Scaled-Down Bluesbreaker

[hawkes1962]

Hi all,

I’m wanting to build a scaled-down Bluesbreaker, hitting around 18-20 watts using KT66 tubes.

Now, I understand some will have the knee-jerk reaction to say, “Why would you waste the potential of big-bottle tubes?!” And if you feel you need to say it…I guess go for it!

The recent release of the Marshall studio series is what has inspired me, which make roughly 20 watts with a pair of cathode-biased EL34s. Though I haven’t had the opportunity to play one in-person, the SV20 has sounded pretty dang good in every demo I’ve heard. My conclusion: one can get a great sound from bigger bottles at lower voltages.

I have an Edcor PT (from a project that never quite materialized) that I would like to use for this project. The specs on the secondary are as follows:

255-0-255 @ 125mA
3.15-0-3.15 @ 4A
(These specs are with the transformer fully loaded. Unloaded, they will be about 5% higher.)

The transformer should have enough heater current, and I’ll be using solid state rectification. I believe B+ will be around 350-355V max.

Can someone help me determine the operating conditions the KT66s need to be at to deliver 18-20 watts, fixed bias? I’d like to maintain a similar tone, if possible, to the earlier Bluesbreaker which used the 6600Ω primary on the output transformer. I know the output transformer primary impedance can effect the distortion characteristics as well as the output.

[66Strat]

With 255 volts AC, diode rectification will result in 360 volts DC under a 125 ma. load.

Push pull KT66 at 360 VDC plate voltage, 300V DC screen voltage, -24 VDC bias voltage will produce about 12 watts with a 6K6 OPT and 14 watts with an 8K OPT. The amp operates in mostly in Class A with either transformer. The harmonics of the KT66 with the 6K6 OPT are predominantly second order harmonics, that are cancelled out in PP operation. The 8K OPT will generate more third harmonics, which are not cancelled out. Whichever you choose, the power output will be constrained by the power transformer.

Look at the model in this link. You can simulate operating conditions and see their effect on performance.

https://www.vtadiy.com/loadline-calculators/loadline-calculator/

[66Strat]

BTW, I like your idea. It sounds like a fun project.

[Willabe]

I think it's a great idea!     

And since the power tubes will be loafing, they should last a LONG time! 

I've been very seriously thinking about either building a K bias dual channel Plexi with EL34/KT77/KT66's or the same in a SE format. They'd both be around the same power, ~12 to 15w for SE and 66Strat you give examples of ~12w to 14w in PP. Just want some low power crunch for at home.

The harmonics of the KT66 with the 6K6 OPT are predominantly second order harmonics, that are cancelled out in PP operation. The 8K OPT will generate more third harmonics, which are not cancelled out.

That's the thing, that could make a lot of difference in the amps tone. But isn't a lot of the Marshall crunch in the 3rd order harmonics? And getting the right balance of some but not too much 3rd order harmonics the key? 

Ken Fisher, Trainwreak amps, said that the OT primaries R is adding the 'spice' in an amp, that would be the amount of 3rd order harmonics generated in the power amp stage.

[66Strat]

I think it's a great idea!     

And since the power tubes will be loafing, they should last a LONG time! 

I've been very seriously thinking about either building a K bias dual channel Plexi with EL34/KT77/KT66's or the same in a SE format. They'd both be around the same power, ~12 to 15w for SE and 66Strat you give examples of ~12w to 14w in PP. Just want some low power crunch for at home.

The harmonics of the KT66 with the 6K6 OPT are predominantly second order harmonics, that are cancelled out in PP operation. The 8K OPT will generate more third harmonics, which are not cancelled out.

That's the thing, that could make a lot of difference in the amps tone. But isn't a lot of the Marshall crunch in the 3rd order harmonics? And getting the right balance of some but not too much 3rd order harmonics the key? 

Ken Fisher, Trainwreak amps, said that the OT primaries R is adding the 'spice' in an amp, that would be the amount of 3rd order harmonics generated in the power amp stage.

