Hopefully the mods make this a sticky. OK - there seems to be a bit of confusion as to the 70% vs 90% rule. I blame everyone who sells "cathode current" DIY bias probes for this. They have made a bad attempt at "dumbing down" the bias procedure to those who have no knowledge of amplifiers and how they work and in the process have created a mass of confusion via lies and misinformation as to how this shit really works. So here I will attempt to explain the "traditional plate current" method, along with the more modern dumbed down cathode current method in hopes of clearing up the confusion and explaining where it comes from. Inside of a power valve you have a - Cathode Control Grid Screen Grid Suppressor Grid Anode (aka "plate") In a Class AB amp, a negative voltage from an external supply is "fixed", or "applied" to the control grid, hence the name "fixed biasing". The cathode is grounded. This makes the control grid more negative than the cathode. The plate receives a positive charge from the positive side of the power supply through the output transformer. Current flows from the negative side of the power supply to the cathode, using the chassis/ground as a path to get there. This supplies the cathode with electrons, which are negatively charged. Opposite charges attract. The positive charge on the plate pulls electrons from the heated cathode to the plate, where the electrons return to the positive side of the supply through the output transformer. The screen grid has a positive charge on it as well. The positive charge on the screen grid helps pull additional electrons to the plate. Since it is a "grid", most of the electrons pass right through it. However, some of them flow to the screen, so effectively you have the screen grid itself drawing current from the cathode as well as the plate. Now...with 450 - 500 volts on the plate, if the negative voltage on the control grid didn't exist, so much current would flow through the valve that it would fry it along with the other components that are connected to it (i.e. socket, screen resistor, OT). The purpose of the negative voltage is to control the current flowing from the cathode to the plate/screen grid. It does this because it is more negative than the cathode. Since electrons are negatively charged and like charges repel, this negative control grid voltage creates a repelling force between the control grid and the cathode. This repels some of the electrons that are flowing from the cathode to the plate, thereby reducing the current flow to a safe level that the valve can handle at the plate voltage it runs at. It is this negative "control voltage" that you adjust when you're setting bias. By varying this negative voltage more & less negative, this directly controls the current flow through the valve by increasing/decreasing the repelling force between the control grid and the cathode. Plate current vs cathode current - As mentioned above, the cathode must supply electrons to BOTH the screen grid as well as the plate. Plate current deals ONLY with the actual current flow to the plate itself. However, cathode current deals with BOTH screen grid AND plate current flow FROM the cathode. What does this mean? This means that cathode current will read HIGHER than plate current since cathode current deals with current flowing to BOTH the plate and the screen, whereas PLATE current only deals with the current flowing to the plate only. Now we get into the interesting part - Plate dissipation vs screen dissipation! "Dissipation" is the term used to refer to the ability of the individual parts in the valve to dissipate power as heat as well as how much DC power it can handle as "constant current flow" during no-signal conditions. As electrons flow from the cathode to the plate and screen, there is a resistance to that flow which creates heat when struck by electrons, which must be "dissipated" by the plate as well as the screen grid. Power (Watts) = Potential (Voltage) x Electron Current Flow (Amps) or P = V x I If we have 475 volts on the plate with a plate current flow of 35mA, the plate is dissipating - 475 Volts x 35mA (0.035 Amps) = 16.625 Watts This would be well within acceptable limits for an EL34 running in class AB. This is because an EL34's maximum plate dissipation rating is 25 watts. 16.625 watts is 66.5% of 25...just under the 70% limit for Class AB. 16.625 Watts / 25 Watts = 66.5% 70% of 25 watts = 17.5 Watts (25 Watts x 0.7 = 17.5 Watts). If we wanted to know what the max plate current for an EL34 running class AB at 475 volts - 17.5 / 475 Volts = 0.0368 Amps = 36.8mA This would be the absolute maximum plate current you would want to see when biasing via measuring plate current. This was the 'traditional' method that us 'old school' techs have used over the years. This was also the way I learned to do it and for reasons I'm about to mention the only way I prefer to do it. Now's where things get REALLY interesting. If I haven't lost you yet...read on. The 70% rule states that you bias your valves in class AB no higher than 70% of the maximum rated plate dissipation. You can bias lower, but no higher than this. Biasing via Cathode Current This seems to be the more modern way to do things as it gives a safe method of biasing to the novice. This is because when biasing via cathode current, you're looking at a voltage drop across a 1 ohm resistor that is in the millivolts range rather than trying to measure current across the OT that has a much higher voltage on it. The theory is simple Ohm's Law - Voltage = Current x Resistance, or