Page 1 of 2

6F5P/ECL85/6GV8 SET Amplifier

PostPosted: Sat Mar 19, 2011 7:00 pm
by Brik
I decided to build a low-cost SET amplifier to play around with.
I had a whole bunch of Russian 6F5Ps (Hi Mu Triode, Power Pentode combo) that I bought on an online-auction site simply because they were dirt cheap ($3.95 each).
As I understand it, 6F5Ps can be poor-man's 2A3s when they are triode-connected, although their power dissipation rating is smaller:

http://goldprivacy.com/6f5p-se
http://www.dmitrynizh.com/6f5p-set.htm

I was not keen on impressing the full 300V on the pentode, so I dialed back on the B+:
Image
The anode power dissipation on the pentodes is a little over the spec, so I will have to reduce the B+ further...

While I was waiting for the irons from Edcor, I decided I would put the whole thing into a 10" x 6" X 2" Hammond chassis. I used the free software from Front Panel Express to plan for the layout of the amplifier. It appears that everything would fit into the chassis, but it would be a tight fit:
Image


Now all the parts are ready to go:
Image


The Front Panel Designer printouts are used as drilling templates:
Image
Image


Wiring in progress:
Image


I was curious about the Plate-to-Plate feed back scheme mentioned in the "6V6 / 6SL7 / 5Y3GT SE Amp" thread, so I tried it:
Image


Here is a FFT spectrum analysis chart of the unmodified amp using another free software, Wave Spectra (downloadable from http://www.ne.jp/asahi/fa/efu/soft/ws/WS140.ZIP) with 1KHz sine wave to produce output at 1W at with THD of 0.1%
Image
You can see the 60Hz hum, 120Hz PS ripple, 1KHz primary, and its harmonics. 2nd harmonic is at -62dB and higher order harmonics declines in magnitude.

Here is the corresponding FFT spectrum analysis chart with the Plate-to-Plate feedback resistor. I expected some decrease in the distortion (it's negative feedback, after all), but the FB actually increased the 2nd harmomic distortion by ~17dB without having much impact on the higher order harmonics! THD increases to 0.5%:
Image
I suppose the addition of the P-PFB resistor would make the amp sound more prominently like a SET amplifier, at least in this application.

BTW, the Edcor OT transformers are superb. I couldn't believe how these relatively inexpensive transformers are so well-made, and sound so good.
The primary inductance was measured at 13H.

I will look deeper into the mechanism of what causes the profound increases in the 2nd order harmonic distortion. (???) :| :(

Image
Image
Image

PostPosted: Sun Mar 20, 2011 2:24 am
by soundbrigade
So nice, so simple so tidy! ;)

Really nice build!

Re: 6F5P/ECL85/6GV8 SET Amplifier

PostPosted: Sun Mar 20, 2011 3:28 am
by Geek
Hi,

Lovely build!

Brik wrote:I will look deeper into the mechanism of what causes the profound increases in the 2nd order harmonic distortion. (???) :| :(


The key to the 2H increase is all due to the phase of the 2nd harmonic. It's summing in the anode-follower topology (every other wavecrest will be additive).

Cheers!

PostPosted: Sun Mar 20, 2011 5:48 am
by azazello
Congratulations! One more "pocket" amplifier with very good tube!

PostPosted: Sun Mar 20, 2011 7:57 am
by Brik
Magnus, Gregg, and Azazello,

Thank you for the kind words!

With the daughter away in college, I have more free time, but *far less* money: This was originally intended to be a 2A3 SET build, but I had to scale back the plan. All the parts came in well under $150...

I was impressed with how linear these TV sweep tubes are, and how dirt cheap they can be, though they seem to prefer lower B+ of, say, around 220V?, with relatively high current levels, which narrows down the selection of power transformers.

Also, I have been impressed with how good and versatile the *free* real-time FFT analysis software is. You can make distortion measurements as well as frequency response charts with nothing other than a decent sound card. In the past, these measurements required $$$ test equipment, but now, we can use a regular PC with a $80 M-AUDIO sound card. (love) $) :))

I salute the TV sweep tubes and free software.

PostPosted: Sun Mar 20, 2011 1:23 pm
by EWBrown
Nice one! The 6F5P is similar to the well-kn own 11MS8 / 10GV8 except for the filament voltage. As I found from hands-on experience, , these particular tubes aren't great performers with Ultralinear, the best options are triode strapped or pentode mode.

You may want to experiment with the value of R8 (1K) in combination with the 330 ohm R7 in the triode cathode circuit, that will generate some seriously strong NFB, I'd go with a 25K trimpot and dial it in for best combination of power and THD.

The plate feedback 220K resistor is good, anything from 220K to 270K is about optimum. The combination of both types of NFB is almost "magical" as I found with my 6F3P / ECL82 SEP amp from a few months ago.

You should be able to get away with pushing the max PD by a little bit, but I wouldn't press it to 15 watts, that would cause the dreaded red plate syndrome =:o The TV vertical sweep service, and inside-the-set environment is a lot more severe than class A sudio. The tubes are cheap and readily available, so it's not a case of melting down a rare and precious vintage tube made of unobtainium ;)

I like the nice compact layout and design, looks great!

I have a few more irons in the fire, but I've been busy with "other stuff" so tube amp building has taken a back seat for a while, though I have several new designs I want to try out. I also need to order some more Edcor SE iron, as I've used just about everything up over the past few months.