There is a difference on paper, but whether or not it's an audible difference is up for debate. The Bluesbreaker amps (Clapton Beano era) had Radiospares OPTs that featured multiple wiring connections for the primary windings. Some amps came from the factory with 6K6 primary connections. Others were connected as 8K. There was at one time a chart floating around the web that cataloged a number of amps as to production date and output transformer wiring configuration. I believe that it was the guys that started Obsolete Electronics. They made clones of the Radiospares transformers used in the Bluesbreaker amps.

[hawkes1962]

And since the power tubes will be loafing, they should last a LONG time! 

Yes, that’s one of my hopes!!

Look at the model in this link. You can simulate operating conditions and see their effect on performance.

https://www.vtadiy.com/loadline-calculators/loadline-calculator/

I found that site the other day, but wasn’t quite sure how to interpret the results I was getting. They show output power, but it seems too high. With 360V plate, 300V screen, 8K primary, I get 27.6 watts.

Using the Valve Wizard’s output power “estimation equation” found at the bottom of the page here, www.valvewizard.co.uk/pp.html I get pretty much the same thing. Using the generated loadline from vtadiy.com with the same inputs listed above:
(360-27)*(0.162/2) = 27 watts

So where am I going wrong?

[PRR]

Remember the 6L6 (real KT66) datasheet has many suggestions and some are near your zone. (You want to avoid the AB2 conditions.)

6.6k load with 395V on plates and cathode bias made right around 23W clean on one project. This worked with metal 6L6 (a bit past rating), 7027, antique 6550, new commie 6550... tube is tube. IIRC I had shared 250Ω cathode bias and shared 2k(?) screen dropper. A skilled player took it from me to his studio and it never came back. Yes part of my goal was loooooong tube life. (This was a rebuild of an amp that annoyed me by burning-up.)

I've never been sure of that VTADIY calculator. It is certainly possible to provoke it into bogus numbers.

[66Strat]

And since the power tubes will be loafing, they should last a LONG time! 

Yes, that’s one of my hopes!!

Look at the model in this link. You can simulate operating conditions and see their effect on performance.

https://www.vtadiy.com/loadline-calculators/loadline-calculator/

I found that site the other day, but wasn’t quite sure how to interpret the results I was getting. They show output power, but it seems too high. With 360V plate, 300V screen, 8K primary, I get 27.6 watts.

Using the Valve Wizard’s output power “estimation equation” found at the bottom of the page here, www.valvewizard.co.uk/pp.html I get pretty much the same thing. Using the generated loadline from vtadiy.com with the same inputs listed above:
(360-27)*(0.162/2) = 27 watts

So where am I going wrong?

I may have been too conservative when I ran the model. Initially, I limited quiescent idle current to 60 ma. I was under the impression that the model was basing calculations upon the quiescent idle current for each tube. I was trying to stay within the power transformer current limit of 125 ma (2 x 60 ma), and also trying to keep the load line from passing through the plate dissipation limit curve.

I've played with the model some more. By increasing the idle current to 75ma., output power increases to:

• 6K6 31.9 Watts Class AB1, 18.6 Watts Class A
• 8K - 27.6 Watts Class AB1, 22.5 Watts Class A

[66Strat]

Remember the 6L6 (real KT66) datasheet has many suggestions and some are near your zone. (You want to avoid the AB2 conditions.)

6.6k load with 395V on plates and cathode bias made right around 23W clean on one project. This worked with metal 6L6 (a bit past rating), 7027, antique 6550, new commie 6550... tube is tube. IIRC I had shared 250Ω cathode bias and shared 2k(?) screen dropper. A skilled player took it from me to his studio and it never came back. Yes part of my goal was loooooong tube life. (This was a rebuild of an amp that annoyed me by burning-up.)

I've never been sure of that VTADIY calculator. It is certainly possible to provoke it into bogus numbers.

