KT120, KT90, KT88, 6550.

This page was first created February 2012.

I recently tested samples of Electro Harmonix KT120, and was keen to see how they
compare with common KT88 and 6550, and less common KT90 also from EH. 

Graph 1. Beam Tetrode Operation.
The graph 1 shows the measured results of tests with TWO tubes in PP with a very
good quality OPT with very low winding losses, and using a 1kHz sine wave.
The power supply used has a regulated B+. Without an OPT with very low winding
losses the maximum power at low anode loads is reduced by power dissipated
in winding resistance. Without regulation the increase in anode and screen DC
currents will cause the Ea and Eg2 to reduce and you will not be able to obtain the
same high maximum power I have found in my tests.
Typical power supplies for anode B+ rails in most amplifiers are not actively regulated,
and although B+ may initially be perhaps +550Vdc at the idle condition, at maximum
possible clipping power output, the B+ may reduce under load to less than +500Vdc,
and thus maximum power levels shown in my tests could not be obtained.
Because different manufacturers have different quality standards for power supplies
which exhibit varying output resistance, the test results have been gained using
a well regulated power supply.
In the real world, with an unregulated PSU, high power is obtainable for a small time
duration, so that drum-beats with high peak signal voltages will produce the maximum
power for the duration of the drum beat. For a more realistic measurement of tube
performance in the real world, I recommend pink noise be used as a test signal with
its level uncreased until occasional clipping is visible on the oscilloscope. The same
generator of pink noise should be used for comparative testing.

For where peak signal Ia in an individual tube reaches to more than 3 x Ia at idle,
the following formula may be used to calculate the total Pda in 2 two PP tubes.

Total Pda = (1.8 x Ea x Va-a ) - ( Va-a squared )

In the above graph 1, to calculate Total Pda for 2 tubes, beam tetrode, Ea = 500V,
RLa-a = 1k5, PO max = 136Watts, Vaa = 451Vrms.
Pda = 451 ( 1.8 x 500 - { 451 } ) / 1,500 = 135Watts.
Therefore Pda per tube = 67.5 Watts. 

The Graph 1 shows the Pda for ONE output tube at high output power.
From the graph the use of loads below 2k0 will endanger KT120. Loads below 3k0
will endanger KT88 and 6550. These load values are only valid for Ea = +500V,
and for where Ea was say up to +700V, the load line analysis and graphs for
PO Vs RLa-a must be drawn and Pda calculated. Such data does NOT exist
for KT120.

A conclusion may be drawn here. From the rate of increase in Pda generated by signal
voltages and currents, the benefits of having a pair of KT120 instead of having say KT88,
or 6550 are marginal at best. Serial amplifier killers who insist on continuous high levels of
signals of Dark Metal or Heavy Metal or Be-Bopitty-shitty will manage a way to make
tubes smoke sooner rather than later.

I tried using a regulated anode B+ of +550Vdc, with screen voltage shunt regulated at
+400Vdc but found maximum output power was limited by the Ra line curve for
Eg1 = 0V having a knee at Ia which was too low to to allow highest output power.

The effect of screen voltage above +350Vdc has a dramatic effect in making high power
possible into a low RLa-a.

Graph 2. UL Operation.
Graph 2 shows Ultralinear ( UL ) operation. The Ea and Eg2 level and 40% screen taps give a safe
maximum PO of 115Watts for KT120 in class AB1, with RLa-a = 2k5. Pda is nearly 50 Watts.
The Vac applied to screens reduces screen current and Pd g2 to fairly safe levels.
The use of cathode feedback windings ( CFB ) on the OPT using between 12.5% and 20% of total
primary turns gives very similar power levels to pure beam tetrode operation where Eg2 = Ea.
With CFB, there is the huge advantage over pure beam tetrode and UL because the CFB % may be
chosen to be anywhere between say 10% and say 50%. With CFB, the screens may be taken to
a fixed Eg2 supply between +330V and +500V, and considerably less than the Ea. The Eg2
level chosen will then limit the maximum Class AB1 PO level with low load values. It is also possible to
connect screens to UL taps where Ea is less than +550V to obtain a combination of CFB and UL
The optimal operation depends on the intended amplifier use. Probably a guitar amp made for high
power may simply employ pure bean tetrode with Ea = Eg2 = +500V and the high THD is of zero
concern. But I find many musicians tell me the best sound comes with less maximum power and a large
amount of class A operation which means that pure beam tetrode operation is still chosen, but Ea and Eg2
will be as low as +320V. Such use of KT120 would be very reliable.

