I tapped the connector and the cable - I could not generate anything I could positively attribute to that as a source for the noise.
On the input there are two 100k devices in parallel - one side of the two resistors goes to 0V. The top ends have the 100uF cap between them. One resistor top side goes to the + input on the opamp and the other side of the cap to the input socket.
When the input is shorted (at the input socket) , so are the resistors so there is no noise pickup when I tap the caps and surrounding area. When the input is open, any pickup across the resistors is amplified. Short the cap out and you get almost the same result.
My hypothesis that the resistive element on the 100k's is stressed like a SG and that's how the noise arises. The other explanation is that its the COG SMD caps - but when I tapped directly above these (they are under the PCB), there as no noise pick-up.
I'll do some more tests tomorrow.
On the input there are two 100k devices in parallel - one side of the two resistors goes to 0V. The top ends have the 100uF cap between them. One resistor top side goes to the + input on the opamp and the other side of the cap to the input socket.
When the input is shorted (at the input socket) , so are the resistors so there is no noise pickup when I tap the caps and surrounding area. When the input is open, any pickup across the resistors is amplified. Short the cap out and you get almost the same result.
My hypothesis that the resistive element on the 100k's is stressed like a SG and that's how the noise arises. The other explanation is that its the COG SMD caps - but when I tapped directly above these (they are under the PCB), there as no noise pick-up.
I'll do some more tests tomorrow.
Don't forget re, the high Ic is a necessity even with no rbb. BTW Richard have you ever come across a true self noise calculation for a ribbon mic? I figure the noise of the air load would limit how far you can go. A perfect 1nV FET amp does nothing for the typical 1/2" condenser (I suppose folks are still cutting up those tiny electrets and putting in SK170's).
It depends on the efficiency which is down to the magnetic circuit. IIRC, BBC Engineering Monograph No4 has an analysis of the 4038 noise including an estimate of the Johnson air noise. Excuse me if I don't dig this up tonight as its been a gruelling day at the beach for this bum
You can roughly say the Johnson noise is roughly proportional to the diaphragm size ASSuming the other stuff is well designed (??!!!) I'll ignore ribbon bandwidth and other trivial stuff
But the 4038 is what I would want to use as IMHO (and that of many Recording Engineers .. including one of the famous Direct Cut Disc people) this was the best ribbon of the 20th and possibly the 21st century. I'm sad no one is updating it with Neo magnets etc. It's main defect today is about 10dB less sensitivity & efficiency than other good ribbons.
But don't forget, Eric & I were only going for 3dB NF. ie 57.7pV/rtHz & 8.14nV over 20kHz (-161.8dBV)
I'd be disappointed if I didn't get 2dB NF with a big Lundahl transformer & 5534. I think 3dB NF was a BBC requirements for mike inputs circa 1980. So our proposal was really just a good wank.
___________________
I'm going to get some cheapo 2n5458 FETs, scrape off the numbers, micro-etch "Hand Carved from Unobtainium by Virgins on the Great Barrier Reef" and sell them for $100 each. Scott, will you recommend them in your next Linear Audio article for a 50% cut of the profits?
Gotta find a SM version with built in bias diode for those electrets. This will actually improve noise a bit if we use your Linear Audio recommendations ... eg BF862.
I've dreamt up a variant of my SimpleP48Vmike to take full advantage of this that might actually fit in eg the bigger Primo capsules.
Great Guru Baxandall (& I) tried these circa 1980 too.
Back then I found the ZTX medium powers were good but had unacceptable 1/f & popcorn. When I found the Hitachi devices, I stopped looking cos they were better and no selection was required.
It looks like Zetex have cleaned up their process after 35 yrs.
I think this is the guy you mean
[url]http://www.synaesthesia.ca/LNschematics[/URL]
This is what prompted me to re-visit noise performance. A cursory look at these very detailed pages suggests he’s finally achieved my 0.28nV/rtHz performance. Closer examination says some of his noise measurements are wonky and I don’t believe his rbb’ figures at all. He does mention some of the things which need addressing though there is no need to adopt his supa methods. The bottom line is his slightly (??) more complex circuit is not as quiet and probably has more THD with real MCs. Overload recovery probably OK.
There's also
[url]http://www.hoffmann-hochfrequenz.de/downloads/lono.pdf[/URL]
AT LAST! A device that might be quieter than mine for real life MC cartridges. Gerhad hasn’t actually tried this … but in theory, his amp will increase the noise of a 5R cartridge like the Ortofon Quintet Blue by 2.1dB while mine will be slightly more than 0.9dB worse, say 1dB.
Even more complex than the synaesthesia stuff above.
Gotta get some sleep.
Last edited:
As someone said MC pick ups have high current output. A transformer of the best possible design could make available lets say 20 mV and still have a low coil resistance ( less than an MM coil I guess ). Who knows where that might go. As far as I know nobody did it that way. I have no idea if the realities of how transformers work would catch up with us. Tim de Paravacini at one time thought transformers should go down to 5Hz. I only mention it as he did.
Here is rough schematic. I gave up on opamps eons ago so didn't keep any records on this but you get the gist. All balanced of course (OPA1632). There is a zero at 50kH, I know there is a lot of controversy about this. But my first preamp was an Allen Wright FVP, and I learned a lot from his preamp cook book, so I always put this in. And anyway, it makes sense, doesn't it. Of course infinite gain is NOT applied to a cutting lathe, there must be a hf roll-off somewhere. Where should be complimented in the playback circuit.Could you post the schematic please. I'm seriously interested
You might want to try my little circuit from my Yahoo MicBuilders Files.
