Phono Pre-Amp design question

[Biasrocks]

I've been working on improving a mid-level phono pre-amp for use in the studio.

I've included the schematic below.

So far I've upgraded C2 input cap to a low esr type and bypassed with a 1uf film type, C8 output cap to a 3u3 film type.

I've removed the 220pf shunt to ground on the input because my cartridge doesn't need that amount of filtering, although I think 100pf is the commonly used value I've seen these left out entirely on other designs and I can't detect any downside.

I'm thinking of bypassing the 1K input resistor R1. I'm not exactly sure what purpose it serves in this design so I'm hesitant to go ahead until I understand what it's doing.

Is C2 the 10uf input cap absolutely necessary?

I'm also going to swap out the wall wart power supply for a linear, regulated power supply.

Thoughts on the resistor removal or mods in general are welcomed.

 

[mgriffith]

Yeah, C2 is essential, without itQ1 will be turned off when you plug in your  cartridge, as  their DC R will pull the base to ground

 

[JohnRoberts]

I've been working on improving a mid-level phono pre-amp for use in the studio.

I wasted a couple decades designing phono preamps back last century.

I've included the schematic below.

Hurts my brain a little to look at that.... very old school. (Last time I saw a discrete preamp like that was in a Bozak disco mixer).

So far I've upgraded C2 input cap to a low esr type and bypassed with a 1uf film type, C8 output cap to a 3u3 film type.

Seriously save the money... the phono cart has a source impedance maybe 1500 ohms and is inductive at HF so low ESR and bypassed input coupling cap is not doing anything useful. Output side has 560 in series so again not very low Z.

I've removed the 220pf shunt to ground on the input because my cartridge doesn't need that amount of filtering, although I think 100pf is the commonly used value I've seen these left out entirely on other designs and I can't detect any downside.

Not a filter but a termination to tame the cartridge HF characteristic.  You can look up your cartridge's nominal termination capacitance, then subtract the phono cable capacitance, and tone arm capacitance from that total, to determine a nominal termination. Back in my kit days I sold a 4 position dip switch with small caps (24pf+47pf+100pf+200pf) so you could dial in a termination over a pretty wide range without having to solder in different caps.)

I'm thinking of bypassing the 1K input resistor R1. I'm not exactly sure what purpose it serves in this design so I'm hesitant to go ahead until I understand what it's doing.

It may be for stability... not sure.  It is an added source of noise, but not sure removing it will make a noticeable difference in that context.

Is C2 the 10uf input cap absolutely necessary?

yes, I am too lazy to do the math but from observation the base of Q1 is sitting up at a positive voltage (low single digit volts), to bias up the other devices.  + side of cap goes toward the base.

I'm also going to swap out the wall wart power supply for a linear, regulated power supply.

again not the weakest link.

Thoughts on the resistor removal or mods in general are welcomed.

Are you married to that approach.... ? (You could brag about it being class A.  ;D ;D).

  I've done some pretty esoteric preamp designs but wouldn't suggest following that path.

Today we can buy some off the shelf opamps that deliver truly excellent performance in that application.

If you are married to that approach, maybe swap in a low noise JFET for the first device, but it will have lower transconductance so you may need to up the loop gain in the second section (not sure how with q2 already common emitter) . Again a bunch of work and probably still not as good as a decent off the shelf op amp.

Here's a discussion with TMI about phono preamps ...from Wayne's forum.  http://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=6&t=423  Warning there are 24 pages of discussion.

JR

 

[Gold]

Considering you can pick up an NAD preamp for less than $50 I don't think it's worth the effort. Unless it's for educational purposes. The NAD preamp is about as good as it gets in my book.

If you want a super kick ass pre check this out.  http://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=7&t=753

 

[PRR]

> wouldn't suggest following that path.

+1

12V supply is too low. The whole thing is muddled. There are much cleaner designs (both to look-at and to listen). A 1-buck '5532 will meet or beat any 3-Q phono preamp, no muss/fuss.

 

[Biasrocks]

While it appears this is a very deep rabbit hole.

While I'm not married to this particular design, it is what I have on hand. I feel like I've made some strides in making it sound better than the stock configuration.

