Output Tube Biasing, An Introduction – Part 2
By Steve Graham
One thing to clarify from Part 1: when we speak of power dissipation, this is not the power output available to the speaker. Tubes are inefficient creatures; the output power will always be significantly less than the tube(s) will dissipate, i.e. consume. Think of dissipation as absorbed power, not all of it can be transferred to the speaker, the same way not all power produced by an automotive engine is transferred to the wheels.
The other common type of output tube biasing is cathode bias. This type of bias is sometimes referred to as self bias, or a few other “marketing speak” terms. I don’t mean to disparage cathode bias, but sometimes the language amp manufacturers conjure up does tend to cloud the fact that this method of biasing does not allow user adjustment.
Cathode bias also tends to run tubes a little bit less than the “magic” 70% dissipation discussed in part 1. This is to provide a margin for tube variability, so that no tube, within production variations, will draw too much current and therefore approach the upper limit of its power dissipation.
However, just because an amp does not have user adjustable bias does not necessarily make it a cathode-biased amp. A few amps, usually very pricey ones from high end manufacturers, have active bias monitoring and adjustment. These typically do not have user adjustments, but are fixed bias amps. The “fixing” is performed by a microprocessor without any user involvement. Other active fixed bias amps that are designed to work with a wide variety of output tubes may include a switch to optimise the “family” of tubes the end user prefers. One position will set the bias adjusting mechanism for EL34/KT77 tubes, whereas the other switch position might optimise for 6550/KT88 tubes.
Now back to cathode biasing.
Though less intimidating than the ARC amp pictured in part 1, the amplifier shown above can still bite if you are not careful. This amp is also pictured in My Audio Life no. 7. It’s based on a design published on the Angela Instruments website some years ago. The output section is shown in the schematic below.
This is a single-ended (non-push pull) amp that uses an EL34. Alhough it’s a pentode, the EL34 here is connected in triode configuration with a 100Ω resistor. If you compare this schematic to the one shown in Part 1, you’ll see that there is no separate bias supply and no variable resistor to adjust bias. The 500 ohm resistor (red arrow) connected between the cathode and ground, controls the current flow through the tube, thereby setting the bias. In this instance, calculating the current flow through the tube and the dissipated power is easily done – the circuit designer has thoughtfully supplied the nominal voltage and component values needed.
An inside view of the amp pictured above and in the schematic can be seen below.
Using Ohm’s law from Part 1, we can calculate the current through the tube:
I (current) = E (voltage) ÷ R (resistance), 28 ÷ 500 = 0.056 amps (56 milliamps)
Using the power formula from Part 1, we can calculate the tube’s dissipated power.
For the voltage we must subtract the voltage (see schematic) on the cathode, 28 volts, from the voltage on the plate, 375 volts = 347 volts.
Power = E (voltage) x I (current) 347 x 0.056 = 19.4 watts
Referring to the same EL34 data we used in Part 1:
19.4 watts calculated dissipation is about 60% of the maximum permissible, of 3 and 4 added together (32.5 watts).
So can we crank it up for better sound? The short answer is no, there’s nothing to crank (adjust). The longer answer is yes, if we want to change out the 500 ohm resistor for other resistors until the desired power dissipation is found. In this case, I wouldn’t start changing out resistors. The reason being, when pentode tubes like the EL34 used here, are triode-wired, they typically won’t dissipate as much power as they will in pentode or ultralinear mode. As well, production variations in tubes, as mentioned above, dictates a little extra margin of safety when cathode biased.
If you are the type that’s not bashful about using a soldering iron and are comfortable around high voltages, then changing out bias resistors is a potential possibility. In the example above, and assuming the power supply and output transformer is up to the job, it might be possible to swap out resistors to increase the current draw to work more optimally with a 6550 or KT120. On the other hand, say you have a favourite KT66 or 6L6 (their max. dissipations are less than the EL34), then by changing out the bias resistor the amp could be optimized for a bit less current draw on the output tube.
Now you might ask, “What’s the point of cathode bias if I can’t easily change it?” Well, cathode bias has the advantage of being plug and play. If you are the type that is disinclined to get involved enough to buy a multimeter and adjust things in a live component, then a cathode bias amp might be your best choice. If you have a cathode biased amp and you are inclined to measure it live, then you can check the cathode voltage (the 28 volt value in the schematic above). If this voltage is significantly different, say more than a 20% deviation from the nominal, then you might have an aged tube that needs replacing. Also, if there’s a large difference from one channel to the other, it’s an indication that one of the tubes might be out of spec. Running the measured voltage through the current and power calculations shown earlier, will give you a feel for the health of your tubes, assuming everything else in the amp is OK.
In the schematic below, both output tubes in this push-pull amp share one cathode resistor. This makes it difficult to determine if one tube is weaker than the other. The current draw through the resistor is the aggregate of the two tubes. In this instance, it’s best to buy a matched pair of tubes, and hope that they age in a similar manner. If you are inclined, measure the voltage across the bias resistor when a new set of tubes has been installed – well actually wait a few hours until the tubes have had a chance to settle in. From time to time check the voltage and when it has wandered off from the initial value by say 20 to 25%, it’s probably time for new tubes.
