Single ended EL84 amp - cannot bias hot?

[marcus262]

Hi guys,

I would really appreciate if someone would help me with this problem.
The amp is single ended el84, cathode biased, made by a local guy who makes point to point tube amps,
and it is working fine with no problems but this one.

Problem: using a bias calculator from robrobinette page, calculation show that it was biased very cold,
so I tried to bias it hot by changing the original 180 Ohm resistor, but if I lower the resistor then the plate voltage and the voltage across cathode resistor also  drop and bias changes but by a very little amount.
So I am wondering if this is normal, because I am afraid to go any lower with the cathode resistor value, also posted the schematic I managed to trace.

1. Value of 180 ohm

voltage drop across resistor: 6.8 V
Plate to cathode voltage: 283 V

using calc: plate current: 35.7,  10.1W ,   84.2% dissipation.

2. Tried 150 ohm

voltage drop across resistor: 6.07 V
Plate to cathode voltage: 272 V
screen voltage: 237 V

using calc: plate current: 38.3,  10.4W ,   86.7% dissipation.


3. tried 133 ohm

voltage drop across resistor: 5.73 V
Plate to cathode voltage: 268 V
screen voltage: 230 V

using calc: plate current: 40.7,  10.9W ,   90.8% dissipation.

4. tried 123 ohm

voltage drop across resistor: 5.44 V
Plate to cathode voltage: 266 V
screen voltage: 230 V

using calc: plate current: 41.9,  11.1W ,   92.5% dissipation.


Afraid to go lower, it seems that something is wrong, usually I see around 150-130 ohm resistors on SE for max dissipation.
Is this normal, should I go lower, maybe something is wired wrongly or I am measuring something differently.
Any suggestion is welcomed.

Thanks

[sluckey]

but if I lower the resistor then the plate voltage and the voltage across cathode resistor also  drop and bias changes but by a very little amount. So I am wondering if this is normal

That's perfectly normal. Yes, the voltages will change each time you decrease the cathode resistor but the power also creeps up. Maybe not as much as you were hoping, but it does increase. Don't be afraid to continue decreasing the cathode resistor until you reach the power dissipation you want. You may have to increase the power rating of the resistor so check that too.

[shooter]

help me with this problem.

welcome to the life an SE builder
Like Steve said, keep going, you might also get a small tweak on the plate by changing your B+ filter R from 470 to 390, or 330.  That might add unwanted ripple, so check that.

I've found there isn't a lot of "sound" increase from high 80 to 100%.  Use a more efficient speaker to "go the extra yard"

[kagliostro]

Use a more efficient speaker  ....

+ 1


Franco

[marcus262]

but if I lower the resistor then the plate voltage and the voltage across cathode resistor also  drop and bias changes but by a very little amount. So I am wondering if this is normal

That's perfectly normal. Yes, the voltages will change each time you decrease the cathode resistor but the power also creeps up. Maybe not as much as you were hoping, but it does increase. Don't be afraid to continue decreasing the cathode resistor until you reach the power dissipation you want. You may have to increase the power rating of the resistor so check that too.

Thank you for your answer. So I would probably need a 100R resistor maybe even lower, got to go shopping again. :)
I find it very strange that i would have to go that low just to get to 95% dissip.
Maybe my multimeter is giving me incorrect readings.

help me with this problem.

welcome to the life an SE builder
Like Steve said, keep going, you might also get a small tweak on the plate by changing your B+ filter R from 470 to 390, or 330.  That might add unwanted ripple, so check that.

I've found there isn't a lot of "sound" increase from high 80 to 100%.  Use a more efficient speaker to "go the extra yard"

thanks 

so,  decreasing B+ filter would increase voltage on the plate if I am understanding correctly, but then the screen voltage would rise also, and I read that it is not a good idea to cross the 250V on the screen, so I would have to tweak the 4.7k screen resistor too?

I don't want to bias hot because of extra wattage, It is just that I noticed when I increase master volume to 11:30 the sound gets very harsh,
at home I usually play at MV 9:00 and it is "normal" loud and sounds ok,
but because  tube amps are supposed to sound better when turned up, not harsh like solid state I raised the MV to hear for myself.
So I am guessing maybe the sound gets harsh because of cold bias thing.