All of the things that I did to the circuit yielded discernible improvements.

I did increase the power supply voltage up to 18volts using an external supply and that seemed to bring some noticeable improvements too.

Thank you for your input on this.

Mark

 

[JohnRoberts]

Removing the input termination cap is likely to make an audible difference, the other stuff not so much (IMO). 

If you increase the rail voltage you can tweak the bias resistors to increase the DC operating point for symmetrical clipping (perhaps increase R9 slightly but a scope would be helpful to confirm this). If the output isn't clipping now, increasing headroom may be of little value (perhaps more linear).

Have fun, do no harm.

JR

 

[Gus]

IMO a waste of time unless you are using this to learn from.

As PRR posted a 5532 will be better for this type of preamp

IMO this has been solved for some time.

You have three basic setups
EQ Feed back "around" the gain stage, 2 transistor to IC opamp like in the threads first post
Passive. A  "flat" gain stage to passive EQ to "flat" gain stage
OR flat gain stage to a passive section to gain stage with the other part of the EQ around the 2nd gain stage

I have not seen a gain,  part of the passive EQ, gain, other parts of the passive EQ, gain.  John, PRR anyone have you seen a design like that


I like the passive EQ between stages.  You need to think about first stage clipping because you often want the most gain in the first stage for lower noise however the RIAA EQ from the record boosts the high end.    When I built a passive phono preamp in 79 I hand matched polystyrene caps and resistors for the builds.  If I did this again I would try class one ceramics.

A link you might want to read
http://www.douglas-self.com/ampins/discrete/2Q-RIAA/2Q-RIAA.htm

The sound change between different phono preamps is partly/mostly due to EQ differences in the preamps.  Also like John posted the interface between the phono cartage and preamp is important.  search for "Apt Holman preamp"

 

[JohnRoberts]

IMO a waste of time unless you are using this to learn from.

As PRR posted a 5532 will be better for this type of preamp

I was leaning toward a modern JFET input op amp but in the thread I linked to, there is discussion of some modern bipolar op amps that seem slightly better than JFET for this application. But this is extreme noise management and not necessary for mid-fi application.

IMO this has been solved for some time.

You have three basic setups
EQ Feed back "around" the gain stage, 2 transistor to IC opamp like in the threads first post
Passive. A  "flat" gain stage to passive EQ to "flat" gain stage
OR flat gain stage to a passive section to gain stage with the other part of the EQ around the 2nd gain stage

I have not seen a gain,  part of the passive EQ, gain, other parts of the passive EQ, gain.  John, PRR anyone have you seen a design like that

I've published phono preamps designs (several in kit construction articles). The schematics are around the WWW.

I've handled the RIAA different ways in each one.  I am not a huge fan of fully passive EQ. If you look at the shape of the RIAA EQ curve the falling closed loop gain with increasing frequency, looks a lot like the open loop response of most op amps so that makes sense to perform as feedback around an simple op amp gain stage (but this is pretty esoteric).

As I've said I skinned that cat several different ways (and every cat was skinless when finished). Only my last preamp to end all preamps, used any passive EQ, a real pole for the 75 usec RIAA time constant (the 318 uSec and 3189 uSec time constants were active negative feedback networks). My passive real pole before the op amp gain stage guaranteed that lightning striking the input couldn't slew rate overload that preamp.

This is all mental masturbation since the rate of change coming from a cartridge even if improperly terminated will not slew limit a modern op amp.

There is more enough discussion in that 24 page thread I linked to, and the shorter one Gold linked to.  I have probably posted most of my old phono preamps design schematics here in other threads over the years.

This is a (very) mature technology.

I like the passive EQ between stages.  You need to think about first stage clipping because you often want the most gain in the first stage for lower noise however the RIAA EQ from the record boosts the high end.    When I built a passive phono preamp I hand matched polystyrene caps and resistors for the builds

yes I like polystyrene... not very process friendly for large scale manufacturing, but a nice dielectric and OK for my kit business. Today I'd search out some NPO/COG caps as more mechanically robust.

