Konnichiwa,
Well, that "DCMB" is actually called "Loftin/White", after the folks who invented and patented it in the late 20's (split supply or not makes zip difference).
Please observe the following caution from the RCA Tube Manual:
"The 50 is not recmmended for resistance coupled circuits."
In fact, the gridreturn resistance to the cathode should not exceed 10KOhm according to some extant specification sheets.
You may wish to replace the Anode resistor by a suitably selected anode load choke to satisfy this requirement.
This will also help attaining sensible levels of output power from the circuit, which is otherwise too limited as the anode current of the Driver Valve is unlikely to even approach 0mA and thus the output valve cannot be driven with anything approaching a peak voltage equal to it's bias voltage.
Finally, in order to eliminate a good quality coupling capacitor you have replaced it with a usually much poorer quality powersupply capacitor which no doubt will be unlikely to aid sound quality.
I would submit that the use of anode load chokes and Interstage transformers or coupling capacitors/gridchokes presents an alternative solution thathas many advantages over a split supply Loftin/White, the only drawback is cost but even that can be addressed.
Sayonara
ari polisois said:
Well, that "DCMB" is actually called "Loftin/White", after the folks who invented and patented it in the late 20's (split supply or not makes zip difference).
ari polisois said:
Please observe the following caution from the RCA Tube Manual:
"The 50 is not recmmended for resistance coupled circuits."
In fact, the gridreturn resistance to the cathode should not exceed 10KOhm according to some extant specification sheets.
You may wish to replace the Anode resistor by a suitably selected anode load choke to satisfy this requirement.
This will also help attaining sensible levels of output power from the circuit, which is otherwise too limited as the anode current of the Driver Valve is unlikely to even approach 0mA and thus the output valve cannot be driven with anything approaching a peak voltage equal to it's bias voltage.
Finally, in order to eliminate a good quality coupling capacitor you have replaced it with a usually much poorer quality powersupply capacitor which no doubt will be unlikely to aid sound quality.
I would submit that the use of anode load chokes and Interstage transformers or coupling capacitors/gridchokes presents an alternative solution thathas many advantages over a split supply Loftin/White, the only drawback is cost but even that can be addressed.
Sayonara
sorry for the intrusion...
Hi to all, guys!
I tought I was the prolific one... but wherever some of you appear, I just have to go to read because it helps me get a better picture of myself 🙂 🙂 🙂
Ari, I could not actually imagine you were Italian?! Adesso mi sei molto ma molto piu simpatico... 🙂 Peccato che questo thread sia in inglese...
Well, after this intrusion, something more constructive: I think this is all about a 50 amp, and a simple one at that!!! The only circuit so far that looks like simple (I mean, simple is low parts count, eventually low budget as well -- without loosing that much potential...) is the Loftin-White (did not look at it, actually... but I know how that might look like, so the values of the components do not make much difference generally).
Isn't a simple amp a driver + output tube? With low parts count, something you can build out of "diletto" and once you get to know it, eventually improve or redo in order to learn something from practice?
Regards to all,
Aleksandar
Hi to all, guys!
I tought I was the prolific one... but wherever some of you appear, I just have to go to read because it helps me get a better picture of myself 🙂 🙂 🙂
Ari, I could not actually imagine you were Italian?! Adesso mi sei molto ma molto piu simpatico... 🙂 Peccato che questo thread sia in inglese...
Well, after this intrusion, something more constructive: I think this is all about a 50 amp, and a simple one at that!!! The only circuit so far that looks like simple (I mean, simple is low parts count, eventually low budget as well -- without loosing that much potential...) is the Loftin-White (did not look at it, actually... but I know how that might look like, so the values of the components do not make much difference generally).
Isn't a simple amp a driver + output tube? With low parts count, something you can build out of "diletto" and once you get to know it, eventually improve or redo in order to learn something from practice?
Regards to all,
Aleksandar
Konnichiwa,
Another thought, if the transformer offers sufficient current you could use a hybrid rectifier and get basically double the HT Voltage, more than enough for a classic direct coupled Amplifier, even with near choke input supplies.
We did all the math involved for the 45 Amp, so it should be easy to adopt this for the 50. If we use the 400V/60mA/-65V (appx) and operate the 417A Driver at 195V/10mA we need an HT of around 670V, just right for the Lundahl Transformer and choke input.
BTW, attached for amusement the design of the original Loftin/White Type 50 Amplifer.
That one gets around the Gridresistance limit by using what in effect amounts to a DC servo loop. Not bad for around 1927.
Power Transformer supplies 820V AC BTW..... The type 81 halve wave rectifier is not recommended in modern circuits, we can do better now. If you build a hybrid bridge a type 80 Rectifier could just about supply a stereo Amp.
Sayonara
arnoldc said:
Another thought, if the transformer offers sufficient current you could use a hybrid rectifier and get basically double the HT Voltage, more than enough for a classic direct coupled Amplifier, even with near choke input supplies.
