RMS output differences in all-tube amps

[SactoBass]

I did a search and couldn't find exactly what I am looking for.

This thread is only about bass amps with tube power sections, and this thread is focusing on just the power output section.

I have seen amps with four KT88s that are rated at 150 watts output power (example: Shaw B150), and I have seen amps with four KT88s that are rated at 300 watts output power (example: New Vintage Undertow). What's up with that? Is it simply a difference in the amount of THD in the output rating?

In addition, I have seen amps with six 6550s that are rated at 300 watts output power (the Ampeg SVT). How can the SVT pump out 300 watts if 6550s are rated at 30 watts output each? (with six 6550s, that should add up to 180 watts, correct??) What am I missing? Is this another case of THD?

 

[okcrum]

You can get more power from a set of tubes by raising the plate voltage, but that New Vintage spec is fantasy unless they're talking about a few milliseconds of 300W.

You can get 100W from a pair of 6550s at 600+ volts, but not much more.

Plate dissipation rating is not the same as power output rating.

EDIT: RCA and GE rated the 6550A at 35 watts "design maximum" and 42 watts "absolute maximum" plate dissipation...KT88 has about the same limits, but can run slightly higher plate voltage and get a bit more power.

 

[Munjibunga]

You can get more power from a set of tubes by raising the plate voltage, but that New Vintage spec is fantasy unless they're talking about a few milliseconds of 300W.

You can get 100W from a pair of 6550s at 600+ volts, but not much more.

Plate dissipation rating is not the same as power output rating.

The Mesa Prodigy is rated at 250 watts RMS from four KT-88s.

 

[timv]

The fender 400ps does 435 watts from 6 6550's!
Make sure to use good tubes.

 

[okcrum]

The Mesa Prodigy is rated at 250 watts RMS from four KT-88s.

That's not RMS, it's "tube watts". Funny, I can't find an actual spec sheet for the Prodigy, even in the manual. From the Mesa site:

"Class A/B, Multi-Watt™ Power Amp provides two power options via a Full/Half Power Switch. Choose from 4 power tubes, producing 250 Tube-Watts* (“Sounds like in excess of 500”) or 2 tubes, producing 125 Tube-Watts*"

 

[okcrum]

The fender 400ps does 435 watts from 6 6550's!
Make sure to use good tubes.

Y'know, back in WWII, they used to get 100W out of a pair of metal 6L6s by running them upside down, in water. Tube life wasn't so good then, either.

It's getting hard to find GE-6550As and the three cabs you need to get full power.

What's the distortion at full power on these things? (It's a rhetorical question, you won't find a spec, just like Mesa). 71% efficiency (as this amp claims) is awfully close to the theoretical class B efficiency of 78.5%.

 

[beans-on-toast]

The 35w per 6550A tube rating is for class AB fixed bias ultralinear operation.

Normally the rating for a pair, push pull, fixed bias for class AB1 is 100W. But keep in mind, the data sheet specs are for a long service life. With the right plate and screen voltages, fixed bias, and cooling, the tube can be pushed beyond the published limits. Is it good to do that is another question.

With a more conservative design, the tubes will last longer. An important design consideration.

 

[teemuk]

I have seen amps with four KT88s that are rated at 150 watts output power ... and I have seen amps with four KT88s that are rated at 300 watts output power .... What's up with that? Is it simply a difference in the amount of THD in the output rating?

THD rating likely can't explain such a big difference. Perhaps it is a difference between PEAK and AVERAGE power ratings...?

Nevertheless, in terms of perceived loudness the difference is nothing but barely noticeable.

How can the SVT pump out 300 watts if 6550s are rated at 30 watts output each? ... What am I missing?

What you're missing is fundamental understanding of the difference between maximum plate dissipation and output power ratings. Let's assume there's an amp that has efficiency of 70%. It draws 100 watts of power from the power supply to produce 70 watts of power to the load. While doing this it wastes 30 watts of power due to its inefficiency to amplify. So crudely you can understand that as follows: When 70 watts of output power is produced 30 watts of power is dissipated somewhere.
So when the SVT produces 300 watts of output power it does that with certain efficiency. 50%? 60%? Maybe 70%...? Given that, what are the power losses in total that output devices must dissipate in this condition? Divide by six (per each tube) and I'm pretty sure the result is well under the 30-watt plate dissipation rating of 6550s.

 

[teemuk]

That's not RMS, it's "tube watts".

Hillarious.

250 Tube-Watts* (“Sounds like in excess of 500”)

Funniest thing in that quote is that while the difference seems huge in watts, in terms of perceived loudness the difference is practically negligible.

So... Even if non-existing thing such as "tube watts" was actually real, and there would be watts that are more powerful than other watts, the whole thing would still mean practically nothing since the difference to generic watts would be barely noticeable in terms of produced SPL.

 

[okcrum]

The 35w per 6550A tube rating is for class AB fixed bias ultralinear operation.

Normally the rating for a pair, push pull, fixed bias for class AB1 is 100W. But keep in mind, the data sheet specs are for a long service life. With the right plate and screen voltages, fixed bias, and cooling, the tube can be pushed beyond the published limits. Is it good to do that is another question.

With a more conservative design, the tubes will last longer. An important design consideration.

Are you mixing up plate dissipation and power output? The amp efficiency (roughly, neglecting OT losses) is going to be the power output divided by the power output plus the total plate dissipation. The latter is what you are actually setting when you set the bias...

EDIT: I should add that MI design is different from baseline audio design. Some nonlinear (distorted) operation is considered normal. The bandwidth is also reduced, which helps with tube bass amp power ratings, since the OT is rated for more power at 30 Hz (MI) than at 20 Hz (hifi).

