If the data sheet doesn't give a recommended resistance for push-pull operation, how does one calculate the correct plate-to-plate resistance?
I'm looking at using a dual triode for low wattage in a push-pull configuration. The tube I'm most interested in is the 6SN7. The 12BH7 is another possibility. (If there's a good low wattage pentode I would consider it, but I haven't identified one yet).
This will be a 1 Watt guitar amp.
The GE spec sheet for the 6SN7 lists plate resistance in a Class A design as 6700 Ohms with 90V plate voltage, or 7700 Ohms at 250V on the plate. Looking at a variety of sources showing resistances for various tubes in single ended Class A vs push pull Class AB operation (but none of which list the 6SN7), there is no clear simple correlation. The push-pull isn't simply a multiple of the single ended number. So it doesn't seem to help having the 6700 or 7700 ohm numbers on the spec sheet.
There are 10k ohm OTs available. Do I just use one of those and assume it is close enough?
If it is any help, here's a link to the GE spec sheet for 6SN7. http://faculty.frostburg.edu/phys/latta/ee/twinplex/6sn7/6sn7gtb.pdf
I'm looking at using a dual triode for low wattage in a push-pull configuration. The tube I'm most interested in is the 6SN7. The 12BH7 is another possibility. (If there's a good low wattage pentode I would consider it, but I haven't identified one yet).
This will be a 1 Watt guitar amp.
The GE spec sheet for the 6SN7 lists plate resistance in a Class A design as 6700 Ohms with 90V plate voltage, or 7700 Ohms at 250V on the plate. Looking at a variety of sources showing resistances for various tubes in single ended Class A vs push pull Class AB operation (but none of which list the 6SN7), there is no clear simple correlation. The push-pull isn't simply a multiple of the single ended number. So it doesn't seem to help having the 6700 or 7700 ohm numbers on the spec sheet.
There are 10k ohm OTs available. Do I just use one of those and assume it is close enough?
If it is any help, here's a link to the GE spec sheet for 6SN7. http://faculty.frostburg.edu/phys/latta/ee/twinplex/6sn7/6sn7gtb.pdf
- You will not get 1 Wa from PP on a single 6SN7 bottle, likely around half;
- You will need one-two front stages for gain, probably both halves of 6SN7 will give enough for a bit of overdrive;
- Yup, get that 10K PtP OPT and hook up the 8R speaker to 4R tap (or a 16R one to 8R tap), this would be acceptable enough.
- You will need one-two front stages for gain, probably both halves of 6SN7 will give enough for a bit of overdrive;
- Yup, get that 10K PtP OPT and hook up the 8R speaker to 4R tap (or a 16R one to 8R tap), this would be acceptable enough.
Help understanding spec sheets
These old spec sheets are pretty lame sometimes! Nothing is defined. I spent a decade as an engineer on digital semiconductor devices (logic, memory), with a big part of the job writing specs. So here I am trying to decipher tube data sheets.
Specifically I am trying to figure out two graphs on the 6SN7. Terms are not defined, and both graphs have exactly the same conditions listed. The graphs are on page 4 of this link to the GE data sheet http://faculty.frostburg.edu/phys/latta/ee/twinplex/6sn7/6sn7gtb.pdf .
Both graphs are titled Average Plate Characteristics. The vertical axis is plate current, the horizontal axis is plat voltage.
The top graph looks like a typical pentode characteristic, while the bottom looks like a typical triode characteristic. This tube is a dual triode.
What is Ef? Both graphs state Ef = Rated Value.
The top graph labels the curves as Ib @ Ec = V, with some voltage specified. Ib is plate current, so presumable Ec is grid voltage?
If so, how is this tube operating with huge positive grid voltages? +35 is the top most curve, with 5 volt decrements down to -10v for the bottom most curve.
These old spec sheets are pretty lame sometimes! Nothing is defined. I spent a decade as an engineer on digital semiconductor devices (logic, memory), with a big part of the job writing specs. So here I am trying to decipher tube data sheets.
Specifically I am trying to figure out two graphs on the 6SN7. Terms are not defined, and both graphs have exactly the same conditions listed. The graphs are on page 4 of this link to the GE data sheet http://faculty.frostburg.edu/phys/latta/ee/twinplex/6sn7/6sn7gtb.pdf .
Both graphs are titled Average Plate Characteristics. The vertical axis is plate current, the horizontal axis is plat voltage.
The top graph looks like a typical pentode characteristic, while the bottom looks like a typical triode characteristic. This tube is a dual triode.
What is Ef? Both graphs state Ef = Rated Value.
The top graph labels the curves as Ib @ Ec = V, with some voltage specified. Ib is plate current, so presumable Ec is grid voltage?
If so, how is this tube operating with huge positive grid voltages? +35 is the top most curve, with 5 volt decrements down to -10v for the bottom most curve.