We did all the math involved for the 45 Amp, so it should be easy to adopt this for the 50. If we use the 400V/60mA/-65V (appx) and operate the 417A Driver at 195V/10mA we need an HT of around 670V, just right for the Lundahl Transformer and choke input.
BTW, attached for amusement the design of the original Loftin/White Type 50 Amplifer.
An externally hosted image should be here but it was not working when we last tested it.
That one gets around the Gridresistance limit by using what in effect amounts to a DC servo loop. Not bad for around 1927.
Power Transformer supplies 820V AC BTW..... The type 81 halve wave rectifier is not recommended in modern circuits, we can do better now. If you build a hybrid bridge a type 80 Rectifier could just about supply a stereo Amp.
Sayonara
What if it was simple?
OK, after my last post here, let me be constructive. I just imagined I was the lucky guy being given a pair of 50s...
Here's an output stage... the transformer might really be that one, but generally speaking I would opt for a smaller one, since I think that this one is too large (oversize primary wire, for more than twice the DC current...).
OK, after my last post here, let me be constructive. I just imagined I was the lucky guy being given a pair of 50s...
Here's an output stage... the transformer might really be that one, but generally speaking I would opt for a smaller one, since I think that this one is too large (oversize primary wire, for more than twice the DC current...).
Attachments
Driver stage
Well, now I need a driver. This attachment is what I would opt for -- and it is indeed simple and effective.
After this one, all you need is a decent power supply -- you can always use i.e. Duncan's PSUD2 to calculate one: I would do something like 10H - 500uF - 10H - 500uF to the output tubes and an additional little choke and some 10-20uF oil cap for the driver tube... with an 80 or 5W4 in a hybrid bridge configuration... but any other arrangement might do just fine...
Simple, OK? I think even a beginner could build this one... unfortunately for the begineers, they seldom have 50s at hand.
Regards to all,
Aleksandar
Well, now I need a driver. This attachment is what I would opt for -- and it is indeed simple and effective.
After this one, all you need is a decent power supply -- you can always use i.e. Duncan's PSUD2 to calculate one: I would do something like 10H - 500uF - 10H - 500uF to the output tubes and an additional little choke and some 10-20uF oil cap for the driver tube... with an 80 or 5W4 in a hybrid bridge configuration... but any other arrangement might do just fine...
Simple, OK? I think even a beginner could build this one... unfortunately for the begineers, they seldom have 50s at hand.
Regards to all,
Aleksandar
Attachments
THOSE TWO, AGAIN...
Guys, will someone tell me what is going on? Who is this Thorsten guy who seems to have a Korean or anyway Asiatic name? And, why is Serengetiplains always after him? Why is it that Serengetiplains always posts one particular kind of post (do a research on his posting, and you will get to that conclusion)? Finally, I myself might have character flaws (who does not?) but at least I try to be constructive (i.e. Arnoldc mentioned TubeCAD) and never mock others!
Regards to all you folks of good will (or should it be "good willed folks" -- maybe someone would like to give me grammar lessons as well, although this is just a forum to exchange tube ideas and opinions and not na IOLTS exam)...
Aleksandar
Guys, will someone tell me what is going on? Who is this Thorsten guy who seems to have a Korean or anyway Asiatic name? And, why is Serengetiplains always after him? Why is it that Serengetiplains always posts one particular kind of post (do a research on his posting, and you will get to that conclusion)? Finally, I myself might have character flaws (who does not?) but at least I try to be constructive (i.e. Arnoldc mentioned TubeCAD) and never mock others!
Regards to all you folks of good will (or should it be "good willed folks" -- maybe someone would like to give me grammar lessons as well, although this is just a forum to exchange tube ideas and opinions and not na IOLTS exam)...
Aleksandar
Arnold,
Have a look at the Loftin White, attached ( easier to read, your is huge, Thornsten).
Now, Thornsten, one could say that they are parents, but I am convinced that mine is much simpler ( I am talking about the adjustments you must make, I guess, quite often, with the L&W, considering the interconnections between all gradients).
You must have noticed that all the resistors are in series, from the B+ of 650 volts to ground and they have to supply the right amount of current to the driver stages, which in this specific case, does not draw much. What if you needed to ?
Also, may I suggest you read again the summary of the performances that the DCMB has, that I sent some time ago.
I don't think you can do the same with the Loftin White ( invented in 1928 ).
You seem to focus on aspects that are liable to be criticised.
This might be constructive, but please, make an effort to see the good sides, also.
Regarding the coupling capacitor that you barter with the electrolytics, please observe that the DCMB attacks straighforwardly grid and cathode of the output tube, transferring every detail the driver load resistor has collected. This is what I claim to be an ultimate transfer connection and it is not the case with a coupling capacitor, right ?
Unfortunately, electrolytics are commonly used around the circuits. If you don't like them and you are wealthy, all you have to do is to replace (all of them) with paper/oil ones or even better quality.
Anyway, to be honest, I don't dislike your objections. You can always learn something, anytime.
