Tube Amps / Music Electronics
|For current discussions, please visit Music Electronics Forum.||New: view Recent Searches.
New: visit Schematic Hell!
The sunn still shines online!
|Listen to great tunes streaming live right now!|
|11/2/1999 6:39 PM|
||U.L. Fenders (calling Ken Gilbert)|
Ken, in a recent post (interesting stuff!) you mentioned that you didn't like the Fender UL designs. I have a 135 watt UL Twin that i'm tweaking and would be interested in knowing what it is you dont like about these amps and any suggestions for improvements.
|11/3/1999 5:57 AM|
Well, it's not like I hate the amp. I mean, at least it's UL, right?
IMHO, the greatest benefit of UL ops is the greatly lowered output impedance... It's leagues lower than pentode mode, and much closer to triode mode. The best thing is that you don't give up much power compared to pentode, and it's about twice the power of triode.
So what the low output Z means is that you can throw away your NFB loop. Just disconnect it entirely. You will probably have to adjust some gain earlier on in the amp to compensate for the loss of NFB, but you can also use this "extra gain" for a presense control or other active EQ types. The Baxandall type of active EQ that Kevin O'Connor uses in The Ultimate Tone is very effective and only uses NFB around a single stage. To my ears at least it is not nearly so offensive as that which is wrapped around multiple stages--and ESPECIALLY not through the OPT and connected to the SPEAKERS.
Once you throw out that traditional NFB you will be welcomed to a place where NO AMOUNT of overdrive or peak signal will sound bad. Well, I take that back--There's still one more place left to go, and that is to drive the output tubes with cathode followers, sourced from a negative rail, and directly coupled to the output grids.
Then you no longer have grid blocking, and you can make MORE power since you can pull some grid current. There is no end to the dynamics here... as you crank the shit out of the output stage it will clip, and since it has no NFB around it it will NOT sound like traditional output tube distortion. It will be much much more euphonic and very sweet. You just CANNOT make it sound bad at all!
It helps to have a nicely distorting output tube--in my case it is the unbeatable and unfortunately NOS EI KT90.
But actually I have patents on all of this shit, so if you use it, I will sue you.
|Book Of The Day||
The Ultimate Tone, Volume III by Kevin O'Connor
Note: The Ampage Archive is an Amazon Associate site. A small commission is paid to the site owner on any qualified purchase made after clicking an associate link such as the one above.
|11/3/1999 1:14 PM|
The UL Fenders have a high B+, so you need to bias them cold so you don't blow the output tubes. In such a condition, simply removing the feedback loop would make the crossover distortion very noticeable.
|11/3/1999 1:49 PM|
What's the B+ like in those amps?
I mean, it can't be TOO high--the plate and screen voltages are by definition the same. That limits it to under 600V or so.
To rely on the NFB loop to eliminate XOD is bullshit, if you ask me. What happens when the output stage clips and your NFB goes out the window?
|11/3/1999 2:25 PM|
Flipping through the Tube Amp Book, all of Fenders UL models use the 6L6GC with about 500V B+. Differing output wattages correspond to fewer or more parallel output tubes. There were three main ranges as far as power output was concerned--75W, 135W, and 200W, with 2, 4, and 6 6L6's respectively.
With 500VDC on plate and screen, you can bias the amp to 20 mA per tube with -60V on the grids. (For the DC case you can use the triode curves for purposes of determining plate current.) 20 mA per tube is 10W Pdiss @ 500VDC. I would expect them to last a good long time at this dissipation, and with 20 mA flowing I doubt you'd have any trouble with XOD.
And if you're up to modding the amp in an UNINVASIVE way, lower those damned 100nF grid coupling caps on the output tubes. When you start drawing grid current those things store plenty of charge and drive the bias more negative. That WILL induce XOD, and what's worse is it's level dependant. Simply make them as small in value as you an stand without losing too much low end, and it will "speed up" the bias circuit.
|11/3/1999 4:44 PM|
If the NFB "goes out the window" when the output stage clips why are you so keen to "throw out that traditional NFB" so that "NO AMOUNT of overdrive or peak signal will sound bad" ? (g). I just wondered because the traditional NFB sounds fine to me with the output stage overdriven.
I don’t think that XOD is a problem at high outputs. It looks bad on the ‘scope but it gets masked by the high volume (and clipping distortion if the amp is overdriven). If the amp isn’t clipping I don’t think the amplitude of the XO notch will change much with signal level so XOD becomes more obvious at lower levels where NFB can be used to reduce it. I like a bit of NFB. There ain’t nothing wrong with the 5F6A. When some marketing type tells me that a "pure class A" (meaning cathode biased) output stage with no NFB is somehow intrinsically sonically superior in a guitar amp that’s what I really call bullshit.
|11/3/1999 5:52 PM|
|ken at work again|
Well, just because the NFB loop becomes increasingly ineffective when large distortions arise in the output circuit does NOT mean that it's the same as having no NFB loop.
