The perfect phono preamp.

[Geek]

Cascode, eh Giaime? We must be drinking the same wine

I'm trying to keep this fair minimalist. Cascode input too, 6AQ4 (EC91) bottom, ECC88 top. Sure it has a lower Gm than using a whole ECC88 on the input, but the extra mu (100) and far lower noise and microphonics is worth the tradeoff. Total gain of the input cascode is 180, so that sure won't be a problem. 3 tubes/channel, unless incorporated into an existing preamp with decent linestage, then there'll be only 2 tubes/channel

 

[SY]

Great minds and all that. The preamp I just retired to make way for the new was a FET-tube hybrid cascode input, the tube being an ECC88 variant, feeding passive RIAA, and then a Berning-type tube-FET gain stage, with a buffer very similar to the Heretical.

Hey, this is going to be a very interesting winter, guys!

 

[SY]

Greg, you made me look up that 6AQ4. Interesting tube, an equivalent noise resistance of 400R, input-referred. That should be fine for MM and higher output MC. Have you tried it "straight"in the first stage or was the ca. 250pF of Miller too much to deal with? The reason I ask this is that at 10mA, the rp is 12k, which seems rather high for the bottom bit of a cascode; that's reflected in the gain, which is only a few dB higher in the cascode than the "straight" tube, CCS-loaded.

 

[Giaime]

Cascode, eh Giaime? We must be drinking the same wine

I think I'm drinking too much of that wine, because I'll go with an hybrid cascode (aaargh heretical! ) and I'm strongly thinking of learning good solid state design to make a good sounding discrete all FET phono amp

 

[Geek]

Giaime, you want me to go on heart pills, don't you


Hi SY,

Greg, you made me look up that 6AQ4. Interesting tube, an equivalent noise resistance of 400R, input-referred. That should be fine for MM and higher output MC. Have you tried it "straight"in the first stage or was the ca. 250pF of Miller too much to deal with? The reason I ask this is that at 10mA, the rp is 12k, which seems rather high for the bottom bit of a cascode; that's reflected in the gain, which is only a few dB higher in the cascode than the "straight" tube, CCS-loaded.

I'm using a Shure cartridge which can handle 250pF OK, but after reading about the responses at various loads, it seems that one is better with as little C as possible. The resultant peak is over 32Khz, which is way out of the EQ's band and is moot. I might try a switchable loading C switch, for full control. Have a Grado F3+ too, but want to find a belt-drive TT to put that one in
(I'm too mechanically inept to build one)

If you have 10mA through a 6AQ4, you've probably killed it. Noise goes up exponentially above 4mA (played with this one for a mic pre.. dyn-o-mite at 1.3-2mA). I'm running it at 1.4mA with an LED on the cathode.

Should be able to post a circuit by tomorrow

 

[SY]

Cool! I've never used the tube, just going off the data sheet. That's a great piece of practical info.

Are you giving the LED a bit of extra current to get its impedance down?

 

[Geek]

Cool! I've never used the tube, just going off the data sheet. That's a great piece of practical info.

Glad you found it useful. Email me your addy, I'll send you some

Are you giving the LED a bit of extra current to get its impedance down?

No. Good idea though. Would I do this by adding a resistor from the LED/cathode junction to B+?

Cheers!

 

[SY]

Yes, that's it. A CCS will be even better than a resistor. What you don't want to do is inject power supply noise, so the high Z of a cascode CCS (I suspect you might have a board or three laying around!) will be beneficial here.

Now, it may well be that decreasing the dynamic impedance of the LED might have the opposite effect- rdf did some interesting measurements and found that there was some distortion cancelling in some setups.

I'll be delighted to take a couple tubes off your hands!

 

[wrenchone]

The 6AM4 has similar specs to the 6AQ4 (though no noise spec), and is intended for the same type of grounded grid RF service. Unike most other tubes, the elements are mounted horizontally instead of vertically, so it looks like a little barrel inside the glass envelope. The glass itself is the small size, like a 6AK5. Since I have some of the 6AM4s around, I'm tempted to build up a phono preamp with them based on the comments here.

 

[Geek]

Hi SY,

Yes, that's it. A CCS will be even better than a resistor. What you don't want to do is inject power supply noise, so the high Z of a cascode CCS (I suspect you might have a board or three laying around!) will be beneficial here.

Now, it may well be that decreasing the dynamic impedance of the LED might have the opposite effect- rdf did some interesting measurements and found that there was some distortion cancelling in some setups.

I'll be delighted to take a couple tubes off your hands!

OK, we'll give it a go and see, thanks!

Got a couple wrapped up and ready to go off to you

The 6AM4 has similar specs to the 6AQ4 (though no noise spec), and is intended for the same type of grounded grid RF service. Unike most other tubes, the elements are mounted horizontally instead of vertically, so it looks like a little barrel inside the glass envelope. The glass itself is the small size, like a 6AK5. Since I have some of the 6AM4s around, I'm tempted to build up a phono preamp with them based on the comments here.

That's the super thing about those V/UHF triodes. Almost all of them have some use in audio and I'd say a good 1/2 are suitable for phono service

I call those TV tubes "hidden treasures", since it's unlikely the audio guru's will pay any attention to them, ever, despite the electrical/physical superiority of many. After all, it's just a lowly, worthless TV tube....

 

[wrenchone]

I have a particular fondness for the lowly vertical oscillator/amplifier dissimilar triodes as well. My idea of a sweet audio power triode would be the big triode from the 6GF7, but with 20W plate dissipation instead of 10W.

