Hi,
I have a strange idea: make a pre - preamplifier with a triode ( instead the normal step-up or solidstate stage).
My target is a linear gain of about 4-6 times with the less noise as possible
Thus I need a lower-mu triode I believe with high transconductance, right?
The circuit will be a basic common cathode. Output impedence is unimportant, microphonics tubes has to be avoided of course.
With these carachteristic I have no idea of what tube I can considerate
to make a little experimentation
Every suggestion will be very appreciated!
Thanks,
Paolo
I have a strange idea: make a pre - preamplifier with a triode ( instead the normal step-up or solidstate stage).
My target is a linear gain of about 4-6 times with the less noise as possible
Thus I need a lower-mu triode I believe with high transconductance, right?
The circuit will be a basic common cathode. Output impedence is unimportant, microphonics tubes has to be avoided of course.
With these carachteristic I have no idea of what tube I can considerate
to make a little experimentation
Every suggestion will be very appreciated!
Thanks,
Paolo
Ciao Paolo,
santa madonna, you have a 'strange' idea indeed. A challenging one and judging from all the commercial offerings, not a good one ;-) Actually I have been thinking about an tube MC-phono stage, and I did some calculations. I concluded it made no sense, whatever tube you use. Let me explain. For MC phono you foremost need low noise. In tube terms this means rugged, low microphonic and especially high transconductance. You could think of e.g. E88CC or E182CC.
Now on to the numbers I came up with. My cartridge is a Benz Micro ACE (low). Its source resistance is 12 ohm, and the output @ 5 cm/s movement of the tip is 400μV (400 millionth of a Volt). With the appropriate bandwith (a bit more than 20 kHz), the source resistance can be 'translated' into 67 nV of noise (67 billionth). This means the signal to noise distance is 20*log(400,000/67) dB, or 75.5 dB. Now with an ideal amplifier you don't add any noise... However, Antonio, if you like to amplify the small MC signal with tubes, some noise is introduced. With an E188CC biased at 10 mA, I calculated a noise figure (NF) of 13.2. This means the S/N distance would be decreased by that number. OUCH, let's face it 62.3 dB S/N is not enough!!
You could improve a little by paralleling four E188CC triode sections, the NF becomes 7.8. With a hybrid approach you have much better starting point: for an JFET (e.g. the popular 2SK170 @ 10 mA) the NF would be 3.9. The best result however can be obtained with a special low-noise BJT: with the 2SC3329, again biased @ 10 mA, I calculate a NF of 1.4.
If you really hate solid state, you can do worse than opt for a built-in step-up transformer in your phono-amp. I calculated (for my microbenz ACE) a NF of 3.4 if you were to use a Lundahl LL1681 (1:13).
I hope you can now understand why I decided to do myself a favour, and go the hybrid route with the 2 SC3329.
happy listening!!
Mark
santa madonna, you have a 'strange' idea indeed. A challenging one and judging from all the commercial offerings, not a good one ;-) Actually I have been thinking about an tube MC-phono stage, and I did some calculations. I concluded it made no sense, whatever tube you use. Let me explain. For MC phono you foremost need low noise. In tube terms this means rugged, low microphonic and especially high transconductance. You could think of e.g. E88CC or E182CC.
Now on to the numbers I came up with. My cartridge is a Benz Micro ACE (low). Its source resistance is 12 ohm, and the output @ 5 cm/s movement of the tip is 400μV (400 millionth of a Volt). With the appropriate bandwith (a bit more than 20 kHz), the source resistance can be 'translated' into 67 nV of noise (67 billionth). This means the signal to noise distance is 20*log(400,000/67) dB, or 75.5 dB. Now with an ideal amplifier you don't add any noise... However, Antonio, if you like to amplify the small MC signal with tubes, some noise is introduced. With an E188CC biased at 10 mA, I calculated a noise figure (NF) of 13.2. This means the S/N distance would be decreased by that number. OUCH, let's face it 62.3 dB S/N is not enough!!
