Calculating audio impedances

[W8ACR]

This topic has probably been discussed here in the past, so my apologies if this is flogging a dead horse. Audio impedances have always been a mystery to me. I seem to do OK when I work with RF plate impedance calculations such as figuring out which component values to use in a plate tank to get a desired Q value, or which taps to use on the secondary of a modulation transformer to properly match the modulator to a final amplifier.

But calculating the audio impedances to select a proper modulator driver transformer completely baffles me. I am completely dependent on transformer manufacturers' tube/transformer tables. Is there a straightforward way to calculate impedances from the audio driver plates to modulator grids at varying plate voltages?

In the same vane, is there a straightforward way to calculate plate to plate impedances of class B modulator tubes at varying plate voltages? Again, I am completely dependent on published tube data, and I'd like to be able to make these calculations myself.

Guess I'm a bit slow on this subject.

Thanks in advance,

Ron W8ACR

[k4kyv]

The optimum plate to plate impedance of the modulator is not really something you "calculate".  The figures given in the tube charts are manufacturers' recommendations for a given tube. The P-P load is simple to calculate, using the final amplifier plate voltage ÷ final amplifier plate current, as transformed by the impedance ratio of the modulation transformer.  The load can be varied quite widely as long as the maximum plate voltage, plate current and plate dissipation are not exceeded. This can be determined by trial-and-error, but you are probably safe if you stay within 25% or so of the recommend values.

It can be calculated using the characteristic curves published by the tube manufacturers, but to me that is far too tedious, even with the aid of a good scientific calculator. 

I have always worked backwards when homebrewing my stuff.  Instead of designing a circuit (or following someone else's design from a handbook or magazine article) and then acquiring the necessary specific parts to follow the design, I accumulate a bunch of parts or dig them out of my junk box and then try to design the best circuit that will work with the parts on hand. Again, a lot of trial and error involved.

[WD5JKO]

Ron,

Let me take a stab at this. Then let the experts straighten me out. I welcome the input since this is a discussion that could benefit many on this forum.

  This is always a topic that we trip over. The charts are confusing to me also. I prefer to look at turns ratio first, and then figure the impedances from that. Remember that the impedance varies as the square of the turns ratio. So a transformer with a 2:1 turns ratio has a 4:1 impedance ratio.

   So take a modulation transformer with push pull audio drive. A 2:1 turns ratio assuming perfect tubes (plate swing goes to 2X B+, and to ground over audio cycle) and the same B+ on the modulator as the RF final, and we can just make 100% modulation at the clip point. Then combine real tubes, transformer losses, and maybe the ability to modulate to 125% on upward peaks, and we need a lower turns ratio, such as 1.4:1.

  You can take those multi-tap modulation transformer charts, and since the values given are impedances, you can calculate turns ratio from that. Then calculate the RF PA load impedance RL=B+/plate current. Then back calculate the modulator load impedance from that AFTER you figured out what turns ratio you wanted to have. When looking through the terminal diagrams, try to find a solution that uses all of the windings whenever possible. This will give you better transformer than one that only uses a portion of the winding.

  So, lets say your modulator tubes according to the tube book say 6600 ohms plate to plate load impedance. So what if the calculated Rl to the modulator is 10K. Is this bad? No, but realize the tubes will provide lower power into a higher impedance, but the benefit will be lower distortion. Taking the same tubes and loading them at 3300 ohms is bad because the modulator plate current will rise, the plate downward swing will be diminished, and the tubes will produce more distortion. There is nothing magic about the tube manual specification; it is just to illustrate a given RMS power, at a given load impedance, and at a given amount of distortion.

  If you need the modulator to drive a lower impedance that what is published for the tubes, maybe start over with the variables to find a solution. Taking the RF power amplifier and lightening the loading might do the trick, or possibly adding bigger modulator tubes, or use 4 tubes of the original type in push pull parallel, or back off on the turns ratio a bit, etc..

Clear as mud right?

