Biasing certain tube amps....

[sumgai]

(Dedication - this is for Runewalker, it's actually a carryover from another thread.)

Tubes operate in a vacuum. They're about the only things that do, these days, but they still have odd requirements in order to make them work properly, or at least, decently.

Proper: It doesn't blow up.

Decent: It sounds pretty good.

I had started out to give a 200 to 300 word synopsis of what bias is and does for you. HAH! That idea went to hell in a handbasket, I stopped at nearly 1800 words, and I still wasn't done. Oy vey, maybe some other day, if conditions warrant! Here's the condensed version, no theory, just the practical how-to.

I use the more complicated "designers method" of adjusting bias, but I'll start with the simple version, and then go into the way I do it.

The easiest method is to look at the schematic, find out what the bias voltage is supposed to be (-45 volts would be my guess), and then measure it at the appropriate point in the circuit. Unfortunately, this requires you to remove the chassis from the cabinet (Unplug the bleeping thing, right now!), the test point is not accessible from outside. Once the chassis is out, rest it on something that will protect the power tubes. You'll also be getting them hot in a moment, so you don't want them touching anything anway.

Locate the point in the circuitry where you're gonna find the negative voltage and hook up your meter here. Don't forget to connect the other lead to ground, and set the meter to DC volts (auto range). Finally, locate the bias pot itself - it'll be between the solid state rectifieres and the main circuit board, somewhat behind your master volume control. Procure a screwdriver that will fit this control.

Now turn on the amp. Check the bias voltage, is it close? Hook up a signal source, preferably a signal generator, but your guitar wil do. Strike a note, and hold it. Does the bias voltage change by very much? It shouldn't, if it does, then we have other issues. If it stayed steady, then set it to the prescribed voltage as found on the schematic, and you're good to go. Tear it all down, and put everything back the way you found it. Go get a beer, then come back and rock out!

Method Two:

You need an oscilloscope, a frequency or signal generator, and a dummy load. Got all of that? Good, then reply back here when you want these instructions, because it's a tad bit longer than what I've already typed in tonight! ;D Me, I'm off to see if I can't get in some jamming at a local joint. Catch you all tomorrow!


sumgai

[ChrisK]

Well, ackshally,

One might surmise that the best way to set power tube bias is for equal quiescent current thru each tube. This effects DC balance in the output transformer.

This is done by measuring the resistance of each leg of the output transformer for resistance. Then, under power, measure the voltage across each leg. The voltage, divided by the respective leg resistance then equals the plate current for each tube.

Since tubes are transconductance modulating devices, a current model approach may be in order.

Equal bias voltage on dissimilar devices (all tubes ever made are at least somewhat dissimilar) does not effect equal plate current.

[sumgai]

Chris,

First, the operative word of my initial post was easy. I assumed that measuring voltage is well within the capabilities of everyone here, and that we all have such an meter sitting in our toolboxes. Anything beyond that runs the risk of leaving some folks behind. Or so has been my experience, anyway.

EDIT
Second, I'm modifying this post after a few hours. Upon reflection, I've come to realize that the tone I've taken here is more than just pedantic, it's downright uncouth towards the members of this forum. With that apology in mind, ahead of time, please allow me to let the thing stand as is - it took a lot of work to get this down in black and white. While it is "in your face", I do go into full detail in explaining myself, about why I believe as I do. Let's just chalk it up to this being one of my very few pet peeves, and hope that nothing else comes along that triggers the "soap box" mode in me again. ;D

Thanks for understanding, and cutting a guy some slack.
/edit

To add spice to this thread, I'm gonna go out on a limb and do two things. One, I'm gonna debunk your quiescent current theory, and two, I'm gonna irritate most everyone that doesn't have a lot of somewhat expensive test equipment laying around the workbench. But I will be thorough in showing how to bias the venerable Fender power tube stage. (Just be aware that this adjustment can't be done to every Fender ever built, there are some exceptions out there.)

Bias Adjustment Methods, Part Two.

Quiescent current is defined as that plate current which flows under no-signal conditions. If there is no signal to be reproduced at the plate (amplified), then for our purposes, it does not matter how much current is flowing, we don't care. The only exception to that statement is when there is a fault somewhere, and the quiescent current raises to such a level as to harm the tubes themselves. Usually when that happens, the plates turn cherry red, and some sort of Gawd-awful sound comes out of the speakers. But take careful note, that is not the normal case.

Now, if there is no signal flowing from grid to plate, and we don't care about the quiescent current, then when should we care about the amount of plate current? Only when we're checking to see that we are within the design's maximum permitted values. Other than that, we don't care how much current is being drawing by any tube, singly or in pairs, matched or otherwise.

Rash statement, eh? Well, as you said yourself, no two tubes are identical anyway, so why worry about how much current flows through either one of them, in balance or otherwise? Because we aren't interested in how much current is flowing, we're interested in the sound that arrives at the speaker! Notice that I didn't say anything about the output transformer. But that's in the circuit too, you say. Right, I'll mention it now: No transformer known to man has ever been wound with identical parameters on each side of the primary winding center tap.

In point of fact, where all this is headed, is that we need only concern ourselves with the proper reproduction of the waveform as it appeared on the input grids, and compare that to how it appears at the speaker - what's happening on the plates, current-wise, is mox nix. Bottom line in English: if the signal's waveform doesn't match (except in amplitude), then there's been some distortion, and that's exactly what we're trying to avoid.