 

[beans-on-toast]

Are you mixing up plate dissipation and power output? The amp efficiency (roughly, neglecting OT losses) is going to be the power output divided by the power output plus the total plate dissipation. The latter is what you are actually setting when you set the bias...

Not mixing dissipation and power output up. Efficiency is something else as you pointed out.

See this data sheet for the 6550A: http://drtube.com/datasheets/6550a-ge1972.pdf.

 

[SactoBass]

Great replies so far! Thank you everyone! It sounds like THD is only a small part of the picture, and what seems to be more significant is the amount of B+ plate voltage that is being fed to the power tubes.

So, that leads me to my next question: does a power section sound (and feel) better to the player if the power tubes are receiving higher B+ plate voltages? I ask that because, in my experience, preamps that have high plate voltages (say, around 300v being fed to an 12AX7) have always sound much better to me than starved plate preamp designs. So does that also apply to the power section of a tube amp?

 

[okcrum]

Not mixing dissipation and power output up. Efficiency is something else as you pointed out.

See this data sheet for the 6550A: http://drtube.com/datasheets/6550a-ge1972.pdf.

Yes, I have that data sheet. Svetlana also spec'ed their 6550 at 42 watts. Tung Sol and RCA (rebranded Tung Sol) spec'ed theirs at 35 watts. You see both values for some modern tubes, depending where you look.

I suppose you could get a pair of 6550s down to 35W if you ran them cathode biased and ultralinear at 400V.

 

[okcrum]

Great replies so far! Thank you everyone! It sounds like THD is only a small part of the picture, and what seems to be more significant is the amount of B+ plate voltage is being fed to the power tubes.

So, that leads me to my next question: does a power section sound (and feel) better to the player if the power tubes are receiving higher B+ plate voltages? I ask that because, in my experience, preamps that have high plate voltages (say, around 300v being fed to an 12AX7) have always sound much better to me than starved plate preamp designs. So does that also apply to the power section of a tube amp?

Not so much - B+ will affect the power more*. The interaction between the driver tubes, power tubes and output transformer is what gives some distinction to the sound.

*Here are some comparisons for a pair of 6550 from the RCA tube manual RC30:

Fixed bias
600V B+ 100W, 2.5% THD, bias 58 ma. per tube
400V B+ 55W, 3% THD, bias 90 ma. per tube

Cathode bias
400V B+ 41W, 4% THD, bias 83 ma. per tube

NOTE: bias is set around 35W per tube in each case

 

[SactoBass]

B+ will affect the power more. The interaction between the driver tubes, power tubes and output transformer is what gives some distinction to the sound.

Hmmm.....very interesting! So it sounds to me like the amount of B+ voltage being fed to the power tubes will affect the "loudness" of the amp, whereas the interaction between the driver tube(s), power tubes, and the design (and quality) of the OT will affect the "tone and feel" of the amp.

 

[okcrum]

Hmmm.....very interesting! So it sounds to me like the amount of B+ voltage being fed to the power tubes will affect the "loudness" of the amp, whereas the interaction between the driver tube(s), power tubes, and the design (and quality) of the OT will affect the "tone and feel" of the amp.

Yes, though it's important to remember the most distinctive "power amp sound" comes when overdriven, so the power supply and cabinet are also in play. The basic response of the amp, plus preamp voicing and settings are more important factors when running clean. The OT will usually affect both of these more than any other single component in the amp, since it's one of the biggest design tradeoffs.

 

[beans-on-toast]

Yes, I have that data sheet. Svetlana also spec'ed their 6550 at 42 watts. Tung Sol and RCA (rebranded Tung Sol) spec'ed theirs at 35 watts. You see both values for some modern tubes, depending where you look.

I suppose you could get a pair of 6550s down to 35W if you ran them cathode biased and ultralinear at 400V.

You are right, there are many different ways to operate a tube.

Ultra linear operation results in less power output in order to gain less distortion. Always tradeoffs. I find that 6550 tubes in a vintage V4B have a very very long service life. The advantage is that under the normal voltages, the amp even puts out more power than with the standard 7o27A's. See page 2 of this V4 brochure, options: https://www.talkbass.com/attachments/ampeg-brochure-v4-pdf.741460/. They specify 130W @ 3% THD. In this amp, the 6550's are plug and play. I'd want to optimize the bias. More power, longer tube life are win-win.

I find that the 42W plate dissipation in this data sheet can be exceeded if you want to push the tube to produce more output power wattage. In the end, how much the tube's plate can dissipate helps define how much output power the tube can put out. As I said above, the specs are for a long service life. Fender pushed tubes beyond their specs in some of their designs. Tubes were a lot less expensive back then so a shorter life was not such a big deal.

 

[CL400Peavey]

The Mesa Prodigy is rated at 250 watts RMS from four KT-88s.

At 7.5% THD, where as Ampeg measures at 3% for the SVT. So the measuring sticks are a little different.

 

[SactoBass]

Ampeg measures at 3% for the SVT.

So does an SVT really put out 300 watts RMS power (at 3% THD) with only six 6550 power tubes?? That would be 50 watts RMS per 6550. That seems high to me.

 

[CL400Peavey]

So does an SVT really put out 300 watts RMS power (at 3% THD) with only six 6550 power tubes?? That would be 50 watts RMS per 6550. That seems high to me.

That is their published spec, and the amp has been around a long time with little discussion about it being under powered. My amp is rated at 420 watts with 8 6550's, so I have no reason to doubt ampeg.