Tube power amp sounds better than solid state power amp for modelers

Jarick

Shredder
Messages
2,269
Over the last few months, I've been experimenting with running my Fractal Axe FX 3 through a power amp and speaker instead of studio monitors or headphones. I tried a few different solid state amps with flat power amps, and found that while I liked the sound in some ways more than monitors, it wasn't great.

So I tried a few different amps and speakers. I started with the Line 6 Catalyst which sounded good but a bit muddy. Then a Boss Katana 50 which was noticeably brighter. Then a Celestion V-Type which improved the Boss a bit but not the Line 6. Then a Boss Katana Artist Mk 2 which was much bigger and more balanced sounding.

Still, it wasn't blowing me away. The sound was too bright from what I would expect when running cab sims off, and turning them on just meant I was running a guitar speaker emulation into a guitar speaker, which made no sense.

So I ended up buying a Blackstar HT-5 Mk 2 tube amp that featured an effects loop, and I really liked the sound of the Fractal into that a lot more than the Katana, even running into the same speaker and cab. Then I got a Two Notes Torpedo Captor which is a loadbox and DI, and made some measurements using Room EQ Wizard.

First off, the Torpedo Captor has a speaker impedance curve that kind of simulates a guitar speaker, although it's not really close to a real curve. There's a smooth scoop in the mids rather than a big low end spike and steeper rise in the treble frequencies. Still, it's a baseline.

52945901570_4dc8f196b4_b.jpg


Next, here's the measurement of the Boss Katana Artist. I used the Power Amp In for both 0.5 watt and 50 watt modes, plus the effects loop return for 0.5 watt mode. You can see there's no frequency response difference, just a little bit of level differences as I didn't fully normalize volumes. And in general, it looks really close to the Torpedo Captor by itself.

52945656639_edc6feb85e_b.jpg


Then, I tried all of the different Contour and Cab Resonance settings, and found they had zero impact on the tone when running through the back panel and bypassing the preamp section. Here you can see everything overlaid and it's literally the same line (again slight volume difference between power amp input and effects loop return).

52945976408_63a9ac75fb_b.jpg


Everything's looking pretty flat so far. Now let's look at the Blackstar HT-5, which is a five watt amp running a single 12AX7 in the preamp and 12BH7 in the power amp. There's the standard 5 watt mode and a reduced volume 0.5 watt mode.

52945656634_eac52f9d4a_b.jpg


That's wildly different than the Katana! We'll compare those in just a minute, but one interesting thing here is that, aside from the ~ 10 dB in volume difference, there's less high end in the 0.5 watt mode, with a lower cutoff frequency and steeper slope. Anyways, here's the Blackstar in 5 watt mode compared to the Katana:

52945656614_e70ce5002a_b.jpg


Huge difference in frequency response. The Katana, considering the Captor's impedance curve, looks really flat. The Blackstar has a high pass filter on the lows, a low pass filter on the highs, and a more pronounced midrange scoop.

So I switched to running the test signal through the Axe FX 3, put in a parametric EQ curve, and tweaked it for a couple minutes until it looked pretty close.

Low cut at 80 Hz (slope = 24 db/oct)
High shelf +4 dB at 2.0 kHz (Q = 0.7)
High cut at 7.0 kHz (slope = 6 db/oct)

52945656609_57455c1209_b.jpg


At this point I compared the Blackstar against the Katana and still preferred the tube amp. I tweaked the EQ by ear and that got me to a better place, with a little more low end and a little less highs.

Low cut at 75 Hz (slope = 18 db/oct)
Low shelf +1 dB at 100 Hz (Q = 0.7)
High shelf +1 dB at 2.5 kHz (Q = 0.7)
High cut at 7.5 kHz (slope at 12 db/oct)

52945656594_081decbda5_b.jpg


What's interesting to me is that my EQ settings to match the solid state amp to the tube amp are pretty damn close to what you would normally use on an impulse response. 80 Hz and 8 kHz are extremely common starting points.

Anyways, this was a really interesting and fun experiment. I wish I had more amps that I could measure, but I already spent $500 just to satisfy my curiosity.

Also, I've got a Kemper Profiler being delivered tomorrow and I've got no plans to get rid of my HX Stomp and Axe FX 3. Any requests for the next episode?