In my 8585 hi-fi amps, initially I tried Ea = Eg2 = 400V with fixed bias, with 12.5% CFB windings,
and screens taken to taps on the anode windings to give Vg2-k signal voltage = 37.5% of the
Va-k signal voltage. In this case a quad of KT88 or 6550 and Ia = 60mA per tube with RLa-a = 4k4
give about 50Watts maximum with THD < 0.7%, and Rout of the amp without any additional global NFB
is less than triode. At lower levels the THD is less than triode. I later changed the 8585 to using Ea = +480V,
and connecting screens to a shunt regulated fixed Eg2 = +330V. UL taps were abandoned.
There was a very slight increase of THD but Ia at idle could be far lower at 33mA and the sound
quality remained quite outstanding. I see no reason why the use of KT120 would give excellent
sound in an "8585" amp. At present, one 8585 has 4 x KT90, and easily gives 100W per channel,
with RLa-a = 4k4. A quad of KT120 might provide 110Watts with the 4k4 and there could be higher
Ia, but in practice the use of higher Ia to obtain more pure class A does not appear to translate into
better sound where the power levels used always remain below the class A limit level with low Ia.

In 1957, The General Electric Co of the UK produced a book "Audio Frequency Amplifier Design with
17 schematics for amps from 5W to 1,100W. PP KT88 is shown with 43% screen taps and Ea = 550V
giving 100 Watts max with RLa-a = 4k5, but not much mention is made of screen dissipation. We may
assume the GE amp recipe would be safe with a sustained sine wave at clipping at 100W with RLa-a load
of 4k5, but with say 2k5, I would suspect the KT88 would overheat and destroy themselves, as they do,
when there is sustained excessive Pda. KT120 would work OK with lower loads of 2k5 and without
exceeding Pda rating of 60W.

For hi-fi, it is far better to use much lower Ea and Eg2, lower Ia, and have a quad of tubes instead of
only a pair.

Graph 3. Beam Tetrode Loadlines.


Graph 3 above shows the Ra "diode" curves for Eg1 = 0V at Eg2 = +350V and +500V.
The Ra curve for Eg2 = +350V has been taken from older data sheets, and not all
present 6550 and KT88 or KT120 can be expected to give exactly what I show here.
Usually the diode line for present tube productions will lean more to the right, and be a
line equivalent to a higher Ra resistance. However, the Ra line I do show is approximately

Where Eg2 is +500V, and for EH KT120, EH KT88, and EH6550, I found the
Ra "diode line" being what it is above.

The diode line shows the limit of negative going Va swing, and it is impossible to force
Class AB1 Va swing on any load extending to the left side of the Ra diode line.
The diode curve knee, ie, Ra curve for Eg1 = 0V is through 175V and 860mA.
Such high Ia was calculated. 146W into 1k5a-a means Vaa = 468Vrms, and the
peak Va swing at each anode =  330Vpk, and class B load = 375 ohms per tube,
so peak Ia = 330V / 375 ohms = 880mA.  This neglects the winding resistance. In practice,
at high PO levels, winding resistance can become a substantial fraction of the load seen by
the tube, and to get the tube to see a load of 375 ohms, RLa-a created by the OPT would
have to be made lower than the theoretical 1k5 by using an OPT secondary load about 10%
lower. The OPT I used for tests weighed 10.5Kg, and and was meant for 8 x 6550.
It has lower winding losses than nearly all other OPTs I know, so there was no need
for me to make tests .
Peak current production capability of tube is limited by the Eg1 = 0V diode line, and not
by the cathode itself, which probably could produce far more current if Eg2 and Eg1 were
both raised to high enough levels. This is avoided lest the tube destroy itself all too easily.
KT120 has a data specified Pg2 limit of 8 watts, and if Eg2 = +500V, then average
Idc current into screen must not exceed 16mA. I have not made average screen current
in my tests, but I did not see any red glow of screens. A Vdc meter held across
a series 10 ohm current sensing R to screen indicated 4mA dc at idle, and 17mA at high PO.
But this does not tell the whole story because the screen current at high PO is a non
linear signal current with wave form like the Ia current wave in each tube which is like
a 1/2 wave rectifier during hi PO class AB operation.

Alarming and possibly dangerous levels of peak anode and peak screen current can be
generated when Eg2 is high.

