• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Possibility of 100 Watts?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.


Joined 2003
Paid Member
Default Possibility of 100 watts? ... 450 vdc...., 4K P-P .... Tubes that I have ...

Simple math tells us that 450V peak in 4kCT (1k each side) makes 101 Watts, assuming the tube gives *zero* voltage drop at 450mA.

No tube does that. No sane tube comes close.

The 6550 is no slouch. http://www.mif.pg.gda.pl/homepages/frank/sheets/135/6/6550A.pdf
Page 3 shows 450V and 4k UL making 70 Watts; 450V and 3.5k making 77 Watts (so maybe 68W into 4k).

Some of your sweep tubes may do a shy hair better, with awkwardly low screen voltage (which must be stable), and unknown Pdiss up-rating. (The sweep ratings are conservative because it was difficult to measure directly; audio amps are easier to measure so can usually stand more, but how much more?)

Also - very few tubes really deliver 100W per pair. While 6550 KT88 and 8417 show 100W/pair suggestions (at near 600V), you may note that commercial products rarely promise more than 75W/pair.
assuming the tube gives *zero* voltage drop at 450mA. No tube does that.

I stated that in my first post, 102 watts assuming a perfect tube AND a lossless OPT, neither of which exist.

Some of your sweep tubes may do a shy hair better, with awkwardly low screen voltage

I plotted some curves for the 6LW6, but I did not take the plate voltage below 50 volts. At 50 volts the plate will draw 650 mA at VG1 = -5 volts, 850 mA at VG1 = 0 volts, and 1040 volts at VG1=+5 volts. This is with G2 = +150 volts, which must be stable. I use a zener feeding a mosfet follower setup for all my sweep tube amps.

The data sheet curves show that the plate voltage will be about 25 volts for a plate current of 450 ma and 0 volts on G1......however the screen grid will be very unhappy, so operation in this area must be limited to brief music transients. To deal with this issue one limits the screen current with a resistor in series with the drain of the screen supply mosfet, and a suitably sized capacitor after the resistor to feed the screen long enough to sustain those big bass drum hits at full power, but not enough total energy to melt it should someone attempt to do a full power test with say a 10KHz square wave.

None of this is necessary for a 100 WPC amp on 450+ volts, but I am planning to build the BIG ONE before I get too old to lift it. I have all the expensive stuff for a 500 WPC tube amp, but I don't really need one, so it hasn't been a high priority project. Something like this does need to be well thought out and planned to avoid making a BIG box full of fried stuff.
Regardless of the tube used, the power output is basically the AREA under the Load-Line.. Using first order ideal equations..By simple examination you can show that the Load-Line is 1/4 at 1K Plotted at 450V /1K it will intercept the current at 450mA @ 0 plate volts.. The area under the curve is (450*.450) /2 = 101 Watts... However, ideal equations are simply " fairy-tale" in the non-linear world of tubes...

The B+ will droop to some degree lets say 425V for S###s and giggles... The Load-Line will actually intersect the "triode" region of the curves..figure 60V to 100V range ... If your using a sweep tube with low screen voltage limit in the 150V to 200V range...Now your load-line intersect with 0 BIAS curve of G1 will be a fairly low current...This will require higher B+ .... If you do not have the higher B+ you can extend into the AB2 range , which would require a low Z driver... Now you know your load-line area...next fit a suitable tube....or tubes...


Joined 2003
Paid Member
...you can extend into the AB2 range..

Most of the tubes in rsumperl's list will not make more current at positive grid voltage. (Anyway most are already bottomed.)

In triodes Mu is a compromise between max current and insane drive voltage. Many power triodes, if we accept insane drive including current, can push to AB2 with some benefit.

Pentodes with high Mu(g2) and low grid voltage have a similar possibility. High Mu allows Vg2 to be similar to Vp, so not requiring a separate screen supply for good regulation.

Pentodes adapted for LOW Mu(g2) can make gobs of current; and with low Vg2 do not have obscene drive demands. Going AB2 is not much use until we drop the load impedance much lower. (But rsumperl only has what he has.)
A quick review of all the tubes on your list... Given the 450V and the 4K load... Figure 125V to 150V regulated on the screens.. for just a P-P pair... For maximum power output your best bet is M2057 and a close second is the 6LB6.... Figure in the 80 watt range.. assuming your B+ is stiff as a rock...not realistic of course.. You did not mention if you have enough tubes for Parallel P-P operation... This way you can offset the secondary impedance selector to make 2K ohm plate load... cheater connection.. You low frequency response should be good however the leakage will play greater role with the 2K with high frequency corner... As long as your power level are with 100W ....you primary wire should handle the current density...
Last edited:
8950/6LX6 has low knee voltages and is quite brawny too, 125 Watt possible. Low knee V's will maximize power output for available B+, but might want to save these for a higher B+ design.

21LG6A or 6LB6 will easily do the 75 Watts (B+ constrained) and have low screen current distortion, as well as low knee voltages.

M2057/6LW6 suggest saving these for a design with more B+ available, lots of power, 150 Watt possible.

For the 75 Watt possible with the available B+ limit, a cap-less 22/17KV6 or 6HJ5 could do just fine.
Using parallel tubes might be the easiest way to get you the most power, since the tubes will see twice the load resistance and the load line will hit saturation at a lower plate voltage at that lower peak current. It certainly won't double your power or anything, but it should give you more.

