(Details and explanations are below)
Step 1: Measure the idle current of the power tubes.
Step 2: Look up a recommended idle current for your power tubes.
--> GREAT ONLINE CALCULATOR <--
General range for most power tubes = 10 to 40mA
Tube specific:
6V6/EL84/65BQ = 22-27mA
6L6/5881/KT66 = 30-35mA
EL34/6CA7 = 35-40mA (sometimes more)
6550 = 40-50mA
Step 3: Adjust the control grid voltage until the desired idle current is obtained.
That's it!
Q: What does it mean to bias an amp?
A: To bias an amp is to adjust the idle (quiescent) current draw through the power devices (tubes or transistors).
"How to bias a tube amp" seems to be a simple enough question, but it can elicit a confusing array of answers. This is my attempt to clear up that confusing for someone who knows what Ohm's Law is, is ready to tinker with their tube amp (and well aware of the danger inside!), but isn't quite sure what to make of all of the suggested methods available.
The procedures or methods of biasing can be separated into three groups:
1) Measuring (directly or indirectly) the idle current of the power tube(s) and adjusting this to a recommended number of mA.
2) Measuring the grid bias voltage (or "C-") of the power tube(s) and adjusting this to a recommended number of volts.
3) Observing the output waveform of a sine wave of with a certain recommended audio frequency (often 1kHz) and adjusting the bias until the output is just short of crossover distortion.
Measuring Current
The most meaningful measurement is the amount the current flowing flow the tube from cathode to plate. This combined with the plate voltage allows you to calculate the power dissipating in the tube. Excessive power results in a broken tube, and a broken amplifier follows that. Checking that you aren't exceeding the maximum recommended power dissipation is a very good idea. Unfortunately, measuring current is usually much more inconvenient than measuring voltage, and this method does not check for crossover distortion. Adding to the inconvenience is that you also need a chart of recommended mA numbers for type of tube and/or specific amplifier. But with your mA chart in hand, this is my recommended method to use. Recommended idle currents are found below in the detailed procedure section.
Measuring Control Grid Voltage
Cathode to plate current is controlled by the control grid voltage. Voltage is easy to measure, and the bias adjustment potentiometer is most likely adjusting this negative voltage. A higher negative voltage will bias the amp "colder" by lowering the idle current. A lower negative number will bias the amp "hotter" by raising the idle current. If your grid voltage is at ground (0V) or positive, you're either working with a self biasing amp, you're using your test equipment incorrectly (is the black meter lead to ground?), or your power tubes are glowing bright red and either are damaged or will be damaged soon due to excessive current flowing inside.
Trying to set the bias via the control grid voltage alone means you must have pre-selected tubes and a circuit in good working order otherwise. This comes up sometimes with Mesa Boogie amps where you buy pre-selected tubes that have been tested to work correctly with a specific grid voltage in a specific (Mesa Boogie) amp. Many schematics will specify a particular voltage for the grids. You should not set the voltage to this number blindly. The plate to cathode current measurement will always trump the grid voltage measurement.
Observing Crossover Distortion
If you have a class AB amplifier, you have to watch out for crossover distortion. If your output sounds especially distorted at low volumes, or during the decay of a sound, it is likely crossover distortion or a faulty speaker. To observe crossover distortion, the output is connected to an oscilloscope. The input is connected to a low distortion sine wave at a fixed audio frequency. 1,000Hz or 1kHz is often selected as a test tone as it is a rough midpoint for the human hearing range, and it is a nice round number. If you have a variable bias control, turning the amp maximally cold should result in the output sine wave developing flat spots where the wave crosses the zero voltage point. During the flat spots the output devices are all "off." Increasing the idle current should decrease the flat spots until a continuous sine wave is observed at the output. This method is the most test equipment heavy, requiring both an oscilloscope and an oscillator. Most technicians would also use a "dummy load" resistor so as to not actually drive a speaker with the annoying 1kHz tone. However, once your ear is trained to what crossover distortion "sounds" like, you can get adequate results by simply listening to the output and adjusting "by ear," only needing a simple meter to ensure the final bias setting doesn't run the amp too hot (see above).
