Fender 90s "Sunn Model T" - 6L6 power stage headroom & bottleneck?

[jamesmafyew]

I've got a Fender made "Model T" from the 90s. I know it's not the same amp as the real 70s Model T. My research leads me to believe that its preamp was an evolutionary stage in between the red-knob "The Twin" and the current "Twin Amp", with some extra cathode-followers thrown in. I like it a lot.

The power stage of this amp is part-for-part exactly the same as the red-knob "The Twin". Same PI, same 6L6 output stage with 470R screen resistors and 1.5k grid stoppers. The PT is different so the voltages could be different --- this amp puts 480-490V on the plates and feeds the screens from a 3H, 90-ohm choke. The OT is the same part number as the red knob twin - 026478.

My problem is that I like to play really really loud, and the amp seems like it's breaking up too early. Not so early that there's clearly something wrong with it, just too early for something that's supposed to do 120W. I'm wondering where the bottleneck is. This is not a speaker efficiency issue, I have a 4x12 with 102dB Eminence speakers in it. Power stage is biased cold - 52% idle dissipation. When it breaks up the tone gets darker and loses bass and it's like the amp just rolls over.

Right now I'm thinking it could be the OT. I can't find specs on the 026478 OT, but Hammond just came out with a drop-in replacement (the 1750D) which says it has a "2350-ohm C.T. primary" (as compared with the usual 2000 for the Twin Reverb 100-watters). It does seem on the small side, compared with other 100W amps I have around.

I have a Marshall TSL that I had a similar problem with, and put in the Mercury upgrade OT, a choke, and put in E34L tubes instead of EL34s, because they're supposed to have higher headroom. One or more of those changes made that amp able to go significantly louder before it got "raspy". Interestingly, I emailed Mercury and asked about their red-knob twin upgrade OT for my Model T, and they told me to get the Mercury OT made for the 70s Model T! I suppose an OT that works with 6550s would work with 6L6s, but I don't need ultralinear taps or a 2-ohm secondary tap.

Any opinions on where the bottleneck is, and how I can address it? If it is likely the OT, what are my options? Can I actually put a MM 70s Model T OT in this amp if wanted to (aside from mounting considerations)?

schematics:
http://local.anchorstates.net/images...reissue-p2.jpg
http://local.anchorstates.net/images...reissue-p1.jpg

EDIT: If the OT really is 2350-ohm, I did a load line on the 6L6 curves and it looks like the arrangement will only do about 85W. Not sure if I did it correctly, but that's definitely not the quoted 120W.

[KevinOConnor]

Hi Guys

Do you have a scope? If so, you could easily see if you are clipping the output stage.

One mod you might wish to try is to change the feedback loop values by an order of magnitude or so. The usual 820+100 gives a gain of 9.2, but the main thing is the really low values here interact with the parasitic elements of the circuit - especially the OT - in a "congestive" way. Increasing the values to 8k2 + 1k will open up the sound a bit without changing the gain. Going further to 82k and 10k will open it up further. Note articulation will be better.

One concern you may have is that the 100R is the lower part of the splitter connection to ground. Increasing this to 1k or 10k shifts the splitter voltages up a little bit, but this is inconsequential to the splitter's function.

Also, even though you feel that 120W will satisfy your needs, you admit to wanting to play loud. This much power is easily eaten up by guitar players and a bass will over-run it a lot quicker.

A 12AT7 usually provides the highest clean output in standard-value splitters.

B+ in "The Twin" was 500V, which is no problem for 6L6GCs. 100-135W is easily attained. You might want to check the screen and grid-stop resistors for the output tubes. Maybe one of the 470s is open. This would easily cut output and produce an asymmetric distortion that you may find doubtful with bass.

Have fun
Kevin O'Connor

[jamesmafyew]

Hi Kevin,

I do not have a scope, unfortunately, or that would have been my first step. I have tried a 12AT7 in the PI and did not detect a noticeable difference. I guess I'm confused as to why my EL34-based 100W Marshall is perceptibly quite a bit louder than this amp before it gets ragged (that's the amp with the large Mercury OT in it), side-by-side with an identical 4x12 cab. I've checked the grid and screen resistors in the Model T and they all check out fine.