Now would be a good time to get down to how bias should be measured and adjusted. In fact, we don't even care about measuring it, per se, we just want to adjust it properly. (Remember Lecture #1 - proper = tube doesn't blow up. : What you need for this is one each of the following: a signal or frequency generator in the audio range, an oscilloscope, and a dummy load.

The dummy load is easy, in fact the best kind for our purposes is nothing more than an ordinary household light bulb. Just be sure that it can handle the projected wattage, even though we won't be encountering 110 volts at the output of the amp. You can also use one or more power resistors making up the proper ohmage, but that isn't the same thing, it's not inductive like a speaker is. Still, when all you've got are resistors.....

A signal generator can be cheap, but it has to have a fairly clean output, distortion wise. If you can get specifications for it, it must have less than 1% distortion, anything greater than that is gonna make life miserable. And the o'scope can be almost anything, even cheap used ones will do. Just have good quality leads, and you're set to go.

Now, from what all I've said above, it should be easy to guess that we're gonna set our 'scope up at the output of the amp. Right. Hook up the dummy load, and place the scope leads across that. MUY IMPORTANT! Make sure the scope's ground lead is hooked to the same side of the dummy load as the one going to the amp's chassis. It wouldn't do to short the amp's output to ground at this point in the game, now would it? The scope should be initially set to 5 volts per cm, and the time base to 0.5ms.

All the amp's controls should be turned down. Feed a 2KHz signal at about 1/2 volt from the generator into the amp from the generator, and turn the amp on. Crank the treble control half way up, and start cranking the volume up, slowly. See anything on the 'scope yet? Slowly work the controls up until the signal on the 'scope takes up the majority of the screen. Adjust the sensitivity and the timebase to suit your preferences.

You're looking for two things. Make sure that at all times, the signal peaks remain curved, not flattened out. As you progress along, cut back on the volume, or the generator, to keep the peaks curved. The other thing we want to see is that the crossover point is not distorted. More on that in a moment. At this point, if you've not yet seen anything, then you've got a problem with your lash-up, or your equipment. Check it out.

Adjust the volume and the generator output for the maximum signal you can get, without flattening the peaks. Keep doing this as you adjust the bias, the two will changes will affect each other. Which also decries the ability to use a current measurement, particularly the quiescient current at idle. Understand this: input voltage greatly affects output current. That's the whole operating principle behind how tubes operate. Just ask Lee de Forest. (What was that sound? Could it be a punctured balloon?)

And voila, here we are, at the crux of this whole discussion. Bias is what affects the crossover point. All the math and design parameters aside, if you have too little bias, you'll see this point on the waveform seemingly "go horizontal". I mean, it will start to twist away from what it should be, in the horizontal direction. Not by much, but noticeable. Too much bias, and the waveform will twist toward the vertical. Why? Don't worry about it right now, what's important is to make that crossover as smooth as possible, given the rest of the waveform's overall curvature.

Now, the why of it is this: You are adjusting to prevent over-amping the waveform (both tubes conducting on the same portion of the signal at the same time, which for our purposes is bad juju), and to prevent cutting off the signal before it has indeed been amplified at all by either tube. Sounds easy, eh? It's not. It takes a little practice to see the exact effect of changing the bias on that little portion of the waveform. Perhaps your 'scope has a magnifier function, that'll help.

Now, the 64 dollar question. How important is all this exactness? Short answer: It ain't. You can be off by maybe 5%, and still be so close that you surpass hand grenades for accuracy. Here's another thought. (This isn't my original thought, sorry to say.) The amp belongs to the owner, and that owner can have whatever he or she wants. If they want to blow up tubes every week, so be it, set the bias the way they like it, and replace the tubes whenever they darken your doorway. The moral here is, the check's the same at the end of the week! ;D Ditto for biasing the other way, but those folks won't be back as often.

Now, who's up for a beer, all this typing has made me thirsty! ;D


sumgai

p.s. 1366 words!

[Runewalker]

(Dedication - this is for Runewalker, it's actually a carryover from another thread.)

I'd like to thank the academy and ......

This is pretty deep for a finance guy. Can we talk about net present value and efficient markets?

This is going to take some study over some serious coffee.

[sumgai]

Rune,

(Dedication - this is for Runewalker, it's actually a carryover from another thread.)

I'd like to thank the academy and ......

This is pretty deep for a finance guy. Can we talk about net present value and efficient markets?

This is going to take some study over some serious coffee.

Sorry, coffee not included in lecture fees. Please deposit an additional one dollar.

;D

.
.

OK, here's what we've covered so far. The easy way, just set the voltage to whatever the schematic reads. The current draw methods, both quiescient and full load, and the one true method of using an oscilloscope to inspect the shape of the waveform. You choose which seems to be easiest for you, or the best, if not the easiest. The bottom line in my previous post reveals that I'm being overly hard-assed about the whole thing, and I admit that. But now that I've done my duty (exposed the BS about current measurements), I can stand down and let you take whichever course catches your fancy. Just get within reasonably close, and you shouldn't have anything to worry about.

