Dropping String / Power Tube Current - 7591A

[Sonny ReVerb]

Trying to calculate the current requirements for a PP pair of 7591A's to determine the power section dropping resistors.

The datasheet says the max plate dissipation is 19W. If I plan to use 450V on the plates that would be a max plate current of about 42ma per tube (19W/450V)?

For the screens, max dissipation is 3.3W. With 400V on the screens that would be a max screen current of about 8ma per tube (3.3W/400V)?

For the dropping string, the current at node A would be 42 * 2 + 8 * 2 + (PI/preamp current) = 100 + (PI/preamp current)?

Current at node B would be 8 * 2 + (PI/preamp current) = 16 + (PI/preamp current)? Etc.

Does the datasheet Zero Signal Plate Current, Max Signal Plate Current, Zero Signal Screen Current, Max Signal Screen Current have any correlation to these numbers?

Any edification is greatly appreciated.

[sluckey]

I have a real working 7591 amp with voltages close to what you want. May be worth a look...

     http://sluckeyamps.com/hammond_2/AO-63.pdf

[Sonny ReVerb]

Very interesting. Thanks for sharing your work. I'm still not getting the dropping string calculation. Your amp essentially has seven 12AX7 sections (~7mA), two 12AU7 sections (~21mA), and the power tube screens (~16mA) across the first dropping resistor. That should give a 220V drop (.044*5000)???

I went back and calculated the actual plate current for each section using your measured voltages. It's only 5.5mA plus the screens. If the screens only pull 12mA it works out very close to your 87V drop over the 5k resistor. Not sure how you would design to those numbers though.

[PRR]

The data sheet says screens pull ~~10mA at idle and 30mA at full power.

A typical amp is maybe 4 to 6 sections of 12AX7 in front, say 6 to 10mA.

So preamp+screens is 20mA to 40mA.

If your first (OT) node is 450V exact, and you would like to observe the 440V limit on G2, then you need 10V drop at 20mA, so a 500r resistor. Check at full power: 40mA*500r is 20V drop, to 450V-20V= 430V. Higher than the book. Try 1K. 430V at idle, 410V at full power, ready to rock.

Screen current can be VERY variable. Especially in the Aligned Grid types 6L6 and kin including 7027. I have no doubt Slucky's values are correct for the tubes he has. It is not precision shooting. Use rough math, round-value resistors, and see what you get.

[sluckey]

Sonny, I wanted you to look at the original schematic. Plate voltage is 450, screen voltage is 400, each tube cathode is idling at 40mA, so 18W idle dissipation. The first dropping resistor is 800Ω and drops 50v, so screen current plus rest of the amp current is 62.5mA. Second dropping resistor is 1K and drops 50V, so the rest of the amp pulls 50mA. That means total screen current is 12.5mA. These numbers are in line with what you are calculating from the data sheet.

The numbers are different on my converted amp mainly because I wanted lower screen voltage/current to get long life from the 7591s. My dropping resistor values were determined experimentally with a handful of resistors but you can apply the math after the fact. Plates at 470V and hotter top tube cathode current is 24.4mA, so only 11.5W idle dissipation. First dropping resistor is 5K and drops 87V, so screen current plus rest of the amp current is 17.4mA. Second dropping resistor is 2.7K and drops 38V, so the rest of the amp pulls 14.1mA. That means total screen current is only 3.3mA. Very conservative, but that was part of the plan. I know I could decrease that first dropping resistor to run the tubes hotter, but I don't have a need to do that. The amp sounds very good as is and hopefully I'll never need to buy another pair of 7591s.

[Sonny ReVerb]

Thanks for the assistance gents.

Sonny, I wanted you to look at the original schematic. Plate voltage is 450, screen voltage is 400, each tube cathode is idling at 40mA, so 18W idle dissipation.

I missed that at first glance. The 18W idle is interesting with a max of 19W. A hot 70% would be 13.3W.

[Sonny ReVerb]

So how does this look for a starting point? Each screen will have a 1k / 5W resistor, as well. I'll have some extra resistors on hand for tweaking and I'll whack it with a 20% hammer before I'm done ;)

Is there any problem with adding the D* node without a smoothing cap? I've seen it on other schematics, but not sure of the ramifications.

[sluckey]

Why would you even consider no cap? When you don't put a filter cap at node D* the prior dropping resistor becomes part of the plate load resistor.

[Sonny ReVerb]

ok

[92Volts]

If they're out-of-phase gain stages (i.e. consecutive ones in a normal preamp design) D and D* can probably be connected to the same node, as you did with the 6SN7 and 12AX7 further down the string. sluckey's right, there's no point to separating with a resistor and no cap, but they may not need separation if you want to lower the parts count.