5686 BPP

[DummyLoad]

cool little tube... it's kind of kinky, but that should add character. 

http://tubedata.milbert.com/sheets/093/5/5686.pdf

load 4-5k; 300Va and Vg2; Ib 44mA (2 tubes); Rk ~400ohm; estimate VRk around 16-18V - would that be a resonable starting point?
estimate about 5-6 watts, downhill, with electrons with a tailwind.

--DL

[tubeswell]

Doesn't it say 9k load resistance (for Class A?), so wouldn't 7-9k be more reasonable for a pp pair?

(BTW - Interesting how the suppressor/cathode has 3 different pin connections (and the screen has 2 pin connections))

[DummyLoad]

(BTW - Interesting how the suppressor/cathode has 3 different pin connections (and the screen has 2 pin connections))

tube was designed for class C PP 200-400MHz RF - duplicate tie points are to simplify circuit assembly to minimize interconnection capacitance.

Doesn't it say 9k load resistance (for Class A?), so wouldn't 7-9k be more reasonable for a pp pair?

convert from PP to A1 single tube; double the load - A1 single tube to PP; 1/2 the load - assuming B+ & Vg2 are maintained. what PRR learnt me anyho. ;) in reality, a wide range of loads would work - trim bias to taste without seeing red, of course.

 

--DL

[HotBluePlates]

(BTW - Interesting how the suppressor/cathode has 3 different pin connections (and the screen has 2 pin connections))

tube was designed for class C PP 200-400MHz RF - duplicate tie points are to simplify circuit assembly to minimize interconnection capacitance.

Given that the tube is designed for RF, and has nice low input capacitance, you'll probably want to slug it with a nice big grid stopper to reduce the chance of oscillation.

Doesn't it say 9k load resistance (for Class A?), so wouldn't 7-9k be more reasonable for a pp pair?

convert from PP to A1 single tube; double the load - A1 single tube to PP; 1/2 the load - assuming B+ & Vg2 are maintained. what PRR learnt me anyho. ;) in reality, a wide range of loads would work - trim bias to taste without seeing red, of course.

We oughta draw some loadlines.

Probably best to use Va, Vg2=250vdc for 2 reasons: it is the maximum rating cited, and the curves are drawn for Vg2 = 250v.

I don't know if I'm misquoting PRR or if I've got it wrong. But knee-jerk guess, a likely good load for push-pull (plate-to-plate) is numerically equal to a good SE load. So if 9k is a good SE load, 9k plate-to-plate is a good starting point for a push-pull load.

For calculation, you take the plate-to-plate load, cut that in half. That's the load a single tube sees when both tubes are passing current. When one side cuts off, the remaining side sees plate-to-plate/4.

Also, you can do initial figuring by using the plate curves for a single tube. Instead of drawing the load line using the total (plate-to-plate) impedance and running through the operating point (as you would for SE), you draw a line representing 1/2 plate-to-plate load and running down to the ordinate (Va,0mA). That is, one point of your loadline is on the plate voltage axis at the proposed plate voltage and no current.

That technique allows you to figure power output quickly, but not necessarily distortion. It assumes you stay basically class A. If you have to do "true figuring" (and we usually don't), you draw a composite set of curves. Can't say I've ever seen anyone really do that.

Most of the time, people figure using some math formula, assume the tube to be perfectly linear, with equidistant grid curves and able to pull the plate to 0v, and then realize that their power calculations will overestimate what they'll really get.

[DummyLoad]

A1 single tube to PP; 1/2 the load - assuming B+ & Vg2 are maintained. what PRR learnt me anyho

correcting myself - sorry that's wrong. for class A you double the load - becomes Ra-a.

[kagliostro]

  

I was sure if 9k is a good load for SE

for AB¹ PP a good load will be 4.5Kohm ~ 5.00Kohm

     

Kagliostro

[DummyLoad]

I was sure if 9k is a good load for SE


it is. read the datasheet. class A1 amplifier 250V; 9K load; 2.7W. - i was correcting myself in one of my posts.

i should read it too... specifies up to 160MHz in class C not 200-400MHz.

to make a PP - you'd double Ra for an Ra-a of 18K. closest thing i can find is 10K hammond 1609 and probably what i'd use with a deeper bias.

ignoring the prideful on fleabay, several sellers have listings for 4-6 bucks ea., so it won't bust the bank to experiment. they seem plentiful - nebraska surplus sales lists them for 10 bucks ea., ESRC lists them for 5 bucks ea.