Any questions? Oh, that was a silly question on my part, right? Someone here will shed further light, if I don't get back to you immediately. (I anticipate another good day for outdoor work tomorrow.)


sumgai

[Runewalker]

I appreciate this effort. Running accross the ocilloscope, dummy load and tone generator (which no one is going to accuse my playing of being) requirements means no go. Guess I have to rent amp repairman storage fee. D@mn.

I actually think I could do this, but being fresh out an occilloscope kinda stymies this ambition.

I'd like to get her going again (Super 6). She is missed. I guess I bite and putit in the shop --- since I am paying for an hour of shop time anyway, have the dude put a half power switch in. This sucker is loud.

On a half power switch, what actually happens? Do the other 2 power tubes cut out simply idle, or they pumping just as hard but shooting blanks, so to speak.

Is the wear on the inactivated tubes the same. Is there benefit or a way to alternate which 2 power tubes are set aside.

I know some amp dudes admonish against this because the transformer need all the current to drive to break up, or something like that. But this amp is used as clean overhead to be driven by a modeler, so I don't seek normal breakup.

RW

[sumgai]

Rune,

There are several methods of "cutting power" with a switch. The fastest, and cheapest, way is to do as you suggest, somehow cut out two of the four tubes. That can be done by "lifting" the cathode (disconnect it entirely) on two tubes, or else disengage the grid drive signal from two tubes.

More thoughtful designs will let you designate which pair to disable, but the vast majority don't go that far. I've even seen one design that automatically alternated between the two pairs each time you flicked the switch!

But better yet, however more expensive, would be to tap the power transformer at a lower voltage level. Some folks will install power resistors of large wattage and somewhat low resistance - not elegant, but in some cases, it works. The idea here is that all four tubes remain in the circuit, and operating at the same levels. Which brings up the hidden problem with the easy methods I listed above.

The problem with cutting off the grid drive to two tubes is similar to the above. Even though plate current is flowing, it's quiescent, not full-out. That also creates an imbalance, as the full-power tube is now loaded down by the idle tube, which translates to less power available to go to the xformer. (A bit closer to the arguments of the "maximum drive" crowd.)

Now, taking everything into consideration, is there truly a downside to using the easy methods? Well, beside the tone issue, you run a low risk of shortening the xformer's life. Quite low, but not zero. The tubes themselves, of course, are now being heated up, but with no work being done, that's a waste of potentially useful tube life. And of course, there is a small possibility that running those extra wires to the grid might somehow introduce a stray capacitance, thus causing parasitic oscillation, a condition that plagued Leo in his early days. (Now you know exactly why Fender lays out their (hand)wiring like they do, to avoid this kind of thing.)

Downside for the tapped transformer? None that I know of. It's a properly designed part, and providing it's installed correctly, then I don't see any shortcomings. BTW, this is why you see amps like Mesa and MusicMan use a 3T switch, they force you to turn the whole plate supply off for a moment (passing through the center-off position). This prevents nasty shocks to the tubes themselves as they transit from greater or lesser voltage, and it also protects the xformer from any backlash from the tubes as they fight to recover from that shock. Complicated logic, but it works better that way.

Writing a book again, ain't I? Time to let it go. Hope all this helped.

And yeah, get the thing biased while it's in for the upgrade. If it's been sitting for some time, either check the tubes, or else just replace them summarily. Best to start out with fresh stuff, particularly in light of what you're about to do (modify the supply voltage to those power tubes).

OK, time to go, ciao.


sumgai

(Edited to remove slightly faulty math, which in turn was more fully explained in several posts below.)

[ChrisK]

Yow.

As to

Well, ackshally,

One might surmise that the best way to set power tube bias is for equal quiescent current thru each tube. This effects DC balance in the output transformer.

The was my best guess at a puzzled look.

I was the one surmising.

I'd read a post about the Fender Blues JR that talked about the factory bias being so hot (not negative enough) that the EL84's were idling beyond their rated power dissipation. Not only is tube life affected by excessive power, but the output transformer response is affected (in a push-pull topography) by unequal (non-net of zero) winding currents that cause asymmetrical core saturation. That's all.

For a push-pull topography, buy a matched set, set the bias voltage as recommended. You're probably close enough.

Times are difficult, I haven't seen a drug store tube tester for a couple of years now.

I'd check the biased plate current first, but that's just me wanting to know things. (We do care about the plate currents, but only indirectly, as far as it being "centered" such that our response is fairly linear.)

A method of bias adjustment that was centered on current measurement is fairly difficult (and dangerous, you're messing w/plate voltages and not"just" bias voltages).

But, in the end, some are also (somewhat) interested in testing where we are on the linearity (current transfer function) curve eh?


As to reducing the output power, Mesa has a few ways such as pair cathode switching, triode/pentode switching, mixed classes (tapped output transformer, mixed biases), and pair substitution. Fender, in the Pro Tube series switches the power amp supply (from a doubler? to a straight rectifier) and the bias network.

unklmikey,

Thanks for the bump, I'd forgotten about my innocent surmise....

[sumgai]

Chris,

Yep, we are interested in how much current is being drawn at full blast, for just the reason you mentioned, making sure that we're somewhere near the center of the projected transfer curve. That becomes much more than "just interesting" when we can't adjust the bias to give us a correct-looking waveform.

Your safety precautions are well worth repeating. I hereby incorporate them by reference! ;D Reader, go look at them again, they're not that far up the screen!