You could improve a little by paralleling four E188CC triode sections, the NF becomes 7.8. With a hybrid approach you have much better starting point: for an JFET (e.g. the popular 2SK170 @ 10 mA) the NF would be 3.9. The best result however can be obtained with a special low-noise BJT: with the 2SC3329, again biased @ 10 mA, I calculate a NF of 1.4.
If you really hate solid state, you can do worse than opt for a built-in step-up transformer in your phono-amp. I calculated (for my microbenz ACE) a NF of 3.4 if you were to use a Lundahl LL1681 (1:13).
I hope you can now understand why I decided to do myself a favour, and go the hybrid route with the 2 SC3329.
happy listening!!
Mark
Hi Mark,
you are lucky: our Asaka Linn is only 0,25mV !!
I understand your point, numbers are crude but realistic!
Now hear this:
suppose I take a E288CC, two sections paralleled.
Expectded gain will be about 12-14 ( without cathode bypass).
Now I become heretic and apply 12 dB of feedback!
new gain will be 3-3,5 times and input impedence will be 1/4 , am I right?
S/N improve...........
What do you think about it?
Cheers,
Paolo
Commitment
Hi Paolo,
you show real commitment to the tube cause! And hey, the E288CC is a very, very nice tube. I am sorry to let you know, this one won't save you either.....
Just crunch the numbers: a Linn Asaka is low output. I found confusing numbers ranging from 0.1 to (your) 0.25 mV. I guess the 200μV @ 5cm/s in some test is the most dependable (?). The 5 Ω source resistance seems OK. We can now calculate the basic S/N of the source as 20*log(0.2mV/43nV)= 73.3 dB.
Then on to the amplification options: using a parallel doublet of E288CC triode sections, biased at 20 mA each, gives a noise figure of 11.3. This would result in 62 dB of S/N distance: TOO little! You propose to amegliorate the situation by using negative feedback. But to obtain the (local) feedback, you need resitors that, in practical terms, are bigger than the source resistance.... I am afraid this route won't bring you much good. In general, for low noise it is realy much better to have the largest amplification in the first section of the amp.
The other alternative inputs for the Asaka:
a singlet (1) of SK170BL @ 12 mA bias current: Noise Figure (NF) 7.3 dB
a triplet (3) of SK170GR @ 4 mA each: NF 5.3 dB
This JFET (or multiple) can very nicely be used in a circuit like Allen Wright's Vacuum State FVP5:
http://www.vacuumstate.com/
You could also opt for a step-up transformer with relatively low internal resistance. As an example the Lundahl LL 9206 (1:20) would give a NF of just 2.8 dB. Respectable at such input level! You would just have to loose some amplification further downstream.
And if you want to go all the way, consider the circuit of (agian Allen's, also to be found on the same site) RTP3C, using a special low-noise bipolar junction transistor such as the 2SC3329. The biasing requirements of bipolars make a differential (in this case balanced) circuit preferable... More complicated, but it pays off, look at the numbers:
A single 2SC3329 per leg of the differential pair @ 12 mA per leg: NF = 4.0 dB
a triplet 2SC3329s @ 4 mA each: NF = 3.0 dB.
The phonostage I built uses the RTP3C scheme, with a quartet of mentioned BJTs per differential leg. With a regulated rail voltage of 330 V this circuit gives me 57 dB amplification in one stage. Oh yeah, I apply 3 dB of emitter degradation (local feedback) to improve linearity and to get the right amplification to go from 400μV to -10dBV (standard for consumer line level). This is about as far as I could go
remember to listen to lots of music!
kind regards, Mark
Hi Paolo,
you show real commitment to the tube cause! And hey, the E288CC is a very, very nice tube. I am sorry to let you know, this one won't save you either.....
Just crunch the numbers: a Linn Asaka is low output. I found confusing numbers ranging from 0.1 to (your) 0.25 mV. I guess the 200μV @ 5cm/s in some test is the most dependable (?). The 5 Ω source resistance seems OK. We can now calculate the basic S/N of the source as 20*log(0.2mV/43nV)= 73.3 dB.