Sayonara.
Have a look at the Loftin White, attached ( easier to read, your is huge, Thornsten).
Now, Thornsten, one could say that they are parents, but I am convinced that mine is much simpler ( I am talking about the adjustments you must make, I guess, quite often, with the L&W, considering the interconnections between all gradients).
You must have noticed that all the resistors are in series, from the B+ of 650 volts to ground and they have to supply the right amount of current to the driver stages, which in this specific case, does not draw much. What if you needed to ?
Also, may I suggest you read again the summary of the performances that the DCMB has, that I sent some time ago.
I don't think you can do the same with the Loftin White ( invented in 1928 ).
You seem to focus on aspects that are liable to be criticised.
This might be constructive, but please, make an effort to see the good sides, also.
Regarding the coupling capacitor that you barter with the electrolytics, please observe that the DCMB attacks straighforwardly grid and cathode of the output tube, transferring every detail the driver load resistor has collected. This is what I claim to be an ultimate transfer connection and it is not the case with a coupling capacitor, right ?
Unfortunately, electrolytics are commonly used around the circuits. If you don't like them and you are wealthy, all you have to do is to replace (all of them) with paper/oil ones or even better quality.
Anyway, to be honest, I don't dislike your objections. You can always learn something, anytime.
Sayonara.
Attachments
Konnichiwa,
Actually the LW circuit with the Pentode driver uses the screen grid to introduce a DC servo loop which auto-adjusts the output Valve Bias and is quite resilient to valve parameter variations.
Correct. They do, BTW. Allways.
You re-arrange. the same principle scales to any sensible driver stage current levels. There are also variations with a similar stabilising DC loop for triode drivers, but there the DC control loop has much less gain and is less efficient. Well, Pentodes sound a lot better than triodes when driving triodes (the reverse is true the other way around) so nothing to worry.
ANYWAY, even with all these caveats, a LW using a single supply actually has a very stable operating point under all condtions, if the input valve sets the bias correctly (adjustment to suit different driver valves may be needed), simply because it has a ton of DC local feedback in the output stage.
Bottom line, classic LW needs normally no adjustment even with output stage valve change (input usually requires adjustment).
You only start to get into trouble once you use a split supply LW circuit (such as the one I showed in another thread), then the output valve has no DC dengeration adn a DC servo loop becomes neccesary.
The specific DCMB Circuit(s) from Audio Express and Valve have some additional features you omitted in your drawing, mainly the presence of a feedback loop acound the output valve.
Surprisingly (perhaps), you can.
My problem is that I do NOT see any good sides. Non at all, honest. Once you draw out the current loops all aparent good sides readily disappear and we are left with drawbacks only. You cannot even claim lower cost as a good powersupply for the driver and seperate supplies for the Output stages eat up the money "saved" by not using expensive gridchokes and coupling capacitors.
The LW with split supplies (aka DCMB) is one of the many possible ways to arrange circuits, I find that it merely trades one compromise for another.
You always need to account for the current return. It must return somewhere. In your case it has nowhere to go but the powersupply. Please draw out all the relevant current loops and you realise that instead of coupling between driver anode and output grid with a capacitor you now couple between driver cathode and output cathode through the powersupply.
You have not eliminated the capacitor (you cannot anyway), you have merely moved it around.
Yet it is not the ultimate transfer as the loop is closed through the supply.
I actually tend to do that for all my amplifier powerspplies anyway.
However, let us consider two options:
1) 417A Anode loaded by an 18k resistor and coupled via a capacitor to the 50 with a gridchoke (to satisfy the requirement of a gridcircuit return DCR < 10KOhm) using fixed grid bias.
We use a 0.22uF industrial Aluminum Foil & Polyroplene Film coupling capacitor between 471A anode and 50 Grid, cost appx $2.50...3.
2) For arguments sake we also apply fixed grid bias to our 417A driver valve and ground it's cathode. We use the same sort of 0.22uF coupling capacitor and a 100K gridresistor.
3) Our supply for driver and output is the same 400V Supply obtainable from one generic 350-0-350V Main transformer. We decouple the 417A Supply with 3k9 & 10uF MKP from the main supply. We use a capacitor of 47uF MKP as final supply filter capacitor for the main HT, I'd probably go CLC using electrolytic capacitors if I want to be economic, coupled via 100R to the 47uF MKP per channel.
We obtain the gridbias voltage also from rectifying the 350-0-350V winding using solid state diodes and very large series resistors. By making sure to load the bias supply (we must keep the gridreturn < 10K) we stabilise it, we require around 8mA per channel for 9k1/75V.
This supply is strongly insulated from the signal circuit by the gridchoke (say 1500H into a 2K Anode impedance of the 417A) for the output stage and by 100K into the source impedance for the input supply, however given that we tap off our input grid bias from a "long" voltage divider chain we can filter the supply further at little expense, eliminiating any supply noise from the bias supply and equally preventing audio signal leaking into the supply.