I'm not too sure yet just how all of these new-fangled "power adjusting" circuits will work out in the long run. You can pair things up, split things up, change rectifier types, all kinds of things, but I pretty much believe that the best method so far is to change the voltage available to the power transformer's input side. Mind you, I'm not 100% sold on this being the only worthwhile method, but my "think by exception" mindset says that all the other methods are either hokus-pokus, or else just hype. The number 1 question in my mind is "how are these methods gonna affect the life of the components"... ultimate tone is number 2 for me. (Again, I'm thinking like a technician who has to keep his customers happy, so this is a part of that equation.)

I also believe that to do the job correctly, there should be a coordinated effort at controlling not only the voltage, but keeping the current draw under control (reference your post above), and keeping the transformer within specs vis-a-vis the impedance it sees. All of these things can be done, but not for free. The user will either pay the price at the cash register, or else in not-quite-100% satisfaction with the resultant tone.

But that's just my fuddy-duddy way of thinking. I'm hoping to learn more about how we can get something, apparently for nothing, new out of our equipment. Keep those cards and letters coming, folks! ;D


sumgai

[Runewalker]

sumgai said:

Rune,

There are several methods of "cutting power" with a switch. .... More thoughtful designs will let you designate which pair to disable, but the vast majority don't go that far. I've even seen one design that automatically alternated between the two pairs each time you flicked the switch!

But better yet, however more expensive, would be to tap the power transformer at a lower voltage level. Some folks will install power resistors of large wattage and somewhat low resistance - not elegant, but in some cases, it works. .....

sumgai

Well Since I am unlikely to do this myself unless I had extensive pix or drawing and a how to, What specific in structions would I give the amp tech on installling a half power function on my Twin Brain?

I am not sure how you would do that second option of the tapped pTrans. What would I tell the tech to have him execute this specific change? Realistically, what is a reasonable $ for this change?

Is there a downside to the "...I've even seen one design that automatically alternated between the two pairs each time you flicked the switch!" option. That does not sound too terrible, and may be more affordable. I don't gig so I get a long life out of tubes.

again on that option, what would be my specific instructions.

Torrez has had this little switch and instruction set for years:

store.yahoo.com/torresengineering/50watswit.html

So I was thinking in that direction. But your idea sounds more complicated, and of course accompanied by more expense.

RW

[ChrisK]

I do have the "Variac" function available in my Mesa Mark IV. Somehow, reducing the overall supply voltage just rubs me the wrong way. The purist in me wants to reduce the plate supply (and bias voltages as the model scales), and not the filament voltages.

But, how to scale the impedance. Fenders approach in the Pro Tube series seems more "sane" than just switching out the cathodes on a pair of push-pull drivers. They reduce the plate supply (by 1/2), the plate current (by 1/2), the plate headroom, and the bias voltage. Hmmm, for half of the original supply voltage with half of the original plate current going thru the SAME load impedance (realizing half of the original voltage swing), it sure seems like 1/4 power to me!!!!!!!!

Aaahhhh!!!!, an engineered approach. ;D ;D (I'm soooo happy.)


Ya know, I almost bought the Pro Reverb 50/12 Watt model last month for $850 new. It's a discontinued model (I must test their limits/understanding of debt service on a few year old amp that they've carried at 1/2 to 1 points a month...). Somebody's always the bank.... But, somehow, a subset of a Mark IV for the same coin (to me) and the same weight......

I really liked the amp's concepts. On the Pro Tube series, one can use the tube effects loop as an additional high gain stage. But, it didn't have a tube rectifier (somebody "shun the vintage out of me" please). I'm this moron that thinks that he needs a '65 Deluxe ('cept it has no middle tone or presence controls, and runs at 22 Watts, so it's still too loud).


Oh wait, that's why I bought the Vax Valvetronix AD50 (OK, now where did I leave it?). And, I just found a very lightly used Fender Pro Tube Concert Reverb for $700. Wow, 'cept it's 100/25 Watts (and that tube rectifier thing).

Hmmm, don't I have a Mark IV? What am I looking at tube amps for anyway?

Oh yeah, Bright Shiny Objects. And then there's the Mesa Lonestar Special. (What, $1,800 for a 5/15/30 Watt amp? This definitely isn't "for the house or kid".)


Anyhoo,

Go to ww.uspto.gov and search for Smith, Randall (the Mesa guy). Many other works (including Leo's) will show up as cited art. There will be a link for a couple of free tif viewers (all uspto online docs are such format). Get one and begin to read away. In many cases (Mesa), simple is.

I really want to find the application (if published) of Mesa's single-ended/push-pull patent filing, 'cuz if it's as simple as I think (based on their history), I've got a really neat one...


But, back to topic, this recent thread has been MOST informative for me. I "let go" of tube technology back in the late 60's, and have been "astudying" my "recent' (1971) RCA tube manual that I'd picked up whilst boringly self-incarcerated in the Army at that time.

My surmise again was related to my ancient understanding of tubes as transconductance modulating devices (current "valves", sometimes the Brits are right on descriptively) and missed the practical bent of this thread.

Remember, at the time, the focus was on actually "improving" tube amplifiers for actual fidelity (never mind efficiency). We all stopped when the Solid State's came along 'cuz they was better at it.

