Time for a tube discussion...

[slev]

Alright, I'm about to put some Svetlanas 6l6GCs in my amp for the first time and it has me wondering what guys are using in their power an per-amp positions and what differences you see among tubes an between amps. I'm putting the Svetlanas in my Mesa 5:50, it has the stock tube set in there now. I have EL84 JJs in the power section of my Mesa .22+ and it really smooths out the Lead channel. What are you PRS amp owners running your amps?

 

[]-[@n$0Ma☩!©]

I like old Mullard tubes in my amplifier's pre-amp section. They are hard to find these days - and expensive when you do.

For most other applications I like JJ's (Formerly known as Tesla). I've heard that [FONT=Arial, Helvetica]JJ tubes are made on some of the original European equipment that was used back in the day.[/FONT] Whatever the truth might be, they're affordable and they sound good (to me).

I stay away from Chinese tubes - though I'm told they're improving.

 

[slev]

I've used JJs a lot in both the pre-amp and power section of an amp running EL84s. I agree they are a nice all around tube. Never tried Mullards. I did have a few old GE 12AX7s that were awesome when I mixed them with some JJs.

 

[Em7]

Alright, I'm about to put some Svetlanas 6l6GCs in my amp for the first time and it has me wondering what guys are using in their power an per-amp positions and what differences you see among tubes an between amps. I'm putting the Svetlanas in my Mesa 5:50, it has the stock tube set in there now. I have EL84 JJs in the power section of my Mesa .22+ and it really smooths out the Lead channel. What are you PRS amp owners running your amps?

I would advise against installing non-Mesa-branded power tubes in a Mesa amp. While there is nothing special about Mesa-branded power tubes from a tonal point of view, they are special from a bias point of view. Mesa selects their power tubes to bias correctly in their amps. All non-Mesa-branded power tubes require one to re-bias the amp for proper operation. As the bias voltage on Mesa amps is permanently set at the factory, re-biasing a Mesa amp means modifying the bias supply voltage division network.

 

[László]

I stay away from Chinese tubes - though I'm told they're improving.

Actually, for several years after they ran out of NOS USA tubes, Two-Rock, who make fantastic amps, was shipping them with NOS 1980s Chinese preamp tubes along with NOS USA power tubes, and I have to say they sounded great. In fact, I bought some NOS smooth plate Telefunkens and NOS RCAs and tried them in my Onyx Signature, and wound up putting the Chinese tubes back in the amp.

I'm not saying that the NOS Chinese tubes were somehow better or worse, just that the amp was voiced for these tubes, and I wasn't as happy with the amp with the others. The moral here is that it can matter what the manufacturer chooses, especially in the case of smaller boutique shops. When Two-Rock ran out of the NOS Chinese tubes, they shipped their amps with more readily available tubes, and they still sounded every bit as good (I had a Custom Reverb Sig v.2 and v.3).

The Telefunkens and RCAs also sounded appreciably less good to my ear in my Roccaforte Rockie that shipped with JJs at the time. I wound up giving them to friends who HAD to have NOS glass in their amps.

Then again, I had a Bogner Metropolis that sounded lots better with NOS tubes. So it's not a bad idea to experiment, as obviously there are no rules!

 

[FrankiePRS]

I would advise against installing non-Mesa-branded power tubes in a Mesa amp. While there is nothing special about Mesa-branded power tubes from a tonal point of view, they are special from a bias point of view. Mesa selects their power tubes to bias correctly in their amps. All non-Mesa-branded power tubes require one to re-bias the amp for proper operation. As the bias voltage on Mesa amps is permanently set at the factory, re-biasing a Mesa amp means modifying the bias supply voltage division network.

Sorry, but that's as ridiculous as Randall Smith's "White Paper". It's strictly a marketing tool by Mesa. The bias on any tube amplifier changes from the moment you flip the power switch. Components degrade, plate voltage wanders, tubes wear down. Some Mesa amps do have a bias adjustment. For those that don't, particularly the MK IV, installing one can make a huge difference.

 

[slev]

So this is somewhat repetitive and less technical than Em7's post but it's my "tube amps for idiots" version of the same thing. Although, I agree that Mesa amps are meant to be used with power tubes that draw in a certain current range but, don't most of the more common brands of power tubes we all use fit that bill? I haven't used anything exotic and have never had an issue with non-Mesa tubes. Any feedback on that thought?

