6L6 types VS. 1614 ratings and idle thoughts about the 6L6GC

[Opcom]

Comparing the 6L6 and 1614 in push pull AB1 audio service, it looks like the biggest difference is the 1614's ability to take up to 375V on the plate and 300V on the screen, compared to the 6L6's 360V and 270V respectively.

The 1614 also has ICAS ratings of 550V on the plate and 400V on the screen. Not having seen the insides of those metal tubes or the explanation in literature, I can't know what any other differences might be to allow for the higher voltage rating.

The 1614 has a 21W CCS plate rating, and the 6L6 a 19W rating, but in some books the non-GC types of 6L6 are rated 21W. The 2-Watt difference is not so much compared to the voltage rating anyway. The 1619 has an ICAS plate rating of 25W.

My idle thought is what voltages above the 6L6 ratings have been run on 6L6s? Is there any real reason the 6L6 can't be run at the higher 1614 voltages as long as it is considered ICAS duty? What is the limiting factor? The base and a ceramic socket should be good up to 700-800V so why not 550V?

Today 1614 are rare and non-GC 6L6s are also not so common. Not common enough to consume in experiments.

Next thought is about the 6L6GC. It's rated 500V plate and 450V screen, but additionally it carries much higher current per pair in AB1 service, 210mA at 450V vs the 1614's 160mA at 530V ICAS rating.

I'm thinking that the 6L6GC plate voltage could be pushed to 550V, and in the same AB1 circuit be loaded for 210mA, and make more power, maybe 70W, as long as it is understood that it is ICAS type service due to plate dissipation issues.

[John K5PRO]

Somewhere recently I read about the RCA 1614 but nothing stuck between my ears about it. The 6L6GC is definitely an improvement. I read somewhere that enterprising hams stuck the metal 6L6 cans into oil and pushed them to 70 watts in ICAS.

Eric Barbour wrote about the 6L6 and its variants in 1996 Issue 4, Vol 1 Vacuum Tube Valley. A later evaluation of the sound of 20 6L6 types was in issue 13 in 2000. He had a photo of a 1614 but not much description of the difference.

Another chap has a website on some 6L6's:
http://www.eecs.berkeley.edu/~loarie/blues.html

In this issue of Radiotronics archived online, there is a decent technical description of beam power tubes like 6L6GC and Pentodes like 6BQ5 for hi fi use:
http://tubedata.itchurch.org/other/AWV_Radiotronics/Radiotronics_1962/1962_02_AWV_Radiotronics_27_02.pdf

[WA5VGO]

The 1614 is nothing but a hand picked 6L6 that was run through some extra factory tests to check its suitability for rf service. In fact, if you look at the crimp on the base, you'll see they're stamped "6L6". Personally, I've never seen a place a 6L6 wouldn't work in the place of 1614.

Darrell

[w4bfs]

I assume you want to run 550V or higher .... look at Sovtek 5881 spec sheet ....6l6gc at up to 550/600V .... used to be cheap but not any longer.... guitar amp boyz like em ok

[WBear2GCR]

If it is POWAH that you want, then 6550.

WRT these ratings, my view is that the import tubes must be derated.
Expect ur talking NOS tubes?

An urban legend I heard had someone running the 1614 upside down in a bath of coolant and squeezing out
an inordinate amount of power!

The other thing is that if you want voltage and ICAS the 807 is hard to beat. I have a schematic for
4 x 807 in PP parallel at ~700-750v B+ in AB2 that claimed 200 watts. So a pair would do 100watts, if that
holds... of course people say the 807 is "just" a 6L6 with the plate out the top...

                 _-_-

[KD6VXI]

The biggest 6l6 replacement (with mild to no mods)  is the kt120.  I've used a kt88 as a replacement in a single ended 6l6 mod x 6l6gc.

The kt88 would load up as the rf tube,  but did not give me any additional power (it was lower output,  actually).   I didn't pursue any additional mods / tuning other than dipping and loading with the stick circuit.   But,  I was able to get it to amplify on 29.1.