Jim
WD5JKO

[W8ACR]

OK Don thanks for the input. Perhaps an example or two might give some further insight into what I'm asking.

example 1
Suppose two modulator tubes operating at 1500V plate voltage have a plate to plate impedance of 12000 ohms (according to the tube data sheet). The RF amplifier is operating with 1500V @ 200mA. The mod transformer must therefore match 12000 ohms primary to 7500 ohms secondary. OK simple enough.

Unfortunately, the only plate transformer I have available will provide only 1000V for the plates of the modulators. This means I must match a different impedance to 7500 ohms. The tube data sheet does not give the plate to plate impedance at 1000V. How can I determine the plate to plate impedance at 1000V so that I can match the tubes properly to the load?

example 2:
My 1940 multimatch driver transformer data sheet gives me the info to correctly match 2A3 driver tubes to 203Z's, 838's, and a whole host of other tubes which are now impossible to find. Furthermore, I don't have $500 to buy a pair of 2A3's. So I'd like to use this driver transformer to match a pair of triode connected 6F6's to a pair of 811A's. How do I determine the plate and grid impedances so that I can choose the proper taps?

Am I being too picky about this idea of impedance matching? My reading has led me to believe that it is important to match impedances properly to reduce audio distortion and protect tubes and transformers. Maybe trial and error is good enough when dealing with audio circuits.

Thanks, Ron W8ACR

[WD5JKO]

Am I being too picky about this idea of impedance matching? My reading has led me to believe that it is important to match impedances properly to reduce audio distortion and protect tubes and transformers.

  Ron,

  I believe your fixating on impedance alone is clouding your vision. As I stated earlier,

"There is nothing magic about the tube manual specification; it is just to illustrate a given RMS power, at a given load impedance, and at a given amount of distortion."

  With that said, the tube published specs are on the low side of the acceptable range. So if they state 6600 ohms P-P, then I wouldn't go too much lower. I would not hesitate to go higher, maybe double that. Going higher makes the tube swing easier, less plate dissipation, and less distortion. The consequence though is lower power output since P = E^2/R.

   In the tube manuals and the W6SAI books there are step by step instructions for calculating the RL for a given set of conditions. Also Patrick Turner of Turner audio has a nice write on the topic as well:

See Load Matching 1,2,3,4:
http://www.turneraudio.com.au/education+diy.htm

  Maybe Don or someone else can verify this. With a multi-match modulation transformer chart, the impedances given only apply for a peak modulation capability of 100%. If you want to modulate 125% on a positive peak, then you must switch to turns ratio.

Jim
WD5JKO

[W8ACR]

Thank you Jim also for the reply.

I understand (I think) the concept of how the modulation transformer works, but on a practical level, when my selection of available plate transformers forces me to use plate voltages that are not listed in the modulator tube operating data, it seems that there should be an easy way to determine the unknown impedances so that I can properly modify the circuit.

Or mebbe I need to loosen my rigid concept of impedance A and impedance B must be matched exactly.  

Ron W8ACR

[WD5JKO]

Ron,

 Before you rule out 2A3/6A3/6B4 driver circuitry, take a look at the triode connected 6AV5:

http://www.tubelab.com/6AV5.htm

Jim
WD5JKO

[W8ACR]

OK Jim,

Thanks again. Your second post helped a lot. Yes, I guess I was fixated on the (false) concept that modulator tubes at a given plate voltage MUST see the particular load impedance specified by the tube data sheet to operate properly. If I understand you correctly, this is looking at it backwards. It would be better to determine the modulator load as seen at the primary of the mod transformer, and then determine if this load is close enough to the optimum load for those particular tubes at that particular plate voltage. If not, then operating parameters must be changed.

Does this sound correct to you? Is my thinking closer to reality now?