Then on to the amplification options: using a parallel doublet of E288CC triode sections, biased at 20 mA each, gives a noise figure of 11.3. This would result in 62 dB of S/N distance: TOO little! You propose to amegliorate the situation by using negative feedback. But to obtain the (local) feedback, you need resitors that, in practical terms, are bigger than the source resistance.... I am afraid this route won't bring you much good. In general, for low noise it is realy much better to have the largest amplification in the first section of the amp.
The other alternative inputs for the Asaka:
a singlet (1) of SK170BL @ 12 mA bias current: Noise Figure (NF) 7.3 dB
a triplet (3) of SK170GR @ 4 mA each: NF 5.3 dB
This JFET (or multiple) can very nicely be used in a circuit like Allen Wright's Vacuum State FVP5:
http://www.vacuumstate.com/
You could also opt for a step-up transformer with relatively low internal resistance. As an example the Lundahl LL 9206 (1:20) would give a NF of just 2.8 dB. Respectable at such input level! You would just have to loose some amplification further downstream.
And if you want to go all the way, consider the circuit of (agian Allen's, also to be found on the same site) RTP3C, using a special low-noise bipolar junction transistor such as the 2SC3329. The biasing requirements of bipolars make a differential (in this case balanced) circuit preferable... More complicated, but it pays off, look at the numbers:
A single 2SC3329 per leg of the differential pair @ 12 mA per leg: NF = 4.0 dB
a triplet 2SC3329s @ 4 mA each: NF = 3.0 dB.
The phonostage I built uses the RTP3C scheme, with a quartet of mentioned BJTs per differential leg. With a regulated rail voltage of 330 V this circuit gives me 57 dB amplification in one stage. Oh yeah, I apply 3 dB of emitter degradation (local feedback) to improve linearity and to get the right amplification to go from 400μV to -10dBV (standard for consumer line level). This is about as far as I could go
remember to listen to lots of music!
kind regards, Mark
Paolo,
I used a tube MC pre-preamplifier with "success". It consisted of both halves of an E88CC parallel, no cathode resistor, Vp about 50V. The input grid resistor was 100 ohms for my cartridge, and the anode resistor in the kohm order. It was a bit noisy but very nice sound... I also tried an SRPP arrangement without resistors, it worked also well. I measured about 60 dB S/N (A-weighted), and I think it is quite difficult to obtain > 60 dB with tubes if the input signal is in the order of 0.2 ... 0.3 mV. I think the transformer or transformer+tube is the way to go.
Laszlo
I used a tube MC pre-preamplifier with "success". It consisted of both halves of an E88CC parallel, no cathode resistor, Vp about 50V. The input grid resistor was 100 ohms for my cartridge, and the anode resistor in the kohm order. It was a bit noisy but very nice sound... I also tried an SRPP arrangement without resistors, it worked also well. I measured about 60 dB S/N (A-weighted), and I think it is quite difficult to obtain > 60 dB with tubes if the input signal is in the order of 0.2 ... 0.3 mV. I think the transformer or transformer+tube is the way to go.
Laszlo
Re: Commitment
Hey Mark,
Your math is like terminator here!
Very helpful, thanks!
I have to reflect a lot now before make some "disaster"!
P.S.: Indeed, my problem is that I listen too much hours/day with outrageous hi-end system. I am totally obsessed by hi-fi and need always better...
Add the fact my ear is prodigious and I am not rich, so you can imagine because I must DIY only and avoid accurately audio-shop!
Cheers,
Paolo
bambo said:Hi Paolo,
you show real commitment to the tube cause! And hey, the E288CC is a very, very nice tube. I am sorry to let you know, this one won't save you either.....
Just crunch the numbers: a Linn Asaka is low output. I found confusing numbers ranging from 0.1 to (your) 0.25 mV. I guess the 200¼V @ 5cm/s in some test is the most dependable (?). The 5 © source resistance seems OK. We can now calculate the basic S/N of the source as 20*log(0.2mV/43nV)= 73.3 dB.
