RH84 power output
- Thread starter mr2racer
- Start date
Thanks guys. I had a feeling someone had confused it with a triode amp. But the original schematic is not strapped. I was trying to use a 6P15P instead of a EL84. But 6P15P is too sensitive for this amp. It lays really loud and is seriously distorted. It needs to be redesigned for 6P15P with a different preamp tube. I was thinking of 6922.
If you are going to do that consider substiting a small signal pentode. Using a triode in this circuit creates distortion. In the original RH84 it seems this is balanced sonically reasonably well - but if you attempt to change the output tube its a whole different design.
Something like the excellent 6AU6 fits the bill nicely.
Shoog
Kosmin, Plate voltage 305 and screen 150. Originally I was using a pair of Motorola vintage OPT's. I tried a pair of Eico AF4 OPT's recently made by Heyboer. Same outcome.
I went back to a pair of EH EL84's and the original values and the amp is fine now.
Shoog, I'm planning another amp using 6P15P, have a bunch of them. I'm going to use a 6922 to drive two 6P15p SE in parallel. I had Arlen at Heyboer build me a pair of ten watt 4k OPT's with a 6ohm secondary. I plan to use fixed bias and hope for between 8-10 watts.
I went back to a pair of EH EL84's and the original values and the amp is fine now.
Shoog, I'm planning another amp using 6P15P, have a bunch of them. I'm going to use a 6922 to drive two 6P15p SE in parallel. I had Arlen at Heyboer build me a pair of ten watt 4k OPT's with a 6ohm secondary. I plan to use fixed bias and hope for between 8-10 watts.
Hey planet10, I've seen a lot of your posts on these amps and other amps. I just built the rh84 and it sounds really clean but put it on the scope today and with a 2ppv signal in, and only get about 1.7 watts rms out. I do have a 100k pot on the input so maybe that's the issue?
List the EL84 plate to cathode voltage, list the EL34 cathode self bias resistor, and the cathode voltage. Then list the output transformer rated primary impedance, the primary DCR, and the secondary DCR of the 8 Ohm tap.
I will estimate the power from that.
If your power is less than the estimate, then suspect the driver circuit, or the shade feedback values driver Rload and feedback resistance.
I will estimate the power from that.
If your power is less than the estimate, then suspect the driver circuit, or the shade feedback values driver Rload and feedback resistance.
This is just for a el84 no el34. plate to cathode is 280 volts, screen is only 2 volts less than the plate. Cathode voltage drop is 10v across a 260 ohm resistor. OT primary impedance is 5k to a 4 ohm secondary. It's a edcor gxse so should be efficient. All my values are almost spot on to the original rh84 schematic.
10V across 260r is 38.5mA.
The "5.7W" in the EL84 datasheet calls for 53.5mA idle and 60.3mA cathode current at full power. -7.3V bias. Which looks like 130r cathode resistor.
38mA and 270V P-K suggests a 7K load, not 5K.
You are idling too cool. Come up to 50mA total cathode current. At 280V supply you are still fine on Pdiss.
I assume you have the RH84 triode driver. Without the driver it will take many volts to whip the EL84 to full fury.
True, I was pretty tired when I measured it so I'll double check. I know I also checked with a sine wave amplitude of 5v and when I got to around 3 watts of output the sine wave started looking like a mix between a sine wave and a sawtooth wave at both frequency ends. From what I've read that seems to be from 2nd harmonic distortion, is this correct and normal for this amp. Guess the way people have described this amp I thought I would get closer to 4 watts of clean power. I'm surprised that anyone adds a volume pot if this really needs a preamp to drive it to full output. Adding more in the signal chain usually degrades the sound.
If by triode driver you mean the 12at7 then yes. My screen voltage is only a couple volts lower than the plate so is it safe to run that much current? I've never seen anyone use a 7k load on one of these amps. I figured my voltage was only 10 volts off from the schematic if the 300v marked is from ground and not cathode and I thought people run these with about 40mA? I'm already around 11 watts of dissipation so another 20 mA would put me over the 12 Watts max per El84 data sheet.
(280v)(.05A)= is 14W. Thought EL84 max should be 12W?