Ron

[WD5JKO]

Yes, I guess I was fixated on the (false) concept that modulator tubes at a given plate voltage MUST see the particular load impedance specified by the tube data sheet to operate properly. If I understand you correctly, this is looking at it backwards. It would be better to determine the modulator load as seen at the primary of the mod transformer, and then determine if this load is close enough to the optimum load for those particular tubes at that particular plate voltage. If not, then operating parameters must be changed.

Does this sound correct to you? Is my thinking closer to reality now?

Ron

   Ron,

  This is closer to the way I think about it. Let me try to make an example to see if this helps. You built a home brew transmitter using a pair of 6146's that you want to plate modulate. Your power supply provides 600V, and is beefy enough to supply both the RF amplifier, and the modulator.

  So let me start this process for the plate modulator. You want the ability to modulate fully, and with a little extra headroom. So lets say you want to be capable of 110% on positive peaks.

   So to get to 110%, multiply 1.1 X 600 = 660v. The modulated B+ goes from 600 to 600 +660 = 1260v peak at 110% modulation. On the other way down it will go to -60v (over mod!).

   For modulators you have a pile of 807's, so lets explore them. You also want high quality audio, so you want to bypass the interstage transformer idea, and instead you want to build a phase inverter, and R-C couple the audio to the 807 grids. This means class Ab1 operation since you cannot provide any grid current. With class Ab1, the plate voltage minimum occurs when the peak AF grid voltage matches the grid bias amplitude. If the bias were set to -33v, and peak AF swing brings the G1 to 0v. You cannot go positive since grid current would flow.

   Next look up the 807, and look at the plate characteristic graph. At zero bias, the plate voltage saturation is about 50v, but that is the hard clip point. Back off from that, and lets say the plate voltage minimum is 150 volts. So this means the 807 plate can each swing from B+ (600v) to 150v (Eb min) to 1050v (Eb (max), or +/- 450v. Since this is Push Pull, one 807 at Eb min of 150, and the other will be at Eb max of 1050. So we make 900v peak across the primary of the modulation transformer.
 
  We said earlier that we need 660 v peak across the secondary for 110% modulation, so the voltage ratio is 900/660 = 1.36:1. The voltage ratio is the same as the turns ratio.

  So lets say the 6146's run at 600v @ 200ma. The RL = 600/.2 = 3000 ohms.
With a secondary load impedance of 3000 ohms, and a mod tranny with a 1.4:1 turns ratio, the primary plate to plate load impedance is:

1.4^2 X 3000 = 6000 ohms.

It looks like a pair of 807's will do fine. But what if it worked out to 3500 ohms instead. that is too low for those 807's, but two pair in push pull parallel would be fine.

It's late, and I don't know if this is fully correct or if it can be followed. But notice how far I went into the design steps before getting into impedances.

Jim
WD5JKO

[AB2EZ]

Ron

Adding to what Don and Jim have said... which I agree with...

It may be helpful to keep in mind that the traditional rules (circa 1930-1970) for matching a modulator to a plate modulated rf output stage arose from a set of assumptions that may not apply to what you are trying to do when you design and build a homebrew transmitter in 2011.

What's different?

If you were going to design a transmitter to be manufactured (e.g. a Johnson Ranger) in 1960, you would be heavily focused on the cost of the parts. In that era, the cost of the parts (the modulator tubes and possibly the power supply transformer) would be higher if you designed the modulator to be capable of producing more audio output power. Since you would be able to specify the modulation transformer to be used (and purchased from whomever you selected as the supplier of transformers), you would be inclined to select the turns the ratio, after you selected the tubes you were going to use in the modulator, to optimize the  audio power you could obtain from whatever modulator tubes you decided to use. Likewise, you would select the modulator tubes to optimize the audio power you could produce... for a given tube cost... subject to whatever distortion requirements you had to meet, and subject to other cost-driven contraints (like using the same value of B+ on the modulator tubes as on the rf output tubes).

