Folks,
This exercise is interesting, not the least because it shows our collective experience. However, I won't pursue the car analogy further.
I had another look at Frank's clever balanced/SE circuit. As it happens it does not preserve gain, since the output tracks just one input, although the other seamlessly combines with it to further lower Zout. It strictly would only be useful if the output requirement were SE, for those unfortunate mainstream punters, no less.
However, the requirement is to drive a PP tube amp, and like Brett, I'd use the 813 in triode with zero negative feedback. This would obviously require further voltage amplification. Let's see, with 50W into 8R, we need about 58Vpp. Not sure of the turns ratio on a couple of 813s, probably around 17.5 (assuming 5K plate to plate, that about right Brett?), so this would mean a combined voltage swing of 1015, viz 507Vpp at each output tube plate. If the outputs each have a mu of 8, operating into 8R loads we should be able to get by with 70Vpp at each grid.
So, from our MC cartridge at 300uV, a balanced hybrid JFET/ECC88 cascode gives us a gain of 180, RIAA costs about 20dB, down to 18, we are up to 6mV. I will keep to octals. We need a further gain of 300 (49.5dB) to reach line level. I select a 6SL7 (Av of 43 with 100K load at 340V B+ running 1.6mA per section) followed by a 6BX7-GT (Av of 7.5 with 4K7 plate resistor, 150V/32mA across/through tube, -9V5 bias). This then gives us a grunty line output straight from the plates, with Zout around 1K03, which should be sufficient to drive cables to a power amp.
The power amplifier could actually get away with just an EL84 in pentode as a driver, coupled capacitively to the 813 grids. With a gain of 100, this easily gives us full drive. Truth is, you could probably scale the gain down a little, to say 45, using feedback from the plate, to drive the outputs even harder.
The primary advantage of this topology is a near constant power draw from the various supplies owing to the push/pull nature of the balanced configuration. This has huge advantages for the bass particularly, as it stiffens the power rail, holding much more stable voltage with minimal regulation and filtering. In fact, hum present on the plate loads is automatically cancelled all the way along since the outputs in every case are differential.
Hope this pushes the thread along.....
Notice that little attempt is made to balance the gains of each section. In my experience, and as I understand it that of others, gain asymmetry does add a little H2 but it's not really an issue for the sonics.
Cheers,
Hugh
This exercise is interesting, not the least because it shows our collective experience. However, I won't pursue the car analogy further.
I had another look at Frank's clever balanced/SE circuit. As it happens it does not preserve gain, since the output tracks just one input, although the other seamlessly combines with it to further lower Zout. It strictly would only be useful if the output requirement were SE, for those unfortunate mainstream punters, no less.
However, the requirement is to drive a PP tube amp, and like Brett, I'd use the 813 in triode with zero negative feedback. This would obviously require further voltage amplification. Let's see, with 50W into 8R, we need about 58Vpp. Not sure of the turns ratio on a couple of 813s, probably around 17.5 (assuming 5K plate to plate, that about right Brett?), so this would mean a combined voltage swing of 1015, viz 507Vpp at each output tube plate. If the outputs each have a mu of 8, operating into 8R loads we should be able to get by with 70Vpp at each grid.
So, from our MC cartridge at 300uV, a balanced hybrid JFET/ECC88 cascode gives us a gain of 180, RIAA costs about 20dB, down to 18, we are up to 6mV. I will keep to octals. We need a further gain of 300 (49.5dB) to reach line level. I select a 6SL7 (Av of 43 with 100K load at 340V B+ running 1.6mA per section) followed by a 6BX7-GT (Av of 7.5 with 4K7 plate resistor, 150V/32mA across/through tube, -9V5 bias). This then gives us a grunty line output straight from the plates, with Zout around 1K03, which should be sufficient to drive cables to a power amp.
The power amplifier could actually get away with just an EL84 in pentode as a driver, coupled capacitively to the 813 grids. With a gain of 100, this easily gives us full drive. Truth is, you could probably scale the gain down a little, to say 45, using feedback from the plate, to drive the outputs even harder.