I do like tube amps, even though they're still lousy. It's like buying a '56 Chevy 'cuz ya like to adjust valve lifters.....

[Runewalker]

CK:...."... It's like buying a '56 Chevy 'cuz ya like to adjust valve lifters....."

More like, because you can. Buy an Impala today and try adjusting the carb. Oh, wait, there isn't one.

[ChrisK]

I learned most of my "French" from watching my father adjust lifters......

[ChrisK]

Soooo,

The Pro Tube Twin (100/25 Watts) DOES NOT share a common output transformer with the Concert or Pro Reverb (both 50/12 Watts).

The Pro Tube Twin has its 2 each (8 Ohm) speaker load connected in parallel to a 4 Ohm tap (uh, like a twin or something). It has four 6L6's.

The Pro Reverb has its 1 each (8 Ohm) speaker load connected to a 8 Ohm tap (uh, like a single or something). It has two 6L6's.

If'n I recall, the load presented to the plates by the output transformer is Zspeaker * SQRT[turns ratio]. Since the turns ratio doesn't change much dynamically ;D, I might surmise that simply by removing two 6L6's and one (assumed parallel) speaker (using the SAME 4 Ohm transformer tap), I gots a Pro Reverb (more or less). The ratio's the same, the impedance presented is twice, but the drivers are reduced by two for the SAME individual tube plate load/current. Zingo bingo, a (near) new 50/12 Watt Pro Reverb Twin, reversible back into a 100/25 Watt Pro Twin.

If I like having two speakers (uh, a twin), I could place both 8 Ohm speakers in series and connect them to the 8 Ohm tap.

Are drugs in play here?

[sumgai]

Chris,

Nope, no drugs were allowed to come out and play here, your math is correct. I even did some old-fashioned calculating before I started to answer this post, just to be sure. (I'm learning not to get too jumpy here. )

My only question is, where did/do you see a Pro Reverb with only one 8 ohm speaker? My book, and the AmpwaresFFG backs it up, says that the Pro Reverb is a 2x12", 40 watt begger that is essentially a Twin Reverb on anemic pills (it's the same weight, 60+ lbs., but only half the power). The speakers therein are the exact same as the Twin's, 8 ohms apiece, so the output is nominally 4 ohms, also just like a Twin Reverb. I'm curious to find out if Fender has come out with yet another model name that subsumes the old model design.


sumgai

[ChrisK]

The Pro Tube series is mostly discontinued now. I don't see the Pro Reverb (50/12 Watt 1 Jensen C12N speaker) or the Concert Reverb (50/12 Watt 4 10" Eminence Special Design speakers) in the 2006 Fender catalog. Only the Pro Twin (100/25 Watt 2 12" Eminence Special Design speakers) remains (page 190).

The 2002 and 2003 catalogs have them therein. Since things "evaporate" so rapidly from the Fender site (and catalogs year to year), I've learned to keep the paper stuff.

I suspect that the Super Sonic is meant to replace them.

Mr. Gearhead has the schematics for all three. It's the same pdf, just with different names. Essentially, tube reverb, tremelo, effects loop, and a gain channel. Solid state rectifier (and some other stuff), and modern engineering (I hope).

[UnklMickey]

OK guys,

you knew i would eventually ring-in on this thread.

unfortunately, it necessarily has to be in wet-blanket mode.


did anyone even bother to look at the diagrams for a super6 ?

(you know, the amp RW is using. the guy who this thread is "dedicated to")

it would appear Fender is less concerned with the absolute bias currents, and more with relative currents. there is only a "bias balance" control.

don't get me wrong, i don't always share Fender's design philosophy. but unless RW is going to be doing some serious modifications, that's the only adjustment available.

the "crossover-notch" method of bias adjustment is commonly used in the earlier versions of the "twin" topology. (AB-763^* or AC-568). although some very respected amp gurus recommend against it:

www.aikenamps.com/CrossoverNotchBiasing.html

as i mentioned before, that isn't even possible in RW's amp, unless some mods have already been done.

so RW, what will your amp tech be charging you for, other than that "amp repairman storage fee"?

worst case, he'll set the pot so that each pair of output tubes will have the same bias voltage on the grids.

best case, he'll adjust the bias balance so that whichever pair naturally has less current in it, will have a lower voltage on it so that the currents will be equal in both sides of the output transformer.

which or these courses will he take? i don't know, but i'd certainly want to know what i was paying for. Questions ARE free. (although answers sometimes are available only for a fee.)

if you have him install separate bias controls for all 4 tubes, (and a small resistor in series with each cathode)you could do a more accurate (and difficult) job of biasing.


as far as ' "lifting" the cathode (disconnect it entirely) on two tubes', this could lead to emission from the heaters to the cathodes. "bad juju" indeed!

if you choose to disconnect two tubes, a high-quality switch or relay, to disconnect the plates and screen grids of two tubes (1 from each side of the output transformer) would be possible. just make sure that the voltage ratings on the switch are far in excess of the supply voltage.

decreasing the B+ voltage (but not the filament voltage), whether by resistors, zener diodes, transformer taps, or using a tube in series (as a voltage regulator) seems like a better idea. BUT, a large reduction in B+, will require a lower bias voltage on the grids. else, the tubes will be operating too cold, and might even be operating in class C mode. (nobody wants that!)