For our purposes, cost is still an issue, but the constraints and objectives are different. We have to use whatever modulation transformer(s) we can get our hands on... but typically we only need one, because we are only going to make 1 rig. So we start out being constrained by the turns ratio(s) of the available modulation transformer(s). We can design the modulator to be capable of producing significantly more audio power than we need (into whatever load optimizes its audio output power), and then operate it with a different load that is convenient for whatever reasons. One reason is that, given the turns ratio (output turns / input turns) of the modulation transformer we have on hand, and given the modulation resistance, R, of the rf stage... the impedance that the modulator will look into will be R/(N x N).

For example:

If we have a few turns ratios to choose from... and lots of available audio power from the modulator... the low frequency performance of the modulated transmitter may be better if the turns ratio is lower... even though that will place a higher impedance load on the modulator than is required to produce the best match from the perspective of maximum power transfer from the modulator to the rf output stage.

My suggestion is that you design the modulator to have significantly more audio power output capability than 50% of the rf stage's input power... and therefore give yourself the flexibility to use a variety of modulation transformer turns ratios (either to accommodate the modulation transformers you can get your hands on, or to optimize such things as the low frequency performance of the transmitter).

Stu

[k4kyv]

My suggestion is that you design the modulator to have significantly more audio power output capability than 50% of the rf stage's input power... and therefore give yourself the flexibility to use a variety of modulation transformer turns ratios (either to accommodate the modulation transformers you can get your hands on, or to optimize such things as the low frequency performance of the transmitter).

Also, remember a modulation transformer is not 100% efficient. I have seen it stated, although I never tried to measure it, that even a good quality modulation transformer is only about 90% efficient.

We have to consider peak power vs average power.  It takes 500 watts peak power to modulate a KW DC input 100%.  With a sine wave tone, that will measure out to be 500 watts continuous power.  But with the typical human voice, the peak-to-average ratio is very high, so it's not uncommon to be able to voice modulate a KW input to 100% with less than 150 watts of audio, even though the modulator still must be capable of delivering the full 500 watts on peaks.  I recall seeing ads by RCA in late 40s and early 50s QSTs and CQs, advertising the 211 triode, touting that it would "voice modulate up to a kilowatt", even though the tube charts rate it for only something like 270 watts audio output in class-B modulator service.

Most commercial ham rigs, and many broadcast and communications type AM transmitters are designed for 100% modulation maximum.  In reality that was likely a cost-limiting measure, although the manufacturers would never admit to such; they claim it to be a "safety" feature to make it impossible to overmodulate the transmitter.  Problem is... if the audio is driven to maximum, the  modulator clips right before 100% producing a flat-topped waveform, which causes exactly the same splatter and distortion as that caused by overmodulation.  Sometimes a high-level low pass audio filter is inserted between the modulation transformer and final to limit the splatter, which may eventually result in modulation transformer failure, a frequent problem encountered with the Collins KW-1.

It is better to design the modulator for output capability substantially beyond 100% modulation for two reasons: (1) to take advantage of the natural asymmetry of the human voice and achieve more sideband power by allowing the negative peaks to more closely approach 100% before flat-topping occurs on the positive peaks, and (2) to allow some unused head-room for the positive peaks, which results in less distortion and cleaner audio even when 100% positive modulation is not exceeded.  

With expanded positive peak capability, we  have to consider one inconvenient truth: the audio power required to modulate a transmitter is a function of the square of the modulation percentage. To modulate up to 125% positive requires 156% the audio power required to modulate 100%.  150% modulation requires 225% of that audio power, more than twice the DC input to the final! OTOH, 60% modulation requires only 36% of the audio power needed to modulate 100% - a whopping 18% of the DC input to the final!  Of course, since we are talking about peak power here, the average audio power will be much less, but the modulator still must be capable of delivering full peak power for brief periods of time without distortion.