Then on to the amplification options: using a parallel doublet of E288CC triode sections, biased at 20 mA each, gives a noise figure of 11.3. This would result in 62 dB of S/N distance: TOO little! You propose to amegliorate the situation by using negative feedback. But to obtain the (local) feedback, you need resitors that, in practical terms, are bigger than the source resistance.... I am afraid this route won't bring you much good. In general, for low noise it is realy much better to have the largest amplification in the first section of the amp.
The other alternative inputs for the Asaka:
a singlet (1) of SK170BL @ 12 mA bias current: Noise Figure (NF) 7.3 dB
a triplet (3) of SK170GR @ 4 mA each: NF 5.3 dB
This JFET (or multiple) can very nicely be used in a circuit like Allen Wright's Vacuum State FVP5:
http://www.vacuumstate.com/
You could also opt for a step-up transformer with relatively low internal resistance. As an example the Lundahl LL 9206 (1:20) would give a NF of just 2.8 dB. Respectable at such input level! You would just have to loose some amplification further downstream.
And if you want to go all the way, consider the circuit of (agian Allen's, also to be found on the same site) RTP3C, using a special low-noise bipolar junction transistor such as the 2SC3329. The biasing requirements of bipolars make a differential (in this case balanced) circuit preferable... More complicated, but it pays off, look at the numbers:
A single 2SC3329 per leg of the differential pair @ 12 mA per leg: NF = 4.0 dB
a triplet 2SC3329s @ 4 mA each: NF = 3.0 dB.
The phonostage I built uses the RTP3C scheme, with a quartet of mentioned BJTs per differential leg. With a regulated rail voltage of 330 V this circuit gives me 57 dB amplification in one stage. Oh yeah, I apply 3 dB of emitter degradation (local feedback) to improve linearity and to get the right amplification to go from 400¼V to -10dBV (standard for consumer line level). This is about as far as I could go
remember to listen to lots of music!
kind regards, Mark
Hey Mark,
Your math is like terminator here!
Very helpful, thanks!
I have to reflect a lot now before make some "disaster"!
P.S.: Indeed, my problem is that I listen too much hours/day with outrageous hi-end system. I am totally obsessed by hi-fi and need always better...
Add the fact my ear is prodigious and I am not rich, so you can imagine because I must DIY only and avoid accurately audio-shop!
Cheers,
Paolo
I have another idea for buiding a MC amplifier. The idea is that an MC cartridge needs low load impedance. This is why a step-up transformer is used, but what about this topology:
http://www.diyaudio.com/forums/showthread.php?threadid=100297
I tried it with a current-output DAC, but haven't tried with an MC cartridge yet. Also the turns ratio must be figured out.
http://www.diyaudio.com/forums/showthread.php?threadid=100297
I tried it with a current-output DAC, but haven't tried with an MC cartridge yet. Also the turns ratio must be figured out.
Re: Re: Commitment
Hi Paolo and Laszlo (and all other phono-lovers),
I have been thinking a little about Paolo's situation. What would be a satisfying way forward?
Laszlo, I left out any resistors for stabilisation.. It would have spoiled Paolo's dream a lot faster and more than my (unweighted!) noise calculations
Hope I did not scare anyone, and: you are welcome! Further, I completely agree with Laszlo: the best (simplest) way forward for you, Paolo is to include a step-up. My recommendations:
1) Sowter 8055 1:10 with OCC wire (goood!) @ €80,= approx
2) Lundahl LL1681 1:13 @ €85,= approx
3) Lundahl LL1931 (real high end! ) 1:8 @ €155,= approx
The lundahls could be used parallel: parallel to better match the very low output resistance of the asaka cartridge. This is normally not rcommended practise, because of the limited magnetic coupling. At very low inputlevels though, it's useable with fine effect: wired parallel to parallel, the 1681 would be low noise (noise figure 2.4 dB) while the the 1931 would be stellar (noise figure 1.3). Of course one would need two transformers per stereo pre-pre
Cheers,
Mark
Hi Paolo and Laszlo (and all other phono-lovers),
I have been thinking a little about Paolo's situation. What would be a satisfying way forward?
oshifis said:
Laszlo, I left out any resistors for stabilisation.. It would have spoiled Paolo's dream a lot faster and more than my (unweighted!) noise calculations
inertial said:
Your math is like terminator here!