IMO, there's not a lot to be gained from operating tubes at the maximum limits, except shorter tube life. How many here drive their cars with the tach pegged at the rev limit?

[Willabe]

There is a difference on paper, but whether or not it's an audible difference is up for debate.

Now your talking about the difference between 6K6 load and 8K load?

And do power tubes only start to generate 3rd order harmonics when the amp crosses over into class AB1?

If so then there would be very little or no difference in sound between a SE amp and a PP true class A amp with the same power tubes, same preamp?     

[66Strat]

There is a difference on paper, but whether or not it's an audible difference is up for debate.

Now your talking about the difference between 6K6 load and 8K load?

Not any sound difference between PP's 3rd harmonics and a SE's 2nd harmonics with the same power tubes?   

I'm talking PP.

In SE operation, the 2nd harmonics do not cancel. In SE operation, 2nd harmonics and 3rd harmonics will modulate to create higher order harmonics. In PP operation, the 2nd harmonics cancel, leaving predominantly 3rd harmonics. Attached is a chart from the Ken Rad 6L6 Datasheet for use as an example. This chart is for a 6L6 in SE operation with both plate and screen supply voltage of 250 VDC. The distortion curves will shift to the right with higher operating voltages.

[shooter]

How many here drive their cars with the tach pegged at the rev limit?

I had 2 rally cars, they LIKED it that way 

[pdf64]

…
Look at the model in this link. You can simulate operating conditions and see their effect on performance.

https://www.vtadiy.com/loadline-calculators/loadline-calculator/

I don’t find that intuitive to use, eg what are the fields
‘Next stage AC Impedance (Ohm):      HD%’
and
‘Out. headroom (+/-V):’
for?

Whereas the black magic amps one is intuitive and seems to work well http://bmamps.com/ivds.html

[66Strat]

…
Look at the model in this link. You can simulate operating conditions and see their effect on performance.

https://www.vtadiy.com/loadline-calculators/loadline-calculator/

I don’t find that intuitive to use, eg what are the fields
‘Next stage AC Impedance (Ohm):      HD%’
and
‘Out. headroom (+/-V):’
for?

Whereas the black magic amps one is intuitive and seems to work well http://bmamps.com/ivds.html

I have not had much experience with the model. I've only played with it a few times. The model does have fields that are not used in every situation. The next stage AC impedance is used in modeling pre amps and is not applicable for power amps. The Out. (output) Headroom appears to allow one to input the estimated plate voltage swing of the tube(s) to model harmonic distortion percentages. At least that's the way it seems to work. Some basic instructions for input fields would certainly be useful.

I think the best way to use any of these models would be to start out with published operating points and compare the modeled results with the published data to establish the level of accuracy in the modeled results. Desired operating parameters could then be modeled and evaluated with some expectation as to accuracy.

[jordan86]

Following this thread because I have looked at the Marshall studio heads for a while and wondered 1) how they can honestly market them as 20(ish) watt amps having two EL34’s and 2) if they are really that much quieter.

Am I correct to assume these are basically “standard” plexi’s except they have underrated features throughout the power section?

1)smaller pt
2)smaller ot
3) lower voltage and cooler bias in power tubes

All these changes together would produce a very similar amp to the “big boy” just in a lower (or limited) wattage package? Is it really that much quieter though when all is said and done?

[pdf64]

I think that those 20W Marshalls tend to use cathode bias, so the idle point is pretty hot.
The lower power is achieved by a lower HT and a (probably) higher OT primary impedance.

[hawkes1962]

Remember the 6L6 (real KT66) datasheet has many suggestions and some are near your zone. (You want to avoid the AB2 conditions.)

6.6k load with 395V on plates and cathode bias made right around 23W clean on one project. This worked with metal 6L6 (a bit past rating), 7027, antique 6550, new commie 6550... tube is tube. IIRC I had shared 250Ω cathode bias and shared 2k(?) screen dropper. A skilled player took it from me to his studio and it never came back. Yes part of my goal was loooooong tube life. (This was a rebuild of an amp that annoyed me by burning-up.)