We now have made an amplifier that has exactly the same number of coupling capacitors as your DCMB circuit presented in this thread, namely 3 pcs. Of these one couples the output transformer +B to the output valves cathode, one couples the 417A Anode (output signal) to 50 Grid and one couples our signal to the 417A Grid.
In your circuit we find equally 3 Coupling Capacitors. One, just like in our circuit coupled the output transformer +B to the output valves cathode. Where are the other two coupling capacitors then, after all, the circuit is direct coupled? Well, one of them couples the signal (ground) to the cathode of the input valve and the other couples the signal (ground) to the output valve cathode.
Now how we optimise each coupling capacitor is our choice. However, given that in your circuit we require three rather large value coupling capacitors I would submit that the alternate circuit I suggest may be easier and ultimatly cheaper to optimise.
Plus, let us not forget the other advantage of the alternate circuit proposed, namely that it can actually drive the output valve to full power.
In your case we would have (for arguments sake) a 6SN7 with a 10K or lower anode load. In order to obtain 75V bias from a 10K resistor we require 7.5mA Anodic current.
In order to not overtax our 6SN7 we will use a 300V negative supply giving us a 6SN7 operated at 225V/7.5mA and approximatly 7.5V bias. We draw in our 10K load line. If we now take an anode voltage 75V below 225V (eg twice the bias of the 50). We find this at appx -1V grid bias, requiring a grid swing of 6.5V.
What is our current at -14V grid bias (6.5V grid swing from the quiescent point) for our 10K loadline? We find 1.5mA. Therefor we can only swing around 62V above the quescent point of the Anode. And we have just noticed that our driver stage produces a lot of distotion as well, while we are at it....
Of course, we can solve the gridswing problem and the distortion problem and we can also solve the problem of coupling capacitor quality. If we do, what do we end up with?
Something very much like what I suggested earlier, choke loaded Gridchoke/Capacitor or IT coupled circuits.
So, there is no free lunch (as usual) and once we apply a rigerous analysis to it DCMB does not seem to have any redeeming features.
Of course what will sound better to any given listener is a whole other ballgame. So by all means keep building and enjoying DCMB Amplifiers.
Sayonara
ari polisois said:
Actually the LW circuit with the Pentode driver uses the screen grid to introduce a DC servo loop which auto-adjusts the output Valve Bias and is quite resilient to valve parameter variations.
ari polisois said:
Correct. They do, BTW. Allways.
ari polisois said:
You re-arrange. the same principle scales to any sensible driver stage current levels. There are also variations with a similar stabilising DC loop for triode drivers, but there the DC control loop has much less gain and is less efficient. Well, Pentodes sound a lot better than triodes when driving triodes (the reverse is true the other way around) so nothing to worry.
ANYWAY, even with all these caveats, a LW using a single supply actually has a very stable operating point under all condtions, if the input valve sets the bias correctly (adjustment to suit different driver valves may be needed), simply because it has a ton of DC local feedback in the output stage.
Bottom line, classic LW needs normally no adjustment even with output stage valve change (input usually requires adjustment).
You only start to get into trouble once you use a split supply LW circuit (such as the one I showed in another thread), then the output valve has no DC dengeration adn a DC servo loop becomes neccesary.
ari polisois said:
The specific DCMB Circuit(s) from Audio Express and Valve have some additional features you omitted in your drawing, mainly the presence of a feedback loop acound the output valve.
ari polisois said:
Surprisingly (perhaps), you can.
ari polisois said:
My problem is that I do NOT see any good sides. Non at all, honest. Once you draw out the current loops all aparent good sides readily disappear and we are left with drawbacks only. You cannot even claim lower cost as a good powersupply for the driver and seperate supplies for the Output stages eat up the money "saved" by not using expensive gridchokes and coupling capacitors.
The LW with split supplies (aka DCMB) is one of the many possible ways to arrange circuits, I find that it merely trades one compromise for another.
ari polisois said:
You always need to account for the current return. It must return somewhere. In your case it has nowhere to go but the powersupply. Please draw out all the relevant current loops and you realise that instead of coupling between driver anode and output grid with a capacitor you now couple between driver cathode and output cathode through the powersupply.
You have not eliminated the capacitor (you cannot anyway), you have merely moved it around.
ari polisois said:
Yet it is not the ultimate transfer as the loop is closed through the supply.
ari polisois said:
I actually tend to do that for all my amplifier powerspplies anyway.
However, let us consider two options:
1) 417A Anode loaded by an 18k resistor and coupled via a capacitor to the 50 with a gridchoke (to satisfy the requirement of a gridcircuit return DCR < 10KOhm) using fixed grid bias.
We use a 0.22uF industrial Aluminum Foil & Polyroplene Film coupling capacitor between 471A anode and 50 Grid, cost appx $2.50...3.
2) For arguments sake we also apply fixed grid bias to our 417A driver valve and ground it's cathode. We use the same sort of 0.22uF coupling capacitor and a 100K gridresistor.