--DL

[kagliostro]

Sorry DummyLoad

or I misunderstand you or you misunderstand me

I've read the datasheet (the part I post here) as if A₁ was intended for the use in an SE audio amp

so I was thinking it is necessary a 9Kohm load (at the stated voltages) to have best performance with 5686 in an SE audio amp

than I assumed if 9Kohm is good for SE, for an audio use in PP in class AB¹ the correct load is ~4.5Kohm

But now I'm not sure I've understand  

Kagliostro

EDIT: May be I understand where I was wrong

Assuming SE optimum load is 9Kohm for PP the impedance must be doubled

the impedance become HALF if I double the tubes in a PP amp (as example from 2 x el 84 to 4 x el84 the impedance change from 8Kohm to 4Kohm)

in an SE amp 1 x el84 = 4Kohm load - in a PP amp the number of tubes doubles and also the impedance doubles 2 x el84 = 8Kohm

[DummyLoad]

in an SE amp 1 x el84 = 4Kohm load - in a PP amp the number of tubes doubles and also the impedance doubles 2 x el84 = 8Kohm

i believe that is correct - and as long as you maintain A1 operation (360deg current conduction).

--DL

[tubeswell]

For a good discussion of the load resistance of output tubes, see JM Fahey's post in this thread from the ampage forum.

http://music-electronics-forum.com/t23120/

[DummyLoad]

The sound of a shellac record transmitted on AM with a 3 kHz audio bandwidth
and 5% THD was a cause to get drunk. Certainly if Marlena Detriech was singing.

ref: http://www.turneraudio.com.au/loadmatch4-pp-beamtetrodes.htm

truer words are yet to be written. :-\

--DL

[HotBluePlates]

For a good discussion of the load resistance of output tubes, see JM Fahey's post in this thread from the ampage forum.

http://music-electronics-forum.com/t23120/

There are errors in what he presents. The most significant is that if you are discussing a class AB design, and using a set of curves from a data sheet, the operating (idle) bias point will almost never lie on the load line. If you're discussing a class A design, the operating point must lie on the loadline (at least before the line is corrected for rectification effects).

The bottom line is that for pentodes or beam power tubes, you will never find a simple formula for maximum power output, the way you can find one for triodes. All authoritative texts suggest sketching load lines arbitrarily, then evaluating each for output power and distortion; you then select the load providing the characteristics you need.

[PRR]

> convert from PP to A1 single tube

Convert from Class A PP to A1 single tube....

The further you go into AB, the less you can direct-translate from pair to single.

There's an additional fudge factor. In SE, distortion gets high before the tube is slammed. In push-pull, much of this distortion is cancelled, distortion may stay low all the way to clipping. That doesn't affect load-line slope (load impedance) very much, but does mean two tubes in PP will usually claim more than double the output watts of one tube SE. (6V6: 5.7W SE, 14W PP)

> pentodes or beam power tubes, you will never find a simple formula for maximum power output, the way you can find one for triodes

Disagree.

There's at least four cases for triodes. Several are "open-ended". For maximum power for a given dissipation, no limit on supply voltage, you raise the voltage and impedance and keep getting improved distortion, efficiency, and output. The gotcha is that improvement is slow, high-Z transformers are problematic, and high-voltage plate-stuff is costly. Knowing this, the tube designer specced a plate-voltage limit which "seems reasonable".

For hard-knee SE pentodes, with Vg2 near plate supply, you pick a load at-least twice the tube's triode plate resistance. We also like under-10K for best transformer construction. Then you select voltage and current so the ratio is the load impedance and they multiply to the dissipation.

For 5686: triode Rp at high output is 3.2K. Select SE load greater than 6.4K.

Max Pdiss is 7.5W.

Assume various available supply voltages:

150V: to get 7.5W Pdiss we need 50mA. Ratio 150V/50mA is 3K. Load is too low. (Pentode won't swing the 100mA peak with this G2.)

500V: to get 7.5W Pdiss we need 15mA. Ratio 500V/15mA is 33K. This is not a happy winding: wire too thin, capacitance sucks. Combined with plate-stuff costs, the designer felt you didn't want to go here.

