6V6s on their sides, and negative-feedback.

[G6Tanuki]

I'm planning on building a simple 5-Watt-per-channel stereo amp - using a 6SN7 dual-triode as the driver and a pair of 6V6s (one per channel) as the output.

Silicon diodes for the power-supply - something like 270V as the HT for the output stage, lots of microfarads for smoothing - the power-transformer's been run for a week at 100mA HT and 3A heaters, and handled it with ease. Good audio starts with a "stiff" power-supply!

There's not enough room in the case I intend to use for the 6V6s to sit upright.

Is it truly a sin to mount them horizontally?

Also, looking at a lot of the circuits for this sort of 'single ended' amplifier, they use negative-feedback from the secondary of the output transformer to the cathode of the driver. One thing I don't get is that the resistor from the output-transformer to the driver-cathode often has a parallel low-value [500-1500pF] capacitor - which surely will provide more negative-feedback at higher frequencies so compromising the amplifier's HF response?

I don't want to lose any HF response in my design! Could I perhaps wire the NFB resistor and capacitor in series, so the NFB only operates at the lower-end of the audio spectrum? Would this result in oscillation or odd phase-change issues?

[I have ******-all auditory response below 200Hz - hearing-damage from a major explosion in the 1980s]

[glowinganode]

Would a 6V6GT fit, it's not much taller than a 6SN7?

Rob.

[joebog1]

The cap is correctly wired in parallel !! Its to improve the high frequency. The cap is chosen using square wave testing at about 10 khz. The best square wave at that frequency is the correct value. The actual value will depend on the quality of the output transformer.

6V6's will be happy on their sides. There are literally hundreds of designs around using the bottles you have selected, and I must admit that a 6V6 with a nice output transformer can sound absolutely superb.

Cheers
Joe

[G6Tanuki]

Originally Posted by glowinganode

Would a 6V6GT fit, it's not much taller than a 6SN7?

When I say "6V6" I am using the term generically and mean the stubby tubular GT variety [the 6SN7 is also the GT variety].

Neither will fit upright in a 3U-high rack-chassis.

[G6Tanuki]

Originally Posted by joebog1

The cap is correctly wired in parallel !! Its to improve the high frequency. The cap is chosen using square wave testing at about 10 khz. The best square wave at that frequency is the correct value. The actual value will depend on the quality of the output transformer.

The output transformers I'm using are salvaged from a scrap early-1960s radiogram where they were driven by two EL84s. I'm wondering if I could optimise the design by selecting the NFB capacitor for something closer to 15KHz rather than 10KHz? I like my high-frequency response to have Bat-disturbing impact.

Would I do this by loading the output with a simple non-inductive resistor, or does it need to be done driving the actual speakers it'll drive in real-life? [which could be tricky because I haven't decided on these yet!]

NFB in audio-output stages is something I'm not that familiar with - all of my previous serious audio power-amps have been AM modulators with the "load" a Class-C RF amp not a speaker.

[joebog1]

I use a dummy load resistor to set up the feedback. I dont actually use a lot of feedback, probably 8-10dB as I like that "valve" sound ( read: a bit of distortion) The transformers will be fine, although in my experience ( in Australian radiograms, anyway) they cheated on the iron. I would like to see at least 1 inch cores, instead of the 5/8" or so that were fitted here. 15khz I think will be optimistic as far as the transformers go, but you can tune them at that frequency if you wish.
When you are doing your searches for designs, dont overlook those that use 7C5's, 6BW6's ( identical to 6V6 ) or 6AQ5,s although they are a tad smaller in voltage ratings and output power. The circuits will work without any component changes though, as they are "identical" except for dissipation and the previously mentioned down scaling.

One of my "gunna-dos" is exactly as you have suggested, to use as my PC speakers, just to have a valve audio stage ( I hate digital recordings with a passion, and dont have any music on board, so they will be dedicated to the occasional "ding" when I make a typo )
I have about 50 designs for such an animal if you need more designs, including complete projects from Radio,Television and Hobbies from the 40,s and 50,s( Australian EXCELLENT magazine)

Joe

[joebog1]

Modulation transformers are pretty close to speaker loading apart from the impedances involved, er and the weight, unless you are contemplating 24" subwoofers, er and the 30% modulation depth.

j.b.

[Diabolical Artificer]

As for OP valve choice don't forget the 6N7 which is capable of 5w. From what I've read recently a triode will produce mainly 2nd order harmonics and less distortion, but a pentode may be the valve for you as they produce more 3rd 5th etc order harmonics making the OP sound shrill, which is maybe the sound you want due to hearing loss. Pentodes usually have a R and C in parallel with the OPT pri to reduce this HF.

I can't say much on NFB (am reading up on it ) apart from to say that some early designs I've seen use a 2k pot only with no cap. AFAIK and please put me right, the cap in the NFB circuit is there to partly compensate for phase issues, EG in a amp that uses 3 stages or more.

A two stage SE design ( see PDF ) I have shows the cathode on the input stage partially biased by a 100u cap and the NFB connected at this junction and only has an R in series with this cathode bypass cap. It says " The NFB is in parallel with the load and should be not less than 10 times the load" so for a 8 ohm speaker a 80k R will do. See The Teqhnique of Sound Reproduction - edited by John Borwick- various authors, it's online as a free PDF.

Andy.

