oh nice! this is all very good. I am slowly running out of unknowns and things to ask. I just put in the orders for most of the parts and I am really excited to start this guy!
ONE last thing, I am looking for a reasonably affordable stepped attenuator and was wondering what are the point of having a 3 pole stepped attenuator to ground?
I ask this because on Ebay I found a handful of nice 2 pole stepped atteunators. Is it possible to cheat a bit and make a 2 pole shunt to ground in a way that a 3 could?
ONE last thing, I am looking for a reasonably affordable stepped attenuator and was wondering what are the point of having a 3 pole stepped attenuator to ground?
I ask this because on Ebay I found a handful of nice 2 pole stepped atteunators. Is it possible to cheat a bit and make a 2 pole shunt to ground in a way that a 3 could?
Any reason not to use Edcor XSM10K-50 ?
It seems to fulfill all the requirements. With 10k primary impedance, you can use a lowish-mu twin-triode like 5687 in a long-tailed pair with single-ended input, which should give you a voltage gain of about 6 or 7.
So, how do I figure what power output that will yield?
With 1.5V peak signal input, and with that one 5687 LTP, I figure a voltage gain of 7 will yield 10.5Vpk out to the transformer.
It's an 8.2:1 pri:sec stepdown ratio. so would that mean the volts out would be 10.5/8.2?
Supposing the answer is yes, I get 1.28V peak out. That's 0.9V rms.
To find the power output, Vout squared / load in ohms
So with 60 ohm headphones, I get only 13.5mW (0.0135 W).
Did I do that wrong?
Should the power output in watts be Vrms pk-pk / load Z?
--
It seems to fulfill all the requirements. With 10k primary impedance, you can use a lowish-mu twin-triode like 5687 in a long-tailed pair with single-ended input, which should give you a voltage gain of about 6 or 7.
So, how do I figure what power output that will yield?
With 1.5V peak signal input, and with that one 5687 LTP, I figure a voltage gain of 7 will yield 10.5Vpk out to the transformer.
It's an 8.2:1 pri:sec stepdown ratio. so would that mean the volts out would be 10.5/8.2?
Supposing the answer is yes, I get 1.28V peak out. That's 0.9V rms.
To find the power output, Vout squared / load in ohms
So with 60 ohm headphones, I get only 13.5mW (0.0135 W).
Did I do that wrong?
Should the power output in watts be Vrms pk-pk / load Z?
--
Yes RMS Volts^2 / Load Ohms := Watts.
13.5 mw may be a lot or not depending on your particular head phones. Low sensitivity headphones will require more power.
Assuming parafeed (capacitor) input to the transformer your selected Edcor transformer, it is a workable, affordable choice.
Go ahead and build it, perhaps on a breadboard to see how it works and how you like it.
Other builders have used a Hammond 119DA (600 ohm : 4 ohm and with a 8 ohm tap). The 8 ohm tap has a similar turn ratio to you selected Edcor.
Also for your next breadboard you may want to try a cathode follower output stage for lower output impedance.
DT
13.5 mw may be a lot or not depending on your particular head phones. Low sensitivity headphones will require more power.
Assuming parafeed (capacitor) input to the transformer your selected Edcor transformer, it is a workable, affordable choice.
Go ahead and build it, perhaps on a breadboard to see how it works and how you like it.
Other builders have used a Hammond 119DA (600 ohm : 4 ohm and with a 8 ohm tap). The 8 ohm tap has a similar turn ratio to you selected Edcor.
Also for your next breadboard you may want to try a cathode follower output stage for lower output impedance.
DT
Yes, good question, what's the sensitivity of these 'phones?
My headphones are Fostex T50RP.
Type : Semi-open RP dynamic
Frequency Response : 15 Hz - 35 kHz
Sensitivity : 98 dB/mW
Max Input : 3000 mW
Impedance : 50 ohm
Maximum input 3 watts?? Wow. That's a lot! That must be peak power in. That would be about 2.1W rms. I have to assume these 'phones are low-sensitivity, power hungry. My living room amp (push-pull 2A3's) puts out 6W rms per channel, driving floorstanding 3-ways. Crazy.