The 6550 is another hot tube,  and is a bit smaller than the kt88, if space is a big deal.   I've used it in places where the kt88 wouldn't fit,  or just barely fit.

--Shane
KD6VXI

[KD6VXI]

Oh,  and on the subject of voltage.   My TX ran 400 volts,  and was capable of 100 pct modulation.   Maybe slightly more,  but I'll stipulate a hundred pct....   Lol

--Shane
KD6VXI

[ka1tdq]

I've seen some articles in past ARRL publications using the 6L6 as a PW final RF stage. A person could make a still-in-production tube transmitter using a 6L6GC modulated by a pair. I'm guessing that anything above 75 meters would need neutralization.

You could drive the grid with a watt or two of 50 ohm RF through a tank coupling network, like on the AMWindow.

Jon
KA1TDQ

[Opcom]

This is long because y'all made me think a lot more about it and really hit the books.

well sure, I think the 6L6 including its modern form has plenty of potential. To make a really modern tube transmitter, a switching supply like the boost converter used in PFC controllers could be used. Those are insanely small, 4x4 board, some making 400V@1A via FET and toroid choke for various gear. Sony used them in that general size for some 3-CRT projectors. I used to fix them all the time. One could be made for any reasonable voltage as boost circuits are very simple.

An all-6L6 rig would be really interesting, but would it do 160-10M? The 6550 was shown to do 21MHz in an ARRL handbook as a SSB linear final. High-capacitance tubes like the 813 work on 10M by using a tapped pi coil. So why not the 6L6GC unless basing and lead wires inside the tube make it a nightmare of instability.

There have been articles about economical modulators in which higher than normal plate voltages were used. It looks easier to use higher voltage than to try to demand higher current. The dissipation averaged over time was not exceeded due to speech use without a lot of compression. In comparison to a tone, speech of course has a lower average power for the same peak power.

This might seem like a digression due to the different tube structure but the first article I know of that was about getting more than rated power, twice the regular amount, from a set of tubes in modulator service was in 1935 August QST.

It used type 46's to make 50 Watts, which is more than twice the datasheet's 20W output rating. The 46 is described in the RCA RC-13 as a 400V, 10 watt, directly heated 2.5V@1.75A filament, Tetrode with a peak plate current rating of 200mA. It is rated 20W output into 5800 Ohms CT. I'm not sure if it has a modern counterpart.
Few modern tubes seem to have peak plate current ratings in the data sheets, and this is a critical point when trying to get as much as 2X the power from a given tube.
Do directly heated tubes having high filament currents permit much larger peak currents than indirectly heated ones of the same general ratings?
(side note: The QST article and the RCA RC-13 data for the 46 both use the "special class B" connection. Strangely, I did not find a 'regular' class AB1/AB2 push pull circuit where only the control grids are driven and screens are used at DC.)

The QST modulator ran them on 500V with cathode bias resistor of 1500 Ohms. The premise is that the pair of tubes could make 50W in voice modulator service if the pesky plate dissipation and tube current issues didn't get in the way. With the high voltage, the plate to plate load impedance for the operation was 4800 (lower distortion) to 7600 (high efficiency) Ohms for a 50W output. The article claims the peak current rating was not exceeded and the average current was 70-80mA with a voice waveform. It did not state the plate current with a sine wave so there isn't a direct comparison on this.

That 46 Tetrode article seems to give similar results to the modern, in-ratings 55 Watt, 5600 Ohm p-p load operation of the 6L6GC at 450V.

Maybe a modern 6L6GC with those advertised guitar amp voltage ratings could reliably make 100W a pair in voice service with 600V on the plate without exceeding the cathode current, by using an appropriately higher plate to plate load impedance. The only requirement is the tube socket and anode has to hold off the high B+ voltage. Maybe the anode-to-beam electrode distance can be an issue for arcing in this kind of tube. Maybe not.

It would be very interesting to find out just how much audio power could be had from a pair of 6L6GCs at high voltage. I would almost bet money 100W can be had in voice service.