My read on the Mesa bias question is that the brand doesn't matter as much as that they are a matched set. Meaning the difference in the measured plate current for the power tubes is a small as possible (with in a couple of tenths is usually going to get it done). When you adjust the bias on an amp you are compensating for differences in the tubes. Hence, why most power tubes are sold in "matched" pairs or quartets and you are better off replacing them all at once in stead of one at a time. This is what Mesa is counting on and it actually makes a tube swap pretty simple for most of us. This doesn't apply to the pre-amp tubes for most amps including Mesas. This is why it's fine to use a cocktail of pre-amp tubes in different positions with in the amp. If you read what Mesa has put out on why they build a "fixed bias" amp it is essentially to make it low maintenance and robust, a Mesa hallmark, so that any nub can do the routine maintenance of a tube change with out the need of equipment to accurately measure plate current on a tube and then adjust the bias accordingly. This is why I really like Mesa. I've had several and I like to experiment with different tubes, never had any issue with any brand of power tube as long as I get a matched pair/quartet from a reputable tube retailer. Mesa tubes, like many others, are generally re-branded Russian tubes. Most manufacturers have a preferred tube type, Mesa likes Russian tubes, Bogner for example likes Svetlana power tubes and Chinese pre-amp tubes. I have the "Bogner" mix for their 6L6 Shiva in my Mesa 5:50 right now and I get a very passable Bogner hi-gain tone on the burn channel. Again, this is what I love about Mesa, change the tube mix and tweak away. I can also get a descent Single Recto tone by engaging the contour with my "Shiva" settings dialing. There are a lot of very useable tones in most of the newer amps that Mesa has produced lately. The multi-watt option adds to the mix as well.

I'm a Mesa fanboy so take all this for what it's worth.

 

[Em7]

Sorry, but that's as ridiculous as Randall Smith's "White Paper". It's strictly a marketing tool by Mesa. The bias on any tube amplifier changes from the moment you flip the power switch. Components degrade, plate voltage wanders, tubes wear down. Some Mesa amps do have a bias adjustment. For those that don't, particularly the MK IV, installing one can make a huge difference.

The bias points for self-biasing tube circuits change from the moment that one powers on an amp. They continue to do so until reaching equilibrium. The preamp tubes in most tube amps are configured to be self-biasing. The bias voltages for preamp tubes are created across cathode resistors. However, not all power amp topologies are self-biasing. In fact, most of the power amp circuits that are found in guitar amplification are not self-biasing.

Cathode-biased power amps are self-biasing because the bias voltage is developed across one or more cathode resistors. The control grids in a cathode-biased power amp are at ground potential, which makes them negative with respect to the cathodes because the cathodes are sitting at a positive voltage that is equal the resistance of the cathode resistor multiplied by the current being drawn through it.

For example, the typical Mullard Application Note-based cathode-biased 2xEL84 amp uses a 130-Ohm cathode resistor. EL84s are 12-Watt tubes. With a plate and screen grid voltage of 300 Volts DC (VDC), this topology will settle in with a cathode voltage of around +10.6 VDC. This voltage is the result of a combined plate (37 milliamps per tube) and screen (4 milliamps per tube) current draw of around 82 milliamps for two EL84s through the cathode resistor, which is wired between the cathodes and ground. Since the grids are at ground potential, they are effectively at -10.6 VDC with respect to the cathode.

If we examine the plate current curves on the EL84 data sheet shown below, we see that a bias voltage of -10.6 lies between the curves for -10 and -11 VDC. If we move to 300 VDC on the X-axis (Plate Volts scale), we see that the plate current curve for a single EL84 with a bias voltage of -10.6 pretty much matches are our calculation.

With real world tubes, the bias voltage will adjust up or down based on the transconductance of the actual tubes that are installed in the circuit (the transconductance of a tube is measured as the change in current-out over the change in voltage-in). Any attempt by a power tube to draw more current will be met with resistance because the power tube cathodes are connect to ground via a resistor. The voltage drop across the cathode resistor will rise with any rise in current, which, in turn, will cause the control grids to become more negative with respective the cathodes, resulting in a more negative bias voltage. If one examines that plates current curves on the EL84 data sheet shown above, one clearly sees that plate current decreases as the bias voltage is made more negative.


With a fixed-bias power amp, the cathodes are wired directly to ground and a negative voltage is applied to the control grids to set the quiescent current drawn by the power amp. There is no automatic adjustment up or down, which means that bias voltage has to be adjusted for different sets of tubes, or one risks running the tubes too hot (a.k.a. "cherry plating") or too cold. Most sane amp manufacturers insert a potentiometer into their negative bias voltage supply division network. This potentiometer allows a service technician to easily set the bias voltage for proper quiescent current draw when installing a new set of power tubes. However, Mesa has chosen to make the negative bias voltage supply fixed in their amps, which means that one must either use tubes that are selected to bias at the proper quiescent current with Mesa's selected bias voltage, or one must replace one of the resistors in the bias supply voltage division network with different value that results in the proper negative bias voltage being fed to the control grids. Another alternative is to replace one of the resistors in the bias voltage supply division network with a potentiometer that allows one to set the proper bias voltage without having to replace one of the resistors in the bias supply with every tube change. Replacing one of the resistors in the bias voltage supply network with a potentiometer is a very common modification for Mesa amps because it frees one from having to use Mesa-selected tubes or tubes that have been selected by another vendor to bias correctly at Mesa’s chosen bias voltage.