If you visit a recording studio and look at an old style analogue VU meter with voice program content, you will notice that the pointer rides around 30% or less most of the time, with occasional excursions up to the red mark.  This indicates that the average percentage of modulation with the human voice is about 30%, which represents nine percent of the audio power required to modulate 100%, or 4.5% of the DC input to the final. Of course, the audio power fluctuates at a syllabic rate to much higher levels, but still, much of the time the sideband power of an AM signal is but a tiny fraction of the carrier power.  The same goes for SSB; much of the time the output power is but a tiny fraction of the peak output power.  That's why a lot of slopbucketeers run such mushy sounding audio. They use a lot of  processing or simply overdrive their rigs to bring up the average power in order to "see the monkey swing".

This also explains why so many "legal limit" SSB linears have such anaemic power supplies and flimsy tank components.  As long as the capacitors and other components are rated to take the peak voltage, under-sized components will work fine at the rated peak power as long as the modulating signal is speech only.  But try to whistle into the mic, hold the key down on CW or run a steady tone, or operate without the final properly tuned to resonance or run too much "speech processing", and watch the thing melt down into a puddle of molten glass and metal or simply go up in smoke! This also explains the popularity and usefulness of the so-called "peckers" employed by SSB ops as a tuning aid.  

[W7TFO]

Hi Ron,

There is a lot to digest here, perhaps I might try and simplify.

Think of those transformers doing the "power throughput" job.  Will it handle what you want to send through it with minimum saturation and phase shift?  The inherent quality (read core material and winding technique) come to play here.

Also, as an example for years the audio industry was fixated on hard 600-Ohm sources and terminations for linear transfer of signal.  It became plain later that a much better way to do that was to have as low a source impedance and as high a load impedance as possible.

Applying this to tube-based amplifiers and modulators, specifically gave us first low-mu tubes like the 2A3 et al, chosen for popular driver service.  A few years later with the use of beam pentodes, the cathode follower design emerged. 

If you want to drive your modulator with some iron in it (and I do, too), you will have to figure the resultant load of that class B stage as a 'non player', by making the driver immune to the fact that the load impedance will be going all over the place whilst in use.  That is the tradeoff for the enhanced efficiency over class  A or AB.

If you look closely at the old data charts, load impedance for class B audio drive was not stated in hard fact as it was impossible to predict.  So, they picked one spot in the particular tube's curves a gave that as a point to hang on.  That was fine for an extremely narrow audio passband.

What worked out best over time was, as I said, was good iron, and a circuit that used a set of tubes that worked in the region of best linearity and not working too hard when at it.  For instance, that pair of 2A3/6B4 tubes would do 15-Watts all day long and have little distortion--into a fixed impedance load, like a good speaker system.

Vary the load within the audio waveform, and things get dicey.  So work them at 6 or 7 Watts, and they can handle that varying load much more easily owing to the increased headroom.

One can add negative feedback to somewhat counteract that impedance changing, or go to cathode follower topology and dispense with the driver transformer entirely.

Neg feedback wasn't the best plan with the low-mu triode.

Synopsis: It all boils down to finding the vintage parts and using them again, as none of it is made anymore, IMO.  The tubes are actually the easy part, the iron is a lot more elusive and that discourages many HB'ers.

73DG

PS...Is that rig of yours running one of the 254W's you got from me?

[k4kyv]

Beam power tetrodes with negative feedback can be used as class-B drivers in place of low mu triodes.  I once saw some curves published somewhere that showed a pair of 6L6s with feedback to be nearly identical in performance to a pair of 2A3s when working into a variable impedance load like class-B grids or a dynamic speaker.

This might be an attractive alternative to the transformerless direct-coupled cathode follower circuit used in some broadcast transmitters.

OTOH, a pair of beam power tetrodes without negative feedback makes an extremely poor class-B driver.

[KM1H]

You would think by now that all this impedance matching and ratio manipulations would be reduced to an on line program that would allow it all to be shown graphically as you manipulate parameters.

This would also include the same fuzzy math the tube designers used to come up with the P-P values but allow them to be manipulated or even developed for various unlisted tubes.

I can run any calulator until my eyes blur and head aches and it is always a royal PITA!

