Meet the "Little QTs"

[tube fanatic]

I was going to concoct a story about spending a few days with Marty and the twins (the real "QTs") but figured no one would believe me anyway........

A while ago I was asked to prototype a pair of high power (a relative term- in my world that means 3-5 watts) mono integrated amps on the smallest possible chassis (7 X 9 was the best I could do) using output tubes which have lots of possibilities for rolling, and an ability to be driven by sources with lower output than the usual CD player.  I took the bait and the guy, who's old enough to have owned a Grommes Little Genie and Little Jewel, named these the "little QTs."  What can I say?

The biggest issue when building an amp with a self contained power supply on a small chassis is hum from the AC wiring getting into the signal circuitry.  So, the electromechanical layout can be a bit critical.  AC wiring must be kept as far as possible from the amp's audio input jack and volume control- easy to do with a larger chassis, but not so easy when going this small.  The filament wiring is most important given its proximity to the tube sockets and associated components.  One solution is to go with a DC filament supply; however, that's a pain to implement when there are chassis size issues.  The most popular method involves tightly twisting the filament wires (you can research why that works) and grounding one lead, but even that isn't a perfect solution.  So, I used a tried and true method: bringing the filament wiring to the output tube socket opposite the grid #1 pin, using a voltage divider across the filament winding of the xfmr with the midpoint "dc elevated" (by borrowing voltage from the cathode), and using shielded cable to bring the filament supply to the voltage amplifier (i.e. driver).  The result is that there is only a slight hum audible from the speakers when an ear is within an inch or two of the woofers.  Regarding rolling output tubes, I decided to use 6V6s since there are so many types from which to choose.  Contrary to my usual use of a grid leak biased triode, the driver is a 6SJ7 pentode configured with fairly high gain.  Since it's available in both a metal and glass version, the former was used to prevent any hum issues from the magnetic field of the power xfmr.  Power output at the onset of clipping is around 3.75W at 1 kHz into a resistive 8 ohm load.  These amps will literally blow most listeners out of a moderate sized room when used with the usual Heritage series speakers, and many of the Reference series speakers as well.  Sonically, it sounds terrific with a very potent low end, and a midrange/high frequency smoothness which rivals some of the SETs.  The latter is a result of adjusting the internal, variable, high frequency filter accordingly.  It can be adjusted to suit almost any listening taste.  After hours of use, the amp is only moderately warm (this in a non-air conditioned room).  I've attached a number of pics and the schematic (let me know if anything doesn't look right- I drew it out quickly) in case anyone wants to build their own.  For a modest expense you can have a pair of mono amps (which have no equal when it comes to providing a huge sound stage), built like battleships, and which offer features and a listening experience which is probably impossible to duplicate anywhere near its cost.

Feel free to PM with any questions if you decide to build your own!

Maynard

little-qt-9.pdf

[CECAA850]

You know I'M not building any

 

Those are beautiful Maynard.  Do you still have them or are they "home"?

[erik2A3]

Maynard,

Raising filaments can be effective. Also, since chassis space limits a fully wave bridge rectifier and ripple filter circuit, another way to do this (which could be useful if one's filament supply is not center-tapped) is to make a sort of pseudo center-tap by way of connecting each leg of the filament leads to one end of a resistor in the range of 100 ohms or so. I know you know where I'm heading with this!

The opposite sides of the two resistors are grounded.

If using a separate filament transformer with a center Tap, the classic way to do it would be to just ground the center tap.

Very nice little amp!

[tube fanatic]

Thanks Carl.  The amps are still in their "birth home" burning in for a while (I need to have my fun too!!!).  And, Erik, I'm familiar with your method.  Depending on my mood I use a number of variations on the same theme.  As you know, the smaller Hammond power xfmrs don't offer a center tapped filament winding.  So, I either use the method shown in the schematic above, or use a 100 ohm/3 watt wirewound pot with the wiper either grounded or DC elevated (allowing for adjustment for minimum hum with changes of tubes), and so on.  The DC elevation seems to be a bit quieter than just grounding the junction of the 2 resistors or the wiper.  I still prefer DC filament supplies over anything else.  Nothing can touch it for creating a dead silent amp, especially if the ripple is brought down to less than 1%.  Now I'm hearing some rumblings about a SET version of these on the same size chassis.  We'll see where that goes!