I've never been sure of that VTADIY calculator. It is certainly possible to provoke it into bogus numbers.

Actually, it looks like the vtadiy calculator is pretty in line with the sample operating points listed on the 6L6 datasheet, so maybe it’s not too far off after all.

I may have been too conservative when I ran the model. Initially, I limited quiescent idle current to 60 ma. I was under the impression that the model was basing calculations upon the quiescent idle current for each tube. I was trying to stay within the power transformer current limit of 125 ma (2 x 60 ma), and also trying to keep the load line from passing through the plate dissipation limit curve.

I've played with the model some more. By increasing the idle current to 75ma., output power increases to:

• 6K6 31.9 Watts Class AB1, 18.6 Watts Class A
• 8K - 27.6 Watts Class AB1, 22.5 Watts Class A

Ahh gotcha. That makes more sense. I wasn’t sure, either, if the mA were for both tubes or one. Looks like it’s both combined.

So a hotter bias will have more of the tubes’ cycle operating in class A, but won’t necessarily make more overall output power?

I believe I’ve heard, though, that class A amps (or amps running closer to class A) are perceived as being louder, all else being equal. Is there any truth to this? Maybe I’m confusing class A and single-ended…

-------

Alright, I think perhaps I’ll choose the following operating conditions (at least as a starting point, then tweak from there):

Va - 360V
Ia - 60mA (30mA/plate)
Za-a - 8000

Is it best to have your screens at roughly the same voltage as the plate? This is the typical Marshall way, yes? But with a choke to help keep the screens’ voltage steady?

[hawkes1962]

I think that those 20W Marshalls tend to use cathode bias, so the idle point is pretty hot.
The lower power is achieved by a lower HT and a (probably) higher OT primary impedance.

Yes, I believe I’ve heard they’re at about 260V, 50mA/plate at idle.

[jordan86]

The marshalls are cathode biased. Are you saying they could be even less wattage if biased cooler?

One could also hypothetically fake a higher primary by just using a different speaker tap on a multitasker OT?

[shooter]

are perceived as being louder, all else being equal. Is there any truth to this? Maybe I’m confusing class A and single-ended…

Single ended (SE) is class A.  You can drive them into A/B but results are not easily predictable
Most all my builds are self biased, running between 85 - 105% Max plate dissipation at idle (no signal)
almost every guitarist that's test drove them think I did something wrong because they say "this thing is NOT a 15W amp, it's really LOUD"


SPL and tube heat are 2 very different things.

[hawkes1962]

Single ended (SE) is class A.

Indeed it is. Sorry, what I wrote didn’t make a lot of sense.

I was more trying to ask if biasing a push-pull class AB amp closer to class A vs closer to class B might be perceived as louder, even though overall output won’t change (since the amp will ultimately venture into class B to achieve maximum output). The tone is certainly different, as countless have noted.

I suppose you have answered my question with your anecdote!

[pdf64]

The marshalls are cathode biased. Are you saying they could be even less wattage if biased cooler?

Everything affects everything else.
Cathode bias pretty much necessitates a wide conduction angle, ie close to class A. Otherwise the bias shifts significantly with signal level.

One could also hypothetically fake a higher primary by just using a different speaker tap on a multitasker OT?

That would really be a different primary impedance, not fake.
However, the treble or bass extension would be affected.

Stage gain tends to increase with anode current. So all else being equal, as a stage’s idle current increases from (near) class B to class A, its gain will increase.

[66Strat]

Remember the 6L6 (real KT66) datasheet has many suggestions and some are near your zone. (You want to avoid the AB2 conditions.)

6.6k load with 395V on plates and cathode bias made right around 23W clean on one project. This worked with metal 6L6 (a bit past rating), 7027, antique 6550, new commie 6550... tube is tube. IIRC I had shared 250Ω cathode bias and shared 2k(?) screen dropper. A skilled player took it from me to his studio and it never came back. Yes part of my goal was loooooong tube life. (This was a rebuild of an amp that annoyed me by burning-up.)