3) Our supply for driver and output is the same 400V Supply obtainable from one generic 350-0-350V Main transformer. We decouple the 417A Supply with 3k9 & 10uF MKP from the main supply. We use a capacitor of 47uF MKP as final supply filter capacitor for the main HT, I'd probably go CLC using electrolytic capacitors if I want to be economic, coupled via 100R to the 47uF MKP per channel.
We obtain the gridbias voltage also from rectifying the 350-0-350V winding using solid state diodes and very large series resistors. By making sure to load the bias supply (we must keep the gridreturn < 10K) we stabilise it, we require around 8mA per channel for 9k1/75V.
This supply is strongly insulated from the signal circuit by the gridchoke (say 1500H into a 2K Anode impedance of the 417A) for the output stage and by 100K into the source impedance for the input supply, however given that we tap off our input grid bias from a "long" voltage divider chain we can filter the supply further at little expense, eliminiating any supply noise from the bias supply and equally preventing audio signal leaking into the supply.
We now have made an amplifier that has exactly the same number of coupling capacitors as your DCMB circuit presented in this thread, namely 3 pcs. Of these one couples the output transformer +B to the output valves cathode, one couples the 417A Anode (output signal) to 50 Grid and one couples our signal to the 417A Grid.
In your circuit we find equally 3 Coupling Capacitors. One, just like in our circuit coupled the output transformer +B to the output valves cathode. Where are the other two coupling capacitors then, after all, the circuit is direct coupled? Well, one of them couples the signal (ground) to the cathode of the input valve and the other couples the signal (ground) to the output valve cathode.
Now how we optimise each coupling capacitor is our choice. However, given that in your circuit we require three rather large value coupling capacitors I would submit that the alternate circuit I suggest may be easier and ultimatly cheaper to optimise.
Plus, let us not forget the other advantage of the alternate circuit proposed, namely that it can actually drive the output valve to full power.
In your case we would have (for arguments sake) a 6SN7 with a 10K or lower anode load. In order to obtain 75V bias from a 10K resistor we require 7.5mA Anodic current.
In order to not overtax our 6SN7 we will use a 300V negative supply giving us a 6SN7 operated at 225V/7.5mA and approximatly 7.5V bias. We draw in our 10K load line. If we now take an anode voltage 75V below 225V (eg twice the bias of the 50). We find this at appx -1V grid bias, requiring a grid swing of 6.5V.
What is our current at -14V grid bias (6.5V grid swing from the quiescent point) for our 10K loadline? We find 1.5mA. Therefor we can only swing around 62V above the quescent point of the Anode. And we have just noticed that our driver stage produces a lot of distotion as well, while we are at it....
Of course, we can solve the gridswing problem and the distortion problem and we can also solve the problem of coupling capacitor quality. If we do, what do we end up with?
Something very much like what I suggested earlier, choke loaded Gridchoke/Capacitor or IT coupled circuits.
So, there is no free lunch (as usual) and once we apply a rigerous analysis to it DCMB does not seem to have any redeeming features.
Of course what will sound better to any given listener is a whole other ballgame. So by all means keep building and enjoying DCMB Amplifiers.
Sayonara
i'm overwhelmed 🙂
so much information to digest, but i already made a decision not to go DC.
FWIW, my power transformer is a Lundahl LL1648. if this will become a limitation, i am very much willing to have to this changed.
so much information to digest, but i already made a decision not to go DC.
FWIW, my power transformer is a Lundahl LL1648. if this will become a limitation, i am very much willing to have to this changed.
Alex Kitic said:
So did I. They are globes as well.... 🙁
Alex Kitic said:
Unfortunately that circuit would be dangerous to the 50s since it uses a 200K gridleak. The 50 is only spec'd for 10K between grid and filament (it's an old type valve... might have a soft-ish vacuum or something)
Alex Kitic said:
Do a google search for "Thorsten Loesch" and you will have a pretty good idea. On why he uses a name which sounds Chinese, but signs on and off with 'Konnichiwa', and 'Sayonara' (which are Japanese) - you'll have to ask him.
I have no idea who Tom (serengetiplains) is.
Jason
Arnoldc,
I hope you will choose the classic and proven and simple way I suggested. Do not be scared by the large grid-leak in the schematics, it is something in SE CAD that the user has no control of... I'll try to find your Lundahl trannie in SE CAD and prepare another output stage. Anyway, if the trannie is not there, it can be made "custom" -- just post the values for DCR, transformation ratio, etc.
Arnoldc and Jason,
I am not that familiar with the 50 except that I know it is one of the "rare league" od DHTs. The 10k resistor seems a little too low -- I believe that must be the max. value for negative bias circuits (it is 50k i.e. for the 2A3, if I did not get that one wrong) -- so the actual max. grid leak resistor is most probably somewhere around 100k. That is no problem, since it just means a larger cap will be needed between the stages, like 470nF to 1uF. Although larger caps sound worse than smaller ones (popular myth), I sincerely believe that having a 1uF cap between the stages will not geopardize the sound of the amp 🙂
Jason,
Thanks for the comments and the explanation. Yes, I've heard of Thorsten Loesh, and when I first read the name, I tought of him... he seems to have lots of time to spare? Anyway, is he not the guy who likes to put potentiometers behind the gain stage in an active preamp, instead of putting it before the gain stage? That alone tells a lot about him... and his choices (including mixing Japanese and Korean and Chinese).