As far as the NFB/presence circuit, the Model T's does differ from the red-knob Twin's. The red-knob Twin takes NFB from the 4-ohm tap of the OT, whereas my Model T takes it from the 16-ohm tap. The red-knob Twin has a 1.2k & 120R divider on the negative PI input, but my Model T uses 47k and 4.7k, which actually seems in line with what you've suggested above. Even before the divider there is a 150k resistor in the feedback line bypassed by a 47nF cap (R107, 108, and 109, and C41 in page 2 of the schematic above). The Model T has a Marshall-style presence control with 25k pot and 100n cap across the 4.7k resistor, which is absent in the red-knob Twin.

EDIT: thinking further, seems like the biggest thing here is that there's almost 4x as much NFB in the Model T circuit as in the red-knob Twin. Might it help to move the NFB wire to the 4-ohm tap of the OT? Or am I asking for stability problems doing that?

[Enzo]

Check the phase inverter. Is the plate voltage missing from one side? Suspecting open plate load resistor there.

The problem won;t be the impedances are wrong or anything else cosmic. If that were the case, then ALL of them would act this way.

[KevinOConnor]

Hi Guys

I would short out the 150k and its parallel cap. This is like a preset 'resonance' control that decouples the feedback loop at low frequencies. It always sounds "flabby" and will not help with the tone if the OT is under-sized.

Have fun
Kevin O'Connor

[KevinOConnor]

Hi Guys

Moving the feedback loop to a different ta is unlikely to upset stability, especially with the larger-than-Fender resistances used.

Note that when you look at the 6L6 curves, those are generated with a lower screen voltage than is present in your amp - usually 250V. You have to scale all the values up by the ratio of actual (480V) versus listed (250V) screen voltage. Then you will find that the 2k34-aa load is well handled by four tubes.

Have fun
Kevin O'Connor

[jamesmafyew]

Thanks Kevin and Enzo. This is helpful. Enzo, I had the same thought and checked those resistors, they're fine. Maybe I've overstated the case... the amp sounds fine up to a fairly loud volume, very clean. Maybe they do all sound this way, I haven't played but one other one and I didn't push it this hard. It just seems to get raspy at a lower volume than its 120W rating would suggest, when compared with my Marshall. Though it's not the "real" Model T, it does have a reputation for being a loud amp.

Kevin, point taken about the datasheet curves. I stand corrected, that would explain things. Mercury says their Model T OT has a 2200-ohm primary, and since I have a quantity discount with them I may actually splurge for it and see if it helps this amp. If it doesn't, I can pass it along.

About the NFB loop, I did some math on it. At frequencies below about 650Hz, the NFB voltage to the PI is roughly the same as in the red-knob Twin. The 16-ohm tap and the larger voltage divider cancel each other out. At frequencies above 650Hz, the divider is the same proportion as the red-knob, but there's 4x the voltage because the 16-ohm tap is used, darkening the sound (?). This seems like a potentially crucial part of the "dark voicing" of this amp that I actually like at lower volumes. I suppose I could wire it to a switch so that I have both options.

EDIT: here's a spice AC analysis of the Model T's NFB/presence control range. This is the NFB voltage as it appears at the negative input of the PI. Looks like you can get roughly the same NFB as the red-knob Twin on the Model T by setting the presence control to 9. This sim accounts for the 16-ohm vs. 4-ohm tap NFB hookup, as well as the resistor values.

[jamesmafyew]

Just wanted to follow up on this, since I had a chance to play the amp loud again at practice tonight.

I had the presence set to minimum before, because I thought that was "neutral." I assumed that turning up the presence cut down the NFB in the highs, to raise the brightness. I don't like overly bright amps at high gain and high volume. But it turns out that this presence controls removes highs instead of adding them. Playing the amp with the presence at 9 (roughly neutral, same level of NFB across all frequencies, like the red-knob Twin), it was perceptually significantly louder and less farty. Kinda dumb that before I was running it with all frequencies above 650Hz significantly attenuated. I just rolled down the treble a bit and actually turned the volume down a bit tonight. How about that?

