| | |
|
| |
| | |
|
|
Home > Simple SE 45/2A3 Amp for First Time DIYers
| | |
|
A Simple SE 45/2A3 Amp for First Time DIYers
(with tips on how to "guesstimate" output transformer primary impedance)
VOLTAGES FOUND INSIDE TUBE
AMPLIFIERS CAN KILL YOU! IF YOU DON'T UNDERSTAND PROPER ELECTRICAL
SAFETY PRACTICE, DON'T BUILD THIS AMP!
Late last winter, I started cleaning the basement and sorted out all my junk in preparation for Spring, my favorite season (lots of hamfests). I found a few pairs of push-pull output transformers that I had pulled out of decrepit receivers and integrated amps a couple of years ago. Before deciding to load them in my car's trunk for a drive to the nearest recycling center I thought that perhaps I should try and recycle them for its original purpose.
Taking the cue from Eric Barbour's "Single-Ended Glory.." article in Glass Audio 3/95,I picked the heftiest pair of iron in the lot, approximate dimensions - 2 3/4" W X 3 3/8" L X 1 3/4". In it's former life in a Harman Kardon A500 integrated amp, it was being driven by a pair of 7355
outputs. Not having further data, I had to devise a simple way to calculate the primary impedance to see if it will be an adequate match for a 45 (recommended load is around 4600 ohms).
Fortunately before I pulled the iron out of their chassis, I had the presence of mind to label the various leads (easily done with white sticker labels from your nearby office supply store). By creating a block diagram of the original circuit, I figured the orientation of the primary relative to
the secondary for proper phase (remember that every time you have a gain stage the signal inverts phase) and then noted the plate connections and center tap. The secondary was much easier, I just followed the wires from the original taps on the speaker output terminals. If you are going through a "mystery" transformer, the procedure of identifying the various leads becomes more involved. For more information on how to do this, read through Norman Crowhurst's article, "Transformer Hunting" in Glass Audio 2/89.
For "guesstimating" the primary impedance, I dug out an early issue of Sound Practices (Fall 1992) wherein Mike LeFevre wrote about output transformers. On page 23 I noted:
turns ratio = voltage ratio
impedance ratio = voltage ratio squared
impedance ratio = turns ratio squared
A few paragraphs ahead in the same article, Mike says, "Note that you can always calculate the primary impedance by multiplying the load placed on the secondary by the appropriate impedance ratio......." Since in the above ratio relationships, impedance ratio = voltage ratio squared, I decided to set-up a "crude test jig" using a DVM (Fluke 75) and a Tenma hand held Audio Generator (a neat battery operated item available from MCM for around $70). The manufacturer rated 1.2 Volts maximum sine wave output with a source impedance of 600 ohms. That's not much but I figured if it can put out enough voltage to measure at least .01 volts at the secondary, I'm in business...... You can try using a Variac but even with my years of experience fiddling with tube equipment, I still have high voltage phobia.
With a couple of alligator clip leads and an 8 ohm, 20W non-inductive resistor and the above mentioned test instruments, I started sleuthing. I attached the output of the audio generator to the primaries (phase is not critical for this test just make sure the unused leads are not touching
each other) and then connected the 8 ohm resistor between the 8 ohm lead and 0 (in almost all the output transformers I've seen, this is color coded-black). I set the generator for a 1KHZ sine wave signal and increased the output until a stable .01 volt reading was seen from the DVM
measured from the two chosen secondaries, then without touching anything, measured the voltage between both ends of the primary. Now I have my voltage ratio. In this particular transformer I got approximately .25 volts between the primaries for an output of .01 volts
between the 8 ohm and 0 leads of the secondary.
Reading through the article, I presumed that the proportions of the example (a Dyna A470) was referenced to 1V output at the secondary, so I just multiplied the ratio, .25/.01 by 10 = 25/1, squared = 625, then, multiplied this product by the load resistor value of 8 ohms and voila =
5000. If logic serves me right, 5K is the primary impedance of this OPT. Due to the simplicity of my "test jig" and my simple mind, notwithstanding, I cannot vouch for the total accuracy of my test procedures. But I was satisfied enough with my results to go on with the project. Do not be deterred if you find 6-8K OPTs (I've measured these values from Fisher and Scott receivers and integrateds using hard to find and expensive 7591s), as long as they are hefty (around 6 lbs., each), they can handle the DC unbalanced current.
The type 45 triode is well revered amongst single ended aficionados and designers. JC Morrison and Herb Reichert showed lyrical enthusiasm for its sonic merits in their past articles in Sound Practices. However most of the circuits published in this great magazine are not really simple
enough nor cheaply acquired for first time DIY constructors. I was sure that right after Eric's article came out, the price for even good used 45s will sky-rocket, therefore in 1997 SE Glory may not be achieved under $100, unless you have lots of junk in your basement like me. But the
idea of using a hefty PP iron was a great tip nonetheless. As far as I know only Hammond and UBT have SE OPTs at 5K within the "affordable" range.