300V: to get 7.5W Pdiss we need 25mA. Ratio 300V/25mA is 12K. This would be a fine load except it exceeds the 250V rating. In this case, a "dependable" tube at warehouse prices in a nobody-dies application, 300V may be worth trying.

250V: to get 7.5W Pdiss we need 30mA. Ratio 250V/30mA is 8.3K. This is a fine load, no ratings exceeded. Two ratings (Vp and Pdiss) are "hit", which is bad for "reliable", but acceptable for many other uses.

This is near the datasheet 250V 27mA 9K load condition. 250V/27mA= 9.3K ballpark load.

(The datasheet 250V 27mA 9K load condition takes 100% rated Vp but derates Pdiss and output power to about 83%. In an airplane, it is better to have 83% output all the time than 100% output 83% of the time. {The tradeoff is not that exact.})

The same will apply for push-pull. Yes, the total winding goes up to 16K (not 4K!) which is awkward.

[DummyLoad]

so then...

hammond 1608 - 8k:4/8/16

250V + hammond 1608 w/ 8/16/32 secondary loads - 16k reflected.

...or, damn the torpedoes, we don't care 'bout no hi-rel crap...

300V + hammond 1609 10k:4/8/16

run 8/16/32 load - 18k reflected.

high gain 6AU6 pre & 6AU6 splitter, might make a nice little P-P A1 @ 5W practice amp.

couple of 8ohm 6"x9" w/ "whizzer cones" in series. build the whole enchilada built into an old black samsonaite brief case - might just be quirky enough to be entertaining. the cold war rocker

--mu magnus.

[DummyLoad]

was thinking something along the lines of what's attached for small SE amp suitable for bedroom use.

5755 are a hyG, low microphonic, industrial/military grade dou-triode mfg. by raytheon. they unfortunately have a high plate resistance for a hi mu tube and a different pinning than the 12An7 family. tube was based on a western electric 420A. still available for a reasonable amount of scratch (about 5buks ea.). 

i think what is shown may be a fair work-around to the 5755's high rp. it will have somewhat less gain then a champ, but you could always add another bottle - perhaps a variable gain stage and CF. the tranny and inductor i spec'd will handle the load of another preamp bottle without issue. or you could simply work with 12AX7 and 12AX7 RP and RK values. i attached .SCH for editing.

--DL

[RicharD]

Here's what I came up with.

[DummyLoad]

you can run 250V on the screens of 5686. your PS schema may work better for 6CD6/6BD5 etc.. throw the whole B+ enchilada at the 5686. love the preamp.

--DL

[RicharD]

I want to experiment with the screen voltage.  The 6AU6 values are right off the RC chart.  The other idea is to use a 12BZ7.  I'm just doodling.

[DummyLoad]

The 6AU6 values are right off the RC chart

are you going to prototype it?

--DL

[RicharD]

Ayup.

[DummyLoad]

Ayup.

you have 5686?

[RicharD]

uh.... maybe.  Gotta bust out a shovel and dig for 1.

[DummyLoad]

i have 3 on hand. if you can't find one, you're welcome to one of mine.

-- DL

[DummyLoad]

digging through the stash, i located a pair RCA hi-fi single tube transformers salvaged from a stereo console. they have a funky tap for screen hum-bucking - just use the red and blue wires as you normally would with SE and tape off the brown screen wire. IIRC they measured around 7K : 8ohms.  they should work just fine for this application. message me if you'd like to give one a spin.

--DL

[RicharD]

I dug through several boxes.  No 5686's.  I do have like a kazillion 6CL6's.  Looks similar.

http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6CL6.pdf

[DummyLoad]

I dug through several boxes.  No 5686's.  I do have like a kazillion 6CL6's.  Looks similar.

http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6CL6.pdf

i got some of those as well from steve on my last trip to junk land. he had a drawer full and at first glance of datasheets they seemed interesting, but screen is delicate like sweep tube. should work well with your doubler PS.

10K load, 300Vp, 150Vg2, Ib 25mA, Vgk ~-3.5, 7.5W Pdiss - Rk 140ohm 2W - that's where i'd start. :-)

i HAVE 3 5686 if you want one, you can have one for 2 bux or a cup of coffee.

--DL

[RicharD]

Here's what I protoboarded.

Works pretty well.  Used a reverb transformer.  I calculated 100 ohms, ran with 120 ohms and it's spot on 25mA.  Works for me.