In the end, the reason why we have to re-bias the power section of a fixed-bias amp is because the transconductance of any given tube type differs slightly from tube to tube. If that were not the case, vendors would not sell "matched" pairs. The matching in this case refers to tube transconductance. The reason why me use matched pairs and quads is because we want the tubes in a push-pull amp to draw roughly the same current for any given bias point.

Finally, tubes as well as most amp components do drift over time. However, drift in a power tube is more likely to make it run cold than to make it cherry-plate. Tube life is substantially reduced when the plates are glowing.

 

[Em7]

My read on the Mesa bias question is that the brand doesn't matter as much as that they are a matched set. Meaning the difference in the measured plate current for the power tubes is a small as possible (with in a couple of tenths is usually going to get it done). When you adjust the bias on an amp you are compensating for differences in the tubes.

The bias pot on blackface Fender amps adjusts the bias voltage that is fed to both halves of the push-pull circuit (i.e., it is a negative bias voltage level control). CBS Fender converted the bias voltage level pot to a bias balance pot to accommodate unmatched tubes. However, it never really worked very well. The CBS Fender bias balance pot is the circuit that Randall Smith often references because the Mark Series Boogies were derived from the Fender AB763 circuit. With the CBS Fender bias supply modification, there is no way to adjust the bias voltage level for both tubes up or down, which is why one of the most common blackface mods for silverface amps is to convert the circuit back to being a bias voltage level adjustment. A proper bias voltage level adjustment pot allows one set the quiescent current point for any given set of tubes within the same family. It also allows one to fine-tune the bias point.

With that said, Fender and Marshall include a bias pot on all of their fixed-bias amps. Most of the boutique builders go a step further and supply bias test points in addition to one or more bias pots on their fixed-bias amps. If fixed-biased amps did not need to be biased, these companies would eliminate the extra cost.

In closing, let me tell you about a DC-10 that I fixed a few years back. The owner of that amp also thought that he could use the "cram and jam" method with another vendor's power tubes. He was lucky in that a screen grid resistor failed before the tube that arced-over could take out the output transformer. I brought the amp up on my metered autotransformer, and the plates on three of the four power tubes were glowing red before I hit 100 VAC volts. I installed a bias pot in the amp after repairing the problem, so that he could safely use non-Mesa tubes.

 

[slev]

In closing, let me tell you about a DC-10 that I fixed a few years back. The owner of that amp also thought that he could use the "cram and jam" method with another vendor's power tubes. He was lucky in that a screen grid resistor failed before the tube that arced-over could take out the output transformer. I brought the amp up on my metered autotransformer, and the plates on three of the four power tubes were glowing red before I hit 100 VAC volts. I installed a bias pot in the amp after repairing the problem, so that he could safely use non-Mesa tubes.

So the question that remains for me is have I just been luck with the non-Mesa power tubes I've been using?

I've generally used a "branded" tube set (Mullard, JJ, Svetlana, ect.) nothing exotic and my assumption has been the only real difference is whether they are originally Chinese or Russian (for recently manufactured tubes anyway). Again, my read on this from my 5:50 manual is that Mesa acknowledges this fact and is saying that Mesa branded tubes are truly matched sets and they have gone through the trouble of selecting only matched sets of high quality. If that's the case and I take care to only buy matched replacement sets of the usual origin what's am I doing wrong?

Just want to be sure about what I'm putting into my amp and trying to absorb everything that's been written here.

 

[Em7]

So the question that remains for me is have I just been luck with the non-Mesa power tubes I've been using?

If you check quiescent current (current drawn by each power tube with nothing plugged into the amp) with a set of Mesa tubes and the quiescent current with another vendor's tubes, I guarantee that the readings will be different. Luckily, you have not received a hot set of tubes. You should still check for cherry plating. This test can be performed simply by turning the amp on, letting it warm up, and tuning off the lights. You should see no signs of glowing on the plates if the amp is running at a safe bias point. Often, guitarists will fail to recognize that the tubes are cherry-plating because only the very edges of the plates are red.