you can avoid that by changing to cathode bias, instead of grid bias, but this thing is starting to look less and less like a super6, the further we go.



unk




* the AB-763 designation was also used for some Deluxes (6V6 pair) and Supers (6L6 pair), as well as the Twins and Dual-Showmans (6L6 quartet)

[ChrisK]

Oh, and BTW,

On the Pro tube amps, when you adjust the bias, since you're looking across the cathode resistors to common, you're setting the bias current (w/o caring about the bias voltage). Knowing the cathode resistances (which ARE NOT on the schematic) one can choose idle current points. This is a real safe way to do it in that the voltages presented at the external bias measurement jacks are clamped to a maximum of one power diode drop positive (and whatever the cathode resistance is as a negative impedance to common [ground is that round lug in the wall plug]).

Fender want one to adjust for 60 mV on the Twin (4 each 6L6's) for each side pair. Sounds like aboot 1 Ohm to me.

I will have to address the 100K vs 220K bias resistor issue in the conversion to 50/12Watts (seems like some (?dynamic) grid current consumption issues, eh?), and possibly the cathode resistors.

[Runewalker]

unklmickey said:

OK guys,

...did anyone even bother to look at the diagrams for a super6 ?
...
it would appear Fender is less concerned with the absolute bias currents, and more with relative currents. there is only a "bias balance" control.

...unless RW is going to be doing some serious modifications, that's the only adjustment available.

so RW, what will your amp tech be charging you for, other than that "amp repairman storage fee"?

worst case, he'll set the pot so that each pair of output tubes will have the same bias voltage on the grids.

best case, he'll adjust the bias balance so that whichever pair naturally has less current in it, will have a lower voltage on it so that the currents will be equal in both sides of the output transformer.

...if you have him install separate bias controls for all 4 tubes, (and a small resistor in series with each cathode)you could do a more accurate (and difficult) job of biasing.

if you choose to disconnect two tubes, a high-quality switch or relay, to disconnect the plates and screen grids of two tubes (1 from each side of the output transformer) would be possible. just make sure that the voltage ratings on the switch are far in excess of the supply voltage.

decreasing the B+ voltage (but not the filament voltage), whether by resistors, zener diodes, transformer taps, or using a tube in series (as a voltage regulator) seems like a better idea. BUT, a large reduction in B+, will require a lower bias voltage on the grids. else, the tubes will be operating too cold, and might even be operating in class C mode. (nobody wants that!)

you can avoid that by changing to cathode bias, instead of grid bias, but this thing is starting to look less and less like a super6, the further we go.

unk

Thanks Unk.

The guy with the 3 week-wait-unless-you-want-to-expidite-$$ was going to convert the ground toggle in the back to the half power switch. I am sure he was going to just do it the easy way, and the proper ways you describe.

That 1 to 4 bias switch idea may be overkill for this since it worked fine and would that produce "hearable" differences?

if you choose to disconnect two tubes, a high-quality switch or relay, to disconnect the plates and screen grids of two tubes (1 from each side of the output transformer) would be possible. just make sure that the voltage ratings on the switch are far in excess of the supply voltage.

When you do this are the two disconnected always the same, and what are they doing -- idling, running just as with no throughput contribution. Do the disconncected tube wear at the same rate. How does one rotate pairs?

BUT, a large reduction in B+,

What is large? Half power, 3/4s power reduction?

The end point objective started simple here:

• New power tubes
  
• Necessary bias adjustment for the new tubes
  
• a Half-power switch.
  

No I don't want to go nuttzzo on modifications. I am not sure the amp is a good enough base for that. It does what it needs to do, and is a clean loud driver. The very thing that a lot of guitarists don't like about Silverfaces, but is perfect for my application.

Were I to throw money towards an amp I would move to a lower power, more transportable amp. I wish frankly that I could find a solid state amp that did not add that wierd crackly, Beatles-Revolution-drive-the-sound-engineers-nuts-by-pegging-the-line-in-guitar sound to the modeler.

The tubes sort of embrace the modeler kid, sand its edges and warm it up nicely.

None the less the thread has been instructive, but not yet directive in a way that someone without this electronics education (awwwwe hllll, without any education) can take and make operative, yet.

No one had a comment on the Torrez $15 half power mod.

I see myself going to the tech and reading a specific list. Is there a way to give specific instructions that does what Unk said, "if you choose to disconnect two tubes, a high-quality switch or relay, to disconnect the plates and screen grids of two tubes (1 from each side of the output transformer) would be possible", and has one of those switches that rotate the pair dissconnected?

[UnklMickey]

"That 1 to 4 bias switch idea may be overkill for this since it worked fine and would that produce "hearable" differences?"

it would be the best way to extend tube life, while maintaining low distortion.

it would also be a royal pain to adjust.



"What is large? Half power, 3/4s power reduction?"

at half power (but that's way louder than half volume*), there would probably be a need for re-biasing.



"No one had a comment on the Torrez $15 half power mod."

sorry, don't know what they're doing. could be same as i suggested, could be disconnecting just the plates, could be disconnecting the signal to the grids of 2 of the tubes, could be disconnecting the signal AND tying the grids to the -63v point of the negative supply. "there's more than one way to ........."


if you have a matched quad, i'd probably do what 80% of the owners do.