[marcus262]

just two more questions 

1. I always read that on cathode biased amp, as you apply more signal, it biases itself colder, but when I play and raise
master volume, I don't notice any change in cathode or plate voltage.

2. also, I read that changing cathode bypass capacitor to 1000uf would stiffen the amps response, that the bass wouldn't flab so much,
that was for amps with 2 power tubes, would that work for single ended?

[sluckey]

Put your 180Ω and 150Ω in parallel. That's 82Ω. Does that get you to 100%?

The EL84 is a bright sounding tube to me (Some may even call that harsh). And it's easy to overdrive, especially with the preamp you have.

[SILVERGUN]

.... also, I read that changing cathode bypass capacitor to 1000uf would stiffen the amps response, that the bass wouldn't flab so much,
that was for amps with 2 power tubes, would that work for single ended?

You wouldn't notice much difference from the 47uf that's there now (i.e....it's not the magic bullet)
You 'could' try removing the bypass cap to reduce input sensitivity (but again, not the magic bullet)

I don't want to bias hot because of extra wattage, It is just that I noticed when I increase master volume to 11:30 the sound gets very harsh,
at home I usually play at MV 9:00 and it is "normal" loud and sounds ok,
but because  tube amps are supposed to sound better when turned up, not harsh like solid state I raised the MV to hear for myself.
So I am guessing maybe the sound gets harsh because of cold bias thing.

You could be pushing the 3rd stage too hard. (I would add a grid stopper on V2a )
What is your V2a plate voltage?
You may have just built a preamp that doesn't sound that great into an EL84. (please don't take offense, we can fix it)
This would be a great use for a scope...scope the signal at the grid and then plate of each stage and see where it's going wrong.


The EL84 has the qualities sluckey mentioned above, and looking over your schematic shows some serious bass cut and mid/high boost. So the amp has cascaded gain and is boosting highs and cutting lows into an output tube that is bright and easily overdriven....sounds like a recipe for "harsh" tone. Plus, the low plate voltages makes for tubes that clip earlier.
Where did you get this schematic?....not all of 'em are "winners".

[Ed_Chambley]

I have a few BIG wattage wirewound Pots that are 500 ohms.  I got tired of swapping resistors since I am not quite good enough to calculate what the plate and screen will do when changing the resistor.  I hook that thing up and play and dial down and watch plates.  I move it until I get the response I am looking for.  Then I take measurements to check dissipation.  If I land at 95% I would use it.


I have found not very much difference above 80% or so with EL84.  As Sluckey mentioned they are bright and easy to overdrive.  My favorite set to overdrive are Amperex, but I do know how much they do now cost so may be prohibitive to the ole pocketbook.


For all around the RCA "pan getter" have some warmth and all I have ever used bias up readily using 125 to 130 ohm.  One thing I always build into a EL84 amp is a cut control to trim the highs and therefore that harsh, raspy deal.  I have never built a EL84 SE amp tho, but I would think you could get a single unmatched Amperex and RCA fairly cheap.  Tube rolling in a K bias is sometimes a pain, but not with the pot and I mean potentiometer.  I might say pot and the inside of a running amp may not be a good idea.

[jjasilli]

just two more questions 

1. I always read that on cathode biased amp, as you apply more signal, it biases itself colder, but when I play and raise
master volume, I don't notice any change in cathode or plate voltage.


2. also, I read that changing cathode bypass capacitor to 1000uf would stiffen the amps response, that the bass wouldn't flab so much,
that was for amps with 2 power tubes, would that work for single ended?

1.  Actually it biases both hotter & colder.  The input signal is an AC wave and has a point to point range (low point -to- zero -to- hi point, reverse & repeat).  Above -0- adds bias voltage to G1, hence colder bias. Below -0- subtracts bias voltage on G1, hence hotter bias.  But this oscillating bias signal voltage  wiggles at many times per second, hence it tends to cancel itself from an operational perspective.  Before the plate can get too hot, it's already cooling & vice-versa, usually.  But always check for redplating, 'cause it can happen with hot bias combined with overdriving signal strength.  (Can also be an issue if bias vary tremolo is used on the power tube(s), which also adds and subtracts voltage on G1, but at a much slower oscillation rate than signal).


2. Increasing bypass cap value increases bass response, usually around 100uF is the max.  For testing I built a box with dual Weber 500R rheostats (as variable bias resistors)and switchblade bypass caps, to A/B cathode biased power stages.