When it comes to Tom, he's one with an axe to grind... although it seems that he is not out of the picture when it comes to knowledge.
Knowledge and prolificity... truth is, I've heard a saying that you can recognize a genius by the number of idiots who have gathered against him 🙂 Who knows, maybe it's us who just cannot understand the genius behind such people.
Regards to all,
Aleksandar
PS.
Lundahl makes interesting transformers -- it's not listed, but as I recall, their transformers can be wired in may ways, so they adapt to various impedances. If it is a model that can withstand the 50mA DC in the primary without saturation (as they seem to differ in gap, maybe wire as well), it will do nicely... and you can use the schematic I already posted for the output stage directly, just taking care of a few details:
- DCR resistance of the Lundahl primary (will affect the available voltage to the plate)
- the above mentioned grid leak resistor (I believe 100k to 200k would not damage the tube, but I must check it on the datasheet, i.e. download it first)
- consequently, I believe that a cap between the stages in the range 330nF to 1uF would be just fine.
Of course, the Lundahl trannie should be wired in such manner that your 50s "see" 4.8k-5k approximately while "looking" at your speakers thru the output transformer (as if it were a magnifying glass of sorts). 🙂
I hope you will choose the classic and proven and simple way I suggested. Do not be scared by the large grid-leak in the schematics, it is something in SE CAD that the user has no control of... I'll try to find your Lundahl trannie in SE CAD and prepare another output stage. Anyway, if the trannie is not there, it can be made "custom" -- just post the values for DCR, transformation ratio, etc.
Arnoldc and Jason,
I am not that familiar with the 50 except that I know it is one of the "rare league" od DHTs. The 10k resistor seems a little too low -- I believe that must be the max. value for negative bias circuits (it is 50k i.e. for the 2A3, if I did not get that one wrong) -- so the actual max. grid leak resistor is most probably somewhere around 100k. That is no problem, since it just means a larger cap will be needed between the stages, like 470nF to 1uF. Although larger caps sound worse than smaller ones (popular myth), I sincerely believe that having a 1uF cap between the stages will not geopardize the sound of the amp 🙂
Jason,
Thanks for the comments and the explanation. Yes, I've heard of Thorsten Loesh, and when I first read the name, I tought of him... he seems to have lots of time to spare? Anyway, is he not the guy who likes to put potentiometers behind the gain stage in an active preamp, instead of putting it before the gain stage? That alone tells a lot about him... and his choices (including mixing Japanese and Korean and Chinese).
When it comes to Tom, he's one with an axe to grind... although it seems that he is not out of the picture when it comes to knowledge.
Knowledge and prolificity... truth is, I've heard a saying that you can recognize a genius by the number of idiots who have gathered against him 🙂 Who knows, maybe it's us who just cannot understand the genius behind such people.
Regards to all,
Aleksandar
PS.
Lundahl makes interesting transformers -- it's not listed, but as I recall, their transformers can be wired in may ways, so they adapt to various impedances. If it is a model that can withstand the 50mA DC in the primary without saturation (as they seem to differ in gap, maybe wire as well), it will do nicely... and you can use the schematic I already posted for the output stage directly, just taking care of a few details:
- DCR resistance of the Lundahl primary (will affect the available voltage to the plate)
- the above mentioned grid leak resistor (I believe 100k to 200k would not damage the tube, but I must check it on the datasheet, i.e. download it first)
- consequently, I believe that a cap between the stages in the range 330nF to 1uF would be just fine.
Of course, the Lundahl trannie should be wired in such manner that your 50s "see" 4.8k-5k approximately while "looking" at your speakers thru the output transformer (as if it were a magnifying glass of sorts). 🙂
hi aleksandar,
my Lundahl LL1648 is for power supply duties and is limited to 350V, hence my hesitation about the higher voltage recommendations. but why compromise, right? so i'm willing to have another transformer to do power supply duties.
the output transformer is a Hammond 162x (the one that is 5K) or potentially a Tango U-808 in the future.
my Lundahl LL1648 is for power supply duties and is limited to 350V, hence my hesitation about the higher voltage recommendations. but why compromise, right? so i'm willing to have another transformer to do power supply duties.
the output transformer is a Hammond 162x (the one that is 5K) or potentially a Tango U-808 in the future.
output transformer
Hi, Arnold...
ANY output transformer will do for your project, provided it can do the following:
1) present a 4.8k-5k Ra to your output tubes (50)
2) not saturate at 50mA DC in the primary
3) be able to withstand the highish voltage (take a good look at the schematics for the output stage -- the isolation in the transformer should withstand not only the B+ but the voltage swing as well... since modern isolation materials usually withstand peaks of 3500V, you do not need to worry about that...