Compare that to the 750V, 25W, 807 (so-called 6L6 with a plate cap) which has a current rating about 100mA. It stands 1200V at the peak of the modulation wave. -which also leads me to the 6CD6, a tube with the same plate power but twice the current rating and easily more HV capability.

per pair:
46: Eb 500, Ib 150?mA, R_L 5600 Ohms CT,  75?W DC in, 50W out. <--this is from QST AUG 1935
6L6: Eb 400, Ib 230mA, R_L 3800 Ohms CT,  92W DC in, 60W out. <--this is from QST JUN 1936
807: Eb 750V, Ib 240mA, R_L 7300 Ohms CT, 180W DC in, 120W out. <--this is from RCA TT-5
6L6GC: Eb 750V, Ib 240mA, R_L 7300 Ohms CT, 180W DC in, 120W out. why not?
6CD6: Eb 1500V, Ib 460mA, R_L 7300 Ohms CT, 690W DC in, 460W out.

Might be a daydream, but let's see if the 6L6GC plates get red on the "P" in 'pucker'! Testing would most easily be done with a sine wave in order to find the real RMS power available for speech peaks, but it would have to be intermittently operated, perhaps a few seconds at a time. A digitizing or storage CRT scope would be wanted for this.

If I had a quad of 6L6GCs, I could throw them into a four-6550, 150W amp and see what happens to them at 600V. Would need to buy some replacement fuses first.. but the amp is a speech amp in voice service and heavily swamped for driving class B modulator grids.
I have no proper simulation stuff for this, and have too much paying work to do on my work bench to hack up an existing chassis and try these ideas. First got to put beans on the table, ehh.

[w4bfs]

that is certainly pushing things .... they may do it for a while but wpuld need extreme care... I placed some 6l6gc in an B18 Ampeg amp which is a 550V amp using 7027 tubes .... they worked and sounded good but were unstable and oscillated at ultrasonic freq .... not good for longevity...

sounds like 3-500 and 572b toob tales running around in ham circles nowadays .... unstable and needing extra care or circuit adjustment .... I think this is one of the reasons the military liked using a big toob in a small application at low power .... very tough

definitely like the beans on the table

[N8ETQ]

  Surprised the EL34 has not popped up in this thread.

73


/Dan

[Opcom]

The EL34 at 100W (11K p-p load) ratings already has 800V on the plates. An old-school amp building friend said he had to use ceramic sockets to avoid arcing in Marshall 'clone' amps he built in Czechoslovakia the 1980's in which he ran 800V. The Octal base or socket is not really a HV connector.

The current capability may not have been taken full advantage of though. In the Amperex 'HF656' catalog data for 100W operation it is 91mA per tube. The data sheet gives 100mA plate current in "Characteristics", but also a "design maximum" of 150mA cathode current, which would be mostly plate current. In the 100W data the plate current is 91mA and G2 current is is 19mA. Maybe 140-150W could be had from a pair by pushing the plate current up to the limit?

About oscillation in the B18 Ampeg amp, a question for me is whether the B18 includes suppressor resistors sometimes found in amps running high voltages or parallel tubes, such as 47-100 ohms in the plate leads, 100-470 Ohms in the screen leads, and 4700-10K in the grid leads.

[Opcom]

that is certainly pushing things .... they may do it for a while but wpuld need extreme care... I placed some 6l6gc in an B18 Ampeg amp which is a 550V amp using 7027 tubes .... they worked and sounded good but were unstable and oscillated at ultrasonic freq .... not good for longevity...

sounds like 3-500 and 572b toob tales running around in ham circles nowadays .... unstable and needing extra care or circuit adjustment .... I think this is one of the reasons the military liked using a big toob in a small application at low power .... very tough

definitely like the beans on the table

well yes got to have them beans! I am much more nervous about trying shenanigans with a 3-500 or 572 than with tamer stuff under 1KV or so. The explosions are smaller on the low voltage gear.

[WD5JKO]

   Pat,

   You started this post using class Ab1 in your first post. I wonder why? Let me explain why I ask that question.

First is the pin 3 problem where running the plate at high B+ can lead to arcing to pin 2 (filament), pin 4 (screen), or chassis ground. A lower B+ with the pin 3 issue will certainly be more reliable unless you can get some of the special tube sockets that have extra insulation between pins, kind of like adding little walls
around each pin. Switching to ceramic is not good enough since the arc path distance isn't lengthened.