The primary advantage of this topology is a near constant power draw from the various supplies owing to the push/pull nature of the balanced configuration. This has huge advantages for the bass particularly, as it stiffens the power rail, holding much more stable voltage with minimal regulation and filtering. In fact, hum present on the plate loads is automatically cancelled all the way along since the outputs in every case are differential.
Hope this pushes the thread along.....
Notice that little attempt is made to balance the gains of each section. In my experience, and as I understand it that of others, gain asymmetry does add a little H2 but it's not really an issue for the sonics.
Cheers,
Hugh
Hi,
Gain between left an and right channels should track perfectly though.
Gain asymmetry does not ADD H2 distortion IMHO, it just doesn't suppress any distortion already present in a SE design.
Instead of concentrating on the amp part, I'd rather concentrate on the more delicate parts of the balanced preamp where things are much more delicate.
The RIAA correction in particular is no piece of cake...neither is the MC stage ( if you want to go all tubes, that is).
No wonder this thread's so quiet..🙂
And then, there's the PSU...boy, this is going to be a true gasfactory....
I'll present my MC part based on my MC Hammer circuit somewhere nextweek...I hope...
Guys, this is an interesting project that can benefit you all...I'm puzzled by the lack of interest.
Don't tell me you're all just lazy?
Cheers,😉
Gain between left an and right channels should track perfectly though.
Gain asymmetry does not ADD H2 distortion IMHO, it just doesn't suppress any distortion already present in a SE design.
Instead of concentrating on the amp part, I'd rather concentrate on the more delicate parts of the balanced preamp where things are much more delicate.
The RIAA correction in particular is no piece of cake...neither is the MC stage ( if you want to go all tubes, that is).
No wonder this thread's so quiet..🙂
And then, there's the PSU...boy, this is going to be a true gasfactory....
I'll present my MC part based on my MC Hammer circuit somewhere nextweek...I hope...
Guys, this is an interesting project that can benefit you all...I'm puzzled by the lack of interest.
Don't tell me you're all just lazy?
Cheers,😉
fdegrove said:
Well, it's gotta use tubes and it can't use transformers. And there went my interest. 😴
se
Frank,
Perhaps your last was on the money? Hijacked by prejudice and active disinterest?
Ah, well, perhaps I will just quietly develop it myself out of interest......
Cheers,
Hugh
Perhaps your last was on the money? Hijacked by prejudice and active disinterest?
Ah, well, perhaps I will just quietly develop it myself out of interest......
Cheers,
Hugh
Balanced phone preamp
Some links where a similar topic was discussed :
http://www.tubecad.com/october99/page14.html
http://www.tubecad.com/october99/page10.html
Some links where a similar topic was discussed :
http://www.tubecad.com/october99/page14.html
http://www.tubecad.com/october99/page10.html
There IS real interest.
Frank, I'll gladly contribute to this fine thread when time allows. Life is full of other things that must be done at present.
Cheers,
Frank, I'll gladly contribute to this fine thread when time allows. Life is full of other things that must be done at present.
Cheers,
Hi,
Rest assured other people must be interested as well.
It not so hard to do but it seems you have a major headstart here, Hugh.
I judge from your past experience with circuits you contributed to...no need to pull up any names here.
To Steve Eddy,
Using xformers is relatively easy...but only relatively so.
We decided to avoid the easier way out and explore new ways..., if that bores you then let me remind you that there's still active balanced cicuitry required between stages.
I don't consider that boring...
Hugh,
What you have in mind for the MC stage is a balanced hybrid stage, probably using a 2SK...something with a valve on top as a cascode.
I would parallel a couple of 6DJ8s or anything else low noise and linear with Vg=0 and high transconductance.
I could do the whole thing with 6DJ8s and the main reason for doing so is that I won't have to worry about availablity.
We could use fancy tubes such as the 3A/167 but noise would be an issue.
The 6Gk5 is a fine candidate for MM levels, not suited for MC though, IMO.
Nobody seems to be posting circuits so I'll stick my neck out...all valves are 6DJ8s.
Since I'm only human, there must be mistakes...
John,
No problem...we all know you have got your hands full now...
Cheers,😉
Rest assured other people must be interested as well.
It not so hard to do but it seems you have a major headstart here, Hugh.