[Ed_Chambley]

.... also, I read that changing cathode bypass capacitor to 1000uf would stiffen the amps response, that the bass wouldn't flab so much,
that was for amps with 2 power tubes, would that work for single ended?

You wouldn't notice much difference from the 47uf that's there now (i.e....it's not the magic bullet)
You 'could' try removing the bypass cap to reduce input sensitivity (but again, not the magic bullet)

I don't want to bias hot because of extra wattage, It is just that I noticed when I increase master volume to 11:30 the sound gets very harsh,
at home I usually play at MV 9:00 and it is "normal" loud and sounds ok,
but because  tube amps are supposed to sound better when turned up, not harsh like solid state I raised the MV to hear for myself.
So I am guessing maybe the sound gets harsh because of cold bias thing.

You could be pushing the 3rd stage too hard. (I would add a grid stopper on V2a )
What is your V2a plate voltage?
You may have just built a preamp that doesn't sound that great into an EL84. (please don't take offense, we can fix it)
This would be a great use for a scope...scope the signal at the grid and then plate of each stage and see where it's going wrong.


The EL84 has the qualities sluckey mentioned above, and looking over your schematic shows some serious bass cut and mid/high boost. So the amp has cascaded gain and is boosting highs and cutting lows into an output tube that is bright and easily overdriven....sounds like a recipe for "harsh" tone. Plus, the low plate voltages makes for tubes that clip earlier.
Where did you get this schematic?....not all of 'em are "winners".

Good call.  I did not notice the plexi values of the first 2 stages.  Even a Push Pull El34 Plexi is harsh to me unless channel jumped.

[PRR]

> I always read that on cathode biased amp, as you apply more signal, it biases itself colder

You can read ANYthing on the interwebs.

A properly biased and loaded self-bias amplifier runs essentially the *same* current idle or loud. In practice a 10%-20% rise is often done.

The plate HEAT does drop with signal. The self-bias amp starts by taking all the DC power it could ever need. Then it diverts some of that power to load as commanded. So the plates run max-hot no-signal, and less-hot with signal.

I had an interesting proof. Self-bias 6550 at 90% Pdiss idle. Dummy load so it ran silent. I had a vintage 6550 which "creaked" like an old VW Bug when it cooled down. When I applied full-power test tone, the tube creaked. I could also observe the cooling with an IR thermometer. Some calcs suggest that 39W Pdiss idle became 22W Pdiss when delivering 17 near-clean watts to dummy load, and under 20W Pdiss when over-driven to near square-wave.

But we can't play LOUD or ALL the time to reduce tube stress.

> not a good idea to cross the 250V

Bah. Just do it.

It might be interesting to do the Full Design. The power supply voltages are lower than we sometimes see in guitar-amp work, and the load impedance is unknown?

The conditions you have tried "optimize" to different load impedances.
283V 36mA = 8K
266V 42mA = 6K

Your power supply impedance is near 333 Ohms (5V drop for 15mA more current). (Oddly this is not what is on the back-trace.) 50mA should leave >250V across the tube and touch 12.5W Pdiss (which I would not worry about if it improved the tone), 5K "optimum" load. This seems to be 100 Ohms for Rk. I would try that to get numbers and beat on it a bit for tonal opinion. If you are out of 100r parts, recall that three 330r in parallel is 110r, and a 1K parallel to that is 100r for practical purpose.

180 and 150 parallel may get you over 100% Pdiss. The EL84 will tolerate this for hours so it is worth a try.

[marcus262]

Put your 180Ω and 150Ω in parallel. That's 82Ω. Does that get you to 100%?

The EL84 is a bright sounding tube to me (Some may even call that harsh). And it's easy to overdrive, especially with the preamp you have.

Thank you for your suggestion, I didn't have the guts to try it for myself, was worried that something would burn up.
So, I tried 180R and 150R in parallel and it works great.

I get:
voltage drop 4.09 V
plate to cathode 250 V
screen 211 V
= 11.8W,   47.3 mA,   98.3% dissip.
Hope these figures are ok?

Will try to crank it tomorrow. to compare how it sounds.
Currently I have EI el84, will buy russian 6p14p, read that it has much higher headroom.