Cleaning out the basement yielded other goodies. A 13x 7 x 2" aluminum chassis with bottom plate, a few 4 pin and octal sockets, a Hammond 272BX power tranny that I had intended to use as a separate driver stage power supply in a PP parallel 2A3 amp that never materialized, a
10H, 100ma. choke donated by an HP 200CD oscillator with a blown power tranny, Thordarson 2.5 V, 5A filament transformer and a few other miscellaneous items that can complete a neat and compact SE 2A3/45 amplifier that can alternate in my bedroom system so that my Acro
20/20 can rest.
For chassis work, I rely mostly on a 3/8" variable speed drill and a number of Greenlee chassis punches (3/4", 1 1/16", 1 1/8", 1 1/4" and 1 3/8"), I know they are expensive but well worth it if you are seriously contemplating on rolling your own gear. I also find it handy to have a reamer, assorted files and, 1/2" and 3/4" stepped unibits.
When it comes to layout, the simplest way to do it is to gather all your components (transformers, tube sockets, choke(s), etc.) and just prop them up on the chassis in anyway that aesthetically pleases you, just remember to keep the low level stage (most especially the input jacks) farthest away from the power supply. In the power supply, try and orient the trannies and choke so that the laminations are crossing each other to prevent hum induction, then start making drill hole markers. Nothing beats goode olde terminal strips for hard wiring, very convenient and it gives me enough freedom to wire the circuit to almost mirror a schematic.
You can look at various glossy audio magazines for ideas. The Japanese publication, MJ Audio Technology, is my prime source for layout ideas and they come up occasionally at Tower Records. I usually like to finish my chassis with Hammerite spray paint gray or silver but in this project I just went all over the chassis with a rotating wire brush after all the holes have been cut to finish it off.
Since I had a pretty good stash of Chinese 2A3s acquired from JoLida when they liquidated their tube inventory to make way for their current tube amplifier line-up, I utilized them until I got hold of affordable 45s. Knowing that the 2A3 owes much of its lineage to the 45, 1 decided to operate the 2A3 at 45 operating points: plate = 275 volts and DC grid bias = 56 volts for a maximum output power of 2W at the OPT secondary. Perhaps not the most linear way to operate the 2A3, but it got me started and a 45 can just drop into place when I get them.
When you compare the specs of the two tubes, plate current and filament current for a 45 is around half a 2A3 but the plate resistance doubles to 1600 ohms and DC grid bias = 56V as opposed to 45V for the 2A3. The rule is to load the plate of a triode tube with at least 3 times the plate resistance value or more for less distortion, so my 5K OPTs are optimal for 45s. Also, I noticed by looking at the load lines that the plate resistance of a 2A3 rises using 45 operating points.
I matched two 1.5K, 10W wirewound resistors, connected them in parallel bypassed with a generic 100uf/100V cap attached to the center tap of the 2.5V, 5A filament trannies to ground for cathode bias shared by the two output tubes. Actually the ideal way of doing this is to use separate biasing for the output tubes which calls for two separate filament transformers. Just keep in mind that the 45 only needs 1.5A of filament current compared to 2.5A for a 2A3, a good thing to keep in mind when shopping for 2.5V transformers. Since I had 5A unit, I utilized it and the added benefit of alternating between the two types of tubes.
Having quite a bit of experience with 6SN7s, I knew that 1/2 of my favorite preamp/driver (as used by Eric) just would not have enough gain for systems that do not use hyper-efficient speakers (Horns and Lowthers) or you'll need a high gain line level preamp to get enough drive. I will get into how I came up with a solution for a more sensitive driver stage later. Two watts of power is admittedly not much but long before the advent of single-ended amps, I've been an ardent believer that it is the quality of watts not the quantity that matters most.
I have a pair of relatively efficient quarter wave loaded 2-ways with 92 dB/1W/1M sensitivity speakers with an impedance that stays almost within the vicinity of 8 ohms across its frequency bandwidth, I can hear it roll-off around 50HZ in my room. My 300B amps can play Mahler
symphonies with hardly a trace of clipping in my 10x 12x 10' listening space. The theory goes, 1W gives 92 dB Sound Pressure Level, 2W = 95,dB SPL, 4W = 98 dB SPL, and etc...at 1M, that is only on one channel! Keep in mind that each time you double the power (watts) you gain 3
dB. I prefer to listen at an average of 88 dB SPLs sitting 1.5 - 2M from both speakers and reserve the remaining 10 dB for peaks, anyway in a typical seat in a concert hall, I bet you'll never encounter more than 95 dB SPLs. I hope that settles the neurosis that abound regarding flea powered SE amps. Despite the fact that 2W from a 45 does not have as much dynamic headroom vis-a-vis 8W from a 300B, it is still worth the effort to try this project as long as you have an easy load (simple Ist order cross-over 2-ways) 88 dB or higher sensitivity pair of speakers. In fact it gives satisfying SPLs with Radio Shack LX 4s in my room which probably has mid 80s dB sensitivity; Steve enjoyed it with his ProAc Response 2s.