V = I x R Due to this law, any amount of current flowing through a 1 ohm resistor will equal itself as a voltage measured across that 1 ohm resistor, because anything times 1 equals itself. If you have 40mA flowing through 1 ohm of resistance, across the 1 ohm of resistance you will have - 40mA x 1 ohm = 40mV Of course here you are subjected to the resistor being a true 1 ohm and not slightly off. Let's say for instance the resistor was actually 1.2 ohms - 40mA x 1.2 ohms = 48mV There it just threw your reading off by a factor of 8mA, yet the resistor value was only 0.2 ohms off. Let's go the other way and say the resistor is actually 0.8 ohms - 40mA x 0.8 ohms = 32mV In thise scenario, your reading would be LOWER than what's actually flowing through the resistor. Throw in the fact that cheapo multimeters aren't very accurate when reading super low voltages in the mV range and you have a real recipe for disaster. This is reason #1 why I DO NOT recommend the biasing via cathode current method. Now we get into where the guys who developed the "dumbed down" method fucked up and created the mass of confusion. Cathode current will always read higher than actual plate current because again cathode current includes BOTH the screen grid current AS WELL AS actual plate current. Since we have a 70% rule for plate current, we ALSO have a 70% rule for cathode current. This is how it SHOULD'VE been explained. An EL34 has a plate dissipation rating of 25 watts while having a screen dissipation rating of 8 watts. Remember, cathode current deals with both of these, so naturally we would want to add these two figures together. 25 watts + 8 watts = 33 Watts Plate+Screen dissipation Now...the proper way to do this is to take 70% of the above figure - 33 Watts x 0.7 = 23.1 Watts So for a plate voltage of 475 Volts, we should see - 23.1 Watts / 475V plate voltage = 0.048 Amps = 48mA Since we are taking this reading across a 1 ohm resistor, we will see this as 48mV instead. This is a MAXIMUM FIGURE. Again you can go lower, but DO NOT go higher. Biasing for a cathode current mV reading of 48mV will put us right at 70% of both the plate & screen dissipation combined. Here's where the fuckup came in - In an effort to dumb down the procedure, all the valve dealers who favor this method as well as the manufacturers of DIY Bias Kits and Cathode Current monitors took this and tried to relate it to PLATE DISSIPATION ONLY! WTF!? You ONLY deal with plate dissipation by itself when biasing via the PLATE CURRENT method. Cathode current is BOTH PLATE & SCREEN! Well this is what happened - Lets' take our 23.1 watt figure, which is 70% of PLATE & SCREEN, and apply it to PLATE ONLY dissipation, which is 25 watts - 23.1 Watts / 25 Watts = ... YOU GUESSED IT! 92.4% Dissipation This puts us close enough to THEIR recommended "90%", which gives a 2.4% 'cushion'. However, this is flat bullshit. This does not properly explain just how things work inside of an amp and if you're going to teach someone how to properly bias an amp, you should teach them the CORRECT way rather than "dumb it down" in a way to where if they tried to apply it using more traditional methods it won't make sense. Not only that, but applied in the wrong way this can prove to be a very dangerous practice. I just had to help a guy who was trying to make sense of "classes of operation" going by a PLATE CURRENT LOADLINE, yet trying to apply the 90% rule used to apply cathode current to plate only dissipation and it just flat out wasn't making sense to him. AND IT WON'T! Because it's not a correct way to explain it and a VERY bad idea to say the least. Throw in the fact that now you have novice guys biasing their own amps going onto public message boards telling people that it's OK to bias to 90% plate dissipation and you can create a whole epidemic of disaster. This is because not everyone biases via cathode current and lots of us bias via the more traditional plate current method and applying the 90% rule here would fuck a lot of shit up. It is for this very reason why people who don't have any valve amp tech background should not post up tech info without stating that they're not a tech and have no way to verify the accuracy of a certain bit of info they happen to feel like posting up. For the record, the all too common statement "I know it's true because 'so and so' at 'such and such' shop told me so" or throwing up the defense of "Go tell so and so at such and such shop that...I'd like to see what they tell you" when you get called out on it just simply won't cut it. If you can't back it up with electronics theory and hard facts, either start your post off with "I'm not a tech by any means so I have no way to verify this, but I seem to recall reading somewhere that...blah blah blah" or "...but I happened to hear from so and so from such and such that...blah blah blah". This notifies the many people who could be reading your posts immediately upon reading that what they're reading is not yet verified hard fact info rather than reading a bullshit post of "Yeah you bias it to 90% and call it good" as if it's already verified info, and risking the reader blowing their amp up. Even if it's something you've done with no problems...just because you haven't had problems YET doesn't mean you WON'T. It just means that it hasn't happened yet. If you can't do that, then just simply don't post it or expect to get called out. I hope this somewhat clears the air as to the 70% vs 90% vs Plate Current vs Cathode Current debate that's been going on here for quite some time.