A link you might want to read
http://www.douglas-self.com/ampins/discrete/2Q-RIAA/2Q-RIAA.htm

The sound change between different phono preamps is partly/mostly due to EQ differences in the preamps.  Also like John posted the interface between the phono cartage and preamp is important.  search for "Apt Holman preamp"

Sorry I never read that Self piece so have no thoughts on that.

Holman did a fairly strange preamp design in the Apt preamp using a JFET for one side of the input long tail pair, and bipolar transistor for the other side... Unusual to say the least, but that cat was skinned too. FWIW back then low noise JFETs were rare and dual JFETs not very quiet.

Yes cartridge termination matters....  It used to make me nuts how the audiophile reviewers, had systems that were all over the place... Typically their listening room and loud speakers made a much larger difference than any preamp's RIAA EQ curve accuracy.  I even got wildly different reviews for the exact same preamp from two different magazine reviewers, no doubt hearing differences between the rest of "their" playback systems (one reason I escaped from the hifi business back in the '80s... too flaky).

JR

 

[Gus]

John

My post was to Biasrocks.  I can see it was not clear.

You did bring up a number of good points.  When I posted about the phono preamps eq diferences I meant the same room and system with just the phono preamp changed.
Back in the late 70's early 80's when I was building  things like phono preamps and changing parts in bought components I realized after a time the speakers count more than anything else unless a component is designed very wrong.
After that the cable,  amplifier etc reviews meant little to me.
I also think a properly recorded and mastered CD sounds better than an LP.

 

[JohnRoberts]

John

My post was to Biasrocks.  I can see it was not clear.

But I'm not going to pass up an opportunity to pontificate about phono preamps. (I paid a lot of dues)

I also think a properly recorded and mastered CD sounds better than an LP.

Certainly more accurate so if the master et al is good the CD will be good...
=======
Another point about active vs. passive RIAA, if the flat front end gain stage has more than 20dB of gain, that topology will have less headroom @ 20kHz than a typical feedback EQ topology (typical RIAA is only +20dB gain @ 20kHz). That said clipping at 20kHz is not really a problem.    (Sorry I can't stop.)

Lots of different ways to make phono preamps so whatever floats the OP's boat.

JR

PS: I'm pretty sure I got blocked on facebook by one former friend for pointing out the futility in advancing phono preamp state of the art this late in the game.  8) .  Since then  I try to be a little less honest about how I really feel about designing new phono preamps in this century  8) . 

 

[PRR]

> gain, part of the passive EQ, gain, other parts of the passive EQ, gain.  ...have you seen a design like that

Been done.

A design which is not what _I_ would do, yet I respect VERY much, is the Hagerman Bugle.

As a bonus (for this or ANY phono preamp), under "Tools" he has an "Inverse RIAA" which you can use to check the flatness of your preamp EQ. Most commercial preamps are quite unflat. I was astonished at the differences _I_ could hear with 1dB errors. Nothing I could point to, but different feeling. Testing RIAA curve is insanely tedious without an Inverse RIAA.

Measure RIAA with something better than a sound card or H-P 200AB. Some of the Dynas had unsuspected peaking near 0.5Hz, which is "record warp" at 33RPM. Most "simple" topologies "flatten out" above the audio band, which relative to a velocity pickup amounts to a response rising to infinity. The defense is that classic 600mH cartridges with R-C loading fall-off badly past 20KHz. But then there are Grados and MC which don't.
____________________

Jung's Audio book has a very well-considered phono preamp. At the time, it looked extravagant. Today it would be just a few bucks.
____________________

The Self "2Q" link is interesting. I have some more examples from the US side. However with just two transistors it is not possible to provide enough gain to integrate a conventional hi-fi and cover RIAA with "large" NFB over and beyond the audio range. Yes you can get by with penny transistors for a $10 preamp or dollar transistors in a $50 preamp, and millions of these were sold and used with ignorant satisfaction.

Adding a 3rd transistor seems obvious, but often the improvement is very slight.
____________________

> pick up an NAD preamp

I'd listen to Gold.

I would NOT just buy a popular-price phono preamp by a "good brand". A *good* phono preamp is not just matching a gain-curve 50-15KHz.