-----

Measurement details for the curious:

I used a MOTU M4 audio interface for all measurements. Ran the test signal from the MOTU line output, and the return signal into the line input. With the Torpedo Captor, I kept the line out level maxed as I was running it passively. I did have to compensate the output levels and/or input gain, but tried to match the -12 dB return level so no clipping at the interface. Also, I measured the interface and it's basically ruler flat, so no worries there. To measure the Captor's frequency response, I had to really boost the line out level and boost the input gain, but it looked pretty clean without a lot of noise or distortion.
 
Last edited:
Cool! :beer

Kind of confirms what I already have come to believe just using my ears all of these years.

Seems to emphasize that being smart with Hi and Lo Cuts is important with guitar tones, period.
Whether it is done via amp and speaker, or at FOH, or in the modeler is maybe not as relevant as
the fact that it get done somewhere along the chain.

I think this may be due to keeping that power from an amp from being dumped into frequencies not designed
for guitars, and that are our ears are not used to hearing over the past 50 to 70 years of electric guitar music.
I don't want flat. I want those curves. Let the cymbals handle 8kHz+ and the Bass and Kick handle anything
below 80-100Hz. Just makes everything sound better. And I believe it helps to make the guitar more focused
and... . well, punchy.... and keeps the top end from being too harsh and brittle.
 
Cool! :beer

Kind of confirms what I already have come to believe just using my ears all of these years.

Seems to emphasize that being smart with Hi and Lo Cuts is important with guitar tones, period.
Whether it is done via amp and speaker, or at FOH, or in the modeler is maybe not as relevant as
the fact that it get done somewhere along the chain.

I think this may be due to keeping that power from an amp from being dumped into frequencies not designed
for guitars, and that are our ears are not used to hearing over the past 50 to 70 years of electric guitar music.
I don't want flat. I want those curves. Let the cymbals handle 8kHz+ and the Bass and Kick handle anything
below 80-100Hz. Just makes everything sound better. And I believe it helps to make the guitar more focused
and... . well, punchy.... and keeps the top end from being too harsh and brittle.

Yes!

I don't know enough about tube amp design, but I'm guessing that at least the low cut filter was by design, and possibly the high cut filter as well. Possibly to maximize the power and reducing distortion by cutting low frequencies that aren't needed (and would be reinforced by the speaker cabinet anyway)? I mean, they have audiophile tube amps that can go full range pretty flat.

But either way, it's applying a much more pleasing EQ curve to the modeler than the solid state amp is. And from the previous discussion, it sure looks like the Katana amp is really flat, not artificially boosted anywhere. It just doesn't look (or sound) anything like the tube amp.
 
This is PROFESSIONAL COMPARISON!

Dig it man, confirms what I’ve heard all these years is that I prefer tube amps to most SS amps for most applications including modeling.

Best my Axe II ever sounded was running thru the loop of my 6505; makes me wish I would have kept it to try it thru my Herbert just to see.

Thanks for the comparison!
 
I didn't do all the fancy graphs but came to the same conclusion.

4cm isn't really even needed if you have tube power section.

SS can sound really good but you gotta tweak the deeper parameters much more and add more EQs whereas tube just sounds amazing basically on the stock settings.
 
I’m marching with tubepowersections for a while now.
To my ears SS is fine at low volume, but as soon as you raise the volume to even moderate club gig levels…they start to destroy sound.

Using different poweramps…don’t be afraid to use a global EQ. There‘s no value in judging a powersection how it works without compensation for its specifics (or specifics of your presets.) One of my fav powersections bites your head off without a high roll off, but once you done that it’s great!
 
My stomp sounds magnificent through the power amp of my JVM. The Revv gen red is soooo good run this way at gig volume.
 
Over the last few months, I've been experimenting with running my Fractal Axe FX 3 through a power amp and speaker instead of studio monitors or headphones. I tried a few different solid state amps with flat power amps, and found that while I liked the sound in some ways more than monitors, it wasn't great.

So I tried a few different amps and speakers. I started with the Line 6 Catalyst which sounded good but a bit muddy. Then a Boss Katana 50 which was noticeably brighter. Then a Celestion V-Type which improved the Boss a bit but not the Line 6. Then a Boss Katana Artist Mk 2 which was much bigger and more balanced sounding.

Still, it wasn't blowing me away. The sound was too bright from what I would expect when running cab sims off, and turning them on just meant I was running a guitar speaker emulation into a guitar speaker, which made no sense.

So I ended up buying a Blackstar HT-5 Mk 2 tube amp that featured an effects loop, and I really liked the sound of the Fractal into that a lot more than the Katana, even running into the same speaker and cab. Then I got a Two Notes Torpedo Captor which is a loadbox and DI, and made some measurements using Room EQ Wizard.

First off, the Torpedo Captor has a speaker impedance curve that kind of simulates a guitar speaker, although it's not really close to a real curve. There's a smooth scoop in the mids rather than a big low end spike and steeper rise in the treble frequencies. Still, it's a baseline.

52945901570_4dc8f196b4_b.jpg


Next, here's the measurement of the Boss Katana Artist. I used the Power Amp In for both 0.5 watt and 50 watt modes, plus the effects loop return for 0.5 watt mode. You can see there's no frequency response difference, just a little bit of level differences as I didn't fully normalize volumes. And in general, it looks really close to the Torpedo Captor by itself.

52945656639_edc6feb85e_b.jpg


Then, I tried all of the different Contour and Cab Resonance settings, and found they had zero impact on the tone when running through the back panel and bypassing the preamp section. Here you can see everything overlaid and it's literally the same line (again slight volume difference between power amp input and effects loop return).

52945976408_63a9ac75fb_b.jpg


Everything's looking pretty flat so far. Now let's look at the Blackstar HT-5, which is a five watt amp running a single 12AX7 in the preamp and 12BH7 in the power amp. There's the standard 5 watt mode and a reduced volume 0.5 watt mode.

52945656634_eac52f9d4a_b.jpg


That's wildly different than the Katana! We'll compare those in just a minute, but one interesting thing here is that, aside from the ~ 10 dB in volume difference, there's less high end in the 0.5 watt mode, with a lower cutoff frequency and steeper slope. Anyways, here's the Blackstar in 5 watt mode compared to the Katana:

52945656614_e70ce5002a_b.jpg


Huge difference in frequency response. The Katana, considering the Captor's impedance curve, looks really flat. The Blackstar has a high pass filter on the lows, a low pass filter on the highs, and a more pronounced midrange scoop.

So I switched to running the test signal through the Axe FX 3, put in a parametric EQ curve, and tweaked it for a couple minutes until it looked pretty close.

Low cut at 80 Hz (slope = 24 db/oct)
High shelf +4 dB at 2.0 kHz (Q = 0.7)
High cut at 7.0 kHz (slope = 6 db/oct)

52945656609_57455c1209_b.jpg


At this point I compared the Blackstar against the Katana and still preferred the tube amp. I tweaked the EQ by ear and that got me to a better place, with a little more low end and a little less highs.

Low cut at 75 Hz (slope = 18 db/oct)
Low shelf +1 dB at 100 Hz (Q = 0.7)
High shelf +1 dB at 2.5 kHz (Q = 0.7)
High cut at 7.5 kHz (slope at 12 db/oct)

52945656594_081decbda5_b.jpg


What's interesting to me is that my EQ settings to match the solid state amp to the tube amp are pretty damn close to what you would normally use on an impulse response. 80 Hz and 8 kHz are extremely common starting points.

Anyways, this was a really interesting and fun experiment. I wish I had more amps that I could measure, but I already spent $500 just to satisfy my curiosity.

Also, I've got a Kemper Profiler being delivered tomorrow and I've got no plans to get rid of my HX Stomp and Axe FX 3. Any requests for the next episode?

-----

Measurement details for the curious:

I used a MOTU M4 audio interface for all measurements. Ran the test signal from the MOTU line output, and the return signal into the line input. With the Torpedo Captor, I kept the line out level maxed as I was running it passively. I did have to compensate the output levels and/or input gain, but tried to match the -12 dB return level so no clipping at the interface. Also, I measured the interface and it's basically ruler flat, so no worries there. To measure the Captor's frequency response, I had to really boost the line out level and boost the input gain, but it looked pretty clean without a lot of noise or distortion.

When it comes to using my Helix LT through the return of an amp, I have used the following amps generally with a horizontal half open back 2x12 of English V30's.

Mesa DR Tremoverb head
Marshall DSL100H
Marshall DSL40c combo
Line 6 Spidervalve combo or HD100 head
Mesa Mark V 90 combo
Mesa DC-3 head
Jet City JCA100 head
Peavey Bandit Red Stripe combo
Peavey Valveking II 100 head

and a few others that instantly sucked so I didn't spend much time dialing.