[joebog1]

A few articles here that are well designed, and explained for the new amp builder.
http://r-type.org/articles/art-164.htm
http://r-type.org/articles/art-159.htm
http://r-type.org/articles/art-128.htm

Joe

[joebog1]

Here are a couple of circuit diagrams for a "Playmaster" amplifier from my Australian magazine collection. I have built a few of the 1956 version over the years and its a superb little amp. It explains all the design principles including negative feedback. I find the feedback a bit "stiff", by that I mean its perhaps too clean.
You can always reduce the feedback as I mentioned before to suit your personal taste. Its slightly different in valve lineup, and uses a valve rectifier, but that can easily be replaced by a bridge or fullwave sand rectifier. I have the full article should you desire it, this is just the circuit to tweak your interest. The second circuit I have thrown in, is to show how simple a small amp like this can be. Its from January 1952.
Joe

[Aub]

Not quite the same, but I built an amp using an ECL86 for each channel and they are mounted on their sides, so I could fit them in the enclosure I had available. It runs for hours with no problems that I can detect.

Cheers

Aub

[TonyDuell]

The Revox G36 (and I guess older models) mounts all the valves (apart from the ECC82 bias/erase oscillator, which is fitted base-upwards) in a horizontal line at the bottom of the machine. All on their sides. Including the pair of ECL86s used for the monitoring amplifier

[Bazz4CQJ]

Another tried and tested approach to squeezing valves (vertically) into small chassis is to mount the valve holder on spacers below the normal chassis level, as used for the 03-10's in the Pye Cambridge RT.

B

[G8HQP Dave]

Originally Posted by G6Tanuki

The output transformers I'm using are salvaged from a scrap early-1960s radiogram where they were driven by two EL84s. I'm wondering if I could optimise the design by selecting the NFB capacitor for something closer to 15KHz rather than 10KHz? I like my high-frequency response to have Bat-disturbing impact.

Unless it was a very good radiogram the OPTs are unlikely to be doing much above 10kHz anyway. In those days if anyone heard anything much above 5kHz they described the sound as 'tinny'.

There is a reason why genuine hi-fi OPTs are big and expensive, and radio/radiogram OPTs are usually small and cheap.

[Edward Huggins]

Yes, I too was surprised to see the ambition to achieve c. 15,000 cps. Not quite sure what the signal source is to be, but if a ceramic pick up is to be used then there are not many that can reproduce above 12,000 cps anyway. Although not a pentode, the 6V6 was the EL84 of later years. The famous Pye "Mozart" single ended amplifier successfully used its EL34s on their sides too! Edward

[G6Tanuki]

My signal-sources are going to be a couple of CD-players, and the "Line-out" of a homebrew digital-media-centre which I use to play "Ogg Vorbis" losslessly-encoded music (darkwave, club/trance) or similar stuff that's distributed as direct-from-the-digital-mixing-desk in digital form - and there's definitely some components in it that reach 20KHz!

The output will be a pair of smallish 'bookshelf' speakers - this is to be a den/radio-room sound system not a "hifi" in the classic sense.

I've built plenty of similar amps in the past using EL41/EL84/6BW6 but have never done a 6V6 version. Having rediscovered my Octal-base "Q-max" chassis-punch a few days ago, along with some rather nice Amphenol ceramic octal bases (same as used in the AR88) I just thought I'd have a go at a 6V6 amp to see what it sounded like. Like the 6V6, EL41 and EL84 are also too tall to fit upright in the case I intend using - hence the original question. I could use the new-old-stock 6BW6 I have as they *would* fit upright - but I've done a 6BW6 amp in the past.

[Evil thought - I've got a few QQV03/10 here from my days playing with Pye Cambridges: I wonder if the two halves of one of these could be used for the 2 channels of a 'single-ended' stereo amp, as a sort-of homage to the crazy 2-pentodes-in-one-bottle ELL80?]

[G6Tanuki]

Originally Posted by Bazz4CQJ

Another tried and tested approach to squeezing valves (vertically) into small chassis is to mount the valve holder on spacers below the normal chassis level, as used for the 03-10's in the Pye Cambridge RT.

I could do that if I used the 6BW6s I have in stock, but alas a 6V6 on its base still won't sit cleanly in 2.5U of vertical space. On their sides I've got plenty of 'leg-room'.

My plan is to put the 3 valves (a 6V6 at each end, the 6SN7 in the middle) on a rectangular ali plate and mount this sub-assembly on its side to the base-plate of the case using a bit of 15mm square aluminium stock. The pedant in me says I should use the "springs and knitted bobble-hats" style valve-retainers to make sure they stay socketed; the realist in me says "it's a bookshelf amplifier, not a mobile radio" so I probably won't bother.

[Hartley118]

Tape recorders of the 1950s and 60s commonly mounted valves horizontally. These included power output valves such as EL84 and ECL86. As far as I know, none of these valves were harmed by not being vertical. For example, in a Ferrograph, the only vertical valves were the erase oscillator and rectifier.

Martin

[Bazz4CQJ]

Originally Posted by Hartley118

Tape recorders of the 1950s and 60s commonly mounted valves horizontally. Martin

I seem to recall that oscilloscopes of that period frequently had valves in horizontal positions. I cannot recall whether TV's from that era ever had LOP valves horizontal?

Something like 45 years ago I spent several weeks pocket money on a brand new Siemens ELL80 for a modulator for a topband rig ...which never got built . It sits in its original packaging waiting for my attention!

B

[HamishBoxer]

I am sure in my 45years of servicing I came across LOP valves horizontal but which chassis?? Pye , Thorn come to mind.