I have liked many things from ART (especially when market-driven mis-features are avoided). So I got ART's low-price "disco" preamp. Plain simple design; I've lost the details. But when hooked up and brought to "realistic level" (I had 2 grand pianos and 50 piano LPs in the rooms so comparison was obvious), the woofer cone shook madly and then the power-amp shut-down. The 'scope showed huge random sub-sonic noise on the output. A little study showed this was caused by the chip's 1/f bias current noise and the small input cap. It probably worked fine with PA speakers and amps, but our high-class "domestic" system did not like it.

 

[JohnRoberts]

> gain, part of the passive EQ, gain, other parts of the passive EQ, gain.  ...have you seen a design like that

Been done.

A design which is not what _I_ would do, yet I respect VERY much, is the Hagerman Bugle.

Not bad... they spread the 60 dB of total gain across 3x 20dB modern high performance op amp gain stages  (I approve).

I don't see a real pole so there are zeros above 20 kHz... not a big deal since RIAA stops caring above 20kHz, and cutting lathes don't boost all the way up to light.

I am not warm and fuzzy about the price. I never sold a complete preamp for that much (maybe I'm just jealous? and cheap?)

As a bonus (for this or ANY phono preamp), under "Tools" he has an "Inverse RIAA" which you can use to check the flatness of your preamp EQ. Most commercial preamps are quite unflat. I was astonished at the differences _I_ could hear with 1dB errors. Nothing I could point to, but different feeling. Testing RIAA curve is insanely tedious without an Inverse RIAA.

I bought an inverse RIAA network from somebody associated with TAA back in the 70s.  Very handy on the test bench... I could feed regular program or sine waves into a phono preamp prototype for measurement and listening tests,  without extra errors from turntable/cartridge/tone arm, etc. (IIRC that inverse RIAA had a HF zero too, but probably up around 200 kHz, so not a problem in practice.)

Measure RIAA with something better than a sound card or H-P 200AB. Some of the Dynas had unsuspected peaking near 0.5Hz, which is "record warp" at 33RPM. Most "simple" topologies "flatten out" above the audio band, which relative to a velocity pickup amounts to a response rising to infinity. The defense is that classic 600mH cartridges with R-C loading fall-off badly past 20KHz. But then there are Grados and MC which don't.

Back in the '70s I was selling companding tape NR kits.  Too much LF content from turntable rumble, wow, whatever, could modulate the envelope of the compressed signal going to tape, but cassette tape and even open reel recorders were notorious for scraping off LF content.  When the NR playback expander tried to reconstruct the recording with missing LF content , there would be a phantom modulation caused by the missing signal.  Most noticeable during quiet parts.

I eventually fixed that with a sliding pole dynamic HPF built into my NR compressor that rolled off LF only when things were relatively quiet (compressor was at max gain), then restored normal full LF response when music was happening (and compressor was running at lower gain).

I also embraced a additional -3dB @ 30Hz pole added on top of RIAA EQ curve, proposed by the IEC but never accepted by the RIAA, even back in the 80s vinyl was old news and RIAA wasn't interested in tweaking it. It was enough work for them to extend the curve to 20Hz-20kHz 

____________________

Jung's Audio book has a very well-considered phono preamp. At the time, it looked extravagant. Today it would be just a few bucks.
____________________

The Self "2Q" link is interesting. I have some more examples from the US side. However with just two transistors it is not possible to provide enough gain to integrate a conventional hi-fi and cover RIAA with "large" NFB over and beyond the audio range. Yes you can get by with penny transistors for a $10 preamp or dollar transistors in a $50 preamp, and millions of these were sold and used with ignorant satisfaction.

Adding a 3rd transistor seems obvious, but often the improvement is very slight.

  RIAA typically needs 60dB boost at 50Hz, but only 20 dB by 20kHz so not really all that hard to maintain loop gain margin despite falling open loop gain caused by dominant pole compensation, but yes more than two transistors to be comfortable..  An issue with some early dedicated bipolar phono preamp ICs (LM381 etc) is that they weren't unity gain stable so hard to maintain a true 75uSec pole.  (perhaps a personal problem of mine, worrying about what it's doing up at 200kHz).