When it comes to quality -- well, that might be a different issue. I presume most of what you read about output transformers is marketing (Tamura etc.) although after all I would be disappointed to find out that i.e. Tamura or Tango do not sound that well, when compared with cheaper models, like Hammond.
To put it plainly, you could use two (same, of course) output transformers from old radio receivers that had an output stage with SE EL84... the gap, current, power... everything would match just fine (do not worry about the voltage). I know this sounds bad... but only on paper. If you can get two of these, try the design first on a breadboard. After you are satisfied, you can always order some better OPTs, like Hammond 162x... or Tango, even Tamura, if you like 🙂
Best regards,
Aleksandar
Hi, Arnold...
ANY output transformer will do for your project, provided it can do the following:
1) present a 4.8k-5k Ra to your output tubes (50)
2) not saturate at 50mA DC in the primary
3) be able to withstand the highish voltage (take a good look at the schematics for the output stage -- the isolation in the transformer should withstand not only the B+ but the voltage swing as well... since modern isolation materials usually withstand peaks of 3500V, you do not need to worry about that...
When it comes to quality -- well, that might be a different issue. I presume most of what you read about output transformers is marketing (Tamura etc.) although after all I would be disappointed to find out that i.e. Tamura or Tango do not sound that well, when compared with cheaper models, like Hammond.
To put it plainly, you could use two (same, of course) output transformers from old radio receivers that had an output stage with SE EL84... the gap, current, power... everything would match just fine (do not worry about the voltage). I know this sounds bad... but only on paper. If you can get two of these, try the design first on a breadboard. After you are satisfied, you can always order some better OPTs, like Hammond 162x... or Tango, even Tamura, if you like 🙂
Best regards,
Aleksandar
Konnichiwa,
RCA specified the 50 in their early datasheets EXPLICITY as not suited for RC coupling. They where serious. Many 50's exhibit appreciable gridemission even at normal operation. Use a a too large value grid return path resistance and the Output stage will "run away". Here one of the original datasheets:
No. You are misquoting. But I do suggest that the volume (as opposed to level) control should be as far back in the Audio Chain as possible. If JB ever makes a TVC that can stand 50V RMS I'll put the volume control at the output valve grid!
Now level control is a different story, but usually HiFi gear does not split the functions and combines level and volume into oen.
You could just use a cheap torroidal transformer with secondaries the stacked up with the primary. The Lundahl seems to have only one 350V secondary at 0.63A, so if you get a mains transformer of say 25VA with 18V+18V secondaries you can add another 36 - 40V to the HT winding for little money.
With 390V AC and a current consumption of 130mA (55mA per 50, 10mA for 417A and a 5AR4/GZ34 rectifier you should have around 490V HT available, which will do nicely for the 400V/55mA/-70V operating point (use 270R in series with 1K as cathode R) with a few V in reserve for the DCR of the output transformer (especially if you want to use that Hammond one, 440 Ohm Primary DCR!!!!) etc....
That then would be IIRC the one that has the huge treble dip at 13KHz or therabouts. You might be better off avoiding this one.
See also here:
http://sacthailand.com/transformerTest1.html
You might want to consider using these white elephants as Anode load choke and go for parallel feed. They reportedly work very well as anode load chokes.
Sayonara
Alex Kitic said:
RCA specified the 50 in their early datasheets EXPLICITY as not suited for RC coupling. They where serious. Many 50's exhibit appreciable gridemission even at normal operation. Use a a too large value grid return path resistance and the Output stage will "run away". Here one of the original datasheets:
An externally hosted image should be here but it was not working when we last tested it.
Alex Kitic said:
No. You are misquoting. But I do suggest that the volume (as opposed to level) control should be as far back in the Audio Chain as possible. If JB ever makes a TVC that can stand 50V RMS I'll put the volume control at the output valve grid!
Now level control is a different story, but usually HiFi gear does not split the functions and combines level and volume into oen.
arnoldc said:
You could just use a cheap torroidal transformer with secondaries the stacked up with the primary. The Lundahl seems to have only one 350V secondary at 0.63A, so if you get a mains transformer of say 25VA with 18V+18V secondaries you can add another 36 - 40V to the HT winding for little money.
With 390V AC and a current consumption of 130mA (55mA per 50, 10mA for 417A and a 5AR4/GZ34 rectifier you should have around 490V HT available, which will do nicely for the 400V/55mA/-70V operating point (use 270R in series with 1K as cathode R) with a few V in reserve for the DCR of the output transformer (especially if you want to use that Hammond one, 440 Ohm Primary DCR!!!!) etc....
arnoldc said:
That then would be IIRC the one that has the huge treble dip at 13KHz or therabouts. You might be better off avoiding this one.