You could be pushing the 3rd stage too hard. (I would add a grid stopper on V2a )
What is your V2a plate voltage?

maybe it is hard to read, but the 3rd stage has 470k grid stopper, which forms 470k/470k voltage divider, so input is halved.
V2a plate voltage is 105 V.

You may have just built a preamp that doesn't sound that great into an EL84. (please don't take offense, we can fix it)

I bought it from that amp builder, but it didn't sound that great so I started tweaking it, couldn't help myself.
The major problem was, that with higher gain it would sound weird, and change freq. response, so I found out that some signal wires were running too close to high voltage wires and with higher master volume there was high pitched squeel (it turned out it was a v1a cathode wire running too close to OT)...

The EL84 has the qualities sluckey mentioned above, and looking over your schematic shows some serious bass cut and mid/high boost. So the amp has cascaded gain and is boosting highs and cutting lows into an output tube that is bright and easily overdriven....sounds like a recipe for "harsh" tone. Plus, the low plate voltages makes for tubes that clip earlier.

I changed v1a coupling capacitor to 0.0022 because at higher gain the bass would sound fuzzy and not tight, and there is one 470pf/470k booster for the Marshall tone.
traced the schematic myself by measuring all components and connections.

I have found not very much difference above 80% or so with EL84.  As Sluckey mentioned they are bright and easy to overdrive.  My favorite set to overdrive are Amperex, but I do know how much they do now cost so may be prohibitive to the ole pocketbook.

I will try the russian 6p14p which is the same as preferred 7189

One thing I always build into a EL84 amp is a cut control to trim the highs and therefore that harsh, raspy deal.

i know about plate bypass capacitors that can trim high end, but also don't want to trim too much. Don't know which value to use on what triode since I don't wanna lose harmnonics or for it to sound dull.

 

[marcus262]

Actually it biases both hotter & colder.  The input signal is an AC wave and has a point to point range (low point -to- zero -to- hi point, reverse & repeat).  Above -0- adds bias voltage to G1, hence colder bias. Below -0- subtracts bias voltage on G1, hence hotter bias.  But this oscillating bias signal voltage  wiggles at many times per second, hence it tends to cancel itself from an operational perspective.  Before the plate can get too hot, it's already cooling & vice-versa, usually.  But always check for redplating, 'cause it can happen with hot bias combined with overdriving signal strength.

thnx for the info, power section is more difficult for me to understand., currently I don't have any redplating.

2. Increasing bypass cap value increases bass response, usually around 100uF is the max.  For testing I built a box with dual Weber 500R rheostats (as variable bias resistors)and switchblade bypass caps, to A/B cathode biased power stages.

i found this explanation for the huge cap:
"Anyway, the reason you can get boomy bass with a cathode-biased amp is that the bias voltage at the cathode gets modulated by large bass notes. However, using a very large cathode cap (1000uF to 2200uF) forces the cathode voltage to hold itself much more constant and only change slowly with the amplified signal. Therefore the bias voltage doesn't get modulated by bass notes, and follows the overall envelope of the signal instead, which is what it really should be doing. It also helps get rid of transient spikes in the bias voltage which can make the distortion tone sound rough or splattery. The amp will feel a little stiffer than before, and a little more like a fixed-bias amp, but without losing the character of a cathode-biased amp."

Good call.  I did not notice the plexi values of the first 2 stages.  Even a Push Pull El34 Plexi is harsh to me unless channel jumped.

I have on a switch 2.2uf cathode bypass on v1a, to get more fuller sound, also if I remove the 470pf treble peaker capacitor it is missing some of the freq. to get that marshall sound.

Bah. Just do it.

It might be interesting to do the Full Design. The power supply voltages are lower than we sometimes see in guitar-amp work, and the load impedance is unknown?

The conditions you have tried "optimize" to different load impedances.
283V 36mA = 8K
266V 42mA = 6K

Your power supply impedance is near 333 Ohms (5V drop for 15mA more current). (Oddly this is not what is on the back-trace.) 50mA should leave >250V across the tube and touch 12.5W Pdiss (which I would not worry about if it improved the tone), 5K "optimum" load. This seems to be 100 Ohms for Rk. I would try that to get numbers and beat on it a bit for tonal opinion. If you are out of 100r parts, recall that three 330r in parallel is 110r, and a 1K parallel to that is 100r for practical purpose.

I am sorry, having trouble understanding the concept of impedance (not a native english speaker),
is it the same as resistance? I didn't know I have to take that into account.
excuse my lack of knowledge.