I've generally used a "branded" tube set (Mullard, JJ, Svetlana, ect.) nothing exotic and my assumption has been the only real difference is whether they are originally Chinese or Russian (for recently manufactured tubes anyway). Again, my read on this from my 5:50 manual is that Mesa acknowledges this fact and is saying that Mesa branded tubes are truly matched sets and they have gone through the trouble of selecting only matched sets of high quality. If that's the case and I take care to only buy matched replacement sets of the usual origin what's am I doing wrong?

Matched sets do not guarantee that the tubes will draw the proper quiescent current at a fixed bias voltage. What matching does is guarantee that the tubes can be used in a push-pull configuration without losing balance. A push-pull amp will hum if both halves of the circuit get to be too far out of balance. That's why the CBS Fender bias pot mod was called a "hum balance" pot.

The reason why we bias an amp is to set its operational class. Amplifier operational class is one of the things that guitar amp manufacturers like to shroud in mystery, but it is a relatively simple concept to understand. Amplifier class tells us how many degrees of the signal period each amplification device is conducting current. Here is a table that I created for another forum:

[FONT=Courier New][SIZE=3]Operating Class    Duty Cycle 

Class A            100% (current is always flowing through each power tube)
Class AB           > 50%, < 100% (current flows for more than 50%, but less than 100% of the signal period)
Class B            50% (current flows for exactly one-half of the signal period)
Class C            < 50% (current flows for less than one-half of the signal period)

[/SIZE][/FONT]

One often hears terms like "hot biasing" and "cold biasing." In this context, the words "hot" and "cold" refer to the amount of current that the amp draws with no signal. Class A amps are hot-biased, but not all hot-biased amps are class A amps. The VOX AC30 is an example of a hot-biased amp that is not a true class A amp (regardless of what Vox claims). It is merely a very hot-biased class AB amp. In fact, it is so hot-biased that the power tubes over-dissipate when driven hard. That's why AC30s eat tubes. .

Class AB and class B operation are the two most popular operational modes for tube-based push-pull amplification. With class B, one half of the circuit is on while the other is off. On the other hand, neither side of the push-pull circuit is completely off with class AB. The hotter we set the bias point; the longer each side remains on.

If one thinks back to high school trigonometry, one remembers studying sine waves. In class B, one half of the push-pull circuit amplifies the positive excursion of the sine wave while the other half amplifies the negative side. Because each side only amplifies 180 degrees of the signal, we end up with something known as "crossover distortion" where the signal crosses the zero point on the Y-axis.

In order to minimize crossover distortion, we bias the amp such that each half conducts current for more than 180 degrees. This biasing scheme is kind of like a relay race where the runner who is handing off the baton continues to run until the next runner is up to speed. In order to move from class B to class AB, we need to decrease the bias voltage. As illustrated in my earlier posting, a lower bias voltage results in a higher quiescent (zero signal) current draw. Ideally, we want to push the amp as far into class AB as we can without seriously degrading tube life. A rule of thumb for biasing a class AB amp is to bias it such that the amp idles at 70% of the rated plate dissipation. For example, a '65 Deluxe Reverb Reissue has a plate voltage of around 410 volts and 6V6-GTAs have a plate dissipation of 14 Watts per tube; therefore, using our 70% rule, we would adjust the bias voltage until the idle current drawn by each tube in a matched pair is roughly 14 / 410 * 0.70 = 24 milliamps.

 

[slev]

Good stuff, thanks. I actually followed most of your post. I will definitely check for the cherry plating effect you describe just to be sure. Is there any easy way to know that I'm not getting a "hot" set of tubes other than the effect you mentioned?

 

[Em7]

If you already own a multi-meter, know about electrical/electronics safety, and have a few bucks, I recommend purchasing an off-the-shelf bias jig (a.k.a. a bias probe). I built my first bias jig, but I purchased a single socket version of the bias jig linked below several years ago (they sold for $35.00 back then).

http://www.amp-head.com/product_info.php?cPath=21&products_id=70

What I like about this bias jig is that it allows one to safely read plate voltage and cathode current without pulling the chassis. One needs both readings to know if an amp is biased correctly (or at least in the safe zone). On amps that have non-adjustable bias supplies like most Mesas, having one of these bias jigs can give one insight into why one set of power tubes sounds better than another because the difference can often be traced to quiescent current draw at Mesa's fixed bias voltage. If you ever decide to have a bias pot installed in one of your Mesas, this bias jig will greatly simplify the biasing process.

With the above said, tube amps operate at lethal voltages that can reach out an touch you, and the power supply caps in many amps remain charged after AC mains power has been removed. When in doubt, you should have a professional service your amp.