If attempts are made to **sustain** highest output powers then the screen dissipation can
exceed the screen Pd ratings, and care must be taken to ensure the screen voltage is allowed
to reduce when screen current becomes excessive. In practice, when screens are supplied
via an R&C network from the anode supply with a high value C screen bypass cap,
say 470uF, then most music signals will not cause voltage rail change between idle and full
maximum theoretical output power. This is only true where the music level is brought up until
only very occasional clipping occurs on high level transient peaks. This means average power
level may be 1/5 of the maximum sustained maximum. For where an amp is used to amplify
electric guitar or other music instruments to a repeating and frequent grossly overloaded
condition, then the KT120 or other output lesser capable tubes may overheat.

Any conclusions about operating conditions will vary depending on the intended use.
With Hi-Fi amps where no clipping is ever permitted, average power is low,
and its unlikely that tubes will suffer badly unless an idiot connects a load far too low,
or a damaged speaker, faulty/arcing ESL panel, or the tubes age, or fail randomly and
For PA or guitar amp use where gross overload is must me tolerated, coupling caps
to grids should not exceed 0.033uF, with bias resistors not exceeding 68k, and
grid stopper R can be up to 6k8. This gives a LF pole of  71 Hz. It also allows the OP tubes
rapidly begin to operate in class C if there is grid current charging of the caps, and then to
recover back into class AB1 quickly when drive is reduced below grid current producing

Some of my conclusions :-

1. For Ea up to +550V, KT120 gives only a tiny increase in maximum AB1
power over KT90, KT88 or 6550.

2. At maximum possible class AB1 power, and regardless of what beam tetrode
is used, there is a very small amount of pure class A1 power below the A to AB
threshold is reached.

3. The higher Pda rating of 60 Watts for KT120 does allow for a higher amount
of sustained output power into lower load values than tubes with lower Pda ratings.
However, there are limits to what is possible no matter how many output tubes
are used or what their combined Pda max rating. It must be remembered that a
quad of 6550 or KT88 or a six pack of EL34 or 6L6GC will more easily do
what a pair of KT120 will do.

4. High power is always possible with very low anode loads if the screen voltage
is raised to a highest possible voltage. But the highest possible screen voltage
could result in excessive screen dissipation in say a grossly over driven guitar amp.
Thus you rarely ever see guitar amps with B+ rails exceeding +470V.
Where Ea exceeds +470V as in the Ampeg 300W SVT with 6 x 6550s, the screens
are at a much lower voltage produced by a separate lower voltage HT winding.
Ea and Eg2 can sag a bit when high PO is sustained. Each 6550 has a high
value series grid1 "stopper" resistor of 47k from the cathode follower driver to try
to limit overheating if ever grid current flows.
If there is 300W of audio power produced at the anodes with tubes, and efficiency
is 55%, not uncommon, it means the PSU must supply 545W, so anode heat
= 545W - 300W = 245W, so each 6550 has Pda = 40.8 W.  The Ampeg uses
one value of grid bias shared by all 6 x 6550, and if one or more have aged and run
hot at idle then they can be "pushed over the edge" easily at high power output.
No wonder such amps can give trouble when they are over driven. By using KT120
instead of 6550 and with idle current kept below 50mA, an Ampeg would still
produce 300W, but with less tube failures.

5. Consider a pair of KT120 with B+ anode supply = +700V, and screens at +350V.
Let RLa-a = 8,000 ohms. My load line analysis shows in Graph 2 above shows max
PO = 106Watts. This will not sound much different to being able to make 150Watts.
The use of a high load value and high Ea allows two benefits. One is the much lower
peak anode current, and the other is a much lower peak screen current, and the
screens may be at a considerable lower voltage than Ea. But inevitably someone
will connect a low value speaker load, say 4 ohms instead of 8 ohms when the amp
terminals clearly say "8 ohms" and the amp will quickly overheat at high power.

If you want lots and lots of power, and you want reliability, use lots and lots of tubes,
and never expect more than 75Watts from one pair of beam tetrodes no matter what
octal based tubes you have.

