5E3 Compression?

[The Ballzz]

It seems that a common feature of the bone stock Tweed 5E3 Deluxe is compression when turned up fully to 12. What I want to discern is where that compresion is actually happening. Is it in the Cathodyne phase inverter or is it at the power tubes?

Here's the scenario. This amp is the most glorious amp I've ever had and all is fantastic, but......... Plugged into the high of the bright channel, volume at 9 to 9.5, tone at 11.75 and occasional blending in of the normal volume into the 0 to 6 range and it sounds fabulous, though could use just that "tiny bit more." As I bring that Bright volume up past 9-ish to 12, I can hear what appears to be the preamp singing, overdriving and grinding a pleasant, little bit more, but the natural compression negates any benefits from that extra grind. Of course, this same compression could be brought on by putting a clean boost in front of the amp, and it really appears that it's happening somewhere past V2A.

I don't want to alter the overall character of the amp or it's tone, but want to find the exact point that compression is happening and simply attenuate the gain at that point, just enough to allow the preamp to do it's due when fully cranked. I've seen all sorts of mods, changing coupling and/or bypass caps, many other schemes to reduce the low frequencies to in turn reduce the compression, etc, etc, but I want to try to tackle it in a different manner. I've also seen plans for adding either a standard type "pre-phase inverter" master volume, along with at least three or four different versions of "post-phase inverter" masters, as well as several schemes that are called a "cut" by phase canceling after the phase inverter. I'm assuming the choice of pre or post phase inverter will depend on where that compression is occurring?

If I find the right spot for "attenuation," it may not even need to be a pot, but instead a fixed value, as this amp will always get used in the same manner, at the same setting and overall loudness controlled by a good speaker attenuator.

Can I Even Get There From Here?
Gene

[pompeiisneaks]

You should be able to find where that's happening with a sine wave input and a scope. The 'compression' will look like the top of the sine wave flattening out. Basically put the amp to 12 like you mention, input a sine wave then raise it's output level until you hear that compression kick in, or you first see it somewhere. Then you can basically trace it backwards and see where it's appearing, output of PI? Speaker? Preamp node? etc. Or at least that's how I would try it.

~Phil

[The Ballzz]

Well Sir,
I've actually got a scope AND a signal generator, I just wish I better understood how to properly use them! I don't want to blow up or burn up the generator, amp or scope (or myself, for that matter) by improper use. I really don't want to be such a dummy, but I am! I would so appreciate anyone willing to lead me by the hand to start down this road, under the assumption, of course, that I'm an absolute idiot! While I'm not really all that stupid, I've certainly proven myself capable of making "stupid" mistakes. I'm really just a guitarist/singer in search of his "Holy Grail" sound! I feel like I'm so close to achieving my goal, but just not sure how to make these final steps.
Thanks,
Gene

[pompeiisneaks]

I'm no expert on Scope safety, but, the first, most important warning, is that for a tube amp, due to the extremely high voltages, you need to be sure you've got a scope safe for the voltage range. Most scopes come with 1x and/or 10x scopes that are rated for a max voltage of I think around Cat II which is 300VAC (I think similar VDC but not 100% sure). So if you ensure you measure the DC Volts at any point you're going to connect to with the probes, to ensure it's under that point, you're safe, OR buy a 100X probe that fit Cat III which is 600 VAC/DC (Again, don't quote me on that, I don't want you dying or burning up stuff and blaming me, it's my general safety understanding).


The way I ensure this is to only probe AFTER a known good coupling cap. The signal in AC suually is sub 100V at the PI outputs prior to the grids of most power tubes. Again, be careful and triple check everything. The output on the amp side at the speakers is also a spot you can probe as that's also sub 100 if I recall as well (double check with your DMM).

That being said, effectively start with a 400mV 1000hZ sine input it's a common starting point. But turn the volume pretty low so you can bring it up on the volumes until your'e sure it works. Another thing to be prepared for is to have ear plugs, a sine wave into speakers is super annoying, and at the volumes you want to test at, it may hurt too. IF you can, get a dummy load instead so you don't burn up the OT and can protect your ears as well.

Then probe at the post coupling cap area of each stage, just checking voltages first to be safe (DC and AC). You should see a perfect sine wave at the input and the same but amplified output after the first gain stage, just follow along as you go seeing if the top of the sine wave starts to flatten, that first subtle flattening is the compression/drive kicking in.

Then you can start by monitoring at the speaker and raise volumes and the sine input if needed, to get to the point where the output is showing the compression. Then step backwards to the PI and see if it's there as well on either or both sides (they should be identical signals but one inverted from the other, for push pull). Then if that's definitely still compressed, step back to the preamp stage(s) and see if there as well. If not, then it's the stage after (I.e. if the PI is clean but the output isn't, it's the Power Tubes doing it. )

~Phil

[Lauri]

Screen grid supply voltage sagging is what causes quite a bit of compression in a 5E3 circuit.