(or what 25% of the guys who took their amp to a tech got.)

set the bias balance to the middle of it's range, and leave it at that.


you could do the tube rotation manually with any of these schemes.

use an on-on-on switch to select pair A \ both pairs \ pair B.

yes tubes do age faster when you drive them, especially when you drive 'em hard.


you know, Goldilocks, in this case "just right" sounds like a Deluxe or Princeton sitting on top of a 4-12 cab.



unk


* half volume = 1/10 power.

[sumgai]

Wow, what a lot of traffic, of a sudden!

Unk, I didn't notice anything smelly being rained on the parade here! ;D What you did was simply introduce ever more possibilities to the mix, that's all.

As I recall, the designations for Fender's schematics were not based on an amp model (after 1959), they were given basic pattern names and a date code. AA was the first model, AB the first revision, AC the second, and so on. 763 means July of '63, and so forth. You'll find that all of Fenders amps from the 60's use these letter designations, at least, I've never seen anything like AD, or BA, or anything else. CBS Fender stopped using this method sometime after 1970, I'm not sure just when. They simply dated their schematics in the name block.

The Super Six was merely a Twin Reverb in different clothing, to be sure. Fender didn't change over to the "bias balance" system until the mid 70's, as I recall, so since we didn't ask Rune what year his beast is, that bit of info is up in the air.

Be that as it may, whenever I encounter one of these things, I always ask the owner if they'd like me to change the bias pot over to the original Fender design. Most of them, of course, look at me blankly, and I explain to them what you just said. Fortunately for me, if I get the go-ahead, the pot is the same so I merely swap around some connections, remove the excess parts, add a 15K or 22K ohm resistor, and it's business as usual on the ol' workbench.

Beyond, that, I'd like to offer up one solution to the low-power switch scenario. Rune, unk asked you what your tech is gonna do. Since you don't know, it could be something as innocuous as installing the Torres kit, or it could be something really elaborate. But where it got interesting for me was when unk said "reduce the power to the tubes (my favorite method) and if you drop the voltage too far, then you'll have to re-bias to keep everything ship-shape. That's true. But if you're gonna use a relay anyway..... why not have it switch between two separate bias controls?! Low power bias control, high power bias control, what's not to love here? ;D


sumgai

[ChrisK]

Check out the Pro Tube Schematics re

"reduce the power to the tubes (my favorite method) and if you drop the voltage too far, then you'll have to re-bias to keep everything ship-shape. That's true. But if you're gonna use a relay anyway..... why not have it switch between two separate bias controls?! Low power bias control, high power bias control, what's not to love here?

Fender switches the plate supply to half, and the bias voltages to boot.

Switching the plate supply is done using a full wave bridge for high voltage, and using a switch pole (the other pole is for bias) to convert the rectifier topology to full wave center tap (which is less transformer efficient, but ok since the power is reduced).

[JohnH]

When I first scanned through this thread a week or two ago, I got that same feeling as when I attended lectures in non-linear finite element analysis at University - I couldnt understand it at all! The tube part of my electroics course at high school got canned the year I did it - it didnt seem like they would be much use in the 21st century.

However, I've had another go, since I'm about to buy into this technology and I like to understand what is happening under the hood. My fiddling with non-lethal JFET circuits has given me a grounding. once I translated the buzz words.

So if you will indulge the following, which is much more speculative than most of my posts:

RWs super 6 appears to be this schematic:
www.ampwares.com/ffg/schem/twin_reverb_sf_100_schem.jpg

Its dated 1971, so seems to be a good candidate.

Its clear on that , as Unkl points out, that there is no bias adjustment, execpt for a balance between each pair, since the bias is taken from a fixed bias voltage without, adjustment with a pot. That would mean to me that when replacing with new tubes, maybe made on the other side of the world, that there is some risk, even with a matched quad, of not having an appropriate bias current, due to variations between old and new tubes. Would it be good to install cathode resistors to have a easy way to check this? and then if necessay, put some control on the bias voltage?

The four power tubes seem to be two pairs in parallel.

So next question is, to get a half power option, what is wrong with just pulling two tubes? (RW - I suspect that you rarely need 100W - except after a very bad day in the office)
It would seem that all voltages and biasing would remain the same? It looks to me that that is better than partly disconnecting tubes, leaving heaters still running, and some connections still made.

Next - take a look at this. I nearly bought one of these Marshall 9005 power amps yesterday, but someone else wanted it more. What I thought was good with this circuit was the bias setting pot, and also the cathode resistors to allow each bias to be measured. Most interesting though was the triode/pentode switch, which cuts power from 50W to 25W, in a very simple looking way

www.drtube.com/schematics/marshall/9005.gif

Could a switch like that work with the Fender circuit?, hence allowing options down to 25W, given two tubes removed?

Thats all for now

cheers
John

[UnklMickey]

johnh said:

(1)
...Would it be good to install cathode resistors to have a easy way to check this? and then if necessay, put some control on the bias voltage?

(2)
The four power tubes seem to be two pairs in parallel.

(3)
So next question is, to get a half power option, what is wrong with just pulling two tubes? ...

(4)
...It would seem that all voltages and biasing would remain the same? It looks to me that that is better than partly disconnecting tubes, leaving heaters still running, and some connections still made.