I've never been sure of that VTADIY calculator. It is certainly possible to provoke it into bogus numbers.

Actually, it looks like the vtadiy calculator is pretty in line with the sample operating points listed on the 6L6 datasheet, so maybe it’s not too far off after all.

I may have been too conservative when I ran the model. Initially, I limited quiescent idle current to 60 ma. I was under the impression that the model was basing calculations upon the quiescent idle current for each tube. I was trying to stay within the power transformer current limit of 125 ma (2 x 60 ma), and also trying to keep the load line from passing through the plate dissipation limit curve.

I've played with the model some more. By increasing the idle current to 75ma., output power increases to:

• 6K6 31.9 Watts Class AB1, 18.6 Watts Class A
• 8K - 27.6 Watts Class AB1, 22.5 Watts Class A

Ahh gotcha. That makes more sense. I wasn’t sure, either, if the mA were for both tubes or one. Looks like it’s both combined.

So a hotter bias will have more of the tubes’ cycle operating in class A, but won’t necessarily make more overall output power?

I believe I’ve heard, though, that class A amps (or amps running closer to class A) are perceived as being louder, all else being equal. Is there any truth to this? Maybe I’m confusing class A and single-ended…

-------

Alright, I think perhaps I’ll choose the following operating conditions (at least as a starting point, then tweak from there):

Va - 360V
Ia - 60mA (30mA/plate)
Za-a - 8000

Is it best to have your screens at roughly the same voltage as the plate? This is the typical Marshall way, yes? But with a choke to help keep the screens’ voltage steady?

I wouldn't worry about dropping the screen volts. I would just rely on the resistance of the choke to drop the B+ a few (3 to 5) volts. From there, you should be good to go.

I played with the load line calculator some more using voltages in the voltage chart from this schematic.
https://el34world.com/charts/Schematics/files/Marshall/Marshall_jtm45_lead_45w.pdf

With plate voltage = 440 V, G1 = 0V, Ia = 216ma. The Hammond replacement PT for the is specified at 450v345V @ 150ma. 216/150=1.44 or 144% of transformer current rating.
https://www.tubesandmore.com/sites/default/files/associated_files/p-t290maex_specification_sheet.pdf

With plate voltage = 360, G1 = 0 V, Ia =  162 ma. The transformer that you have is specified at 360V 255V@ 125 ma. 162/125 = 1.296 or 130% of transformer current rating.

[pdf64]

...The Hammond replacement PT for the is specified at 450v345V @ 150ma. 216/150=1.44 or 144% of transformer current rating.
...

Just to note that 345V at 150mA isn't the rated limit, rather the max current is 213mA.
And that will be per half winding, ie 0.7071 of the total AC current for a 2 phase arrangement, because current only flows half the time. So the AC current limit in this application would seem to work out to about 300mA.

... Initially, I limited quiescent idle current to 60 ma. I was under the impression that the model was basing calculations upon the quiescent idle current for each tube. I was trying to stay within the power transformer current limit of 125 ma (2 x 60 ma), and also trying to keep the load line from passing through the plate dissipation limit curve.

I've played with the model some more. By increasing the idle current to 75ma., output power increases to:

• 6K6 31.9 Watts Class AB1, 18.6 Watts Class A
• 8K - 27.6 Watts Class AB1, 22.5 Watts Class A

How can increasing idle current increase output power?
Something seems to be wonky.

And another thing with that calculator; if EL34 are 25W, how can KT66 be 30W? That's apples and pears, under the design centre system, KT66 would be more like 23W.

[66Strat]

...The Hammond replacement PT for the is specified at 450v345V @ 150ma. 216/150=1.44 or 144% of transformer current rating.
...

Just to note that 345V at 150mA isn't the rated limit, rather the max current is 213mA.
And that will be per half winding, ie 0.7071 of the total AC current for a 2 phase arrangement, because current only flows half the time. So the AC current limit in this application would seem to work out to about 300mA.

