Calculating SE 6L6GC CLC choke current rating?

[dwinstonwood]

Hello all,
I'm working on a single-ended 6L6GC design that will have a voltage of about 350VDC on the power tube plate.
I want to use a CLC input from the rectifier and use the second capacitor as the B+ "Node A" for the plate. This will be followed by an RC filter for the "Node B" screen supply (see schematic below). The preamp will have a typical current draw of four triodes.

My question is, how do I calculate the current rating for the choke in this configuration?

The 6L6GC datasheet says for Class A1 Pentode:
Maximum-signal plate current is 66mA at 350V
Maximum-signal screen current is 7mA at 350V

Would the Hammond 194B - 4H/90mA be inadequate?
How about the Hammond 158Q - 5H/150mA?
There's also the less expensive Triad C-14X - 6H/200mA.

I hope the math isn't too complicated! Thanks for help with this.

David

[HotBluePlates]

I'm working on a single-ended 6L6GC design that will have a voltage of about 350VDC on the power tube plate.
I want to use a CLC input from the rectifier and use the second capacitor as the B+ "Node A" for the plate. ...
...
My question is, how do I calculate the current rating for the choke in this configuration?

The current of the entire amplifier passes through that choke, so you need to figure the current-draw of the entire amp at max output power.

To do that, I'd need to see your entire planned design (including power transformer specs, estimated supply voltages & the output transformer primary impedance).

Would the Hammond 194B - 4H/90mA be inadequate?

Probably not, but again, I'd need to see the whole plan first.

That part is normally used for only the screen current only of a Fender-style amp.  Those might draw 4-5mA screen current per tube at idle, and rise to 30+ mA at full-tilt (though one side's screen current is max while the other's is at minimum).

Note that if the choke's current rating is exceeded, the core saturates and inductance falls to zero.  So a healthy over-rating is appropriate for most chokes.

[dwinstonwood]

Thanks HBP, I will provide that info later today...

[dwinstonwood]

So, here is the schematic, and links to the PT and OPT datasheets.
I'll be using an NOS 5Y3GT.
Let me know if you need any other info.

https://www.hammfg.com/files/parts/pdf/270EX.pdf?v=1697661948

https://www.musicalpowersupplies.com/app/download/7114418257/Visio-OT10SE_STD_DWG.pdf

Thanks HBP!

[HotBluePlates]

So, here is the schematic, and links to the PT and OPT datasheets.
...
https://www.musicalpowersupplies.com/app/download/7114418257/Visio-OT10SE_STD_DWG.pdf

Your chosen output transformer has a max idle current of 60mA, which limits output power pretty significantly.  Once I started digging into loadlines & likely output voltage from the PT/rectifier, I'm guessing 6-7 watts.  Though perhaps you're fine with that (and any 6L6-type oughta last forever).

The 6L6GC datasheet says for Class A1 Pentode:
Maximum-signal plate current is 66mA at 350V
Maximum-signal screen current is 7mA at 350V

When I started looking at 5Y3 curves with 275vac applied & a probable current draw, it seems likely you'll get ~300vdc out to the 1st filter cap (even lower at the 2nd).

Don't try to adjust idle bias for "max plate dissipation":  ignoring bias voltage (which subtracts from plate-to-cathode volts), 300v x 60mA = 18 watts ---> safe for even first-version metal 6L6s.

I quoted your data sheet conditions, because you need to copy "all of the condition," which means 250v at the screen, and -18v of bias.  The 250v at the screen seems probable/workable.

-18v of bias with 60mA plate (the OT's max) and ~3mA screen is 18v / (63mA) = 286Ω ---> 270Ω or 300Ω cathode resistor would probably work equally-well.

Staying Class A, 60mA idle means a peak plate current of 120mA or less, and based on plate curves this is manageable with as little as 200v on the screen (not a figure to shoot for; it just means sag won't choke power output).

110mA peak under these conditions (with 200v screen) happens when plate is down to ~50v (a 250v drop from an idle of ~300vdc), which implies:
   250v (peak swing) / 50mA (peak swing) = 5kΩ ---> the primary tap you would choose.

The condition you cited does seem correct at no more than ~7mA average screen current at full-tilt, and "66mA average" accounted for a rise in plate-current due to "rectification effect".  Adding another 1mA per tube-section in the preamp (which might be generous), that's <80mA.  So you seem safe to use any of the chokes you asked about, even the one I was skeptical about.