I judge from your past experience with circuits you contributed to...no need to pull up any names here.
To Steve Eddy,
Using xformers is relatively easy...but only relatively so.
We decided to avoid the easier way out and explore new ways..., if that bores you then let me remind you that there's still active balanced cicuitry required between stages.
I don't consider that boring...
Hugh,
What you have in mind for the MC stage is a balanced hybrid stage, probably using a 2SK...something with a valve on top as a cascode.
I would parallel a couple of 6DJ8s or anything else low noise and linear with Vg=0 and high transconductance.
I could do the whole thing with 6DJ8s and the main reason for doing so is that I won't have to worry about availablity.
We could use fancy tubes such as the 3A/167 but noise would be an issue.
The 6Gk5 is a fine candidate for MM levels, not suited for MC though, IMO.
Nobody seems to be posting circuits so I'll stick my neck out...all valves are 6DJ8s.
Since I'm only human, there must be mistakes...
John,
No problem...we all know you have got your hands full now...
Cheers,😉
Attachments
fdegrove said:
I didn't say that the active balanced circuitry between stages was boring.
Here's the deal. I don't know enough about tubes to contribute anything worthwhile to this thread. So I thought I'd suggest some possibilities using transformers to those who do know enough about tubes to contribute to this thread. But transformers have been ruled out so what's left beside 😴
se
Steve,
Don't fence yourself out because there are no transformers! You can pick up on this stuff pretty easily.......
You know, Bob Sugden, of Lancroft Transformers in Sydney, once said to a good friend of mine that the best transformer was no transformer! He reckoned they damaged the sound, and the best way to avoid it was to design 'em out.
Dunno. Maybe he's right. Bob's a clever, methodical guy.
However, I see transformers on tubes as aqualungs on divers. It's still possible, however, and a lot more nimble, to dive with a snorkel.......
Of course, that's not to say trafos don't have their purposes. They do, particularly in pro-audio and studio work. But in hifi I say they are expensive, and probably superfluous with careful design.
For example, use of a known good tube like the 6BX7 as a preamp output gives a Zout around 1K. That's just fine for longish cables. And the balanced configuration, you'll be pleased to hear, is GREAT for tight bass since the power supply feeds a constant current and suffers no voltage sag.
Cheers,
Hugh
Don't fence yourself out because there are no transformers! You can pick up on this stuff pretty easily.......
You know, Bob Sugden, of Lancroft Transformers in Sydney, once said to a good friend of mine that the best transformer was no transformer! He reckoned they damaged the sound, and the best way to avoid it was to design 'em out.
Dunno. Maybe he's right. Bob's a clever, methodical guy.
However, I see transformers on tubes as aqualungs on divers. It's still possible, however, and a lot more nimble, to dive with a snorkel.......
Of course, that's not to say trafos don't have their purposes. They do, particularly in pro-audio and studio work. But in hifi I say they are expensive, and probably superfluous with careful design.
For example, use of a known good tube like the 6BX7 as a preamp output gives a Zout around 1K. That's just fine for longish cables. And the balanced configuration, you'll be pleased to hear, is GREAT for tight bass since the power supply feeds a constant current and suffers no voltage sag.
Cheers,
Hugh
AKSA said:
Well, if it's so easy that someone with no practical experience designing with tubes can pick it up in a short period of time and still be able to contribute something to this project, then those of you who already have practical experience should have had this thing nailed down days ago. 🙂
Personally I prefer to try things for myself and make up my own mind. With all due respect to Bob.
And I see them, in proper application, as gills rather than an aqualung or a snorkel. They can actually relieve a lot of burden. They're passive and don't require a power supply so power supply issues are moot. And because they're not tied to the power supply they can't modulate the supply and interfere with other portions of the circuit. They're inherently stable and cannot oscillate. They don't produce prodigous amounts of thermal noise and don't produce any shot noise. Using just two lengths of wire and some MuMetal, they can provide far more linear voltage gain from the mid-bass on up than any single active device that I'm aware of. This relieves the active devices from having to perform voltage gain allowing them to operate much more linearly as followers where they can compliment the transformers as impedance buffers.