See also here:
http://sacthailand.com/transformerTest1.html
You might want to consider using these white elephants as Anode load choke and go for parallel feed. They reportedly work very well as anode load chokes.
Sayonara
Aleksandar,
Not only the OPT should not saturate at 60 mA, but it should have
enough unmagnetised headroom for the AC requirements.
Thorsten,
The statement concerning the grid leak resistors for the 50 is probably due to the fact that, at that time, they could not imagine that you could use the recommended 10 k, without upsetting the driver stage layout.
Going back to your previous message, I think you are right. Every one of us has his own beliefs ( that's human ). Therefore, you will choose freely, when convinced of your good choice, inductors, capacitors, etc., in spite of their wellknown ( to you also ) drawbacks with respect to frequency and, as far as I am concerned, I will stick to the honest resistors, whenever possible, and avoid the use of caps and inductances.
Not only the OPT should not saturate at 60 mA, but it should have
enough unmagnetised headroom for the AC requirements.
Thorsten,
The statement concerning the grid leak resistors for the 50 is probably due to the fact that, at that time, they could not imagine that you could use the recommended 10 k, without upsetting the driver stage layout.
Going back to your previous message, I think you are right. Every one of us has his own beliefs ( that's human ). Therefore, you will choose freely, when convinced of your good choice, inductors, capacitors, etc., in spite of their wellknown ( to you also ) drawbacks with respect to frequency and, as far as I am concerned, I will stick to the honest resistors, whenever possible, and avoid the use of caps and inductances.
Konnichiwa,
Absolutely. And still, in this day & age you will find that non too many driver stages are overly happy driving 10KOhm at current levels as low as you find to develop the bias needed.
Hmmm. You only omit inductors, not capacitors in your designs. Given that it is fairly trivial to make an anode load choke that does fine for the extended audio band I find that the right inductance at the right place can make all the difference.
I personally like to suggest the following adaptation of the LW Amplifier as one that has optimised the various signal loops and offers the fewest "problem parts" in the signal circuitry, plus is completely safe. The main improvement one might make to this is the use of parallelfeed output circuitry, but the WE Connection is a good alternative.
However you do not have to use inductors. You can use solid state, hybrid or "tube only" current sources as anode loads. If you supply them from a sufficiently high HT Voltage you can drive the grid all the way even positive and you have no "non-linear" devices to contend with.
You can even use a modified SRPP or similar circuit in your DC coupled Amplifier
Don't tell me. You actually found "honest resistors"? My experience with Transformer Volume COntrols suggest that in many positions and cases resistors are at least as dishonest as capacitors or transformers, sometimes more-so.
Component quality is always an issue and there is no simple rule as to what is best. Often the drawbacks incured by substituting a resistor for an inductor are larger than what is gained, I have tried.... Sometimes a suitable Resistor of the wirewound type with an iron core is the better choice. Try it sometimes.
Sayonara
ari polisois said:
Absolutely. And still, in this day & age you will find that non too many driver stages are overly happy driving 10KOhm at current levels as low as you find to develop the bias needed.
ari polisois said:
Hmmm. You only omit inductors, not capacitors in your designs. Given that it is fairly trivial to make an anode load choke that does fine for the extended audio band I find that the right inductance at the right place can make all the difference.
I personally like to suggest the following adaptation of the LW Amplifier as one that has optimised the various signal loops and offers the fewest "problem parts" in the signal circuitry, plus is completely safe. The main improvement one might make to this is the use of parallelfeed output circuitry, but the WE Connection is a good alternative.
An externally hosted image should be here but it was not working when we last tested it.
However you do not have to use inductors. You can use solid state, hybrid or "tube only" current sources as anode loads. If you supply them from a sufficiently high HT Voltage you can drive the grid all the way even positive and you have no "non-linear" devices to contend with.
You can even use a modified SRPP or similar circuit in your DC coupled Amplifier
ari polisois said:
Don't tell me. You actually found "honest resistors"? My experience with Transformer Volume COntrols suggest that in many positions and cases resistors are at least as dishonest as capacitors or transformers, sometimes more-so.
Component quality is always an issue and there is no simple rule as to what is best. Often the drawbacks incured by substituting a resistor for an inductor are larger than what is gained, I have tried.... Sometimes a suitable Resistor of the wirewound type with an iron core is the better choice. Try it sometimes.
Sayonara
Konnichiwa,
Oh well, a copy here then:
BTW, if it interests anyone, the picture is from a small article that looks at a variety of issues around SE Amp's I wrote a few years ago....
http://www.fortunecity.com/rivendell/xentar/1179/theory/seamptheory/SEAmplifiertheory.html
Sayonara
slowmotion said:
Oh well, a copy here then:
An externally hosted image should be here but it was not working when we last tested it.
BTW, if it interests anyone, the picture is from a small article that looks at a variety of issues around SE Amp's I wrote a few years ago....
http://www.fortunecity.com/rivendell/xentar/1179/theory/seamptheory/SEAmplifiertheory.html
Sayonara
- Status
- Not open for further replies.