There's still a bit of rasp though at high volumes. I've been over the load lines several times, accounting for the screen voltages, transformer primary, etc., and I still think the amp is capable of more than it's doing. It still isn't as loud as my Mercury-equipped Marshall. The only thing I can figure is that, like the Marshall originally, the OT core is a bit undersized to save money. I'm going to go ahead and get the Mercury 70s Model-T OT. I'm not ever getting rid of this amp, so I don't mind investing in an OT with a larger core for it. I play down-tuned very low, so improved low-frequency handling would be a huge plus.

[Enzo]

Power isn't loudness, and comparisons like "amp A is louder on 2 than amp B is on 6" are not really of any value. How loud are both all the way up is more of a fair comparison. The difference in loudness between 50 watts and 100 watts is 3 decibels. An amp with a more sensitive input stage will be louder than another amp. And unless you play through the same speaker cab, then the speaker efficiencies also enter the calculus. And just had a discussion with a friend over the contrast between piercing and loud. As in how some little "practice" amp can sound louder than a much larger amp, just because all its little energy is put into one narrow band of freqs that pin your ears back.

One thing about presence controls is they are active controls in the power amp. And they react different at different power levels. In other words settings you like at low levels may not work at high levels, and vice versa.

[Steve Conner]

Well, Enzo has a point with the loud vs. piercing issue. Traditionally, fenders have a bigger mid scoop than Marshalls. The midrange is responsible for most of the perceived loudness because your ears are most sensitive there. So if you crank a 100 watt Marshall, it makes a deafening roar, but the Fender gives more of a flabby, fizzy result, because the bass and treble are more prominent than the mids.

A bigger OT would help, sure, but by how much? The real solution would be to re-voice the amp as a Marshall, but you already have one. Sounds like you really need a 400 watt solid-state rig.

[jamesmafyew]

Yeah, I understand about power vs. loudness. The way I've compared my Marshall vs. the Sunn (which by the way is voiced very Marshall for a Fender) is by "swapping preamps". I have two identical, 102.5dB 1w/1m 4x12 cabs, each head into a cab. I then run the FX send of each head to the other's return. Using footswitching, I can switch either preamp between either power amp for comparison. I've done the math and checked input/output AC voltages to ensure that each amp's FX send level and return sensitivity is set equally.

What I'm trying to evaluate with this setup is the perceived volume that each power amp can get to, before increasing input drive just changes the sound (more "raspy") instead of adding volume. Even when fed from the Marshall preamp, the Sunn is less loud, in the sense that I can't get to the same perceived volume before it starts to distort and the character changes radically.

I now understand that a big reason for that is the presence circuit... as you guys pointed out, before I had it bleeding off frequencies in the most sensitive range of human hearing. I don't think that turning up the presence is creating "more power" or "more headroom", but it has made the amp more "perceptibly loud" before it starts to rasp out, and made the sound much more consistent with the Marshall's power amp (which has the presence set to minimum, incidentally). I do think the other factor is the OT core, because I remember a similar thing happening with the Marshall in question back before I put in the Mercury OT. On that amp I couldn't turn the bass up past 5 o'clock at volume without it getting farty, after the OT change I could turn the bass all the way up and actually hear the bass increasing. So I'm not expecting a new OT to make the amp into a 400-watter, just to increase bandwidth at high volumes and maybe give a touch more leeway before breakup.

Thanks for all the replies!

[jamesmafyew]

Just to follow up on this:

I installed the Mercury 70s Model T OT today. It is about twice the size of the original, though the footprint is roughly the same.

The amp has felt considerably more powerful (perceptually) since I started keeping the presence up at 9. Adding the new OT added a small amount of perceived headroom (breakup is later on the volume control), and the character of the sound at high volumes has gotten clearer... less rasp and fart, more crunch. The biggest change, though, is bandwidth. There's a lot more bass available at high volume than before. Whether it was worth the $$$ investment is going to be up in the air for a while (I lost the XLR line out in the process, and had to file off some of the head cabinet to get the thing to fit), but since I have no intention of ever getting rid of this amp, it'll probably stay this way.

[el_fela]

Hi there, maybe its a litle late, maybe you checked this many times but, how about speaker wiring and impedance mismatch? worn tubes?.
My 2 cents.

Felipe