The 6SN7s had to go. Working off the Resistance-Coupled Amplifier Charts from Ned Carlson (Triode Electronics) cool Dusty Files II, I attempted to concoct a circuit using another favorite octal triode, the 6SL7. It has an amplification factor of 70 compared to 20 in a 6SN7. The drawback is its higher plate resistance necessitating a higher value grid resistor load for the outputs. To keep it plain and simple, I kept the 2 stage philosophy - input/driver stage provided by 1/2 section of a 6SL7 capacitor coupled to the grid of a 2A3/45. With a 100K, 1W carbon resistor for the plate 1K, 1W bypassed by a 100uf/lOV cap on the cathode driving a 470K load (grid resistor for the 2A3/45). According to the chart, this circuit should be able to swing 63V peak with a supply voltage of 300V.
Since I plan to have about 320V supply for this stage, by interpolation, I should have 67V peak output, factor in some losses (15-20%) in practice, it is probably adequate...... the rule being, to at least equal OR ideally, exceed (for more headroom) the bias voltage of the output tube (56V). To be sure that my calculations are within ballpark, I observed the clipping characteristics on my dual trace scope (driver stage on one channel, output stage on the other), the output stage clipped before the driver stage. That satisfies the technical rule for a decent driver stage design, I guess.... most important to me is how it sounds. You may want to lower the grid resistor value to around 220K if you use 2A3s exclusively since 470K is about maximum but on a 45 the maximum recommended grid resistor value is 1M ohm. Although I have not tried it, a 12AX7 (paralleled, is even better) should work and yield slightly more gain in this circuit.
For the power supply, I chose a capacitor input with a choke in between the first two capacitor stages for reduced ripple then the two 1.2K, 3W voltage dropping resistors, being a "split rail" feeding each channel's input/driver stage, for better isolation. The 272BX power transformer has
a 600VCT /100ma., 5V /2A rectifier winding and 6.3V /3A filament. With a 5V, 2A winding the choices for a rectifier tube is rather limited to either a GZ34/5AR4 or a 5V4. I chose an ST shaped 5V4 rectifier to complement those curvy ST output tubes for looks. The luf + .47uf 630V cap combination is a critical value for this particular transformer. It "tunes" the 330 VDC B+ supply on the second capacitor (after the choke). More capacitance at this stage will bring up the voltage and vice versa. The rest of the decoupling caps are 20uf @ 630 Solen polypropylenes. See the schematic.
That completes a simple two stage SE amplifier with adequate 1.2V sensitivity (like an ST70) for full output. The nice thing about this "limited headroom design" is that it clips very gracefully, there is not enough voltage from the driver stage to drive the grids positive, therefore it is pure
Class Al. Admittedly, there are better input/driver topologies around like, variations on the Loftin-White theme (Morrison Micro or Fi 2A3 ), SRPP 6SL7 or the Baby Ongaku by Gordon Rankin, even a 6SJ7 pentode driver ala WE 91, which can all be adapted for 45s. As simple as it is though, I still enjoyed music played through them and even acquitted itself compared to my generic 6SN7 fronted SE 2A3 and 3008 amps. To encourage more people to dabble in DIY, I decided to share this relatively fool-proof circuit. Using Chinese 2A3s will provide the cheapest and easiest route in experiencing the sonic delights of SE/DHT amps.
A few weeks ago my quest for 45s was fulfilled. I attended my first hamfest in 3 years (aside from the treasure trove of rare tubes and associated paraphernalia, I miss the beef barbecue sandwich.....). I got up at six (a hard chore for musicians since we are used to working until late at night and consequently used to getting up late) and armed with a couple hard earned $$$s ventured through the trail of tailgaters. Having had a bit of prior hamfest experience, I knew that by looking out for people who are in the antique radio collecting hobby, I was bound to land myself with a few 45s. Within 20 minutes of searching at the Timonium Fairgrounds, I found an elderly man who had 4! Carefully, I opened the boxes to inspect the contents, the man uttered, "those are used but test good....." Since one of the GE pair had a crack in the base, I opted for a pair of Sylvanias that were soiled, encrusted with dirt and dust but closer inspection gave me enough confidence to take a chance. "How about $25 for the pair I asked. "its yours", he replied. Later I found a dealer who had a used pair of Philcos with engraved bases. This time I knew I had to pay more ($22 each) but I wanted to be sure that in case I got burned with my first $25/pr, I had two more chances of hearing it.