Very helpful, thanks!
Hope I did not scare anyone, and: you are welcome! Further, I completely agree with Laszlo: the best (simplest) way forward for you, Paolo is to include a step-up. My recommendations:
1) Sowter 8055 1:10 with OCC wire (goood!) @ €80,= approx
2) Lundahl LL1681 1:13 @ €85,= approx
3) Lundahl LL1931 (real high end!
) 1:8 @ €155,= approxThe lundahls could be used parallel: parallel to better match the very low output resistance of the asaka cartridge. This is normally not rcommended practise, because of the limited magnetic coupling. At very low inputlevels though, it's useable with fine effect: wired parallel to parallel, the 1681 would be low noise (noise figure 2.4 dB) while the the 1931 would be stellar (noise figure 1.3). Of course one would need two transformers per stereo pre-pre
Cheers,
Mark
oshifis said:
Hi Laszlo,
very interesting argument here!
Mine friend have experimentated various loads on his Asaka( re=5 ohm).
With great surprise, the Asaka sounds natural and full ONLY with low loads, to say 10 ohm or less
(With 470 ohm as suggested by Linn, the sound is fragile, anemic, near digital! No bass, only mid-high !!! Incredible.)
IMO, in this scenario the cartridge start to operate near in power mode instead in voltage mode.
In particular at 5 ohm we have perfect power mode, but the insertion loss is 6dB, no?
I am speculating this is the manner that Dynavector PHA100 and 200 works. Exerpt they claims current mode operation.
But PHA's has a input selector at 3/4 ohm - 10/12 ohm and 35/40 ohm, if I well remember.
Thus I am confused: to work in true current mode, your load have to be zero ohm, am I right? Equal for alls............
Thus, what really makes that selector?
IMO they work in power mode, not current mode, but this is obyovsly only my speculation
Cheers,
Paolo
Re: Re: Re: Commitment
Indeed, mine friend need only 6-9 dB gain, not 20!
Because his A. Loesch gain jet 1500 times !
Unfortunately he is totally unreasonable about transformers, he has fear to try ( purchase) and reject it because bad sound...
I am sure a good Lundhal could be the way to go, but I repeat this man is very ostinate!!
Thanks for very good suggestions, Mark
Cheers,
Paolo
bambo said:Hi Paolo and Laszlo (and all other phono-lovers),
Hope I did not scare anyone, and: you are welcome! Further, I completely agree with Laszlo: the best (simplest) way forward for you, Paolo is to include a step-up. My recommendations:
1) Sowter 8055 1:10 with OCC wire (goood!) @ €80,= approx
2) Lundahl LL1681 1:13 @ €85,= approx
3) Lundahl LL1931 (real high end!
The lundahls could be used parallel: parallel to better match the very low output resistance of the asaka cartridge. This is normally not rcommended practise, because of the limited magnetic coupling. At very low inputlevels though, it's useable with fine effect: wired parallel to parallel, the 1681 would be low noise (noise figure 2.4 dB) while the the 1931 would be stellar (noise figure 1.3). Of course one would need two transformers per stereo pre-pre
Cheers,
Mark
Indeed, mine friend need only 6-9 dB gain, not 20!
Because his A. Loesch gain jet 1500 times !
Unfortunately he is totally unreasonable about transformers, he has fear to try ( purchase) and reject it because bad sound...
I am sure a good Lundhal could be the way to go, but I repeat this man is very ostinate!!