Carl

[W8ACR]

First of all, thanks to everyone for the extensive and detailed replies. I think I have a much better handle on modulator tube impedances and how to design modulators and use mod transformers now. Let's see if I can coherently summarize:
  A class C RF final amplifier circuit will present a certain impedance to a modulator. To work at maximum capability, the modulator wants to see a certain "ideal" load impedance, which varies depending on the tubes and the plate voltages. This is what is shown in the tube tables as the "plate to plate" impedance. The job of the modulation transformer is to show the modulator a load impedance that it is happy with. It is not necessary however for the modulation transformer to show the modulator tubes an "ideal" impedance. A modulator will be happy enough working into a higher load impedance provided that there is enough audio power to mitigate the additional loss. This fact gives the designer some leeway when choosing tubes, plate voltages, and mod transformers. Showing the modulator a higher than ideal impedance is better than showing it a lower than ideal impedance.

Thanks Dennis for addressing the driver question. I'm using 6A3's with a UTC S9 driver transformer. I have gotten good audio reports so far. I was fortunate in that I had manufacturers data to make the proper connections for both the driver and modulation transformers. But are there any general rules when it come to selecting and connecting a driver transformer? When selecting from multiple taps on the primary or secondary of a driver transformer, is one dependent on data provided by the manufacturer, or is there a mathmatical approach to choosing the proper taps, or is it just trial and error?

Don, I have read many times that tetrodes without negative feedback are poor drivers. Why is it then that all of the old handbooks seem to have lots of speech amplifier circuits using 6L6's or 6V6's as drivers connected as tetrodes without negative feedback?

Yes Dennis, I'm using your 254W's. I have two spares. Would you like to send me a couple more?

Ron

[W7TFO]

It is so great when a plan comes together!  If I come across any more 254's, I'll pass them along--you got all of mine previous.

What do I do when I have time to prototype a driver (or any audio amp for that matter)?  Run it and listen.  Blow something you know & love through it and see if it adds any artifacts.

If you want more than the usual PTT audio, try Eurythmics Sweet Dreams for crunching bottom end.  Streisand for vocals that can come apart in the best of modulators.  Any horn concerto on Deutsch Gramophone.  These are some of the audio in which I can pick out design problems by just listening.  Maybe set up your driver transformer test with a switching arrangement so you can change things in real time---you may be surprised by the difference a tap will make.  Also by what you thought the best would be, and might not be.  Play around with B+ & bias voltages too.  Big variable supplies are very useful on the bench....

In order to achieve something I could actually hear before final assembly, I had a 100W output transformer wound with a passband of 16HZ to 20KHZ, and a multitapped primary to give me everything from 1K to 20K CT.  Any tube thing I build, it gets clipped in and then on to an Altec A7 for a real gut-feeling audition.

Even a pair of 450TL's, albeit they were just coasting along.

After I'm satisfied it sounds OK, it might go on to a distortion analyzer to run the numbers.  Mostly it is just a matter of personal taste, and in my case, an effort to rid the airwaves of lousy audio quality.  Maybe it is just my comfort zone, after 42 years in broadcast.  If I want 300-3500 HZ audio communication, I'll call you on the phone.

73DG

BTW, none of this used any math in the actual testing....

[k4kyv]

Don, I have read many times that tetrodes without negative feedback are poor drivers. Why is it then that all of the old handbooks seem to have lots of speech amplifier circuits using 6L6's or 6V6's as drivers connected as tetrodes without negative feedback?

Which handbooks do you have that show those circuits? I think I first read about the problem of beam tetrodes, class B drivers and negative feedback in the 1957 ARRL handbook, the only one I had ever seen during the first few years until I acquired a 1943 edition around 1962. Every other ARRL handbook I have ever perused, at least those published from the mid-30s through the 70s, has the same information, and I THINK the west coast handbooks tell you the same. Are you sure those drivers were intended for a class-B or AB2 modulator?