[The Ballzz]

Some folks are claiming that a lot of the compression occurs in the output transformer? I'm not sure of that, although I could certainly see how it could be either a combination of multiple factors and/or happening as a stack up of it occurring in multiple locations.

After my weekend of gigs, I'll put on my "big boy" pants, drag out/set up my scope and generator and start testing and experimenting! I guess the first place to start scoping is at the output of my signal generator and learn the correlation between the supposed output voltages, wave shapes, etc, to what I see on the scope and then just move along through the circuit, testing each point with the multimeter first. It can't be too difficult?

Thanks Folks,
Gene

[pompeiisneaks]

Sounds like a good starting point. It is possible the OT is linked I hear fender used to under spec those so at max volume it would be a bit pushed, maybe causing that?

~Phil

[xtian]

I can definitely hear compression in my 5E3 build with voltage scaling turned way down--so in my case, at least--it's not happening in the output transformer.

I don't think you can evaluate compression by looking for waveform clipping. Compression is scaling of the signal over time, regardless of whether the signal is clipped or not.

Screen voltage sagging--now that's s sure sign of compression.

[martin manning]

The 'compression' will look like the top of the sine wave flattening out.

That would be second-order harmonic distortion, not compression. Compression is signal level-dependent gain.

Compression is scaling of the signal over time, regardless of whether the signal is clipped or not.

No, compression has nothing to do with time*, but if your signal level decays with time, then compression will increase sustain by reducing the aparrent rate of decay.

Overdriving a cathodyne doesn’t end well. Maybe this is a candidate for a PI mod, such as a large grid stopper (see Valve Wizzard), or fixed bias (Paul C or Ampeg mod)?

*Edit: Except for the time required for the compression to take effect when a high signal level is introduced, and to return to normal when the signal level is reduced.

[Lauri]

I doubt much of the compression occurs in the output transformer. Most of it is because when current increases in the power tubes, cathode voltage goes up and voltages in the power supply go down.

[The Ballzz]

Ah ha! This is just the kind of discussion I was hoping for! And while I understand and agree that wave form clipping is not compression, "per-se" I'm guessing that whatever it is should be visible on an oscilloscope, as some sort of artifact?
Thanks & Please Keep 'Em Comin'
Gene

[xtian]

Compression is scaling of the signal over time, regardless of whether the signal is clipped or not.

No, compression has nothing to do with time*, but if your signal level decays with time, then compression will increase sustain by reducing the aparrent rate of decay.

*Edit: Except for the time required for the compression to take effect when a high signal level is introduced, and to return to normal when the signal level is reduced.

I'm nuts for splitting hairs with you, Martin, but I'll go anyway! The "scaling" I mention is attenuation. If the amount of attenuation does not vary with changes in input volume...over TIME...then it's not compression, it's just fixed attenuation. But I do agree with you that compression is changing attenuation based on input amplitude, not necessarily time. Ha ha! I contradict myself nicely.

[xtian]

Ah ha! This is just the kind of discussion I was hoping for! And while I understand and agree that wave form clipping is not compression, "per-se" I'm guessing that whatever it is should be visible on an oscilloscope, as some sort of artifact?
Thanks & Please Keep 'Em Comin'
Gene

I don't agree, Gene. Compression (as used in the studio) can "squash" (attenuate) the waveform without introducing clipping, distortion, or other artifacts.

[Reeltarded]

Harmonic content in all the right places makes the sweetness that isn't really compression but it's using everything in the circuit up to some point..

Clipping IS tchnically when the signal gets flattened at the points. Compression happens there, but compression hapoens before that too. Compression is why we like tube amps.

Don't continue this conversation, just listen to Martin.

[Reeltarded]

Sorry, was on the phone. Mom says hi.

You can apply a signal generator and check each node along the signal path and at the output. I am pretty sure this design hits is a sweet spot operating at full volume for the output tubes to sing and the somewhat dull eq and lack of tremendous gain before that point helps not to make that fine line turn into scribbles.

When I played early Marshalls I always had to drive them too hard and still hit the input with something to gain them even more. Only after I started building my own amps did I find the best sound is right around where most any amp starts to produce the expected RMS output. I had no idea that there was a balance point about 10-15% inside collapsing the NFB where the amps were fully wrung out and there would be no more excellence on the volume knob. I needed more sustain. The sweet spot is a fairly neutral path to the power tubes and every knob between 5 and 7... if you read this a few times it might start seeming relevant.

It appears the sweet spot on a 5E3 must be wide effin open with the tone control where you like it.