Maynard

[thebes]

Maynard, you never cease to amaze me. You are truly the maestro of the so-called "flea" amps, which is usually intended as an insult but in your case is a compliment. 

 

I really don't think my fellow Forum members realize just what talent and knowledge you bring to our arcane pursuit of musical passion.

 

Folks, I'm here to tell ya that everything Maynard builds is a seriously good sounding piece of equipment.

 

Are they perfect.? No they have limitations.  Tell me this though, can that not be said of everything else in nature, or within our hobby. 

 

I strongly commend you guys to pay attention, and consider building one of his creations for yourself.

 

I've just built my second TubeFanatic design, a SEP, and will have a full report for you in the near future.

 

What I particularly like about this build is the use of those 6V6's.  A great sounding tube with an almost endless cornucopia of tube rolling for those inclined that way,.

 

Good stuff Maynard.

Edited June 6, 2015 by thebes

[Thaddeus Smith]

I know what he's all about, and I agree with everything stated above.

[mike stehr]

Ditto

[erik2A3]

Maynard

If your preference for filaments, in my view there is enough space both above and within the chassis for a separate filament transformer, full-wave bridge and small filter section (10,000uf or so) to smooth DC. As you know, the first filter cap often has the subsequent effect of boosting output voltage, which is desirable on plates, but not with filaments. So, an additional dropping resistor and then another cap in the range of the first would do the trick. Those valves use indirectly heated cathodes, though, which IMO makes the need for DC heaters much less mandatory, particularly in the output stage. Were you designing the circuit around filamentary cathode tubes (eg 2a3, 300b, etc) the use of a hum null potentiometer (in the case of preferred AC) is pretty much required. I have built amps around triodes like the above many times, and have experimented with both AC and DC (including regulated DC) and always seem to prefer what I perceive to be a somewhat more sort of organic quality with AC. Definitely something that is builder/user/listener dependent. No right or wrong way of doing it providing current and voltage demands are met.

Didn't have time before to check the schematic, but it's nicely done. Always good to see a bleeder circuit on the output of the power supply. I agree that 1watt would be very much a minimum power rating. Just for kicks (if you felt so inclined), you could experiment with the use of a lower value (47uf or so) poly film cap as the bypass for the output stage cathode resistor. If you are needing such a high value electrolytic based on measured drop off of low end response, then of course 47uf is not going to do it. I have even used motor run caps in that position with extremely good results, but those obviously take up quite a bit of space.

Observation: The small cap across the volume control: the schematic indicates it's connected between input to the pot and ground. I am assuming your inclusion of that cap is for the same reason I use them on virtually every integrated amp I own, after having found the same neat and simple trick done on many guitar amps. Just more accustomed to seeing the connection made between input and wiper vs input and ground as shown. In any event, are you using a silver mica cap there?

Again very neat little amp! I always like your neatly parallel-perpendicular wiring and placement of parts. Makes for a tidy interior.

Edited June 6, 2015 by erik2A3

[erik2A3]

And Maynard: I appreciate your willingness to share. Since I joined this forum (initially in 2002), that kind of generosity has most certainly not been the norm. Thank you

Edited June 6, 2015 by erik2A3

[tube fanatic]

Erik, your point about setting up a bridge and crc filter for the filaments is well taken.  Actually, when calculating the ripple based on the current draw for these two tubes (0.75A), it would only be necessary to use around 3000 uf on either side of an approximate 4 ohm resistor to get the ripple at the output cap down to about 0.7%.  I may implement something like that for the next variant of this amp, although if its a SET, the lower gain will probably make doing that unnecessary.  Regarding the cathode bypass for the 6V6, the calculated value for 20 Hz is around 320 uf.  I tend to use what I have in stock and often wind up going higher than the calculated value.  It doesn't hurt anything to do that.  But, some believe that the bypass cap should only be selected for the lowest frequency that the amp is likely to reproduce in which case a smaller value would certainly work well.  When power is removed, there's usually enough current draw while the filament of the output tube cools to pull the electrolytic voltage down most of the way.  I just feel better knowing that after a few minutes the caps will go to zero!  The 240k resistor used here only draws a bit over 1 ma, so the dissipation is in the range of a quarter watt.  Lastly, the 2200 pf cap across the volume control is part of the series RC lowpass filter used for the high frequency response adjustment.  But, even when not using that feature, I usually put a 1000-2000 pf cap across the control to act as a parasitic suppressor.  It definitely helps to eliminate the fairly typical damped ringing, and occasional overshoot, noted when looking at the high frequency square waves of a SEP.  Running it from the wiper to ground instead shouldn't make much difference.  Always glad to help and I thank you for the compliments!