[RicharD]

Potty 2.1

Got rid of the low screen voltage.  Played around with the BZ bias.  I tried a choke between the plate and screen but opted for a resistor whereas there wasn't any noticeable difference.

[DummyLoad]

it's one tube class A1 - choke would make most difference between first filter and plate.

[tubeswell]

I tried a choke between the plate and screen but opted for a resistor whereas there wasn't any noticeable difference.

The only difference you are likely to notice between a choke and resistor between the plate and screen supply in an SE amp is when you get to full power. However, to get the hum-cancelling benefit out of a choke in an SE amp, the ideal place for a it is in a (whole-of-supply) CLC filter between the rectifier and the plate supply.

[HotBluePlates]

You gents are certainly doing your part to Keep Austin Weird. Rock on fellas.

[DummyLoad]

You gents are certainly doing your part to Keep Austin Weird. Rock on fellas.

thanks HBP!

stay tuned! more weirdness will certainly ensue...  

if it's odd, deranged, or strange, we'll probably love it!

peace.

--DL

[kagliostro]

May be this Transformer/choke used in an old SE amp has some special purpose about cancelling hum ?

Kagliostro

[RicharD]

Kagliostro, Could you post the rest of that schematic?

I am building the potty mouth 2.1 this weekend into a pair of 2100 electrical junction boxes.  All spare and used parts.  Total expense = 0.

[kagliostro]

No, sorry I can't

Kagliostro

[RicharD]

I forgot to tell ya'll that it's done.

[Willabe]

That is too cool.                  The power supply is even shielded from the circuit.

How's it sound in it's new home?


                  Brad

[RicharD]

I actually built it in 2 sections, married the boxes with chase nipples, and landed the last 5 wire connections.  It sounds..... meh with that soggy old Utah speaker.  I thought it sounded better driving a plain old 4x12", but that defeats the purpose of it being a desktop bedroom amp.  I need to find an old 6 x 9 wizzer cone somewhere.  It does sound much better when the amp is not sitting atop the speaker.  12BZ7's tend to be microphonic, especially considering I tweaked stage 1 for big gain.

[John]

Unreal, I love it!

[RicharD]

Chassis inspired by the Squirrel Monkey, http://www.solorb.com/elect/musiccirc/squirrelmonkey1/ a link DL posted.  I am an electrical contractor.  My 2nd amp (1st ground up) was built in a NEMA1 enclosure.  A serious percentage of this build came out of the back of my truck.  The transformers are MR16 lighting transformers.  They were headed for the trash post lighting retrofit.  Boxes, chase nipples, connectors, and tek screws all spare electrical parts.  Every single resistor and capacitor came out of the Zilla part boxes.  The sockets are reclaimed used parts.  The OPT is a reverb tranny which has seen it's share of abuse on Zilla as well.  I did have my wife spend .99 on some 4-40 nuts, so technically I didn't pull off the free amp build.   

[DummyLoad]

nice... i feel like a proud junk yard owner parent.

It sounds..... meh with that soggy old Utah speaker.

needs NFB?

peace.

--DL

[RicharD]

The 2nd stage is a plate follower therefore it has NFB, just not in a typical geetar amp location.  You've heard that old Utah speaker before and IIRC, your description of it requires language we don't use in this forum.

[DummyLoad]

You've heard that old Utah speaker before and IIRC, your description of it requires language we don't use in this forum.

that one? oh $#!t! - never mind...

still, NFB around output stage lowers the output stage Z and improves damping factor to shitty utah speaker. this can help SE amp immensely. i $#!t, said shitty - are we allowed to say that here?

peace.

--DL

[DummyLoad]

very cool little amp - richard brought it by for demo. it sounds killer for a little ~1w. we found and fixed a minor arcing problem and all was good again in amp-land. very impressive for a parts-bin throw_it_together.

--DL

[murrayatuptown]

Hey, I got a two-for-one good deal here! I followed a 5686 link and also found Squirrel-Monkey.

I saved some Compatrons but never a socket. I don't remember, but think everything I pulled them out of had riveted sockets and the remains went overboard.

Can one still buy Compactron sockets? (Obviously, I have no idea and haven't looked yet. If I go away and search I'll probably forget to come back & post my question later (squirrel half speaking...)

🐿

[sluckey]

Plenty on eBay.

[murrayatuptown]

Very good. Thanks.