____________________

> pick up an NAD preamp

I'd listen to Gold.

I would NOT just buy a popular-price phono preamp by a "good brand". A *good* phono preamp is not just matching a gain-curve 50-15KHz.

I have liked many things from ART (especially when market-driven mis-features are avoided). So I got ART's low-price "disco" preamp. Plain simple design; I've lost the details. But when hooked up and brought to "realistic level" (I had 2 grand pianos and 50 piano LPs in the rooms so comparison was obvious), the woofer cone shook madly and then the power-amp shut-down. The 'scope showed huge random sub-sonic noise on the output. A little study showed this was caused by the chip's 1/f bias current noise and the small input cap. It probably worked fine with PA speakers and amps, but our high-class "domestic" system did not like it.

As I've shared already how you handle 20 Hz and below matters for several applications.

JR

 

[Gus]

A google search found this http://www.analog.com/library/analogDialogue/archives/39-05/Web_Ch6_final_I.pdf

 

[madswitcher]

I built the Doug Self kit from Signal transfer company - costs about 150UKP and is a superb bit of kit - uses bog standard 5534/2. plus some trannies if you want the moving coil option.

Worth a look.

Mike

 

[Gold]

I built the Doug Self kit from Signal transfer company - costs about 150UKP and is a superb bit of kit - uses bog standard 5534/2. plus some trannies if you want the moving coil option.

Worth a look.

The quibble I have with that and many other phono preamps is that they can't be calibrated to a standard reference level. If you are working with records in a professional capacity it's a requirement.

 

[peterc]

I have also looked at the Bugle 1 and have partly done a PCB layout.

Hope to get back to it soon.

Schem:
http://www.hagtech.com/images/bugleschem.gif

 

[madswitcher]

I built the Doug Self kit from Signal transfer company - costs about 150UKP and is a superb bit of kit - uses bog standard 5534/2. plus some trannies if you want the moving coil option.

Worth a look.

The quibble I have with that and many other phono preamps is that they can't be calibrated to a standard reference level. If you are working with records in a professional capacity it's a requirement.

Feed in a 5mV signal at 1Khz  (or whatever the cartridge output is specified to be at 1Khz for a given groove velocity - usually 5 cm/sec). 

I use a calibrated attenuator (2dB per step over 24 steps) and feed that into a line amp with variable/preset gain to get the output of the Self unit up to +4dBm (0VU).  I  set the attenuator to 4 steps backed off from full rotation during calibration to give a bit of gain if required.

The output is monitored by both a pair of VU meters and a twin PPM so I can get the average and the peak values.  This is then a calibrated system since any variation is a dB ratio compared to the calibration level.

Hope this helps

Kind regards

Mike

 

[JohnRoberts]

I built the Doug Self kit from Signal transfer company - costs about 150UKP and is a superb bit of kit - uses bog standard 5534/2. plus some trannies if you want the moving coil option.

Worth a look.

The quibble I have with that and many other phono preamps is that they can't be calibrated to a standard reference level. If you are working with records in a professional capacity it's a requirement.

Feed in a 5mV signal at 1Khz  (or whatever the cartridge output is specified to be at 1Khz for a given groove velocity - usually 5 cm/sec). 

I use a calibrated attenuator (2dB per step over 24 steps) and feed that into a line amp with variable/preset gain to get the output of the Self unit up to +4dBm (0VU).  I  set the attenuator to 4 steps backed off from full rotation during calibration to give a bit of gain if required.

The output is monitored by both a pair of VU meters and a twin PPM so I can get the average and the peak values.  This is then a calibrated system since any variation is a dB ratio compared to the calibration level.

Hope this helps

Kind regards

Mike

I suspect he knows "how", but is critical of the consumer designs that lack a gain calibration trim.

JR

PS: Did you visit the link he posted earlier?

 

[Gold]

I know how to calibrate a phono pre. You need a calibration record. Feeding in a tone from an oscillator doesn't work particularly well IME. I think because the impedance curve of the cartridge coils doesn't  look like the source impedance of the oscillator.

I think a phono pre intended for professional use should have variable level built in.