[jjasilli]

Impedance:  checkout some articles here & you'll get the idea:  https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=impedance%20vs.%20resistance

[SILVERGUN]

maybe it is hard to read, but the 3rd stage has 470k grid stopper, which forms 470k/470k voltage divider, so input is halved.
V2a plate voltage is 105 V.

Not hard to read...
You're correct about the voltage divider, but it won't act as a grid stopper unless it's placed after the 470K grid leak resistor...(i.e. between the 470K to ground and the grid)

With that low plate voltage and a big input signal, that stage is acting like a very narrow, dual headed buzz saw

[marcus262]

Not hard to read...
You're correct about the voltage divider, but it won't act as a grid stopper unless it's placed after the 470K grid leak resistor...(i.e. between the 470K to ground and the grid)

With that low plate voltage and a big input signal, that stage is acting like a very narrow, dual headed buzz saw

so, then I also don't have a grid stopper on v1a, by the same logic?

I am not sure about the voltages anymore, if they are high enough as it stands currently (250v plate el84 and 105V v1a), neither preamp or poweramp.
Would you recommend I change the 10k, 4.7K or 470R resistors?
What would be ideal?

The voltage before 470R is 297V, after 267V
Before 4K7 is 267V, after 219V
10K drops to 187V and 172V, before v1 plate resistors drop that voltage to 105V

Thanks

[shooter]

Your DC volts look close to a lot of what I build, You do have to know what the AC signal is doing stage to stage also, if you're trying to swing a 50V AC signal and your tube can only swing 10Vac, you might be limited to Death-metal

[marcus262]

so, would you all recommend that I lower the 470R and/or 10K resistors,
and put a 470k grid stopper before v2a?

I know my amp wouldn't even be close to 1959 Jmp,
but i tried to mimic the circuit as close as possible.
I have tried wiring v1a and v1b to run parallel into v2a, but there is not enough distortion, had to run both gain pots at max and it lacked something (maybe compression, I don't know).
Guess that is because I don't have phase inverter to contribute to gain or distortion.
And I don't know how to approximate the gain and distortion of PI with the v2a gain stage I have currenty 820R unbypassed.

Sorry if my attempt sounds silly, or that it is even remotely possible,  I am trying to learn. 
The guy who made the amp obviously did a poor job on the B+ voltages, so I am hoping to correct that problem.

Thank you guys for all the help and suggestions.

[SILVERGUN]

so, would you all recommend that I lower the 470R and/or 10K resistors,
and put a 470k grid stopper before v2a?

The 470R is ok, for now (you could drop value)
'Try' changing out the 4.7K and 10K for 1K each(or any low value you have on hand)
You can even try just strapping a lower value in parallel to see where it gets you.
This will provide slightly more 'headroom' through each preamp stage, but will also slightly increase overall gain which will also contribute to overdriving your EL84 even more. (so if that is really your main problem, this could make it worse)
 
The 470K grid stopper cant hurt (270, 220, 100K,,,one of these is better than nothing)
 
These won't completely solve your issue, but you will start to get a feel for what might.
 
You might want to try snipping the .068 bypass cap off of the 2nd stage (V1b) and see if you like that.
 
Do 1 thing at a time and test before making multiple 'shot in the dark' changes
 

[shooter]

I am trying to learn

hang in there, it gets better

If you wanna keep doing n learning, one thing that helped me, take real notes instead of brain notes, next week that comes in handy
I know, paperwork sucks!

[PRR]

> concept of impedance .... is it the same as resistance?

Yes, except we say "impedance" when it is not an Exact Resistance but varies with condition.

And, no, it is not an everyday English word, and even English-speakers who use it are often confused.

We do say a speaker has "8 Ohm impedance", to mean "about 8 Ohms, but really varying from 6 to 60 Ohms at different frequencies". The output transformer is a lever, so we say "5K impedance", which will actually vary like the speaker impedance varies.

All your "wrong B+" worries may be pointless. Is there any way to know what your output transformer nominal impedance is?

If 7K (a popular Champ OT) then you want the lower current, higher voltage, and let Pdiss be whatever it comes to.

If it is 5K (a very popular OT) then 250V 50mA is a very fine operating area.