Bottom line, I see them as allowing for greater simplicity while also maintaining overall linearity equal to or greater than more complex circuits. And I don't see that as any sort of encumbrance. But then that's just me.
Certainly they can be expensive. And I can understand that. But tubes can be expensive too. And just because something is DIY doesn't mean it has to be cheap.
And that'd be fine if sound quality all boiled down to output impedance. We could just use solid state devices and do even better. 🙂
Look, I don't care if y'all don't want to consider transformers. I use them because they satisfy me. And whether anyone else is satisfied with them or not doesn't change that. All I'm saying is that once transformers were ruled out, there simply wasn't much else I could contribute to this thread. That's all. I'm not sitting around feeling sorry for myself or anything. So don't feel that you have to justify your not wanting to consider transformers. Makes no difference to me either way.
se
No problem
like to go balanced all the way, all tube, no transfo, reasonable noise and hiss
Reinhard
like to go balanced all the way, all tube, no transfo, reasonable noise and hiss
Reinhard
Balanced MC
Hi Reinhard
I did stop my attempts to build an all tube balanced MC Phono Stage in 1993 (ten PCC88s in the box)
Guess why: It was too noisy, though I did enjoy the sound my AT-cart was too low on output.
I gave it to a friend then to play his EMT-cart through it.
- better use transformers -
Hi Reinhard
I did stop my attempts to build an all tube balanced MC Phono Stage in 1993 (ten PCC88s in the box)
Guess why: It was too noisy, though I did enjoy the sound my AT-cart was too low on output.
I gave it to a friend then to play his EMT-cart through it.
- better use transformers -
Dear Reinhard
I did three or four versions of the all tube balanced MC-Stage and I´m not sure after all this years what actual was in the prototype box when it went to a good friend.
It would be comfortable to see your design so we can discuss your problems together on the forum this way.
What my designs had in common was the floating balanced arrangement of the passive RIAA network.
I still have some of my old breed-boards sleeping on the shelf but no conctrete schematic to post at the moment.
What I could get from the all tube MC-design was ~ 50db signal to noise with an average (medium output) MC-cartridge.
Since this is above the vinyl surface noise level I stopped to workout and have moved to transformers on both sides of the Phono Stage.
With these there is no need for a balanced circuit layout any more.
A different solution is what Allen Wright did with Fets in his RTP3C design and off concern in a tube topic.
- I would be curious about your ideas -
Klaus
I did three or four versions of the all tube balanced MC-Stage and I´m not sure after all this years what actual was in the prototype box when it went to a good friend.
It would be comfortable to see your design so we can discuss your problems together on the forum this way.
What my designs had in common was the floating balanced arrangement of the passive RIAA network.
I still have some of my old breed-boards sleeping on the shelf but no conctrete schematic to post at the moment.
What I could get from the all tube MC-design was ~ 50db signal to noise with an average (medium output) MC-cartridge.
Since this is above the vinyl surface noise level I stopped to workout and have moved to transformers on both sides of the Phono Stage.
With these there is no need for a balanced circuit layout any more.
A different solution is what Allen Wright did with Fets in his RTP3C design and off concern in a tube topic.
- I would be curious about your ideas -
Klaus
Some food for thought (and the previous blog)
http://www.tubecad.com/2007/09/blog0120.htm
I've build the Wright RTP3 in a couple of versions, and it was very good.
http://www.tubecad.com/2007/09/blog0120.htm
I've build the Wright RTP3 in a couple of versions, and it was very good.
Balanced
Hello,
I have no practical design, maybe Allens circuits are good from a technical viewpoint, I don´t like the sound, cristal clear but no emotion.
Wonder how the guys from BAT get a S/N of 78 dB with a single
6922 and 59 dB gain.
Had a look at Broskies circuit with grounded cathode and floating grids. Don´t know what to think about.
Regards,
Reinhard
Hello,
I have no practical design, maybe Allens circuits are good from a technical viewpoint, I don´t like the sound, cristal clear but no emotion.
Wonder how the guys from BAT get a S/N of 78 dB with a single
6922 and 59 dB gain.
Had a look at Broskies circuit with grounded cathode and floating grids. Don´t know what to think about.
Regards,
Reinhard
- Status
- Not open for further replies.