Push-pull tube amps work best, i.e. have the lowest distortion, when both of the output tubes draw the same current, as this balances the two “sides” of the output transformer’s primary winding. Advantage fixed bias amp, provided you are prepared to do some measuring and adjusting. A true auto bias amp, or a semi-auto one like the Elekit TU-8340, will take care of the balancing for you.
I think you know the type of amp I’d choose to buy, but I’m old-school and not afraid to do a little measuring and adjusting. As mentioned previously, I don’t mean to disparage cathode bias amps. There are very good sounding classic amps, and modern amps too, that employ cathode bias. You must choose according to the amount of involvement you feel comfortable with. Let your ears be the final arbiter when choosing an amp, or any audio component for that matter.
Tube Substitution Considerations:
Generally speaking, it’s much easier to substitute tubes in a fixed bias amp than it is one with cathode bias. The best source of information is the owner’s manual or the manufacturer’s website. The dealer where you purchased your amp might be able to offer a recommendation based on his experience or the experience of other customers.
Failing that, here are a few hints for cathode bias amps; however, I can’t speak to all possible scenarios. If your amp goes into melt down or tubes start going supernova, I disavow any responsibility:
-EL34’s and KT77’s are more or less interchangeable.
-6550’s and KT88’s are more or less interchangeable.
For fixed bias amps, the possibilities are potentially more numerous. The interchangeability list above for cathode bias applies. Additionally, an auto bias or semi-auto bias amp might be able to bias a tube from another “family”. The EL34 versus the KT88 family, for instance. It just might not be able to do it optimally. These amps are “expecting” to see tubes of a certain parameter when setting bias. So the following scenarios could be possible:
- If an amp is designed for the new power tube darling, the KT150, then chances are plunking in your treasured set of NOS Telefunken EL34s could have tragic consequences.
- If an amp is designed for, say, a KT66, then plugging in a set of KT150s will not likely hurt the tubes, but could put serious stress on the amp, especially the power transformer. Even if no damage is done, the bigger tubes might not run optimally and actually sound worse than smaller (and less expensive) tubes. See my power tube shootout for more thoughts on auto bias, fixed bias amps.
Old-school, manually-adjusted, fixed biased amps like my Audio Research Reference 110, or Parks Audio Eiclone, are, generally speaking, more amenable to tube substitution. The reason being that bias can be set to whatever the user wants or thinks will make the output tubes sound their best. Check with the manufacturer, dealer, or a trusted tube seller to tap their experience. Or turn to the internet, there’s always some idiot with an opinion. Oops, I mean, some idiot other than me.
If you are wondering how the pictured single-ended amp sounds, well, so do I. These might, on a good day, put out about 5 clean watts. Using them with my current hungry floor standing PSBs is out. I think I’ll give them a try with my little Wharfies. I’ll get back to you.
Enough talk about tubes. Let’s switch our on stereos and enjoy some music.
Enjoy The Music.com Vacuum Tube Tutorial
Aiken Amplification index to white papers
Send a Donation
If you've enjoyed this content, consider supporting Wall Of Sound with a donation. It only takes a moment and will be greatly appreciated.
Great article. I’ve learned a lot by reading it.
I have an amp using the EL34 and cathode Bias with the diagram similar to first diagram in this page except that it is configured as Ultralinear mode. I calculated the tube’s power dissipation (PD) is about 15W, which is only about 46% of the maximum power. according to you article ideally the PD should around 23W (70% of max).
I am planning to change the cathode resistor to raise PD to close to 23W but I want to ask you if PD at 70% is still a rule for UL mode? (BTW, I am a very experienced electronic engineer with 100s of hours in soldering)
PS: I can send you the diagram and the calculation via e-mail if you need them.
Thank you for your interest in Wallofsound.ca.
The 70% “rule” can be applied to virtually any output configuration. In the case of cathode bias, you might want to run the output tubes a bit below 70%. The reason being that as a tubes age they might draw more current and that will distress a tube unnecessarily. That being said, as tubes age they generally draw less current, but to be on the safe side….. At any rate, the 46% value of your stock configuration does seem a bit too cautious.
My suggestion would be to aim a little lower, 60 to 65%, at least at first. You will need a few different resistor values as it will be a bit of a trial and error process. Even at say 60% you should notice a change – improvement hopefully – in the sound of your amp. All of this presupposes that the power supply in your amp can handle the extra load. Run your amp as it is currently configured and after an hour or so check the temperature of the power transformer. You can use your hand (carefully) or one of the laser IR thermometers available for 20 or 30 bucks from a discount or automotive store.
Change the cathode resistors so that the dissipation is running at say, 55 to 60% and measure again after an hour. If the temp increase is slight, maybe change out the cathode resistors again to draw a bit more power, but I wouldn’t go more than about 65%.