... Initially, I limited quiescent idle current to 60 ma. I was under the impression that the model was basing calculations upon the quiescent idle current for each tube. I was trying to stay within the power transformer current limit of 125 ma (2 x 60 ma), and also trying to keep the load line from passing through the plate dissipation limit curve.

I've played with the model some more. By increasing the idle current to 75ma., output power increases to:

• 6K6 31.9 Watts Class AB1, 18.6 Watts Class A
• 8K - 27.6 Watts Class AB1, 22.5 Watts Class A

How can increasing idle current increase output power?
Something seems to be wonky.

And another thing with that calculator; if EL34 are 25W, how can KT66 be 30W? That's apples and pears, under the design centre system, KT66 would be more like 23W.

My point in comparing the current at voltage ratings of the transformers was to get an idea of the amount of sag present in the power supplies when driven to clipping. Edcor does not publish the Max AC current ratings for their power transformers. Comparison of the current at voltage ratings appears to be a reasonable basis for comparison.

There are bugs and glitches in the model. No argument here. Per GEC, KT66 is 25 Watt under the Design Maximum rating system and 30 watts under the Absolute Maximum rating system. Under Design Center ratings the KT66 would be a 22-23 watt tube. I'm sure that some of the anomalies are due to operator error on my part. I have had very limited experience with the model. The model does appear to track published operating points for at least some tubes. Results do appear to track well for the 6L6G and 6L6GC tubes.

As I said in an earlier post; "I think the best way to use any of these models would be to start out with published operating points and compare the modeled results with the published data to establish the level of accuracy in the modeled results. Desired operating parameters could then be modeled and evaluated with some expectation as to accuracy."

…
Look at the model in this link. You can simulate operating conditions and see their effect on performance.

https://www.vtadiy.com/loadline-calculators/loadline-calculator/

I don’t find that intuitive to use, eg what are the fields
‘Next stage AC Impedance (Ohm):      HD%’
and
‘Out. headroom (+/-V):’
for?

Whereas the black magic amps one is intuitive and seems to work well http://bmamps.com/ivds.html

I have not had much experience with the model. I've only played with it a few times. The model does have fields that are not used in every situation. The next stage AC impedance is used in modeling pre amps and is not applicable for power amps. The Out. (output) Headroom appears to allow one to input the estimated plate voltage swing of the tube(s) to model harmonic distortion percentages. At least that's the way it seems to work. Some basic instructions for input fields would certainly be useful.

I think the best way to use any of these models would be to start out with published operating points and compare the modeled results with the published data to establish the level of accuracy in the modeled results. Desired operating parameters could then be modeled and evaluated with some expectation as to accuracy.

[66Strat]

As an alternative to a model, one could go old school and use the RCA Conversion Nomograph. All one needs are a straight edge, a sharp pencil, and sharp eyes. The degree of accuracy is determined by the fatness of the pencil line and the corrected vision of the line drawer.

[hawkes1962]

As an alternative to a model, one could go old school and use the RCA Conversion Nomograph. All one needs are a straight edge, a sharp pencil, and sharp eyes. The degree of accuracy is determined by the fatness of the pencil line and the corrected vision of the line drawer.

Hi all—been away for awhile. I’ve been working on a schematic, hopefully should have a draft worked out before too long.

I’m also trying to select a choke and output transformer. For a choke, I think I could use a deluxe reverb choke with good success, no? Mojotone Choke

Using the nomograph, it seems I’d be hitting somewhere around 17-19 watts (I’m coming up with a wattage conversion factor of 0.57, so ~32w x 0.57 = 18.24w). I’m thinking I’d either go with Pacific Audio’s Marshall 18w transformer Pacific Audio 18w or this Heyboer tweed deluxe OT upgrade Heyboer Tweed Deluxe upgrade

Do these seem like solid choices?

[hawkes1962]

:bump: