Correct OT impedance for a 2 6L6 cathode biased amp?

[Baguette]

Hello,

I'll be building a 2 6L6GC cathode biased amp soon.
B+ will be around 380V and the tubes will be biased at 25-30W plate dissipation.
What would be the correct theoretical OT input impedance?
I know typical value for Fender amps is 4K but those have fixed bias and much higher voltage than what I'll be using.

Thanks for the answers.

[sluckey]

R = E²/P

[DummyLoad]

5K


4.3K will work


6.6K will work

[PRR]

> 2 6L6GC cathode biased ... B+ ... 380V ... 25-30W plate dissipation.

I'd say, "read the sheet"; but for historical reasons the 6L6 sheet hews to the old 19W dissipation limit.

7027 is the *same* guts on a different pinout.
http://www.mif.pg.gda.pl/homepages/frank/sheets/049/7/7027A.pdf

Page 2 shows a condition for cathode-bias, 380 Volts plate *and screen*, 26W dissipation.

[DummyLoad]

6L6GB RCA sheets are similar as well.


http://tubedata.milbert.com/sheets/049/6/6L6GB.pdf



i'd use 6.6K OT and run g2 at 360-370V. start with 270R cathode R and work down to 220-250R or thereabouts.


--pete

[Willabe]

7027 is the *same* guts on a different pinout.

I know you have said, as far as the marketing, they changed the pin out and then up'd the output dissipation. 

But.....   Same guts and a little bit bigger glass bulb and with a couple extra base pins, both used as heat sinks????   

[HotBluePlates]

7027 is the *same* guts on a different pinout.

I know you have said, as far as the marketing, they changed the pin out and then up'd the output dissipation. 

But.....   Same guts and a little bit bigger glass bulb and with a couple extra base pins, both used as heat sinks????   

Look again at what PRR said first:

I'd say, "read the sheet"; but for historical reasons the 6L6 sheet hews to the old 19W dissipation limit.

The later, non-19w 6L6 variant data sheets often reprint the same/similar conditions as the earlier lower-power original. Compare the conditions across data sheets; the carry-over from early sheets is why PRR pointed to an updated 6L6 type, whose sheet pointed towards the real capability of the later tubes.

[Ken Moon]

Fender used an OT with a 6K primary for several cathode-biased 2x6L6 tweed amps with B+ in the range you mentioned, with a 250R cathode resistor.

ClassicTone and Mojo both sell clones of this transformer, and it's a great little unit that has 2-bolt mounting (3-9/16" between holes), so it fits on Doug's new, highly flexible Stout chassis.

[HotBluePlates]

There is a difference between the 6L6 and 6L6C, and according to GE, the 6L6C has a higher power rating than the 6L6B. 

The original question was optimum OT primary impedance under specific conditions. The point made was to look at the operating conditions listed on the 7027 data sheet rather than those on the 6L6GC data sheet, because historically tube makers updated the maximum ratings for later 6L6 types but often carried over original 19w metal 6L6 operating conditions, even for the 30w 6L6GC.

See the attachment below which compares the 1937 RCA 6L6 sheet to the later Tung-Sol 6L6GB sheet and the (still later) G.E. 6L6GC sheet. Red boxes enclose the (almost exact) same conditions for push-pull class A, while blue boxes do the same for push-pull class AB1. The few differences in idle current and power are largely due to tens-of-volts changes in plate and screen voltage.

The conclusion is that the 6L6GC data sheet operating conditions don't point the user to the plate/screen voltage, bias and load which yield the max power available from the tube; they treat the 30w 6L6GC same-as the 19w 6L6.

  While a reference to the 7027 is helpful, as it provides cathode bias data. 

Reference the original post:

Hello,

I'll be building a 2 6L6GC cathode biased amp soon. ...

The 7027 also has a near-exact condition listed matching Baguette's stated requirements. Thanks to PRR for finding & noticing that!

[PRR]

> tens-of-volts changes in plate and screen voltage.

I think you pulled up a VERY old 6L6 sheet, with higher ratings (later decided to be too generous)?

Maybe a 400V 21W sheet?

Most 6L6 sheets show 360V 19W. So the suggested conditions had to be walked-back, as from 34W to 26W out.

There is a lot of "history" to the 6L6 and its sheets.

And when better techniques came along, a reluctance to tag the good-old 6L6 for improved performance, preferring to baffle designers and users with "new" tube numbers (at "new" prices).

7027 is near-enuff a re-bottled re-pinned 6L6_GC_ guts, but the suggested conditions show fresh-sheet numbers. (Being initially higher-price, the 7027 promoters were eager to show-off.) These conditions *will* work on 6L6GC, no change.

[sluckey]

I trust PRR much more than I trust Wikipedia. Anyone can write anything on Wikipedia.

[jjasilli]

For any resident philosophy majors  , I contribute the following syllogism which appears to be logically Valid, but fundamentally Unsound:


For those who trust wikipedia , a search for "history of 7027 tube",  referenced the 6L6, and stated this was an equivalent tube.
I trust PRR much more than I trust Wikipedia. Anyone can write anything on Wikipedia.
Therefore, PRR should never write anything on Wikipedia!   

[Willabe]

 

[sluckey]

I bet you could argue the other side of that too, counselor.   

[jjasilli]

OK, fine!  At the request of the Contrarian faction, a presentation the Other side:  "Then, PRR should Always write Everything on Wikipedia" 

[sluckey]

Never! I think there should be a Book of PRR. Or maybe a PRRpedia? 

But maybe dr. gonzo could post an article on wiki?

[shooter]

Book of PRR

put me down for the paperback version, I can't afford hardbacks anymore

[tubeswell]

Wasn't the 7027A a 6L6 type tube designed for higher (above 500V) plate voltages? With a plate dissipation rating = 35W on most data sheets.  The modern remakes aren't up to the hype - they are at best a 6L6GC rebranded, or even a 25W tube. And definitely not as reliable above 500V. Might as well chuck EL34s in.


As to the OP question, on a PP OT the PrZ for Class AB1 is measured between the winding ends, so you end up with half the reflected load between each end and the centre tap.

[PRR]

> Wasn't the 7027A a 6L6 type tube designed for higher (above 500V) plate voltages?

No. All 6L6 plates will take high voltage. The early 6L6 had to be walked-back from 400V to 360V because the *base* would not stand 400V long-term. This was entirely solved long ago, but 6L6 was a big hit so there was no commercial reason to change the ratings.

> a primary or secondary source referencing the 7027 as a variant of the 6L6

There certainly would not be a primary source. RCA's marketing would not boast "just like a 6L6 but we screwed the pin-out to mess you up!" Here's how they said it:

[jazbo8]

I expect about 2 to 8 hours of calculations to provide the info, knowing only the B+ voltage and ultimately the 25-30w output.  I will provide the results once I have them.

Why 2-8 hours?! May be I don't understand the question... Is it: to find the ideal screen voltage given the B+ voltage and output power into a certain load?

[shooter]

I tried this method a couple times and it wasn't worth it, but maybe for you?

I have speakers, and dumby  loads that range from 2ish to 16ish ohms.  I start with resistors at what I *think* should be the correct OT impedance, measure VAC-rms at the "load", add 2 ohms, subtract 2 ohms.  ******MAKE SURE***** between changes you power-down, wait, change loads, verify their correct, BEFORE powering back up!.

This current amp "peaked" at 10ohms configured as an 8ohm OT, I cannot tell the difference with old ears, just oldish scope
I use 1khz as my "standard"  as you will probably read an hear, that has little bearing on 9khz, or maybe 100hz
anyway, FWIW

[tubeswell]

You calculate the PrZ using the equation


     Zout = Va/(Pa/Va)
Where:
Va = Plate voltage.
Pa = Maximum plate dissipation.   
In a P-P OT, the PrZ is between the primary winding ends - this is because in Class A operation, you have both output tubes swinging in opposing phase across the primary, making one single primary wave with 2 x the current as you would see if it were a 1-tube SE output stage. The actual reflected load for each tube is from winding end to centre tap - because the CT voltage isn't moving anywhere (except if the power supply sags). So when both tubes are conducting opposing signal phase, each tube is seeing 1/2 of P-P load.


But when one side goes into cutoff, the signal on the remaining 'on' side is the only signal that is being transformed into the secondary winding, and as the turns (VAC) ratio is now 1/2 of the Pr-sec ratio, you get the steeper 'B' load line happening.


That's how I sees it.

[PRR]

Tubeswell's logic is fine as far as it goes.

Hidden in there is the idea that the plate will swing from B+ to *zero* volts.

That does not happen. A "bottomed" tube will pull down to 50V-100V depending on several various factors

First: pick a B+ and find the idle current for the selected dissipation. 380V seems reasonable, and for guitar-market "6L6" (replaces 6L6GC) 30W is nominally safe but IMHO 25W may be a safer bet.

But: is all the 380V available to the tube? Usually, the way we add things up, and depending on G2 bias, we have cathodes standing 25V-30V up from B-. So the tubes only have say 330V to work with.

2 times 25W is 50 Watt idle heat. To get there we need 50W/330V= 0.151 Amps total cathode current. At idle this is 75mA per tube.

For a cathode-biased amplifier the total current should not change much from idle to full roar.

For an ideal amplifier, at FULL power it shifts (as TS said) to 150mA one tube, zero in other.

The tube has 330V to work with but can only pull-down to maybe 50V. So peak voltage swing is 280V.

The peak swing is 280V at 0.150A, which is 1,867 Ohms on the one tube.

We buy P-P OTs by the plate-to-plate impedance. For a CT winding this is *4 times* the one side impedance. So far we pencil 4*1.87K= 7.5K.

However tubes are not "ideal". They work better at higher current (up to a point!). For a happy audio amplifier with the usual audio tubes, we expect the total cathode current to shift-up 15%-20% at full roar; max current around 1.2 times idle current. This works "in our favor", allowing max clean output to slightly exceed half the total idle dissipation. So figure 7.5K/1.2= 6.2K load.

Yes, the 7027 sheet shows a fairly similar bottle working well with 380V, a low 0.138A idle, a high 0.170A full current, and a lower 4.5K load. This is the kind of condition you can only find when you have low-pay junior engineers who can vary all parameters and document every result. It may take them a week to test and then sort the "best" conditions. And "best" is subject to pressure from Marketing department. Those 36W and 44W conditions look suspiciously high to me. I wonder if distortion goes high, but then multiple distortion effects semi-cancel to give a low number right AT that power point. And maybe in guitar we do not care. Also with loudspeakers our load impedance is never any specific value. While 6.2K versus 4.5K looks like a real difference, a specific "8 Ohm" speaker may be 7 Ohms or 10 Ohms.

(Also the published spec-sheet boast-point runs hotter than my assumed 25W Pdiss per tube.)

-----------------
> Zout = Va/(Pa/Va)

This reduces to V^2/P, posted in the #2 message above.

I can't make this work-out for any self-bias amp.

6V6: 280V, 12Wdiss, V^2/P= 3.3K, sheet says 8K.
7027: 380V, 25Wdiss, V^2/P= 2.9K, sheet says 4.5K.

I believe it is missing a root-2 for sine-like waveforms, and several factors of "2" many of which combine. I think the missing factor is ideally 2.828, for real tube nonlinearity about 1.1 to 1.3 times lower (depending a little on Marketing pressure) for 2.6 to 2.2 lower.

[tubeswell]

-----------------
> Zout = Va/(Pa/Va)

This reduces to V^2/P, posted in the #2 message above.

I can't make this work-out for any self-bias amp.

6V6: 280V, 12Wdiss, V^2/P= 3.3K, sheet says 8K.
7027: 380V, 25Wdiss, V^2/P= 2.9K, sheet says 4.5K.

I believe it is missing a root-2 for sine-like waveforms, and several factors of "2" many of which combine. I think the missing factor is ideally 2.828, for real tube nonlinearity about 1.1 to 1.3 times lower (depending a little on Marketing pressure) for 2.6 to 2.2 lower.

Thanks for the insights PRR. RMS -Duh!