(5)
...Could a switch like that work with the Fender circuit?, hence allowing options down to 25W, given two tubes removed?

John,

lots of questions and observations there.

i've paraphrased and numbered the quotes.



(1)
using small cathode resistors to monitor the currents in each valve* is a great way to go.

using 1 ohm or 10 ohms in series with each cathode makes it easy to calculate the current. even at 10 ohms, there is only 1 volt per 100mA, so it won't effect the maximum output much.

i think a good way would be to use 4 resistors and 4 sets of test points.

if you wanted to read all 4 currents simultaneously, you could. (with 4 meters)

using a DPST, you could parallel the resistors on the cathodes of the valves that are on the same side of the ouput transformer.

then you would be read the average current for the valves on each side of the transformer.

having both overall bias voltage and bias "balance" would be my preference.



(2)
yes



(3)
nothing wrong with that, other than: Americans are lazy.

we want all options, and we want them a the flick of a switch.

it also makes some of the "wear" issues moot.

remove all 4 valves (numbered so they go back in the same sockets).

install a separate matched pair, of the same brand.



(4)
those are some very strong arguments in favor of the simple approach.

i assume you are a native Australian.


(5)
makes sense to me.

i also like the idea of being able to switch down the B+ to the output valves and the bias voltage.

that allows a greater reduction in power.

having a 100w / 5w choice would be Super






unk

*language courtesy returned.

[sumgai]

Ya'll know, one of the big reasons I stopped doing all this low-power stuff for myself was, I finally figured out that I was hauling around the world's heaviest 25watt amp! Seemed to me that there was a simpler way to get good tone at low power, without the associated hernia.

The time honored method of doing that, the one most likely to win everyone in this thread's squeal of approval - buy a Deluxe Reverb, and be done with it. Fake it if you have to (off brand, or SF, whatever), but a quarter of the power, less than half the weight, and all of the desired tone!

Done deal.


sumgai

[ChrisK]

Yeppers on the Deluxe Reverb!

Or for half the price, the VOX AD50 w/ the Celestion speaker ain't bad since it models, has an actual tube, and allows fully saturated power amp tone from 50 to 0.15 Watts. It ain't a deluxe, but it ain't bad either.


(Gee, I GeWonder if'n the Deluxe will work w/ EL84 adapters.....)

[JohnH]

chrisk said:

Yeppers on the Deluxe Reverb!

(Gee, I GeWonder if'n the Deluxe will work w/ EL84 adapters.....)

That was another thought. There are THD Yellow jackets, that claim to provide an adapter from the original tube socket, to an EL84, and include self biasing and control circuitry in the socket adapter, so it is just plug and pray. Anyone tried them?

John

[UnklMickey]

i haven't but i can guess what they're doing.


they claim to be converting from class AB to class A operation.

that's a bit of of a misnomer, but i won't go into that.

that suggests to me that they are doing cathode biasing.

the cathode resistors are probably bypassed by capacitors to keep the AC gain high.

this is much more "forgiving" than grid biasing.

it's also often referred to as "self biasing".

likely the control grids are referenced to ground via the cathode pin of the octal plug using a resistor (maybe about 200k).

the control grid bias from the amp would be blocked by a cap in the yellow jacket.

after making that guess i went to the tube depot website and read this:

"Converts from Class AB to Class A operation. Output remains about the same. They add extra current-limiting in the screen circuit - block the bias voltage and set-up a cathode-bias circuit - regardless of the bias configuration of the amplifier. "

which mostly supports my speculation.

doesn't mention anything about bypass caps on the cathode resistors.

also mentions current limiting for the screen grids.


since a matched pair of JJ EL-84s costs $16, and a pair of yellow jackets w/ JJ EL-84s costs $100, that's $84 for the convertors or $42 each.

if you knew what you were doing, you could convert an amp for a lot less than that.

you could even make your own convertors for a lower cost.

but time IS money, so it's not that bad of a deal.



also, by using convertors, you can always go back to the stock 6V6GTs

and since we're talking a Fender amp here, you would have seriously lost resale value, if you did an obvious mod that would have 9-pin sockets where those octals previously resided.

not such a big deal in a re-issue.

worse in an SF.

in a BF? ............... real bad idea!

so using adapters looks better and better, if you want to use EL-84s



unk

[Runewalker]

On the Delux or Vox lightweight options......

This amp is, in flight lingo, Fixed Base Operations Equipment. So transporting is not so much an issue, but functioning ear drums are.

Besides with a 6 speaker + 100W amp box I have my history of lugging large heavy "man's" equipment. Now that I have proven myself worthy, and if I were to do some of the mods discussed, I would pretty seriously consider mothballing the big box, and buiding separate boxed head and monitor. So in that that scenario, hauling parts would not be to difficult, should I haul.

RW

[Runewalker]

This is an old thread, and as usual with this group, ridden all over the range ....... but ......

the original notion of the thread was how to adjust the bias after changing the power tubes.....

so, I found this recently from a tube purveyor I prefer - Bob Pletka of Euro Tubes .....

www.eurotubes.com/euro-video.htm

Makes it pretty accessible. He does, however, like the simplicity of the bias probe style device.

Heres another more elaborate outline for some Carvin models.

members.dslextreme.com/users/rlhasse/CarvinMods/BiasingtheVintageSeries.pdf