If by triode driver you mean the 12at7 then yes. My screen voltage is only a couple volts lower than the plate so is it safe to run that much current? I've never seen anyone use a 7k load on one of these amps. I figured my voltage was only 10 volts off from the schematic if the 300v marked is from ground and not cathode and I thought people run these with about 40mA? I'm already around 11 watts of dissipation so another 20 mA would put me over the 12 Watts max per El84 data sheet.
(280v)(.05A)= is 14W. Thought EL84 max should be 12W?
mr2racer,
I see others have chimed in, before I wrote this:
Lets find out about how much power we might get from your amp, based on some measurements, assumptions, estimations, and calculations:
You have 10V into 260 Ohms; 10/260 = 0.038 A cathode current. Most of the 38 mA is plate current, not screen current.
Lets swing the plate current from quiescent (38 mA) down to 0 mA, and then up to 2 times quiescent (76 mA). Yes, the current can and will go larger than 76 mA, but it can not go less than 0 mA, so those two extremes are all bounded by clipping and/or distortion.
The plate load is 5k Ohms. (0.038A squared) x 5,000 Ohms = 7.22 Peak Watts 7.22/2 = 3.61 rms watts. But we need to check further . . .
0.038A x 5000 Ohms = 190V peak swing We have 10V cathode self bias, the control grid is at 0 Volts. We could swing the control grid up to 10V before drawing grid current (distorting the signal, and or changing the effective grid to cathode bias, the coupling cap will start charging with signal).
190V / 10 V = 19V. That means the output stage with negative feedback applied needs to have a gain of 19. Therefore, the driver stage needs to swing + 19 V and - 19 V from quiescent (with the shade feedback applied). You could easily check to see if your driver swings that far, if you have a scope.
If you can not make that measurement . . . In order to estimate how far it can swing, I need some more information: The driver tube type (and if it is a paralleled pair) The driver tube cathode resistor, and bias voltage across the resistor, and whether or not the resistor has a bypass capacitor. The driver tube plate load resistor. The Schade feedback resistor. The output tube control resistor (control grid to ground).
Then I should be able to estimate the gain, and the dynamic swing, and the signal required at the driver stage control grid.
I see others have chimed in, before I wrote this:
Lets find out about how much power we might get from your amp, based on some measurements, assumptions, estimations, and calculations:
You have 10V into 260 Ohms; 10/260 = 0.038 A cathode current. Most of the 38 mA is plate current, not screen current.
Lets swing the plate current from quiescent (38 mA) down to 0 mA, and then up to 2 times quiescent (76 mA). Yes, the current can and will go larger than 76 mA, but it can not go less than 0 mA, so those two extremes are all bounded by clipping and/or distortion.
The plate load is 5k Ohms. (0.038A squared) x 5,000 Ohms = 7.22 Peak Watts 7.22/2 = 3.61 rms watts. But we need to check further . . .
0.038A x 5000 Ohms = 190V peak swing We have 10V cathode self bias, the control grid is at 0 Volts. We could swing the control grid up to 10V before drawing grid current (distorting the signal, and or changing the effective grid to cathode bias, the coupling cap will start charging with signal).
190V / 10 V = 19V. That means the output stage with negative feedback applied needs to have a gain of 19. Therefore, the driver stage needs to swing + 19 V and - 19 V from quiescent (with the shade feedback applied). You could easily check to see if your driver swings that far, if you have a scope.
If you can not make that measurement . . . In order to estimate how far it can swing, I need some more information: The driver tube type (and if it is a paralleled pair) The driver tube cathode resistor, and bias voltage across the resistor, and whether or not the resistor has a bypass capacitor. The driver tube plate load resistor. The Schade feedback resistor. The output tube control resistor (control grid to ground).
Then I should be able to estimate the gain, and the dynamic swing, and the signal required at the driver stage control grid.
I said 50mA cathode current. Some of that is screen current. Look at the datasheet, estimate Ip/Ig2, you can figure closer than my gut reaction.
Also know that 90% of EL84 sold now go into guitar amps which abuse the poor EL84. A sure 12.5W would be loafing on new production.
7K load is used in Fender Champ. Yes, a different bottle, but the bottle is called a "valve" for a reason. If the valve is big enough, output is limited by load and Vpk, (and in SE, by idle current), and will be similar for all big-enough bottles. Fender used 7K because voltages ran high after field coils went out of style. However your voltages are more suited for 5K; IF you get the plate idle current up near Vpk/5K or over 50mA.