You guys may have a point. I Google searched for comfortable music listening levels, and the consensus is about 70-80 dB. With 100 dB speakers (which don't need to be horns), 8 mW gives you about 80 dB SPL. At this power, a 2A3 SE amplifier will have 400x headroom. If 8 mW is a normal listening level, then slew rate doesn't matter, just RC low pass of 2A3's Miller capacitance limiting HF response. This is how you get away with feeble driver tubes like 6SL7 or 57/6C6. BTW, 57 is better than 6C6 design-wise because, like 2A3, it has 2.5 V heater.
With 90 dB speakers one needs 100 mW of amp output power. Still 35x headroom, not so bad. This is the territory of high quality HF planars and ribbons.
With 90 dB speakers one needs 100 mW of amp output power. Still 35x headroom, not so bad. This is the territory of high quality HF planars and ribbons.
To atone for non-constructive post, here is how I would do a 2-stage 2A3 SET. Driver, triode-connected 57, transformer-coupled to 2a3. 57/6C6 triode has mu of 20 and is quite linear. Due to inductive coupling, gain is equal to mu, and together with step-up interstage transformer, the overall driver gain is 40. Driver is biased with permanently wired 9 V battery, which is better than bypassed or unbypassed cathode resistor. One of the advantages of transformer coupling is driver voltage close to B+ (RC coupling typically only half of B+). Another advantage is low resistance path to 2A3 grid, a safety feature for fixed bias (which has less distortion than cathode bias). Filaments are run on 2.5 V DC with negative side grounded, with CM choke between floating regulator and filaments.
Last edited:
Atonement 2A3: good name for the circuit. I like the complete absence of capacitive energy storage in the driver and output stage current loops. Recovery from overload should be near instantaneous and minimally audible.
In reference to the 57/2A3 schematic above... Since the UTC A-17 is basically unavailable now. Can you specify a currently available Hammond, Edcor, etc. interstage transformer that will work in its place? 8ma 15k to 60k. The A-20 substitute is easy as any 0 DC offset 1:1 input transformer should work there no? The utter simplicity is intriguing here for something to play with on a Sunday afternoon.
A-20 has about 50 H primary inductance, so its Z at 20 Hz is 6 K. 600 Ohms is very conservative rating. Anyway, because secondary is unloaded, transformer input presents almost infinite load to source at all but the lowest frequencies.
Input and interstage transformers sound best with unloaded secondary. 15K:60K rating does not mean that secondary must be loaded with 60 K. It means that the transformer will have specified frequency range (40 Hz to 20 kHz) when it reflects 15 K source impedance as 60 K at the secondary. The lower the source impedance and the higher the load impedance, the better is transformer's performance. In this case, because 75 triode' s Rp is 10 K, and plate current of 6 mA is lower than A-17s allowable 8 mA DC, the low end will be -3 dB at about 25-28 Hz.
15k to 60k is 1:2 transformer.
Line input (high/er/ impedance than 600R) 1:2 transformers are available, but not so common.
Lundahl LL154x transformers complete the task, but many people don't like it.
LL7903 is better, as Sowter 9063 too, but all of it's very pricy toy.
The 1:1 (line level) transformers are more common.
Cinemag CMLI 15/15B or Jensen JT11P-1 are both OK.
Line input (high/er/ impedance than 600R) 1:2 transformers are available, but not so common.
Lundahl LL154x transformers complete the task, but many people don't like it.
LL7903 is better, as Sowter 9063 too, but all of it's very pricy toy.
The 1:1 (line level) transformers are more common.
Cinemag CMLI 15/15B or Jensen JT11P-1 are both OK.
For the interstage what about Lundahl LL2746? I looked up the others LL154x and LL7903 but couldn't find how much DC offset they can handle. The one below says 30ma.
https://www.lundahltransformers.com/wp-content/uploads/datasheets/2746.pdf
https://www.lundahltransformers.com/wp-content/uploads/datasheets/2746.pdf
All currently produced interstage transformers are for lower source impedance, like 5 K. Edcor IMT 5K/20K would be suitable with re-gapping (remove spacer and butt-stack laminations). A-17 is ideal for this circuit, with its permalloy core it is better than modern transformers with steel cores. A-17 commonly appears on eBay with average price of $150 apiece.
I checked all Lundahl interstage transformers, including 2746. They all have lowish primary inductance and are designed for low impedance source. The closest is LL1635/5 mA. This is 1:1 transformer, which is OK for this circuit with 2 VRMS signal source.
Non of them tolerate DC current.
My examples applies to line level (1:2 or 1:4) SUT and common 1:1 IT as anode load of first tube.
Sample:
Ah so 1:1 is ok if there is enough gain? Typically my source is a DAC. Jack at Electra print can wind a 20ma 20k:20k interstate apparently $118.
Would that be ok you think?
https://www.electra-print.com/interstagetransformers.php
Would that be ok you think?
https://www.electra-print.com/interstagetransformers.php
20K:20K should be ok if it can handle 6 mA DC. 2A3 needs 32 VRMS on its grid for full power. 2 vrms input and gain of 20 gives you 40 VRMS. If the source is lower than 2 vrms, this is not a problem since the amplifier is used at a small fraction of its full power under typical listening conditions for this type of amp.
Depending on the driver tube gain.
If you see my sample, this operating point of power tube requires 88Vpp grid swing for A1 (maximum), it's 31.1V RMS.
If the tube stage gain is 8, the requiring grid level is 31.1/8= 3.89V RMS.
Elektra Print doesn't give any necessary data, so this would be OK .. or no ...who knows. Must to try.
Either changing from a 2 stage RC coupled single ended 2A3,
Or changing from a 2 stage DC coupled single ended 2A3, To:
Using an Input Transformer, and an Interstage Transformer for a 2 stage single ended 2A3 . . .
That should be assigned a new thread. Apples, and Steak are not the same. Neither are Oranges and Fried Fish.
Just my opinion, and my experience.
. . . Well, I have never built a DC coupled 2 stage single ended 2A3 amplifier (Chocolate Cake versus Pork Chop).
Or changing from a 2 stage DC coupled single ended 2A3, To:
Using an Input Transformer, and an Interstage Transformer for a 2 stage single ended 2A3 . . .
That should be assigned a new thread. Apples, and Steak are not the same. Neither are Oranges and Fried Fish.
Just my opinion, and my experience.
. . . Well, I have never built a DC coupled 2 stage single ended 2A3 amplifier (Chocolate Cake versus Pork Chop).
Is the driver tube 75 or 57, or 6C6?. Using a 10k Rp tube as IT coupled driver is an odd choice. The Lpri required to give sufficient LF will have a heap of capacitance, and the drive impedance to the output with 1:2 turns ratio will be 100,000 ohms. The 9V battery comes at the price of an input transformer, and the bias supply likely has a capacitor as the final filter, all of which are in the signal path/loop. It might look simple, but its really not.
Sorry, my confusing mistake: the driver is 57 (which is 6C6 with a 2.5 V heater).
Transformer winding capacitance is easily kept in check by winding sectioning, which competent winders do. Top 20 kHz of A-19 is consistent with its low winding capacitance and leakage inductance. Low -3 dB point of IT in this circuit is 25-28 Hz. The transformer-coupled 57 will present 40 K impedance to 2A3 grid, not 100 K.
Input transformer has other benefits: isolation from ground loops, possibility of balanced input allowing longer interconnect, and possibility to tweak overall gain.
The best bias for 2A3 is also from permanently wired battery. Six 9 V batteries. BTW, I am not claiming that this is capacitor-free design. Last filter capacitor is in the signal path in any SE amplifier.