since he consulted for Bozak (speaker design) around the same time I did some consulting (consumer delay) at Bozak too.

But I pretty much agree this P-10 is mostly over engineered, and my later P-100 was over-over-engineered.

The saddest part was the mixed good and bad reviews I got from magazines, in my judgement revealing short comings about their personal listening systems. :lol: But that's why I ran screaming from the high end audio business, decades ago. The sound of the hardware had precious little to do with anything. :roll:


JR
Cancel the "cancel culture", do not support mob hatred.
Post Reply