Assuming "250v screen," we need to drop 50v across what you have drawn as 1kΩ in the power supply.  Preamp current of ~4mA plus another 4mA in the screen implies 50v / 8mA = 6250Ω ---> 4.7kΩ 3w is a bit low, 8.2kΩ 3w is a bit high.  Either would "get the job done."

All this will be a bit off, and voltages are likely to come in a bit higher than these estimates.  That's because the Hammond PT is rated for 144mA output, and this amp is under-loading it in a way that will raise output voltage appreciably.

Most folks these days are looking for "lower output power, quieter home-loudness" so your amp would fit the bill.  If you want more than "a bit louder than 6V6" then you'll want to look at different output transformers than allow much more idle current.  Hammond has a couple of universal single-ended output types that fit the bill (for this low of voltage, you would likely look to use the 2.5kΩ primary arrangement).

[shooter]

fwiw;


I started out like you're doing, started with 6V6, just didn't get where I wanted, moved to 6L6, got better, even hit my "that's it" threshold.
after awhile though, it just wasn't "quite there", moved up to KT88 and found happiness at anything from ~~6W clean in-home noodling to medium venues pushing ~~23Wrms (assuming pedals and guitars with good operators:)


from there I could always "work backward" with PA tubes without changing PT or OT, just some tweaks to the PS circuit, creative math n speaker configs

[dwinstonwood]

I really appreciate all that work HBP! I'm going to read through it slowly a few times so as to not miss anything. There's a lot of great stuff there that wouldn't have occurred to me, stuff that I want to get my head around. 

But, for the moment, I'm thinking I might as well start with a 6V6GT like shooter did, and use my 194B - 4H/90mA choke and the OT10SE OPT (maybe with the 7K primary). That should allow me adequate current margins.

I'll see if I can use you formulas to try and calculate a good cathode resistor value using the 6V6GT datasheet numbers.

Later, I could consider upgrading the choke and OPT to use bigger tubes ala shooter. To be honest, I'm a lot more interested in building my preamp design than I am in using a specific power tube or wattage.

Here's my layout (I'm sure there are errors):

David

[shooter]

I really appreciate all that work HBP

2nd that


 he's the kind of teacher i had both in military schools and tech schools, know the H1B guys i had in college, not so much....

[dwinstonwood]

OK, these calculations are probably off.

Cathode resistor value: -13V bias / 41mA plate & screen current = 317Ohm

I'm assuming that there is a "stable" ratio between plate and screen voltages when working with datasheets.

So, the datasheet offers an operating condition with 315 plate volts and 225 screen volts. If I have 330 plate volts, then 236 screen volts preserves that ratio (if it really does work that way).

Then, I need to drop ~94V across the resistor between plate and screen:

94V / 4.1mA = 2.3K

Close enough for tube work?

Thanks!

[shooter]

something i've noticed with SE amps, the more you "control" G2, the less "guitarness" you get.  and moves closer to audiophile territory.
an SE is operating at "max" with no signal, as you drive/OD with signal the natural NFB of the cathode biasing "cools down" the tube..sorta

[tubeswell]

OK, these calculations are probably off.

Cathode resistor value: -13V bias / 41mA plate & screen current = 317Ohm

I'm assuming that there is a "stable" ratio between plate and screen voltages when working with datasheets.

For any given pentode or tetrode plate voltage, the Ip:Ig2 ratio is relatively constant, and this ratio can be extrapolated to other quiescent plate voltages, which is handy to know when you're working out g2 dissipation. So to be clear, we're not talking about a stable 'plate voltage to screen voltage' ratio (and you can in fact change the screen voltage quite considerably without changing the plate voltage - although you may also need to adjust bias voltage when you do this).

[HotBluePlates]

I'm assuming that there is a "stable" ratio between plate and screen voltages when working with datasheets.

So, the datasheet offers an operating condition with 315 plate volts and 225 screen volts. If I have 330 plate volts, then 236 screen volts  ...

We could "extrapolate if we had only an example condition or two, but we have curves.  So let's just use the curves.  It sounds like I need to show how I used the data sheet to derive the answers I wrote earlier.  So let's look at a 6V6 data sheet.

2 things control plate current:
   1.  G1 volts, which we usually call "bias volts."  It's the voltage from grid-to-cathode.
        Since you are going to cathode bias, the grid is tacked at 0v (via a grid-leak resistor), and the cathode is positive by the voltage-drop across the cathode resistor.