Remember what the NFB attemps to do... it attemps to make the output of the amp match the input of the amp. In an extreme case you'd have a follower, where there is NO voltage gain, the output merely "following" the input. The great thing about followers is that they have very low output impedances. That's why Futterman used a unity gain output stage in his OTL--he gave up voltage gain in order to take advantage of the lowered output Z and decreased distortion of a feedback factor of 1.
But when you inject a signal which the amplifier CANNOT reproduce, because of clipping, bandwidth, or what-have-you, then the NFB loop suddenly attempts to do an amazing amount of work to try to make the output = input. Of course, it cannot do this, but it sure attemps valiantly.
What happens is that the distortive artifacts which are produced in the output stage are fed back, and these artifacts are then re-amplified and fed back once again, and so on. The stickler here is that when you have such a loop of distortion the artifacts (ie the harmonics) MULTIPLY, getting pushed upwards in order. 2nd harmonic distortion of 2nd harmonic distortion is 4th order with respect to the fundamental--your guitar signal.
There is also the issue with the output impedance and the action of the loop. Under normal small signal conditions, the NFB loop is effective and lowers the output impedance. During clipping however, the loop has become "imperfect" and its performance suffers. Whereas the summing junction of NFB and original signal previously had only a small "error voltage" to deal with, suddenly it has become MUCH larger. Now the voltage which is to be amplified will probably clip the output stage HARDER, which results in more of the same problem. This is the loop losing control. As a result, the output Z increases, and the sound becomes subjectively much looser.
The worst thing is, it's all level-dependant. You can't really predict how the amp will react while it's sitting on the bench. Maybe the sound of the loop losing control is what people LIKE... I don't know. I know I don't like it.
This is a very subjective topic to tread--first of all every design is different... if you've got an amp with a lot of HF emphasis going on in the PREAMP, if you push the harmonics even higher in order you may end up with total harshness and fizz. OTOH, if you've got a particularly dark sounding preamp, perhaps with little gain and little upper harmonics, then pushing them up in order might not be a bad thing at all.
You've also got the issue with NFB and the fact that it's effectively connected to the speaker itself. The speaker therefore has a much larger effect on the response of the amp... its back-emf is fed back into a node earlier on in the circuit, and now you've got a whole other signal to contend with. Now plug that Marshall head into a full stack, with 8 12" speakers flapping wildly about, and tell me how complicated the equations become!
The key here is a "bit" of NFB. The effects will become less bothersome with decreasing feedback factors.
But the question I ask is why use it at all? It needlessly complicates the design and obscures the sound of the output tubes themselves. In this case I would prefer none to any, even the smallest amount.
If the only reason is to lower output Z, then UL is the obvious answer. A UL tapped OPT is--at worst case--a few dollars more than a non-tapped OPT. It's cheaper than a choke for your screen supplies, and takes up less real-estate.
If another reason is because you want your presense control, then I say put it somewhere else.
Again, it totally depends on the REST of the amp. I'm sure there are plenty of folks who could fuck up a class A ZNFB output stage. It is certainly not INTRINSICALLY superior.
But if you've ever played a vintage Vox AC30, and cranked it up while plugged into the Brilliant channel, you'd know that a ZNFB output stage has something that no amount of NFB could emulate. There is a cloistering, a blanketing of the sound that NFB throws over the whole tone and it is simply not there when the loop is open. I don't claim to totally understand why it happens--the multiplicative artifacts and load interaction are only parts of the explanation, I suspect--but I do know that if you want the speakers to float in mid-air around your head you'd do best to run the output open loop. It sounds like a giant down comforter has been removed from the front of your cabinet.
I mean, doesn't it make you wonder why the hell EVERYBODY wraps about 15dB of NFB around the outputs of guitar amps? I can safely tell you that's traditional, conventional thinking at work. It's not because they tried it any other way and liked it best. I am absolutely shocked at the dearth of UL guitar and bass amps. At least Dr. Z's on to something, but everybody else (as in Marshall, Fender et al) seems to think an overcomplicated preamp is the answer, and totally neglect the output stage's contribution. IMHO, that's a CRITICAL stage--it gets the final say as to how the amp will sound. Experiment for Christ's sake.
This is the whole beauty of tubes--they're linear enough to use without any NFB. You just can't do that with SS devices. That to me is the largest discrepancy between the two technologies. So if you've got tubes, take full advantage!
Again, YMMV and all that crap. If you've got 3dB of NFB it's obviously not going to change much when you remove it. All I'm saying is that one of my favorite mods--probably the first thing I try--is to switch to either triode or UL and remove NFB. This goes for HIFI as well as MI. And I'll tell you, I've never looked back.
|Page 1 of 5||Next>||Last Page>>|