FWIW, I have the same Rotel RCD-855 CD player, have had it for several years, and it's hard to beat! My backup CD player is an Optimus (rad shack) I bought at a thrift shop a few months ago for $5. It is lacking its remote (hence the low price) , but it is also a fairly decent performer.
I use that for the test sound source, instead of the previously often-used Sony Discman.

/ed B

PostPosted: Sun Mar 20, 2011 1:27 pm
by TomMcNally
Very nice, indeed. Something about that Gentian Blue and black!

PostPosted: Sun Mar 20, 2011 8:21 pm
by Geek
Brik wrote:I was impressed with how linear these TV sweep tubes are, and how dirt cheap they can be, though they seem to prefer lower B+ of, say, around 220V?, with relatively high current levels, which narrows down the selection of power transformers.


Along with what Ed said, TV sweep (especially vertical sweep and video) tubes *have* to be linear, or you get a crummy picture.

Most like low voltage and high current, because they ran at rectified-and-doubled line voltage, rather than use a power transformer (no one should ever do this in DIY).

6AH4GT, 6FD7, 6EM7 and pentodes like EL86 (6CW5) are real sweethearts (love)

7KY6 is a video pentode with 30,000 uMhos transconductance.... pretty much unheard of in a power device until MOSFET's showed up!

Cheers!

PostPosted: Mon Mar 21, 2011 5:10 pm
by Brik
Gregg,
7KY6 is a video pentode with 30,000 uMhos transconductance.... pretty much unheard of in a power device until MOSFET's showed up!

Wow. With that kind of transconductance, you may not need much of a driver stage! (love) :)) Perhaps I could use the Edcor GXSE10-16-1.7K for an 8 Ohm load...

Tom,
Something about that Gentian Blue and black!

I like how they look on a black chassis also. [:)
My personal name for that shade of color is "Norwegian Blue".
I imagine them nailed to their perch atop a chassis, pining for the fjords in New Mexico... :))

Ed,
Thank you for the tips! The amount of NFB is around 8dB right now. It has a bigger impact on the 120Hz ripple than on the harmonic distortion. Also, one of the problems was that the 6F6Ps I have on hand do not behave like the ECL85 specs say they should. I may have to fully characterize them before I can tweak the circuit.
I do like the Rotel CD player a lot. Not that I can hear much difference among very good CD players anyways. [:)
I may have a similar speaker set up to yours as well. Fostex BK-20 back-loaded horns with JBL 8" full range drivers (LE8Ts) and Fostex T90A tweeters (on the right).
I also have a pair of JBL LE8T-Hs in Madisound enclosures (on the left).

Image
These really come alive with the tiny 2.5W SE amplifier.

PostPosted: Tue Mar 22, 2011 9:44 pm
by EWBrown
I was thinking of getting two of the BK20 "flats" and using my FE208Es in them, along with two more of the T90A tweeters, those work fantastically with the BK16s.

The 7KY6 also has a 6.3V version, 6KY6, and their gm is exceedd by the 12GN7 and 12HG7, with gm of 36,000 and 32,000, respectively. These have PD of 10W or better. I;ve seen these in standard and "fat bottle" 9 pin miniature, I'd choose to use the "fat" ones in any SE or PP designs.

These can be run as a true "spud" with no VA stage.

I've also seen reference a simple 12GN7 PP design that used a Hammond 124B interstage transformer (10K:90K CT) feeding the grids, and a Hammond 1609 10K CT OPT, it doesn't get much simpler than that. Of course a signal source capable of driving a 10K load is necessary to get results with this.

/ed B

PostPosted: Sat Mar 26, 2011 8:35 pm
by Brik
I have been playing around with the Plate-to-Plate feedback resistors.
Combined with the Global Feed Back Resistor set at 3.3K Ohms, I tried two different values for the P-PFBR: 220K and 330K.
With the 220K resistor, the second harmonic increases by more than 20dB and third and higher order harmonics increased by 5~12dB:
Image

However, with the 330K resistor, it hits the sweet spot in terms of increasing the 2nd order harmonic, not increasing the 3rd harmonic much, and virtually eliminating harmonics higher than the 4th order:
Image

I am not sure if the dominating 2nd order harmonic distortion makes the sound more attractive. If the sound of the amplifier without the PPFBR is likened to the sound of a jazz ensemble in clear air, the sound with the PPFBR is like the same ensemble in a smoke filled room. :/ [:) :))

PostPosted: Sun Mar 27, 2011 6:03 am
by Shannon Parks
Lovely amp and fantastic documentation, Brik! (y)

Shannon

PostPosted: Mon Mar 28, 2011 4:34 pm
by Brik
Thank you, Shannon.
I learned to like a SET so much that I started saving up for a get*set*go. (love) :))

PostPosted: Tue Mar 29, 2011 9:48 pm
by EWBrown
The get*set*go is a great amp, in its original configuration, and it is also fun to come up with new mutations and permutations, like using 300Bs, 6AV5GAs, EL34s, etc. Having the two power tube sockets off-board makes this really easy to do :)) [:)

/ed B

PostPosted: Mon Apr 25, 2011 3:33 pm
by Tom Bavis
I suspect you're seeing the result of near-perfect distortion cancellation in the two stages... so anything you try makes it worse! Each stage may produce 1% F2 distortion, but they're 180 degrees out-of-phase, so they cancel. The feedback resistor also passes DC, so it shifts the operating point of the first stage, which may make as much difference as the feedback does. Results may be different at different power levels...