As soon as I got home I tested them in my tube tester, whew.....l got lucky, they all tested well above average. After a bit of scrubbing, I inserted them on my amp and sat down to listen. Astounding! Never heard anything like this before. All the glowing praises I've read and heard are well deserved. It is indeed a beautiful sounding tube. Within its limited power range, the last time I heard this kind of sound quality and degree of refinement was when Steve inserted a pair of those expensive Audio Note VV300B Super Linears on my 300B amps.
I would rather not attempt to use "audiophile adjectives" to describe what I heard. Listening through these ancient amplification devices is almost like evaluating the qualities of fine Italian violins which I feel more comfortable with so here it goes.. I consider the 300B as the Guarneri del Gesu of DHTs that is, dark, robust and at the same time lush with an overwhelming power that engulfs, a 2A3 reminds me of most Strads (although a few top-ranked Golden Period and late Strads equal del Gesus) which are sonorous yet sweet with a carrying power that penetrates.
The 45 was most haunting..... like an Amati, very delicate, airy, possessing a gentle warmth and intimacy in the mids that is most alluring. The perfect triode for "chamber music".
That's all for now folks, I'm preparing for a concert tour in the next couple of months, this time it will be real music making......
Joseph Esmilla, May, 1997.
On May 3, 1999, Gary Kaufman e-mailed us with two suggestions,
"First, the Hammond 125E output trannies work very nicely, and are quite inexpensive. If only they weren't so ugly!
"Second, removing the ground from the 2.5V filament supply and adding a 100 ohm/2W hum balance pot between the filament taps and the cathode resistor/cap provided a dramatic improvement in background hum. I use a pair of Altec A-7's and the hum was very annoying previously."
Thanks for a fun project!"
In late October, '99 we received some questions about this project,
Q."What output transformers would you recommend...?"
A.If you will exclusively use a 2A3, you can choose an output transformer with a primary impedance of 2.5K to 3.5K.
As recommended in the project, a "salvaged" hefty (4-6 pounds) push-pull transformer will do, as well as the Hammond 125E (as used by Jeremy Epstein in his "darling project"). I actually prefer the sonics and measured performance fo a push-pull output transformer from an old Scott 222/299, Fisher 400/500B/500C, EICO ST40/ST70 etc. integrated or receiver using 7591, 7868, 7189/EL84/6BQ5 or even 6V6GT. Other good output transformers for this project include the 2.5K Hammond 1627SE, any Audio Note 2.5K-3K iron, Magnequest TFA204 3K, Tango U808 (Best Buy in terms of price/performance ratio and flexibility). For beginners, I would not recommend getting a top flight Tango or Tamura (e.g. XE60 or F5000/7000 series).
Q."Which filament transformer should I use? I couldn't find one that matches the specs given."
A.Here's the deal, get an RCA or GE tube manual and do some reading. The filament voltage AND current draw is specified there. For example" 2A3=2.5V@2.5A therefore you need to get a Hammond or Magnetek 2.5V unit of between 2.5A-3A current rating to light the filament of a SINGLE 2A3. If you're a cheapskate you may get away with using a single Hammond unit specified 2.5V@5A for a PAIR of 2A3s, however even with the use of a hum balance pot, nulling residual hum may not be as effective as using 'dedicated' individual filament transformers for each 2A3. BEGINNERS CAVEAT! There are several surplus electronics stores that sell 2.5V@10A filament transformers for cheap. Sometimes they are even in really nice potted cans. DO NOT ATTEMPT USING SUCH UNITS BECAUSE THEY WILL PUT OUT GREATER VOLTAGES (AS HIGH AS 4-5V) TO YOUR FILAMENTS CAUSING PREMATURE FAILURE! THIS IS DUE TO THE FACT THAT THE TRANSFORMER DOES NOT 'SEE' AS MUCH LOAD (CURRENT).
Q."There is 2.5V@5A coming off of the filament transformer..."
A.This is used to power the filaments of BOTH 2A3s (see above).
Q."...and then there is 6.3V@3A off of the power trannie*. I am not sure exactly where these go to. Do they (one of each) go to V3 and V4?
A. NOPE! The 6.3V line powers the 6SL7 (input.driver) tube. V3 and V4 are your output tubes, which needs a 2.5V filament supply as explained in my answer to your previous query above.
Q. "Also I assume the 330V from the power trannie* feeds both output trannies?"
A.Yes, this is your main B+ line.
Q."And each 325V feeds each side of the V2?"
A. V2 is the 6SL7 input/driver tube. If you do your homework and read the RCA or GE tube manuals you will realize that the 6SL7 is a "dual triode". Each triode in this single envelope tube works as the "input/driver" for EACH channel of the stereo amp as explained in the website article found above.
*YIKES! We deplore the use of the diminutive "trannie" in place of "transformer".
| |
| | |
|
|
|