Thanks for very good suggestions, Mark
Cheers,
Paolo
MC headamp
Hello Paolo,
You may want to look at this MC headamp of Frank de Groove, supposedly pretty quiet: http://www.diyaudio.com/forums/showthread.php?postid=485333#post485333
Hello Paolo,
You may want to look at this MC headamp of Frank de Groove, supposedly pretty quiet: http://www.diyaudio.com/forums/showthread.php?postid=485333#post485333
Hi Inertial,
here you can find a mc step up made with a 417A tube; it is working very good .
With an accurate dc supply it is quite. The reostat in input is to trim the value for mc cartridge.
http://www.multitask.it/step-up/mc-tube.bmp
http://www.multitask.it/step-up/mc-tubepower.bmp power supply
Hiere is the circuit at 1KHz, 3mV input, gain is about 20.
http://www.multitask.it/step-up/tube-1KHz-3mv-in.jpg
Here with 1KZ, 10mV in
http://www.multitask.it/step-up/tube-1KHz-10mv-in.jpg
Here the square wave at 10KHz
http://www.multitask.it/step-up/tube-10KHz.jpg
Bye
Walter Gentilucci
here you can find a mc step up made with a 417A tube; it is working very good .
With an accurate dc supply it is quite. The reostat in input is to trim the value for mc cartridge.
http://www.multitask.it/step-up/mc-tube.bmp
http://www.multitask.it/step-up/mc-tubepower.bmp power supply
Hiere is the circuit at 1KHz, 3mV input, gain is about 20.
http://www.multitask.it/step-up/tube-1KHz-3mv-in.jpg
Here with 1KZ, 10mV in
http://www.multitask.it/step-up/tube-1KHz-10mv-in.jpg
Here the square wave at 10KHz
http://www.multitask.it/step-up/tube-10KHz.jpg
Bye
Walter Gentilucci
Please, try to build it and and listen.
If you look on spectrum you can see where is a noise floor;
my test set is not very sophisticated but I can check the performances of any circuit I build.
Regarding the filaments they must be in dc by a LM317 or similar.
The C-L-C I used is efficent and the results are good for ripple.
I tried also with a tube rectifier, 6X4, with always a good results; the ripple is also low.
If you think that this is not enough, you can use a TL783C, high voltage regulator; with a simply resitor network you can manage until 125 Vdc with 700ma max. of current.
Here is another choice with a very good fet for this job:
http://www.multitask.it/step-up/mc-fet.bmp
The supply is a battery at 9 V
this is a spectrum at 3 mV in:
http://www.multitask.it/step-up/3mVin-63mVout-1KHz.jpg
Bye
Walter
If you look on spectrum you can see where is a noise floor;
my test set is not very sophisticated but I can check the performances of any circuit I build.
Regarding the filaments they must be in dc by a LM317 or similar.
The C-L-C I used is efficent and the results are good for ripple.
I tried also with a tube rectifier, 6X4, with always a good results; the ripple is also low.
If you think that this is not enough, you can use a TL783C, high voltage regulator; with a simply resitor network you can manage until 125 Vdc with 700ma max. of current.
Here is another choice with a very good fet for this job:
http://www.multitask.it/step-up/mc-fet.bmp
The supply is a battery at 9 V
this is a spectrum at 3 mV in:
http://www.multitask.it/step-up/3mVin-63mVout-1KHz.jpg
Bye
Walter
I forgot to say that normally I use transformers (from Sowter mainly).
The 8055 OCC (1:10) is one of the best I ever heard.
Also the 9990 OCC (1:20) is very good.
Regarding the matching with cartridge, with 1:20 (a good ratio for Asak) we have a Zload reflected on primary (if secondary is connected with the standard 47k) of 117 ohm, if we put in parallel another 47kohm we reach 58 ohm that can be good for a low impedance one.
Ciao
Walter
The 8055 OCC (1:10) is one of the best I ever heard.
Also the 9990 OCC (1:20) is very good.
Regarding the matching with cartridge, with 1:20 (a good ratio for Asak) we have a Zload reflected on primary (if secondary is connected with the standard 47k) of 117 ohm, if we put in parallel another 47kohm we reach 58 ohm that can be good for a low impedance one.
Ciao
Walter
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.