It has the lowest noise of any such device in da known universe.
It uses a huge total of 2 devices, so you might rustle it up with parts from your spares bin ... then spend the next 30 yrs trying to do something better.
PS It won't kill your cartridge if you reverse the battery.
These days I use discrete components. Here is description of my current m-c stage. My component values are slightly different to accommodate my SLA battery power supply, and it sounds great with Monster Genesis 1000. This cartridge is only 0.24 mV, but there is plenty of gain. I don't use this circuit with m-m catridges, I am going to build a new one for my Decca - this will feature jfet (2SK369) in the front end, but topolgy will be pretty much the same. So yes, I use a different, optimised ciorcuit for each cartridge, because I also have another pre-amp for my strain gauge.
http://www.diyaudio.com/forums/anal...os-balanced-mc-phono-stage-2.html#post4685885
Attachments
I would be interested in an investigation into the concept of using the preamplifier input impedance to load down the source inductance of MM cartridges, and thereby achieve the high-frequency RIAA roll-off (75uS) at the phono preamplifier input.
The promise is that the capacitance of the arm cable and preamp input is rendered much less critical, and the HF response is extended, thus addressing two aspects of performance where MM cartridges are generally considered to be at a disadvantage in comparison to MC.
Bob Cordell examined this in his article on the VinylTrak phono preamp in Linear Audio Volume 4. He cites previous work by others in his references.
The promise is that the capacitance of the arm cable and preamp input is rendered much less critical, and the HF response is extended, thus addressing two aspects of performance where MM cartridges are generally considered to be at a disadvantage in comparison to MC.
Bob Cordell examined this in his article on the VinylTrak phono preamp in Linear Audio Volume 4. He cites previous work by others in his references.
I wouldn't. This not-so-bright idea goes back to the 1950's and occasionally resurfaces. The snag is that for accurate equalisation you need to know the inductance of the cartridge exactly- and you don't. All you have is a suspiciously round figure quoted by the manufacturer.
The two channels will not have exactly the same inductance.
If you want to change cartridge you have to redesign the preamplifier.
That's enough grief to put most of us off the idea.
Doug's New Book
Amazon already has it listed for preorder ...
Electronics for Vinyl: Douglas Self: 9781138705456: Amazon.com: Books
Amazon already has it listed for preorder ...
Electronics for Vinyl: Douglas Self: 9781138705456: Amazon.com: Books
Sorry, but that's really not a good idea.
How accurately do you think you could return to a flat response?
If you're planning to do this with pots, channel balance is going to be a major issue.
Far simpler to have a separate tone-control stage which can be very simply switched out to return to flat. I'd recommend my variable-frequency 4-knob tone control...
Actually there's a cute trick whereby all one needs to know is the ratio of coil R to coil L - which as it happens is typically constant (ish), enough to be workable. Check out the Aurak designs on the MM/ mechanical resonance thread......... I've been using this topology for a few years now - true transimpedance loading of MM/MI.
It's a different matter if one's aim is to produce a commercial product, since there are bound to be exceptions and bound to be carts that plain don't work.
I suppose most people here are well up to 'select on test' for one resistor value, and there is an upside - check out the thread.
LD
Аll phono stages designed by John Curl and Erno Borbely are HF passive with active bass boost in the feedback of the second stage.Do you think that respected audio designers have a wrong approach in phono preamps topology?Their designs are very well accepted.
If you place a passive 75us pole after a LF boost stage, you are kissing c. 18 dB of dynamic range goodbye. You have no noise advantage over all active with this approach either, and you miss out on the improved HF loop gain of the all active = lower distortion.
Passive EQ is best suited to tube RIAA where you have the voltage swing to deal with the HF dynamic range. All active is best for solid state EQ amps.
Please see Self's SSAD. I have a write up with sims on my hifisonix site as well.
Passive EQ is best suited to tube RIAA where you have the voltage swing to deal with the HF dynamic range. All active is best for solid state EQ amps.
Please see Self's SSAD. I have a write up with sims on my hifisonix site as well.
category error: market success poorly correlates w tech competence
the question is decidable in the realm of preamp EE and measured phonograph grooves and cart response
and sometimes "good enough" is good enough, and better doesn't bring enough more to the table
the question is decidable in the realm of preamp EE and measured phonograph grooves and cart response
and sometimes "good enough" is good enough, and better doesn't bring enough more to the table
Last edited:
John Curl and Erno Borbely phono stages are linear gain first stage- passive HF Eq.-active feedback bass boost in the second stage, not the approach you described.I am using Borbely 419 DIY phono stage with excellent results. Power supply voltage is +/24V, estimated overload with 60db gain is 26db.
If you reverse the EQ process and put the 75us EQ pole first and then boost the LF in a second stage you are sacrificing noise performance. Best way for solid state is all active - best noise performance, best overload, lowest distortion.
Get a copy of SSAD and go through the vinyl chapters cardfully and then do some research on LTSpice - the approach is correct.
Get a copy of SSAD and go through the vinyl chapters cardfully and then do some research on LTSpice - the approach is correct.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.