I have heard rigs on the air using a 6L6 or 6V6 driver without feedback because the builder had an old "hi-fi" or PA amplifier on hand, but in most of those the audio was probably already distorted well before it ever reached the modulator driver stage. Most hams, including AMers back in the 60s and earlier, had been conditioned by certain "experts" not to care about audio quality as long as they could be "understood" and in many cases, their signals just barely met that criterion.

Regarding the modulation transformer impedance and turns ratios, if you are using a common power supply for modulator and final, you will want a turns ratio somewhere between about 1.2:1 and 1.5:1 primary to secondary, regardless of actual working impedance, in order to achieve 100% modulation capability with some head-room.  1.6:1 will just barely make 100% with no head-room. If your only modulation transformer is far away from that, you will need two separate power supplies for modulator and final, and juggle the plate voltages and final amp plate current for 100% modulation capability while maintaining an acceptable load impedance for the modulator.

If the transformer has a low ratio, say 1:1 or even step-up, you may use the common power supply, but in any case you may have to double up to a quad of modulator tubes in order to  get a modulator that works well into the low P-P load impedance the modulator tubes will see through that transformer.

[W8ACR]

You are right Don. The circuits I was thinking of were not using the 6L6's as drivers, but as low power modulator tubes. The only circuit that I could find using the 6L6's as drivers had negative feedback. Almost all the speech amp circuits had triodes as the drivers.

Ron

[KL7OF]

The late 40's and early 50's ARRL handbooks had a speech amp with 6sj7 mic amp, 6sn7 dual triode amp transformer coupled to drive PP 6L6s  The interstage transformer has a split secondary and neg feedback is applied to the 6L6s on the individual legs of the split secondary...The amount of feedback is determined by coupling cap value and a series resistor...The cap and resistor value can be experimented with for best results.....There is a bass and treble boost circuit in the Radiotron Designer Handbook that will plug right in to that 6SJ7 tube for added tone control.....This makes a very nice speech amp although the VRMS output is somewhat lower than a pair of triodes(6B4G)....There is plenty of power to drive most modulator tubes however and the distortion levels are very good....
   I like W7FTOs method (Streisand and Eurythmics) for testing.......I use an electric guitar....Good luck   

[WD5JKO]

I like W7FTOs method (Streisand and Eurythmics) for testing.......I use an electric guitar....Good luck  

That reminds me of the guy from tubelab strumming his guitar to test his prototype 833 tube amp. At the attached picture you can see the 833 plate glowing red.




Full thread here:
http://www.tubelab.com/833SE.htm

End high Jack

Jim
WD5JKO

[WU2D]

The ART-13 uses a single 6V6-GT to drive a a pair of 811's in the modulator. While not ideal, it does show what feed back (in this case shunt feedback directly from plate of 6V6 to Plate of 6SJ7 voltage amp) can do to lower impedance.

[Steve - K4HX]

Yikes! Some of us eat while using the computer. That photo is dangerous.

I like W7FTOs method (Streisand and Eurythmics) for testing.......I use an electric guitar....Good luck  

That reminds me of the guy from tubelab strumming his guitar to test his prototype 833 tube amp. At the attached picture you can see the 833 plate glowing red.




Full thread here:
http://www.tubelab.com/833SE.htm

End high Jack

Jim
WD5JKO

[WD5JKO]

In a nutshell on P-P load impedances, from DIY Audio Forum:

"For most SE amps and many P-P amps you will get more power as the load impedance decreases (to a point). You will get lower distortion and a higher damping factor and a generally tighter sound with a higher load impedance. In general UL mode gives more power than triode with higher distortion and poorer damping factor."

Link: http://www.diyaudio.com/forums/tubes-valves/78462-6l6gc-set-my-first-diy.html

My opinion, the specified ratings in the tube manuals are already on the low side of the impedance range to specify good power at an acceptable distortion level.

This is all about compromises. There is no optimum load impedance that presents all the variables at optimum levels at the same time.

Jim
WD5JKO