How to measure the plate current
There are 3 usual ways to get a measurement of the idle current in mA.
1) If you're lucky, the manufacturer included a 1 ohm resistor between the cathodes and ground. This resistor's purpose is to be a simple current-to-voltage converter so that you can measure the current as voltage. Confused? It's all in Ohm's Law. Simply measure the voltage across the resistor, and take that mV reading literally as the mA reading. (example: 70mV = 70mA)
2) Use "bias probe" tube sockets. The "probes" are inserted between the amp's sockets and the tubes under test. Inside the "probes" are just the 1 ohm resistors between cathode and ground, and then some wire to bring the test points outside for measurement. Usually the probes are attached to a little box fitted with a LCD display that will display the current draw for the selected tube.
3) Use your meter's current measuring mode to short circuit the output transformer, allowing most of the current to flow through your meter. (Alternatively, you can desolder the output transformer from the tube's plate, and insert a meter between the high voltage and plate, but this is much more inconvenient.)
The last method is dangerous, but quite handy. Be extra careful if attempting this method. Usually we only work on our amps with the unit unplugged from the wall, and the filter capacitors drained of any stored voltage. This method requires use to probe the open amplifier while the amplifier is on, which is dangerous and not recommended to anyone unsure as to what they are doing! Only attempt this with alligator clips. You will actually use your meter's mA setting in this method. It is very important to understand that in this mode your meter leads are basically connected directly, especially as far as high voltage is concerned. This means that if you touch 500V with your red lead, and then touch ground (0V) with your black lead, you will be shorting that voltage and sending the 500V to ground. This normally results in the fuse blowing along with a big spark that will probably leave a mark and may even damage your test lead. Worse is that you might connect 500V to yourself, someone else, or parts of the amp that cannot withstand high voltage.
If that didn't scare you off, then we'll proceed. With the amp off, put the meter in current mode (mA range) and attach one end of an alligator clip to the red lead and the other end to the center tap of the output transformer's primary (tube side) winding. You'll need the schematic or the assistance of someone familiar with tube amps to figure out where the center tap of the output transformer is. Next, attach a second alligator clip between the black lead and one power tube plate pin (usually pin 3, consult schematic or datasheet to be sure). If you look at the schematic, you should see that the only thing between the meter leads at this point should be 1/2 of the output transformer's primary winding (or the entire winding if you have a single-ended circuit). The idea is that the power tube current will flow through the meter instead of the output transformer. The amp doesn't work like this (produce sound) but it allows you to measure current without desoldering anything, which is very convenient. Turn the amplifier off and drain the filter caps before disconnecting the alligator clips. Remember, both leads are basically at the same point, so being careless here can easily result in a surprise discharge of high voltage!
mA Chart
Just measuring the current isn't enough - obviously we need to decide if our measurement is too high, too low, or just right.
Ultimately, the greatest resource here is the tube's datasheet, but a beginner won't be able to divine the answers from the datasheet. It is far easier to just take the recommendations of tube amp experts. One such expert was the late Ken Fischer of Trainwreck amplifiers. He suggests in his "Trainwreck Pages" to bias most power tubes to 10-40mA. That's per tube, so be careful when reading 100W with tubes in pairs; you'll need to divide your readings in 2 if you're using the 3rd measurement technique (i.e. the recommendation becomes 20-80mA per pair). Ken advises to keep tubes within that range for safe and reliable tube operation, but let your ear dial in the final number within that range.
Schematic Heaven provides a list of narrower ranges grouped by tube type. The numbers are per tube; double the number for pairs of tubes.
6V6/EL84/65BQ = 22-27mA
6L6/5881/KT66 = 30-35mA
EL34/6CA7 = 35-40mA (sometimes more)
6550 = 40-50mA
Note that neither of these sources take the plate to cathode voltage into account. The voltage drop across the tube is multiplied by the current to get a power dissipation. Ted Weber's speaker company has a great online calculator for taking all parameters into account.
http://www.tedweber.com/bias-calc
If you need to adjust your idle current, what you'll be adjusting is the control grid voltage. This is explained in detail in the next section, if you have any confusion about it.