6. If you want a bullet-proof tube hi power class AB amp, you need several protection
There should be separate grid bias pots for each output tube, 10k, wire wound.
Use red-green LED indication of bias current status at idle, red = too high, green = too low,
so when both LED remain unlit, bias is OK. During class AB power production, red
LEDs will light up indicating increased average Ia in each OP tube.
In PA and guitar amps, it is assumed that DC power drawn from PSU may become 4 times
the idle value. 2 x KT120 might have say 100mA of total Ia and Ig2 at idle. But at full PO to
lowest RL value, Idc may become 350mA. So unlike the situation with hi-fi amps, you cannot
use idle Idc current sensing to protect a tube amp. However with red-green LEDs, if a tube
turns an LED to red at idle while others are unlit, then it indicates a problem, and action can
be taken, which may mean a tube replacement during a break in the GIG. if you Ignore red
LEDs then shit happens.
There have never been any amplifiers made which have a circuit built in to detect when a load
value is too low. Fuses and many other protection circuits developed over the last 50 years
still allow a constant stream of damaged amplifiers to my workshop for what are sometimes
expensive repairs.
To achieve amplifier device protection, the output load signal current must be sensed
and converted to a Vdc and compared to Vdc produced from a fraction of the output load
voltage, so that where current becomes excessive the differential amp used for comparison
switches off the amp, regardless of the level used. Some people will try to use a
2 ohm speaker with an amp meant to have 8 ohms, and try to use low levels so the amp
won't overheat. Sooner or later, too much level is used, and tubes overheat, despite good
intentions of the amp user. The 2 ohm load should immediately turn off the amp within a
second if it detects a low load, even at low levels. Tube amps do tend to last longer
than transistors during times when loads that are too low are used.

Not one amp maker has used such a protection measure; they ALL want you to wreck your
amplifier so you might buy a new one.

7. Real benefits of KT120 would be most possible for where LOW power is desired for hi-fi
and where clipping will never occur, and the power is nearly all pure class A. The maximum pure
class A possible from any two beam tetrodes or pentodes is about 45% of the total Pda at the
idle condition, so that if two KT120 are used with Pda at a safe 40Watts each, Pda total = 80W
and max class A PO = 36W, and this is substantially above a pair of KT88 with safe Pda total = 60W,
and class A PO max = 27W.
For class A, RLa-a is always higher than for class AB, and maximum peak Ea swing will always be a
high fraction of Ea. Peak Ia swing in the class A amp is never more than twice the idle Ia.
With the class A idle Ia set higher than for class AB amps, the gm at the operating point is both
higher and more constant, resulting in much more linear operation. Screen supply voltages seldom
need to be higher than +300V, and this allows lower g1 bias voltages easily generated using cathode
biasing with R&C networks between cathode and 0V, and without wasting much power in hot running
Rk. The class A tube has high voltage gain because gain = gm x RL (approx) and where local CFB is
used as in the Acoustical mode ( like Quad-II ), remarkably linear operation is gained.

For example, consider KT120, Ea = Eg2 = +400V, Ia = 100mA, and RLa-a = 5,600 ohms.
Peak Va swing = 360V, and PO = 46W, AB1, with the first 28 Watts in pure class A1.
The loadline analysis shows the 60W Pda limit line for KT120 well above the class B load line
of 1,400 ohms, so there is no risk of tubes overheating. Even if the load is reduced to 2,800 ohms,
max AB1 PO = 68W, and there is still no risk of tubes overheating. From this it is possible to use
KT120 at an idle Pda = 40W reliably so that Ia = 100mA, Under such conditions, distortion is
minimized, and music may sound much better than if the amp were set up with Ea at +550V, and
Ia at idle of 50mA, and RLa-a = 2k0, and class A max = 2.45Watts.

8. During the history of tube amplifier development, it was possible after 1957 to get 140Watts
of class AB1 audio power from a pair of TT21 which were exactly like KT88 but which had
anode top cap connections which avoided having high voltages appearing at pin 3 of the octal
My copy of the 1976 5th edition of the Radio Communication Book for British amateur radio
operators shows the TT21 having Ea at 1,000Vdc, Eg2 at +300V,
and driven easily with a 12AU7 balanced amp, and one 12AX7 input tube.
If the Ea minimum was say 120V, then Va pk swing = 880V, so Va-a = 1,244Vrms.
This is a soberingly deadly voltage level, and there is a total of +1,880Vpeak volts above 0V
at the anode, and so KT88, KT120, or any other octal tube could not be used because of the risk
of arcing from anode to earthy elements close by.
And if PO = 140Watts, then anode to anode load is 11,000 ohms, and peak Ia = only 318 mA,
and much less than the the anode current to get 140Watts with Ea at only +500V. Pity help
the amp if someone connected a load less than 11k0. Tubes would overheat very quickly.
People using such amps had to be very careful about loads. Guitar amp users and PA operators
are often quite careless about loading, so they walk around with pockets stuffed with spare tubes.
And when a speaker fails and its load value goes LOW, it will damage an amp all too easily,
even when used at low levels. I have had to repair so many amps which died from every
possible fault.