Between resistor changes MAKE SURE THE POWER SUPPLY CAPACITORS ARE COMPLETLY DISCHARGED, i.e. the voltage across them is only a few volts, if not zero.
The advantage of fixed (adjustable) bias is that power dissipation can be set higher, monitored easily and adjusted as required. All the same, checking the voltage across the cathode resistors from time to time will give an approximate indication of tube health.
You don’t mention the age of your amp. If it is more than 20 years old it might be advisable to have a tech (or if you are confident do it yourself) replace all of the electrolytic caps in the amp as they deteriorate with age.
Regards, Steve Graham
Many people ignore the fact that 70% limit applies only to push-pull configs.
Frank, you haven’t mentioned if your amp is single ended or PP. I have learned that there is an important difference between biasing those two. According to robrobinettedotcom for SE amps the safe limit is 90% for fixed and 100% for cathode biased amps. Yes, 100% in idle! Simply because in single ended class A configurations there will be close to zero power fluctuations during operation as opposed to AB push pull.
From my experience there is a humongous difference in tone between 60% and 95% biased tube in an SE amp! Running SE amps too cold is really too much of a shame. Once you hear the creamy lush tone of hot bias , for me there is no going back. Also in SE amps the bass is a bit saggy by definition and this is where the sound suffers the most when running cold, not to say that upper mids and highs sound too thin. Actually with a hot biased SE (90-95%)(not too hot) you will achieve almost as tight a low end as with a PP config.
Also as long as you don’t go over 100% in peak operation dissipation (some even state 100% on average is safe) there will not be that much of an impact on tube life. To put it into perspective running tubes at 60% as opposed to 100% won’t gain you a 40% longer lifespan, taking for instance the fact that filaments are working at constant power and will degrade equally fast.
All taken into consideration and in my opinion. Running tubes too cold is simply not worth it.
Thanks for reading the long post 😀
Great article, glad wiht what you share.
is applying CCS to single ended in el34 to replace the cathode resistor a bad idea?
what impact the application of ccs on the tube has emission is nearing exhausted.
Thank you for your interest in Wall of Sound. Unfortunately my knowledge of, and experience with, constant current sources is very limited. What you propose sounds possible but as for the actual implementation, it would be over my head. I would suggest you post your question on the Tubes / Valves forum at diyaudio.com. Others with more knowledge than I will likely have some good ideas. It would be helpful to attach a schematic of your amp to a DIYAudio post. Even a hand-drawn schematic of the output section with a few voltage measurements would be good.
Regards, Steve Graham
Great article that explains everything in a concise and easy to understand manner. I bought a used 6550 push pull amplifier ( stolen/cloned ) made by Chinese company Bewitch that uses a single common cathode bias resistor. I am currently re-configuring so that it uses 2 resistors. They implemented a single 250 ohm resistor that is passing approx 140mA from both 6550 tubes, voltage at cathode is around 38V. That’s a power consumption of 5.32W in the cathode resistor. They used a 5W resistor to bias 2 6550 tubes. Unbelievable!
I bought it 10 years ago and never got around to fixing it. The smell coming from that resistor was pretty toxic, I knew why the devious seller was selling it when I first plugged it in, but it was cheap. It went into storage until now. Can’t believe some of these Chinese companies cheaping out on critical components like the cathode resistors. When that resistor eventually fails, it can take out the power tube which can fuse the plate to the grids or cathode, causing a melt-down, and potentially fuse the insulation of the transformer winding. I will also be using in-line fuses in the B+ supply to prevent such an eventuality. I would like to re-iterate the message here, using a single cathode resistor to bias high power tubes ( 6550, KT88 etc .. ) is asking for trouble. At the very least, make sure the wattage rating is sufficient to handle the output of both tubes, same for the capacitor voltage rating also.
Thank you for your interest in WallofSound.ca. It is rather strange that a modern or modern-ish amp using 6550 (or KT88) tubes would be cathode biased. Fixed bias, IMHO, would make more sense, though this would, of course, increase complexity and cost. The general disadvantage of cathode bias that uses a common resistor for both tubes in a push-pull output stage, is that the user is unaware of any imbalance in current draw between the two tubes. One way to overcome this (I’ve read of this but not tried it myself) would be to do as follows: 1/ Disconnect the bias resistor (and capacitor) from the cathodes of the power tubes. 2/ Connect a 1 ohm, 1%, 1/4 or 1/2 watt resistor from each cathode to bias resistor and capacitor. The 1 ohm resistors will have negligible effect on the bias but the current draw of each tube can be measured. The other advantage is the 1 ohm resistor acts as a sort of fuse because of its low wattage. So if an output tube decides to go supernova and the current spikes the 1 ohm resistor should overheat and burn out rather than anything else in the amp. At least that’s the theory.
It’s interesting to note that even the classic Quad II had a (slightly) under wattage cathode resistor.
Good luck in your audio adventures.