[HotBluePlates]

... I am sorry, having trouble understanding the concept of impedance (not a native english speaker),
is it the same as resistance? I didn't know I have to take that into account. ...

"Impendancija," opposition to the flow of alternating current (a combination of resistance and reactance, expressed in Ohms).

PRR was walking through an estimate of how the output stage should be operating, since you were adjusting bias of the output tube (to answer, "Does the output stage really need any change?").  He noticed the plate voltage dropped 5v when the plate current increased 15mA, which implied the power supply presents 5v/15mA = 333Ω of impedance.

As for "sounds harsh"... As others said, you have a lot of gain in the preamp, a lot of places bass is cut from the signal, and an output tube which doesn't need much signal to be pushed from "nice" to "nasty" distortion.  And the tube itself can sound brash/harsh.  You may have luck increasing the grid stopper on the EL84 somewhat, or reducing the bass-cut in the preamp.  And you may not be able to turn up the master volume very high when the preamp volume is also turned up high (try master volume full-up and slowly turn up the preamp volume to see where the limit of "nice" is).

[PRR]

> "Impendancija"

You may have meant to type "impedancija" ? (You added an "n"?)

https://hr.wikipedia.org/wiki/Elektri%C4%8Dna_impedancija
"Nalik na električni otpor koji je mjera suprostavljanja prolasku istosmjerne električne struje kroz strujni krug, električna impedancija ... je mjera suprostavljanja prolasku izmjenične struje kroz strujni krug."
https://translate.google.com/translate?hl=en&ie=UTF8&prev=_m&sl=auto&tl=en&u=https://hr.wikipedia.org/wiki/Elektri%25C4%258Dna_impedancija%23Primjena

Yes, like a resistance except different. This article seems (in Google Translate) to cover the simple cases of capacitor and inductor. Good basic engineering. In *our* case, a loudspeaker is a resistance, an inductance, a capacitance, and another inductance, with resistive losses. So not a resistor, and not one of the "simple" impedances. For our sanity, we note that speaker impedance varies, but we reference to a medium-low value near the dip of the curve, and pretend this is the load on the amplifier. The deviations in actual impedance and its effects on amplifiers have mostly been considered by the loudspeaker designer, not our problem.

The English Wiki is somewhat longer but maybe not more useful:
https://en.wikipedia.org/wiki/Electrical_impedance

[marcus262]

We do say a speaker has "8 Ohm impedance", to mean "about 8 Ohms, but really varying from 6 to 60 Ohms at different frequencies". The output transformer is a lever, so we say "5K impedance", which will actually vary like the speaker impedance varies.

All your "wrong B+" worries may be pointless. Is there any way to know what your output transformer nominal impedance is?

If 7K (a popular Champ OT) then you want the lower current, higher voltage, and let Pdiss be whatever it comes to.

If it is 5K (a very popular OT) then 250V 50mA is a very fine operating area.

thanks, the speaker example did helpe me understand better.
Don't know, the builder mentioned he could put another tube 6L6 so I could switch (didn't take that option), maybe that helps in indentifying,
or if I could measure with a multimeter?


Changed today 4.7K resistor to 2.98K, and both 10K to 2.12K .

At first listening it sounds bit louder, fatter, clearer and less fizzy. Currently I can't play at normal volumes to properly do the testing.

for el84:
123R cathode resistor
5.8 V drop on resistor
262 V plate to cathode
242 V screen
= 97.5% dissip.

12ax7 plate voltages:
v1a  171 V
v1b  140 V
v2a  132 V

Still have to play at louder volume to check.
Thank you for helping me,
tube amp diy surely isn't for perfectionists, all that tweaking can drive you mad 

[shooter]

all that tweaking can drive you mad

Ah but if you start out mad.......sanity is just another tweak away

[marcus262]

all that tweaking can drive you mad

Ah but if you start out mad.......sanity is just another tweak away

ha ha   

Tried at normal volume I usually play at home,
it is much louder, clearer and more punchier.
Don't know if I like the loudness though, with gain1 at 13:00 and gain2 full, couldn't turn MV even to 9:00.
(treble was 9:30, mids full, bass 11:00).

After 20 minutes of playing don't know if it is harsh (like bad tonal quality) or just loudness perceived as harshness. Maybe it would sound normal in larger room. What I don't understand is when someone on youtube pushes NMV 100W in a room it still sounds great and not harsh.