 

[Geek]

Ooooo, I hear you there!

Made a beauty of an amp (daily workhorse for a nearly a year) from a pair of 6FD7's. Even my wife, who claims tone deafness said, "That's nice! It can stay.."

 

[Tubelab_com]

That's the super thing about those V/UHF triodes. Almost all of them have some use in audio and I'd say a good 1/2 are suitable for phono service

Shhhhh. I can hear the prices rising!

I call those TV tubes "hidden treasures", since it's unlikely the audio guru's will pay any attention to them, ever, despite the electrical/physical superiority of many. After all, it's just a lowly, worthless TV tube....

I have a warehouse full of "worthless TV tubes" (over 100000) They were given to me because of their lack of market value. Many are indeed worthless in the audio world, and on Ebay. Some however are great audio tubes.

I have a particular fondness for the lowly vertical oscillator/amplifier dissimilar triodes as well. My idea of a sweet audio power triode would be the big triode from the 6GF7, but with 20W plate dissipation instead of 10W.

Triode wire a 6LU8 or a 6LR8 and run it at about 13 to 15 watts dissipation. Not quite 20 but it can take more abuse than the dual triodes. Doesn't sound bad in UL either.

When I was a teenager with zero amp building budget, I could build a decent sounding guitar amp entirely out of old TV parts. Old TV sets could be had for free, at the local trash dump. A vertical output transformer can be used as an OPT. The power transformers put out enough juice to light up a small town. For P-P amps I used small power transformers as OPT's, the HV winding became the primary, and all of the filament windings in series became the speaker winding.

 

[astouffer]

Has anyone tried this phono stage? It looks simple enough but is the 220k resistor on the EF86 cathode a mistake? It seems way too high to be useful.

 

[cerrem]

Has anyone tried this phono stage? It looks simple enough but is the 220k resistor on the EF86 cathode a mistake? It seems way too high to be useful.

The input Z is not steady... It is frequency dependent ...with that feedback loop ....Usually need to add a series resistor to that input stage... To bring the input impedance up...

Chris

 

[Geek]

Here it is, within +0.2dB of the RIAA curve from 20Hz to 20KHz and sounds great!

THD of the unit was below my ability to measure (<0.03%) at 1KHz and the dynamic range is wicked. 7mV in at 30Hz could not overload this thing. Great for those 5mV high-output cartridges.

Actually has less noise than the one I made a few years back based on active FB.

3-tube/channel unit, which if incorporated into an existing pre/line driver is 2 tubes/channel. Design goal was tube unit efficiency and to use standard 5% components for easy repeatability amongst the do-it-yourselfer's.

Output stage shown is my modified White CF with an output Z of 100 ohms, suitable for ignorant cable/amp systems.

Have yet to try SY's boosted LED bias idea to reduce their dynamic Z.

Cheers!

http://geek.scorpiorising.ca/contrib/Geek/phono_preamp.gif

 

[SY]

Thanks, Greg. What's the overall gain at 1kHz?

I'm still scratching my head over that first stage cascode...

 

[Brian Beck]

I'm still scratching my head over that first stage cascode...

I've used the 6AQ4 in an RIAA stage, although not in cascode as here. The 6AQ4 has the mu of a 12AX7 with about 5 times the gm. This combination offers the potential for a low-noise front end. My quick back-of-the-napkin analysis of this cascode design predicts a low-frequency gain of roughly 600X (55dB). So, if there is adequate overload margin, the signal is boosted up to a large amplitude before the passive RIAA filter loss drops it back down closer to the noise floor. However, overload margin and distortion performance may this design's Achilles heel with so much gain.

 

[Geek]

Hi SY,

Thanks, Greg. What's the overall gain at 1kHz?

1mV in = 0.5V out @ 1KHz.

However, overload margin and distortion performance may this design's Achilles heel with so much gain.

Hi Brian,

This is a finicky circuit for large voltages in. That's why in the design process, I used the unused portion of f V2 for the second gain stage instead of a 5751. The input stage, as long as you're within phono levels, will be OK. The most critical component for distortion is the cap from the top tube's grid to ground. Experiments showed 0.1uF would be *just* enough. So I made it 1uF poly and listened to the highest-output albums I had on headphones (Zamfir - any non-linearity will kill those panpipes and an older Boston Pops, if the instrument exists, it'll be in there....). Sounded great!

So, why did I use so much gain in the first stage instead of distributing it? Noise. I hate noise and the only reason I stayed away from passive pre's for so long was because of that. So, we get the gain with the lowest noise components before the EQ. Problem cured.

There will always be tradeoffs. I swear these RIAA stages are best left to philosipher's and theologian's. I do not have a super high output cartridge here to test a straight ECC88 input. Sure the instruments will tell me, but I design for sound

Cheers!

 

[fdegrove]

Hi,

Agree with Brian's analysis.....
Adding an MC stage may have the preamp run into trouble with a number of higher output carts but I guess Greg's happy with his MM and is aware of the shortcomings of his current pre.

Still, it looks fine if you use it for what it was intended.

I do have some qq regarding the decoupling of the Ck of the WCF and why a 10K tail end resistor is being used?

Personally, I'd ommit the decoupling cap, up the Rg for that stage to 1M and use a smaller output cap with a 100K tail instead. Unless I'm missing out on something fundamental perhaps?

The pre probably has enough output to drive an amp straight in, ZOut sure won't be a problem if it's a tube amp we're driving, so somewhere I'd like to see a volctrl. Probably best put right at the WCF's input, right?

Congrats Greg,