   2.  G2 volts.  This is the voltage from screen-to-cathode.
        Changing plate volts barely move plate current, but screen volts definitely changes it (just less than G1 volts).  So moving the screen to a higher voltage will bring up the plate current.

Design is iterative.  I'm gonna start out assuming "320v" plate and "250v screen" mainly because the data sheet has curves drawn for 250v screen.

   1.  Let's assume this 6V6 will idle at 100% plate dissipation (of 12 watts).  12w / 320v = 37.5mA.  So put a dot at "320v, 37.5mA" (this seems to land right about -15v G1 bias) and draw a line from it to 37.5mA x 2 = 75mA on the plate curves for 250v screen.  That's the upper-graph on Page 3 of the linked data sheet.  (Open the image in a new tab to see full-size)




   2.  Single-ended amps must be Class A.  This implies the plate current will peak at 2x idle current on one side of the signal cycle, and drop to almost-zero-plate-current on the other side.  Your chosen OT has primary impedances of 5kΩ and 7kΩ available.  The peak plate current change of 37.5mA will create a voltage-drop across the OT primary and pull the plate voltage momentarily low, so we need to know whether the tube can support the plate-voltage-change.

       A.  37.5mA x 5kΩ = 187.5v -----> Plate pulled down to 320v - 187.5v = 132.5v at 75mA  (Red dot)
       B.  37.5mA x 7kΩ = 262.5v -----> Plate pulled down to 320v - 262.5v = 57.5v at 75mA  (Green dot)

       Plot both these points on the same set of curves, and see whether the tube can manage the plate current at that screen voltage.

       The Red loadline (7kΩ) maximizes plate-voltage-swing, so it delivers higher output power.  And we see that the tube could have pulled the plate down to 35-40v without issue.  So we know the 7kΩ loadline is closer to an "optimum load."



   3.  If we look at the dashed curves on the graphs above, they represent screen current for a few different G1 voltages.  The most-negative dashed curve drawn is for -12.5v bias, and using the scale on the right side we see screen current at our idle plate voltage will be something like 3-4mA.  We could try to plot another curve of our own for "screen current over the signal cycle" but it appears that it won't peak above 18mA; we tack G2 to a filter cap that will supply some of this peak current.

   4.  Now we have some parameters:  320v plate, 250v G2 so 320v - 250v = 70v dropped across the screen dropping resistor.  4mA screen at idle plus 4mA preamp is 8mA, so 70v / 8mA = 8.75kΩ ----> not a standard value, so "8.2kΩ 3w" is "close enough."

[HotBluePlates]

   5.  We need 15v across the cathode resistor, and have 4mA screen & 37.5mA plate:  15v / (37.5mA + 4mA) = 361Ω ---> some places have 330Ω, or slap a 1kΩ in-parallel with 470Ω to get 320Ω.  15v x 41.5mA = ~0.65w ---> 3w total-rating is plenty.

   6.  Power Output will be:
        37.5mA Peak x 262.5v Peak / 2 = ~4.9 watts ---> round down to "4w" or "4.5w"

The 4H 90mA choke you originally asked about will sustain this amp 24/7 for decades.

   7.  "Iteration" would mean taking our assumed current-draw for the amp (37.5mA + 4mA +4mA at idle, or ~46mA) and checking curves from a 5Y3 data sheet (bottom of Page 4) to figure out what the output voltage of the rectifier will be.  We draw a new line for "275vac" between the existing "250v" and "300v" lines, raise a vertical at "46mA" and read the output voltage off the left axis:  "320vdc"

        A. Your actual output voltage will be higher for at least 2 reasons:  you're planning on more filter cap µFs than the "20µF" used by the chart, and you're under-loading the power transformer by more than half.

        B. Remember when we said "grid-to-cathode" and "screen-to-cathode" before?  Well, the "plate voltage" is really "plate-to-cathode voltage" and up until now we haven't added that "extra 15v" across the cathode resistor to the total supply voltage.  So "perfect world" would have been 335v to account for both the 320v plate-to-cathode and the 15v cathode-to-ground.  You'll probably hit "350vdc" or a bit more, and again, we're close enough.

[dwinstonwood]

Much appreciated HBP! All of this has been very useful. I'm going to plot some practice lines for other numbers. The part on choosing an OPT primary impedance was awesome.

For the cathode resistor, I have some 470Ω's and 1.5K's. In parallel, that will give me 357.868Ω.

This is going to be a fun project to build. I can't wait to hear how those less-common preamp tubes are going to sound.

David