Quite a number of people tried to get silly high AB1 power from tubes like EL34 by using
Ea = 900V, and Eg2 at 450V, and load = 11k0. This gave nearly 100 Watts, class AB1.
Anode Pin 3 would develop arcing to earthy heater pin 2 beside pin 1. Tubes failed all too easily.
One could also use KT88 in the same amp for slightly more power, but arcing still happened.
So it was foolish to use such high Ea, where octal sockets were used. Also foolish because
amps were subject to loads which were altered over time, without anyone knowing what might
happen, such as when a priest would decide to add two more speakers to make sure people
at the distant parts the church could hear about fire and brimstone with greater clarity. Well, they
might get a bit of a smell of that as well, and crackle noises as the amp decided to smoke because
the load value had become too low, placing tubes operating outside their Safe Operating Area,
or SOA, which every amp designer might know, but not the priest.

One might possibly use a small rubber O-ring around pin 3 of octal output tubes tube and then tie
the tube bass tightly down onto the tube socket, thus making and air tight join around pin 3 between
socket plastic or ceramic and the tube base, thus preventing corona formation and effects of accumulated
dust, moisture, pollution allowing leakage currents from pin 3. But tying down EL34 with a plastic
base is difficult and the only tubes worth considering  would be the KT120, KT88, 6550 which have
metal base sleeves. These are internally connected to pin 1, and thus pin 1 socket tag MUST NOT be
connected to tag 8 when the sleeve is tied to the chassis with copper wires soldered to the sleeve, and
taken through chassis holes for twisting up tightly and soldering.

There are purchasable retainer clips made for fitting over the top of a tube and which have a spring
each side which pull a tube tightly to the socket but which still allow some movement. These might just
give enough pressure around a small O-ring, as well as stop a tube working its way out of a socket
and losing its connections before falling out of a socket, especially in guitar amps where tubes are
usually mounted upside down on the chassis. Silicone sealant may be applied around socket
tags with HV and thus make arcing less likely. It is an inconvenient measure to have to take, but
one than makers should always practice where Ea exceeds +500V.
For high PO from 6L6, and before the 6L6GC came out with higher Eg2  rating of +450V, the
old 6L6 with Eg2 rating of 300V could be used with Ea = +600V. With Eg2 at =300V, PO max
= 80 Watts using class AB2. Cathode followers were used to drive the grids maybe 15Volts positive.
807 were MUCH better is such amps, because their sockets have wider spacing between pins,
and there is an anode top cap to keep HV away from anything earthy.  It could be argued which
is more reliable, 4 x 6L6GC or 2 x KT120, because either could produce over 100W
if Ea is high enough.

The use high Ea with octal tubes is always associated with low bias current class AB1 or AB2
amps for PA where high distortion is tolerated, and is rarely done in hi-fi amps.

For hi-fi, a quad of 6L6, 6CA7, EL34 can easily give 60W, with 30Watts of possible pure class
A1 with Pda at idle of 18 Watts in each tube. This will be just as reliable as using a pair of KT120,
KT90, KT88 or 6550 to obtain the same power. But KT120 would be the best at safely sustaining
idle Pda of 36 Watts each to match a pair of the smaller octal tubes' total idle Pda.

In some ways class AB2 drive of beam tetrodes may be more reliable than having a high Eg2 to
allow high AB1 PO with a low Ea and low RLa-a. Class AB2 means that the grid input resistance
changes abruptly from megohms while grids remain negative to less than 2k0. Direct coupled
cathode followers are usual to drive grids positively. But AB2 amps are rarely ever made because
of the extra expense of tubes and sockets etc, and because gross overload will cause grids to
overheat, and cause tubes to fail. The AB2 allows for the benefits of low screen power to be used
and all is well until gross overload occurs, and it certainly will where an amp is used by those with
no technical knowledge.

All I can do is give reasons for my preferences, and maybe 1% of the readers will take my advice,
and be very careful to use ONLY the right value load, or one that is a higher number of ohms.
It is tempting to connect a 4 ohm speaker to am amp with output terminals labelled "8 ohms"
Don't do it. But using 16 ohms is fine.