Control Grid Voltage
&
General Bias Voltage Info and Troubleshooting
As a "method" this means consulting the schematic and/or service manual to obtain a "bias test point" and corresponding voltage measurement that the manufacturer recommends. No matter which method you use, adjusting the bias always refers to adjusting the voltage from the control grid to the cathode; what changes is what you measure to check the bias. Measure the control grid voltage by itself is a "blind" measurement, in that we are only checking the controlling element, but not the element controlled. This method assumes that the power tube draws a certain amount of current, and you must have pre-selected power tubes to use "Method 2" alone. This is a relatively rare practice. Mesa Boogie is one manufacturer that uses this scheme so that customers can "pop in" replacement tubes that are tested and selected by Mesa Boogie.
If you've measured the plate current, and discovered an adjustment is needed, then you will be changing the control grid voltage with respect to the cathode. There is more than one way to do this, but the principle is always the same, namely that the control grid must be more negative than the cathode. That means that the control grid voltage doesn't have to be a negative value with respect to ground (0V), but should be a negative value with respect to the cathode. An easy way to check this is to put the black lead of a voltmeter to the cathode, and the red lead to the control grid. What is the difference between black lead on the cathode and black lead to ground? If you have a cathode biased amp, that means there is a degenerative resistor in the cathode circuit that places the cathode at a positive voltage with respect to ground. So if the black lead is slightly positive, and the red lead probes ground, the reading on the meter will be negative, which is what we want for normal tube operation. Fixed bias amps (the ones that usually need a bias adjustment) will normally have the cathodes directly grounded.
Once the goal of normal operation is understood, next you should be able to trace the negative voltage supply ("C-" voltage) from power transformer to control grids. There should be a diode with it's cathode facing the ac from the transformer secondary, and a filter capacitors that may appear "backwards" due to it's positive lead going to ground. This is correct, since the bias voltage will be negative relative to ground (the diode ensures that). From there, you will find some kind of resistive network that divides the bias voltage to the desired number. The resistive network may simply be fixed resistors, a potentiometer, or a combination of the two. This processed voltage is then fed to the control grids via fixed resistors of relatively high value so they do not "load" the ac signal.
If you cannot get the idle current draw where you want it using the bias adjustment potentiometer, then you must use the above info on how biasing works to troubleshoot and/or modify the resistive network to allow more or less negative voltage to appear at the control grids. This is essential when figuring out what to do with Fender "silverface" amps that have a bias "balance" adjustment. These amps require you change a fixed resistor to adjust the total negative voltage available. That is my way of biasing those amps. The alternative is to modify the balance pot into the "blackface" bias pot. If it's your amp, I can see justifying the effort of the complete modification. If you are just setting the bias for someone with new output tubes, changing the one resistor is much easier.
Observing Crossover Distortion
Observing crossover distortion requires the most equipment. You must use an oscilloscope, sine wave generator, and a dummy load in place of the speakers. In a pinch, you can do this with the speakers, but it is much more unpleasant for you and anyone who can hear your amp. Connect the dummy load and oscilloscope in parallel to the speaker output. Inject a 1kHz sine wave (or whatever frequency is recommended) at an appropriate amplitude to achieve the maximum undistorted output (the sine wave shouldn't be "clipped" on the top and bottom when viewed on the scope). Adjust the bias until crossover distortion appears. This appears right at the 0V line on the scope's x axis. The sine wave should develop a flat spot, almost like the wave is splitting into two pieces. Once visible, adjust the bias the other way until it disappears. This is a subjective decision, and does not check that the idle current is within a safe operating area. Regardless of the inaccuracy, this is a popular and often recommended procedure for both class tube and solid state amplifiers.
Crossover distortion is only applicable to class AB amplifiers.