Presumably Ef is the filament (heater) voltage, although that would mean that they used a sensible symbol, which they don't for the grid and plate.
I also guess Ec is the grid voltage, as Ic is the grid current. The tube is indeed specified for positive grid voltages. Looking at the design center maximum ratings, this must be meant for short pulses in horizontal-oscillator service in television sets, it is the only set of ratings that allows such high currents.
I also guess Ec is the grid voltage, as Ic is the grid current. The tube is indeed specified for positive grid voltages. Looking at the design center maximum ratings, this must be meant for short pulses in horizontal-oscillator service in television sets, it is the only set of ratings that allows such high currents.
Why are the curves so different between the two graphs? They have very different shapes for the same grid voltages.
They are a different scale. One goes to 350ma, the other only 28ma. One goes to +35V, the other only 0V.
The top graph, which looks like pentode characteristics, extends pretty far into positive grid voltages. Characteristics of all triodes look like this when grid is driven positive. You don't see a lot of such graphs because in most cases triodes are used at negative grid voltages, and few manufacturers bother to show the positive side.
Last edited:
Plate resistance isn't the load resistance, remember that.
Perhaps these would be of interest.
The Valve Wizard -Push-Pull
Push-Pull Tube Amplifier Simulation
Perhaps these would be of interest.
The Valve Wizard -Push-Pull
Push-Pull Tube Amplifier Simulation
Ef filament voltage - rated value means either 6.3v or 12.6v, because the curves are for both 6SN7 and 12SN7.
As Marcel said, the two charts are for different applications. The huge currents with the high positive grid voltage are not continuous. The typical audio application would fall on the second graph. Imagine trying to put that huge current range and all those grid voltage curves onto one large graph.
You are looking at isolated data sheets. In days of tubes, we had tube manuals, like the GE master book your graphs are from or the RCA tube manual, the RC30 being the most common one on shelves these days. And in the manual, other information is presented like the abbreviations used in the guide. A designer would have been familiar with teh nomenclature at the time. Ef and If are indeed the heater voltage and current.
When they say rated value for heaters, it means heater voltage is on spec as opposed to off voltage. I would equate that to VCC on a TTL data sheet. A 7404 data sheet shows VCC on the power pin, not +5v. That it is +5v is either understood or at least only mentioned elsewhere. A designer would be aware of that.
As Marcel said, the two charts are for different applications. The huge currents with the high positive grid voltage are not continuous. The typical audio application would fall on the second graph. Imagine trying to put that huge current range and all those grid voltage curves onto one large graph.
You are looking at isolated data sheets. In days of tubes, we had tube manuals, like the GE master book your graphs are from or the RCA tube manual, the RC30 being the most common one on shelves these days. And in the manual, other information is presented like the abbreviations used in the guide. A designer would have been familiar with teh nomenclature at the time. Ef and If are indeed the heater voltage and current.
When they say rated value for heaters, it means heater voltage is on spec as opposed to off voltage. I would equate that to VCC on a TTL data sheet. A 7404 data sheet shows VCC on the power pin, not +5v. That it is +5v is either understood or at least only mentioned elsewhere. A designer would be aware of that.
I've read all the books, read the websites. The problem is to find a number for the plate-to-plate resistance in a push pull design. It seems the tube manufacturers will specify that number on their spec sheets for many tubes. But I cannot find such a number for this tube.
Nowhere have I come across a formula or process for determining what is a good number, and/or why it is a good number. Twice the Rp? Some multiple of the Class A single ended load resistance?
Obviously what is needed is a reflected impedance from the speaker load which most closely matches the internal load of the tubes in order to maximize power dissipation in the speaker (and reduce distortion).
Were I to use a 6V6 the number would be found on the data sheet.
Nowhere have I come across a formula or process for determining what is a good number, and/or why it is a good number. Twice the Rp? Some multiple of the Class A single ended load resistance?
Obviously what is needed is a reflected impedance from the speaker load which most closely matches the internal load of the tubes in order to maximize power dissipation in the speaker (and reduce distortion).
Were I to use a 6V6 the number would be found on the data sheet.
I used to write specs for that old TTL stuff as well as newer technologies and memory devices. We'd always have a list of specifications which included the designation, such as VCC, and then some words to describe it, such as Power Supply Voltage. If you looked at the pinout and wondered wtf some designator meant, you could always find it defined in the early parts of the spec.
We dealt with the parts individually, so we weren't putting together a designer's handbook with all kinds of explanations. Each datasheet was expected to stand alone. But that was the military aerospace world where we operated under perhaps different expectations than the commercial world.
No real need "to reduce" distortion in a guitar amp. .. your guitar will be happy with 10%.
Anything 2 x Rp and higher would work for Hi-Fi, but you can go to 1.5 to sqeeze a little more output. IME the less is the tube's Rp the better it tolerates 1.5 and otherwise.