And Marty and Michael, I also appreciate your thoughtful and kind comments.  The only motivation I need to keep going with these amps is knowing that they provide folks with musical enjoyment which, as we know, is what being an audiophile is all about!  Thanks again...

Maynard

[erik2A3]

Maynard

Excellent - A truly well-written explanation. Thanks! I think you are one who can appreciate this: speaking of tubes with metal envelopes, my dad (building vacuum radios since about age 10), was recently making his own metal envelope tubes out of copper tubing. Fascinating process! They were for VLF receivers he used to detect and record, as they are sometimes called, "whistlers," which are potentially musical-sounding by products of lightening.

And absolutely right on PSU bleeder!

Single-ended 6v6 can be incredible! Job well-done!

Edited June 7, 2015 by erik2A3

[Guest thesloth]

Are you positive you are hearing heater hum (60Hz) from the speakers? What are you measuring for 60Hz at the output? Most likely culprit since you went with a SE design would be the 120Hz hum if using a full wave rectifier. Single ended designs require much more filtering going to the plate of the power tube when compared to push pull designs. Push Pull stages cancel noise due to common mode.

 

Do you have a schematic of your power supply?

[tube fanatic]

Sloth, it's nice to have another tube guy on here!  The schematic is posted above as a pdf under the pics.  Let's hold off on the technical discussion for now in the hope that Chad will give the proposed new forum category a trial run.  I'm trying to save up a number of topics to get it going!

 

Maynard

[Guest thesloth]

Oh cool a new forum catagory? Tube catagory?

 

I will wait indeed. Cool to see others are into tubes still too 

[tube fanatic]

Oh cool a new forum catagory? Tube catagory?

 

I will wait indeed. Cool to see others are into tubes still too 

Unfortunately, I have not received Chad's reply to my previous email or to a PM I sent him this morning (I appreciate his being very busy, however.....).  Will keep everyone posted if I hear from him.  If not, I'll try a phone call!

 

Maynard

[Guest thesloth]

Hello Maynard, I just looked at the attached schematic and the PS looks pretty good. Not too sure what speakers you were using but some rough calculations show a possible 40mV ripple at node B. An output transformer w/ a turns ratio of 25:1 (5k:8) should yield around 1.6mV ripple at the speaker maybe audible with very efficient speakers.

 

Another cause of the hum could be ripple at the screen supply on the input pentode. I have not calculated this yet but I would look into these matters before going with DC heaters.

[tube fanatic]

Hello Maynard, I just looked at the attached schematic and the PS looks pretty good. Not too sure what speakers you were using but some rough calculations show a possible 40mV ripple at node B. An output transformer w/ a turns ratio of 25:1 (5k:8) should yield around 1.6mV ripple at the speaker maybe audible with very efficient speakers.

 

Another cause of the hum could be ripple at the screen supply on the input pentode. I have not calculated this yet but I would look into these matters before going with DC heaters.

Hi Sloth!  Just a quick comment on your calculations of the ripple- quite good indeed!  The ripple at node B was around 30-35 mv give or take.  As far as the plate/screen supply of the driver goes, the ripple there is so low as to be virtually non-existent.  The speakers are RF-15s which have a claimed sensitivity of 96db.  Again, let's hold off on more technical stuff until we know whether we can have a dedicated tube section.  I'll give you my thoughts on heater/cathode leakage which, as you know, exists to some degree in all tubes.  And your idea about a community designed amp optimized for Klipsch speakers is terrific (especially if it is simple enough for someone with limited experience to construct).

Maynard