Sure you'll here things like, "I run 800V on my Apache EL-34's and have never had a problem". I say they are lucky, or that the Apache spent most of it's life as a shelf queen.  

Let me use a computer analogy. Back in the dos days with an AT computer, many of them had 1mb of ram. There was a limit of addressing to 640K when 1mb was available to the operating system. This is similar to tubes running class Ab1 where we are limiting the peak grid swing to the threshold of grid current. If the driver can go further, and without distortion, operating into class Ab2 extends the tube dynamic range a bunch. For example, look at the original 6L6:
 
http://frank.pocnet.net/sheets/021/6/6L6.pdf

With 400v plate, class Ab1, a pair in P-P can deliver 34 watts. With the same voltage, boosting the drive into class Ab2, that goes up to 60 watts into a lower P-P load impedance! That is almost 2X the class Ab1 rating.

The trick though with Ab2 is getting a clean low impedance drive source. There are many ways to do this. One way that seems to be virtually unknown is the Power Drive circuit:

http://tubelab.com/articles/circuits/power-drive/

This is rather easy to implement to an existing amplifier that is designed for class Ab1 drive. With a little grid current, a pair of 6L6GC's should easily do 80 watts and do so at a reasonable plate voltage (450v maybe?).

Jim
Wd5JKO

[KD6VXI]

Good old expanded or extended memory addressing choices!   Thanks for that walk down memory lane.   Next you'll be bringing up irq and memory mapping.

:-)

--Shane
KD6VXI

[Opcom]

Jim,

I did start with AB1 for the 1614, but any class is OK by me. The 46's were in AB2 or what some people call B, and I didn't want to go down the road of interstage transformers. The rest of the examples for the 6L6s, 807, 6CD6, etc. were in AB1.

I agree that AB2 can take advantage of excess current capability at lower plate voltages. (I don't know what kind of ceramic sockets my friend used for 800V but they came from a military electronics factory via the black market).

I have seen the solid state Power Drive article. They are perfectly fine but I think it is OK to use small triodes if that is what is wanted. Frank's page was unavailable to me with that link however it seems to be the tubes are now redirected so that I found it at http://scottbecker.net/tube/sheets/021/6/6L6.pdf. I suspect this file is from RC-13.

This is the older Eb=400V@230mA/ Eg2=300V/ 60W conditions that seem to have been eliminated from some the newer manuals.
They lowered the maximum plate voltage at some point. The newer ones give Eb=360V@205mA/ Eg2=270V /47W conditions.

For the 60W conditions, they warn that the grid drive Z should be 500 Ohms or less. I didn't see similar data for the 360V conditions, but would assume 500 Ohms is fair enough.

The 6L6 is not in RC-12 (1934) but appears in RC-13 (1937) with the 400V rating and grid impedance admonitions.
In RC-14 (1940) the voltage was lowered to 360V and the grid impedance admonitions removed.

I found no grid current curves but there is discussion of feedback methods specifically citing the 6L6 in RC-13, page 21.
In RC-14, the notice of grid current is extremely vague, and reference to inverse feedback is as vague, referring to the entire "radio tube applications" section where the reader then has to fish out the schematic, and that same diagram starts on page 20, but the "6L6" designation has been removed.

I do not mean to pick at the tube manuals, but it is as if RCA wanted to make the 400V 60W characteristics disappear.
Could there have been a problem with driving the tubes hard (0.35W) or with the higher plate current resulting from such increased power at the 400V level?

QST April 1947 has a nice article on designing those based on the resistance coupled amplifier data. Hope that's not the annual April Fools joke. It shows a 40W amp with 6L6Gs running on 360V but claims that more voltage and current is available from the 6SN7 cathode followers than needed. There is probably no reason why it could not apply to higher plate voltages.

Now the greater mystery is uncovered. Why did RCA kill the 400V characteristics? Were they optimistic about the 6L6 in the first place? Did war shortages force them to standardize production at a lower performance level?