You can then also use lower screen voltages, since you don't need your knee to be so high, and maybe get nice long tube life from your happy tubes.
Hmmm… here's a not-so-novel thought. How about getting an el-cheapo 2 secondary 48, 48 volt (independent secondaries) transformer, to increase the output of your existing one … which is 'only' producing about 400 volts filtered DC? You'd get another 60+ volts. Doesn't need to be very big. 50 VA plenty big enough.

That'd go a long way to achieving the 100 W output, with the preëxisting output transformers that you already have. With any tube combination you finally settle on.

Just saying,
GoatGuy ✓

Another reminder … 75 W vs 100 W perceptual audible difference will be
dB = 10log₁₀( 100 ÷ 75 )
dB = 1¼ dB difference​
Which is to say not much at all.

I'll keep reminding until someone squawks.
Just saying,
GoatGuy ✓

Finally, tho' it does kind of make me a “KodaBMX shill”, I find a lot of attractiveness in his dual-toroidal-output transformer technique, and amplify-after-phase-inversion amplifier topology. Really straight forward, uses a LOT of output tubes if you have 'em to get quite high power. GNFB resolves most of the not-quite-audio-grade pentode 'knee' problems, too. Just saying,
GoatGuy ✓

And I just want to say that 100W is really only necessary if you have a very large room or very inefficient speakers. My 40W amp I built for my bedroom sounds like it isn't even breaking a sweat at ear-bleeding volume levels.

I didn't really believe people here when they said that 100W was overkill until I made my first prototype triode-connected KT88 amp (25W?) and heard how ridiculously loud it could get.
Just get one of these $99 Hoefer PS 500XT units, and dial in the B+ Volts you want. Even has a timer to shut your Hi-Fi off after you fall asleep in the easy chair. 0 to 500V, 0 to 400 mA, 200 Watt, digital readouts, current limit adjustable, regulated.


  • PS_500XT.jpg
    53.9 KB · Views: 301
Last edited:
Good Wednesday everyone,
I wanted to follow up and thank everyone for their inputs. What I gather is 100 watts may not be possible, but plenty loud should be. Having 5 acres and no neighbors is a blessing!
I included the PDFs of the power transformer and output transformers. My plan was going from a bridge into a capacitor and then into the 21st Century Maida Regulator. I'm thinking I should be able to get a solid 485v out. The "input" will be the "Dyna-70, the ultimate Dynaco ST-70 upgrade kit" from Erhard Audio. The chassis will be a 12" x 18" chassis I had laying around from Landfall Systems. I also attached my layout as it stands today.



  • Output Transformer.pdf
    21.8 KB · Views: 89
  • Power Transformer.pdf
    135.2 KB · Views: 83
  • FPD.pdf
    38.9 KB · Views: 91
I looked at your PDF's......

The power transformer has plenty of grunt, and can be wired for any number of different high voltage levels to beyond 400 volts, so any B+ from 350 to 500 volts is possible. 100+ WPC from this guy should be no problem.

The output transformers were designed for use in a Marshall 100 watt guitar amp. The lowest note in a guitar tuned the traditional way is about 82Hz. No other electrical specs were given in the PDF, so this could be a problem. Their physical size is about the same as the 100 watt Edcors that I have, so they might actually be capable of considerable power below 82 Hz. Only testing will tell.

I don't know anything about the Dynaco upgrade kit, and little info is given on their web site. It was meant for driving a pair of EL34's to about 35 watts per channel. I don't know if it can drive any of the sweep tubes you list to 100 WPC. The bias supply needs to go down to -80 volts or so. EL34's don't need this much.

Again I don't know much about the 21st century Maida regulator board. I have looked at it's schematic some time ago, and people have stuck them into some of my amp designs with good results. I would NOT recommend using it for the main B+ to the output transformers. This is a high voltage point with varying current demands from a few mA to over an amp on transient peaks. The regulator IS an active circuit with an internal feedback loop inside the chip. It can act funny under these conditions. Connect the two OPT's directly to the main filter cap on the B+. Use the Maida to run everything else, and possibly use a second one for the screen voltage supply for the output tubes.

Feed the HV winding from your power transformer through a bridge rectifier, then into a filter cap. I would then use a choke, even a cheap ugly Triad hidden under the deck, then into a BIG cap. Feed everything from this cap. I use an electrolytic of suitable voltage rating, then parallel that with a motor RUN (NOT motor START) cap of around 100 uF. For an amp this big find one rated for 370 VAC or 440 VAC. These are low ESR polypropylene caps without the audiophile price tag.
I also had a close look at the OT's PDF sheet right now again. If this »JCM45-100« indeed has a 4 kOhms primary, this OT by no way is capable of 100 watts output power. Instead it is designed to be used with two EL34's at about 450 Vdc supply voltage, resulting in about 50 watts output power.
If your intention is a HiFi rather than a guitar amplifier, but you're stuck with guitar transformers, have a search for the Hiwatt DR103 OT by Partridge or a good clone of it. This is a very sophisticated transformer, outperforming almost any other guitar amplifier OT.
Best regards!
I'm not familiar with the JCM45, but Marshall used a pair of KT66 in their early JTM45.
As per GEC datasheet it had an 8k OPT and around 500V would get you to 50W. With a quad and a 4k OPT you'll have 100W (and permanent tinnitus :cheeky: ).
Their later 100W models still had the 500+V but used 1k7-2k OPTs and 4 EL34s. That's 'a little' over 100W...
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.