mythbusting

Why are valve amps louder than solid state amps?

by
Are valve amps louder than solid state amps?

Actually, let’s start with a slightly different question…

Are valve amps louder than solid state amps?

Yes, valve amps often are louder than solid state amps. As an example, we could take a Marshall MG30 (30W) and a Vox AC30 (30W) and compare the two. There’s no question at all that the Vox will sound significantly louder to any objective listener. Or take a Fender Blues Junior (15W) and a Fender Frontman 15 (15W). The perceived volume difference is huge.

But hang on, given the price differences in those examples we’re not really comparing apples with apples are we?

How much louder are valve amps than solid state amps? Let’s talk about perceived loudness

It’s often said that a valve amp sounds “twice as loud” as a solid state amp. There’s a whole field of psychoacoustic study to delve into here. For the purposes of this discussion, let’s focus on one psycho acoustic rule of thumb (source).

a power increase of 10dB results in double the perceived volume. 

So if we look at the electronic (and electro acoustic) design of valve amps and solid state amps and we can account for 10dB of difference then we can begin to understand why some valve amps are louder than some solid state amps.

Why are valve amps louder than solid state amps?

1. Player demographic.

  • This is the biggie.
  • Transistor technology is more affordable than the thermionic valve. There’s no doubting that. 
  • So the cheaper amps are often solid state, the more expensive are often valve.
  • Cheaper amps are more likely to be sold to ‘bedroom’ players and children. Someone who’s gigging is more likely to spend more and be able to afford a valve amp.
  • The amp development engineers know this, so they design ‘bedroom’ amps to give sensible volume control in a bedroom range. The user can turn the volume dial to about 12 o’clock before the parents complain. They design ‘gigging’ amps to give sensible rehearsal volumes from about 9 o’clock and full on ‘fight the drummer’ volumes at 12 o’clock. This is just sensible design.
  • If we were to give this concept an ‘electronicsy’ name, we’d call it the amplifier sensitivity. It could be measured by turning all controls to a certain position (eg. 12 o clock) and determining how much signal we need to put into the input jack in order to get 1W output.

How much difference does this make? As very few manufacturers list AC measurements on their schematics it’s very hard to put a dB figure on this without experimentation, so I’m not going to make one up! I’ll be measuring amps as they come it to try and put a figure on it!

2. Speaker sensitivity (efficiency) – for more info

  • See point 1, solid state amps are often used at the price conscious end of the market
  • Speakers have a sensitivity, measured in dB/Watt
  • Let’s look at some real speakers, chosen at random from Celestion’s range:
    • Celestion Vintage 30 – used in mid and high end gear. 100dB/W
    • Celestion 70/80 – used in low mid range amps. £500ish. 98dB/W
    • Celestion Rocket 50 – used in amps around the £200-350 mark. 95dB/W
    • You can’t get manufacturer efficiency data for the guitar speakers used in even cheaper amps, but you can be sure that most will be poorer. Some specs are available for cheap 12in PA woofers such as the MCM 55-2973 at 91dB, roughly 8-10dB lower than comparable high end PA woofers by Fane, Eminence etc.

Already we can see that an amp with a high end speaker has the potential to be 10dB louder than a cheap amp with a low end speaker.

Cheap speakers have low sensitivity
Cheap speakers have low sensitivity

So what happens if we use the same speaker? Read on…

3. Output load rating and power amplifier current/voltage control

  • Valve amps use an output transformer. Although its primary purpose is just to match the high impedance valve output with the low impedance speakers, there’s a side benefit – many transformers have tappings to allow a valve amp to match a 4 ohm, 8 ohm or 16 ohm speaker. This means that the amp delivers roughly the same power into any load.
  • Very very few solid state amps have an output transformer because they’re not needed – it’s easy to create a low impedance output with transistors.
  • So this leads to a scenario in which a solid state amp outputting 15.5Vrms will deliver roughly 60W into 4 ohms, 30W into 8 ohms and 15W into 16 ohms. 
  • When the engineering department tell the marketing department this, guess which figure the marketing team put on the front of the amp?
  • A real world example:
    • a Marshall MG100 solid state amp:
      • 20Vrms = 100W into a 4 ohm speaker,
      • 20Vrms = 50W into an 8 ohm speaker,
      • 20Vrms = 25W into a 16 ohm speaker.
    • Meanwhile a Marshall DSL100:
      • 20Vrms on 4 ohm transformer tap = 100W into a 4 ohm speaker
      • 28Vrms on 8 ohm socket = 100W into an 8 ohm speaker,
      • 40Vrms on 16 ohm socket = 100W into a 16 ohm speaker.
  • It is possible to build a voltage drive (with a switch) OR current driven  solid state amp that will deliver the same power into different loads. But guess what? It’s more expensive due to the higher voltage rails required, see point 1.
  • Depending on the scenario, this can result in between 3dB and 6dB difference in power output between two amps when they’re plugged into the same speaker UNLESS the speaker is 4 ohm

4. Output impedance

  • Most solid state amps are voltage controlled (certain era Roland JC, Marshall 80XX series, Fender ultimate chorus and some Trace gear excepted). voltage control results in a low output impedance
  • Valve amps with no negative feedback around the transformer have a large amount of current control
  • Rod Elliot of ESP (https://sound-au.com/valves/valve-trans.html#s21 ) does a worked example that shows that these differences

As for how much difference this makes, it’s hard to say for certain. Time to experiment, methinks!

5. Valve power amps are designed to distort

  • A distorted signal has a higher RMS (up to 1.414 times higher) than a clean waveform. Valve circuits are often more distorted than solid state, even on the ‘clean channel’. However FETs and diodes can be used to distort solid state amplifiers.
  • But solid state power amps are often not designed to distort. Why, is it the tone? Well it is true that some solid state amps distort in an unpleasant manner, but many distort quite pleasantly!
  • In truth, the primary reason goes back to point one again. Distortion increases the RMS of the voltage signal, which increases the power that the solid state devices must be designed to dissipate, so more money must be spent on heatsinking and cooling them and more money must be spent on providing a power supply that can supply extra the power because…
  • Simplifying the maths here, if a clean signal passing into a power amp is fully distorted into a square wave, it will increase the RMS voltage (or current) by about 1.414x. 
  • So cheap solid state amps (and some cheap valve amps) will normally allow the power to sag significantly when distorting the output stages.  

Best case scenario: The described increase in voltage equates to 3dB more power. However, you’ll only actually hear this level difference when the amps are dimed and the power amps are distorting and when comparing a fully distorted valve amp to a totally clean solid state amp. So it’s not a reasonable comparison. 

6. Distortion adds harmonics

    • This is the first one that’s actually related to technological differences between valves and solid state.
    • A distorted signal has a higher RMS than a clean waveform. Valve circuits are often more distorted than solid state, even on the ‘clean channel’. However FETs and diodes can be used to distort solid state amplifiers.
    • It’s also true that the distortion adds extra frequency content in the most sensitive part of the human ear.

Conclusion

At the start of this article we discussed the claim that valve amps sound twice as loud as solid state amps. We equated this to a 10dB difference in sound level.

  • Volume control sensitivity: To be Confirmed by experiment
  • Speaker sensitivity: up to 10dB difference
  • Output load rating: 0-6dB difference, depending upon the cabinet impedance
  • Output impedance: unknown
  • Distortion: up to 3dB difference, but only when the amp is dimed.
  • Harmonics: unknown.

With various combinations of the factors above, we can easily account for 10dB difference in power and thus a perceived doubling of volume, but what we’ve really seen is that thing that results in valve amps sounding louder than solid state is the target cost price of the unit as it’s designed. If the engineer designing a solid state amp for a pro market were given the same target cost price as the engineer designing a valve amp for a pro market, then a solid state could easily be designed with:

  • a sensitive volume control,
  • expensive 100dB/W or greater speaker,
  • 4x voltage headroom to accommodate different speaker loads
  • Current drive to cancel output impedance effects
  • Multiple power devices and sufficient heatsinking and a large power supply to allow substantial power amp distortion

And we’d end up with a solid state amp that was as loud as a valve amp. Let’s do this!

Orange Tiny Terror Modification – BS watch! – Ampworks, Newark

by

Below is an extract from an email conversation with a customer who asked for some impartial advice on a popular ‘mod’. I’ve redacted names, images and product links. I’ve no wish to slander anyone or their products, but I do hope to bring some clarity in a field awash with mis-information. All opinions are my own, if you disagree with something, please do let me know, preferably in a civil manner! Please view this a general piece of advice regarding amp modification, rather than a specific commentary upon this product.

BS stands for ‘bad science’. If you thought otherwise then my apologies, OK? 🙂

Hi Mike,I’ve just commented on the various claims, referencing these photos. There’s a few interesting potential benefits, […] I’ve gone into quite a bit of detail, so I might pop this on my blog so others can find it. 🙂

My modification rule of thumb is that anything that makes less difference than a new set of guitar strings is pointless. I’d say there are two component changes here that may pass that criteria. Firstly, the bias mod will make a noticeable difference, that you may or may not like. If the designer used C7 to shape the amp’s tone, which is likely from the component position, then removing C7 will bring the biggest tonal change to the whole amp. The boutique components are probably BS. 🙁

[image removed, the following italic text is pasted from the seller’s description, my comments are in colour]

CLAIM: Swapping the valves for a full set of new JJ Electronic ECC83S and EL84

AMPWORKS COMMENT: New valves can of course make a difference, you can try this yourself by buying new valves from a reputable supplier (such as me!) Valve changes generally make less difference than a new set of strings. 🙂

 

CLAIM: The cheap push-on heater connectors were dropping too much voltage between the transformer heater wiring and the V3 heater pins.

COMMENT: Probably BS. It’s feasible that the connector contact resistance may drop voltage and hard wired connections are more reliable, but the sensible approach is to replace only if a valve heater actually stops working (maybe after contact oxidation from storage in a damp garage). This will have no effect upon sound.

 

CLAIM: The rubbish heater connectors are removed and the valve heater circuit wiring is completely hard wired. Wire dress was particularly poor.

COMMENT: This means that the heater connections (yellow and black in the photo) weren’t tightly twisted at the factory. Heaters can be a source of hum, twisting the wires reduces the magnetic field created by the high current in the wires. If you have a problem with hum (with no guitar plugged in) then this is useful. If no hum problem, no point!

 

CLAIM: Bias was set too high. Output tube bias is corrected to produce maximum power output.

COMMENT: OK, this is not BS. The tiny terror is cathode biased (meaning it self biases (see my blog page on bias here). But the designers have chosen to set this self bias point quite hot. Because valves are non linear this does make a difference to the sound the amp produces. Whether it’s a change you’ll like is another question. The output stage will be louder and cleaner. You’ll hear the difference more if you play it pretty loud.

 

CLAIM: Preamp voltages were OK in 15W mode, but a bit on the low side in 7W mode.

COMMENT: Probably BS. This means nothing without context. There’s no such thing as an ‘OK’ or a ‘bad’ preamp voltage, unless the valve is pushed outside of it’s performance limits.

 

CLAIM: Components were ordinary commercial grade, nothing special. Critical components in the signal path are upgraded to boutique grade items, including:

  • Mallory 150M Series capacitors, silver mica capacitors

  •  COMMENT: Someday I’ll do a proper scientific comparison of the ’boutique’ cap brands, I’ll try and remember to let you know, I personally  suspect that any change will be very VERY subtle. 🙂 What’s MUCH more interesting is that on his photographs he’s actually removed C7 completely. This is likely to have a MUCH MUCH bigger effect than any brand change. C7 is between the guitar input and the first valve. It looks like that cap was originally connected in series so removing it will change the low end response (assuming the frequency corner was previously set to over 80ish Hz).

  • carbon composition resistors, for a more authentic vintage valve sound.

  • COMMENT: This is an interesting topic investigated quite thoroughly by RG Keen here. . His conclusions show that there may be some extra distortion created by these resistors that guitarists like. However they have to be used properly, not just replaced as a matter of course. RG recommends that they be replaced every few years as they may become unreliable when used in this way. Carbon comp resistors are generally noisier (hissy) than metal film equivalents.