Graph 4.
Graph 4 above shows the power output obtainable with a pair of KT120 using Ea = Eg2 = +400V
and Ia in each tube = 100mA, and beam tetrode mode.
Notice that there is up to about 35 Watts of pure class A1 available, and usually PLENTY for 95%
of most Hi-Fi enthusiasts who want fine music without ear damage. The Pda of each KT120 at idle
can reliably be 40 Watts. Using any popular off the shelf OPT such as a Hammond 1650P with
6k6 : 4,8 &16 ohm load match, you should get at least 32 Watts of class A1 max at 6k6 load.
If you used EL34, set up as in the classic Mullard 520, and many other amps since 1957, you
would get about 35Watts class AB max, but much less class A. If Pda for each EL34 = 20W,
then mas class A PO = 14 Watts at Ra-a = 14.4k. The 1650P OPT could be used, but there
is only 8 Watts of class A with RLa-a = 6k6.
Clearly KT120 can give better class A performance.

For those contemplating the use of KT120 to replace EL34, 6L6, 6L6GC, 5881, KT66
to obtain more pure class A1 power, they MUST remember that the KT120 will need
twice the anode current and 50% more heater current than EL34 to give the large increase
in class A power. Many amps cannot do this, and you must use the same Ia as used
in the original smaller tubes. This is OK and there will not be any increase in class A1
power but there may be a large increase in class AB power. Whether KT120 may be used
at all instead of EL34 or KT66 depends on the heater power available. KT120 need nearly
twice the heater current of a KT66, ( 1.3A ), and more than twice 6L6 ( 0.9A ). Often,
many amps fitted with EL34, KT66, 6L6 have PSUs designed to provide more power than
they do especially if the amp is fitted with a rear panel octal socket for power to a preamp.
I have found Quad-II amps will happily work with KT88 or KT90, but KT120 would seem
asking too much, and adding an extra 15VA tranny to power one KT120 heater seems
pointless. Anode current from Quad-II PSU is limited to about 150mA max. Of course
in all such "hotrodding" of a standard amp, it is good practice to replace the tube rectifier
and use the available space for large capacitors. Tube rectifiers provide no sonic benefit
at all, and are notoriously responsible for PSU inefficiency, and they cause general
unnecessary heating of all surrounding parts on the chassis. When replaced with Si diodes
the B+ can be +400V instead of +365V at the same Idc, so the power transformer will
not overheat. With higher B+, class AB PO is considerably increased. 32Watts class AB
is easily possible from Quad-II amps using KT88/KT90, and even if the cathode biasing
is retained. KT88 Va swing can be more than KT66, and with higher Ea, there is a higher
power ceiling. I have used fixed bias in Quad-II to get 20Watts of triode power
with KT88 in Quad-II.

KT120 could be well used in Quad-II-Forty amps which were first produced by Quad in
the mid 1990s. KT88 were standard output tubes. But the HT winding for the 5U4
rectifier is 390-0-390Vrms, giving B+ = +428V at 161mA. If Si diodes are used, one could
get B+ = +525V at 161mA. Screen voltage should be kept at +410V, so cathode bias
will be about +53V with Rk = 630 ohms. Pda in each KT120 can be 38Watts, which
is too high for KT88. Since Ea at idle would be about 470V, expect over 50Watts maximum.
The Quad-II Forty in original form has KT88 idling too hot, and I found it impossible to get
more than 32.5 Watts classAB1, and 23Watts max class A with Ia at a sensible 62mA, which
is the sensible value for Ia, and less than in the original amps. With KT120, the Quad-II-Forty
can be persuaded to "mature" into something better, especially if one forgets the 8 ohm outlet
exists, and only uses the 4 ohm outlet with speaker Z always more than 6 ohms.

Graph 5.

Graph 5 shows KT120 compared to KT88 in class AB1 PP triode, Ea = +500V, Ia = 50mA at idle
The amount of power difference is not huge. For either KT120 or KT88, I'd recommend RLa-a =
8k0 to 12k0 to get more than 23Watts of AB1 power with a useful amount of class A power.
One should find that with speakers with sensitivity < 91dB/W/M, the sound for most ppl will
be outstanding, and easily match the use of a pair of 300B, or a lone SE 845 or 211.
With KT120, continuous clipping power with loads below 1k4 will cause overheating.
With KT88, continuous clipping power with loads below about 2k2 will cause overheating.
The curve for KT88 Pda has not been included but will be below that for KT120.