I don't have any hard empirical evidence to prove anything but the very best I have experienced so far is the Line 6 Spider Valve HD100.
Not surprised because the tube power section is designed to work well with amp and fx modelling. Seriously it's really not even close what I prefer. The Spider Valve brings out the best in every modeler I have plugged in to it. Sounds and feels the most natural. I have a small collection of Spider Valves because they can be had so cheap on the used market and I want to make sure that if one dies I have backups.
 
When it comes to using my Helix LT through the return of an amp, I have used the following amps generally with a horizontal half open back 2x12 of English V30's.

Mesa DR Tremoverb head
Marshall DSL100H
Marshall DSL40c combo
Line 6 Spidervalve combo or HD100 head
Mesa Mark V 90 combo
Mesa DC-3 head
Jet City JCA100 head
Peavey Bandit Red Stripe combo
Peavey Valveking II 100 head

and a few others that instantly sucked so I didn't spend much time dialing.

I don't have any hard empirical evidence to prove anything but the very best I have experienced so far is the Line 6 Spider Valve HD100.
Not surprised because the tube power section is designed to work well with amp and fx modelling. Seriously it's really not even close what I prefer. The Spider Valve brings out the best in every modeler I have plugged in to it. Sounds and feels the most natural. I have a small collection of Spider Valves because they can be had so cheap on the used market and I want to make sure that if one dies I have backups.

Interesting!

Have you tried any solid state amps or power amps?

That Spider Valve does look really neat. 40 watt tube power amp and a Celestion V30, often for $300 or less used! I'll have to keep my eyes out.
 
Interesting!

Have you tried any solid state amps or power amps?

That Spider Valve does look really neat. 40 watt tube power amp and a Celestion V30, often for $300 or less used! I'll have to keep my eyes out.

Yeah the Bandit is SS and a couple others I didn't mention were a Samson SS power amp and little Joyo Meteor which is technically a SS amp with a tube pre. I actually didn't mind the Bandit all that much but yeah stiff and and a bit sterile comparatively.

Those Spider Valves are a great value for what you get. A Bogner designed power amp and if you get a combo, like you said stock V30's come in them.
 
Yeah the Bandit is SS and a couple others I didn't mention were a Samson SS power amp and little Joyo Meteor which is technically a SS amp with a tube pre. I actually didn't mind the Bandit all that much but yeah stiff and and a bit sterile comparatively.

Those Spider Valves are a great value for what you get. A Bogner designed power amp and if you get a combo, like you said stock V30's come in them.
The Spider in the name was a grave mistake. Zero chance of getting taken seriously.
 
Interesting!

Have you tried any solid state amps or power amps?

That Spider Valve does look really neat. 40 watt tube power amp and a Celestion V30, often for $300 or less used! I'll have to keep my eyes out.
If your shopping for cheap but great powersections…also check Bogner alchemists…people can’t even give those away on the used market…weird…cause those are great sounding amps.
 
Unless you have access to the schematics of each design, how can you be sure you are comparing tube vs solid state design or simply 2 different circuits?
 
The Spider in the name was a grave mistake. Zero chance of getting taken seriously.

Yeah I agree. Especially at a time when the Spider name was at an all time low. Warranted or not. Me, I witnessed an AC/DC tribute with both guitarist's using Spider III combos sound killer. I mean they were miced and going through a great sound system but I was still pretty surprised at what was coming out of the PA
 
Has anyone tried a dt25 or 50 as a power amp for a modeller? I keep seeing them going for peanuts on ebay.

I did back when I first got my LT. A co-guitarist in the band I am in let me try out his DT25 combo. It was just as good as the Spider Valve easily. The Park 75 model in those was fantastic. I think you can also control the DT series with a HX product if I am not mistaken. Something you cannot do without MIDI with the Spider Valves
 
I did back when I first got my LT. A co-guitarist in the band I am in let me try out his DT25 combo. It was just as good as the Spider Valve easily. The Park 75 model in those was fantastic. I think you can also control the DT series with a HX product if I am not mistaken. Something you cannot do without MIDI with the Spider Valves

Hmm there's a 1x12 combo near me for £300
 
Back
Top