Anything 2 x Rp and higher would work for Hi-Fi, but you can go to 1.5 to sqeeze a little more output. IME the less is the tube's Rp the better it tolerates 1.5 and otherwise.
In a push-pull output stage the two valves are in series as seen from the output transformer's primary, so effective "plate resistance" is twice that of a single valve. But I'm uncertain about what your question is.
All good fortune,
Chris
All good fortune,
Chris
The full book has some clues what the lines mean.
If you grew-up with tubes, mostly it is "common knowledge". Granted, some of this is no longer common.
And as you say, the Military likes its documents very complete, while commercial work tends to be more-of-the-same, few novel designs or questions about the specs.
And as said: those two 6SN7 graphs are for VERY different conditions. Specifically one is for POSITIVE Grid (mostly). In most small-signal amplifiers we do NOT want positive grid. It sucks (grid current). We want the infinite impedance in the negative grid range. However for the last stage of an amplifier chain we might want to use a mid-size tube worked pos-grid, rather than a large tube that can do the job all neg-grid. Pos-grid is less of a drag in RF work where tuned-circuits can average the high peak grid current. Specifically the 6SN7 can be used in (small) TV sets as a sweep tube with high peak currents, plate and grid. Note double-dagger mentions the 15% duty cycle used in NTSC TV sweep work, which matches the 20mA DC 300mA peak claim. (However I don't see how to get a 6SN7 to 300mA without more drive and trouble than it is worth; this is like 2N3055 says 15A max but it takes a big hammer to really exceed 4A.)
Triodes are easy. In SE, loads higher than rp give good THD but loads much higher than rp give lower power (unless the tube allows obscene plate voltages). In P-P, you can go as low as rp each side without much rise in THD (the two hard-worked sides cancel each others' distortion).
If you load in rp each side (about 14K CT for 6SN7), you get about half the power of an ideal (zero rp) tube. You will be lucky to get 20% efficiency; 15% more likely. This gives 1.4W to 1W with 6SN7 working at max Pdiss. Since this will no-way rock a large stadium (we have other amps for that), and 6SN7 no longer grow on trees, we often aim lower than max Pdiss and higher than 14KCT load.
A different approach. Take a tube with Power amp suggested conditions, similar but probably bigger. 2A3 is a 15W Pdiss tube, compared to 3.5W for 1/2 6SN7. Both can be worked around 300V. Since the 2A3 is 4.3 times bigger, we expect the 6SN7 to be happiest with a load 4.2X higher than 2A3. If 2A3 likes 5KCT(?), then 6SN7 wants 21KCT load. *However* 2A3's rp is 800r, which 6SN7 rp is ~~7K. This suggests a factor more like 8.7X. So 44KCT(!) and probably 10W/8.7W and just over 1W output. The problem here is that a 44K winding is hard to make full-audio-band; we typically must cheat down to get a satisfactory OT design.
If you load in rp each side (about 14K CT for 6SN7), you get about half the power of an ideal (zero rp) tube. You will be lucky to get 20% efficiency; 15% more likely. This gives 1.4W to 1W with 6SN7 working at max Pdiss. Since this will no-way rock a large stadium (we have other amps for that), and 6SN7 no longer grow on trees, we often aim lower than max Pdiss and higher than 14KCT load.
A different approach. Take a tube with Power amp suggested conditions, similar but probably bigger. 2A3 is a 15W Pdiss tube, compared to 3.5W for 1/2 6SN7. Both can be worked around 300V. Since the 2A3 is 4.3 times bigger, we expect the 6SN7 to be happiest with a load 4.2X higher than 2A3. If 2A3 likes 5KCT(?), then 6SN7 wants 21KCT load. *However* 2A3's rp is 800r, which 6SN7 rp is ~~7K. This suggests a factor more like 8.7X. So 44KCT(!) and probably 10W/8.7W and just over 1W output. The problem here is that a 44K winding is hard to make full-audio-band; we typically must cheat down to get a satisfactory OT design.
Last edited:
The question is what primary resistance should the OT have? The follow on question is how is that number derived if it isn't on the spec sheet?
As others have already said, reflected primary impedance is chosen based on several rules-of-thumb. Maximum power output comes at a reflected impedance of slightly more, up to maybe twice, effective plate resistance. Distortion rises and maximum output lowers at lower load impedances. Distortion and maximum output both decrease with higher load impedances.
For hi-fi use, a factor of 4 or 5 is ballpark ideal. For a guitar amp (a different forum) a factor of 1 or 2 might work well.
So the number usually isn't so much derived as just rule-of-thumbed.
All good fortune,
Chris
For hi-fi use, a factor of 4 or 5 is ballpark ideal. For a guitar amp (a different forum) a factor of 1 or 2 might work well.
So the number usually isn't so much derived as just rule-of-thumbed.
All good fortune,
Chris
- Status
- Not open for further replies.