[EDITED LATER, after some reading] The *efficiency* rating of 98dB SPL/1mW makes them fairly medium-efficiency. If they yield 98dB SPL at 1mW, and I put my absolute limit at 110dB SPL peaks (and that's really LOUD), then 8mW should equal 107dB SPL. (Each doubling of power yields +3dB SPL.) So that 13.5mW max output looks acceptable, if not optimal. I could tweak it to get more volts swinging into the primary. Maybe a higher mu tube like a 6DJ8 would work better.
--
My headphones are Fostex T50RP.
Type : Semi-open RP dynamic
Frequency Response : 15 Hz - 35 kHz
Sensitivity : 98 dB/mW
Max Input : 3000 mW
Impedance : 50 ohm
Maximum input 3 watts?? Wow. That's a lot! That must be peak power in. That would be about 2.1W rms. I have to assume these 'phones are low-sensitivity, power hungry. My living room amp (push-pull 2A3's) puts out 6W rms per channel, driving floorstanding 3-ways. Crazy.
[EDITED LATER, after some reading] The *efficiency* rating of 98dB SPL/1mW makes them fairly medium-efficiency. If they yield 98dB SPL at 1mW, and I put my absolute limit at 110dB SPL peaks (and that's really LOUD), then 8mW should equal 107dB SPL. (Each doubling of power yields +3dB SPL.) So that 13.5mW max output looks acceptable, if not optimal. I could tweak it to get more volts swinging into the primary. Maybe a higher mu tube like a 6DJ8 would work better.
--
Last edited:
Perhaps the easiest way to get the required 85V peak output to the OPT primaries is to use a classic common cathode to cathodyne (concertina) phase splitter, and cap-couple that to the OPT primaries.
The cathodyne splitter needs cap-coupled outputs anyway. It would also provide more voltage gain and good balance. The icing on the cake is a very low output impedance into the OPT primary.
Does this look workable?
--
The cathodyne splitter needs cap-coupled outputs anyway. It would also provide more voltage gain and good balance. The icing on the cake is a very low output impedance into the OPT primary.
Does this look workable?
--
Attachments
[LATER...] OK, I now see that there are no CCS or chokes on the cathodyne, so that's not going to work as a parafeed configuration. I suppose you'd want a choke in the plate and a matching choke in the cathode of the output tube, and then cap-coupled to the OPT primaries.
Oh well...
So, that Edcor transformer primary can't have the B+ to its center tap and plate current? That means I would need a proper power amp output transformer.
Now I'm back at trying to adapt a standard output transformer with 4, 8 and 16 ohm tapped secondary for speakers. 10k primary to 16 ohm secondary, rated for 10W out?
--
Oh well...
So, that Edcor transformer primary can't have the B+ to its center tap and plate current? That means I would need a proper power amp output transformer.
Now I'm back at trying to adapt a standard output transformer with 4, 8 and 16 ohm tapped secondary for speakers. 10k primary to 16 ohm secondary, rated for 10W out?
--
Ah, another great question! I guess I could go for either one.
I might want to go SE, because I've never built an SE power amp before. That's mostly because I could never justify spending all that money on gargantuan OPT's and overpriced 300B's, only to have to spend even more money on high-efficiency speakers with a flat impedance curve (do those actually exist?). But I figure a little SE triode+OPT 'juiciness' would be great for headphones.
My 50 ohm headphones are rated at 97dB SPL/1mW efficiency. With 1.5Vpk in, about 30x gain from a 6DJ8 or 5842, and a 25:1 turns ratio (10k:16), I get max rms power out of 32mW. Is that enough?
If the 'phones get 97dB from 1mW, then 32mW should yield 112dB SPL. That would be where it clips (theoretically, at least). I guess I'd want more headroom than that.
I could try a triode-wired 6J52P (mu of 60, but probably closer to 50 in real life, low rp under 1500R). That would give me 57mW with 1.2Vpk input. That should yield a couple dB of headroom. It would be tough to get any more than that out of a single triode stage. I don't listen to headphones loud, being a jazz and classical type.
So what do you think? 6J52P-triode into an Edcor GXSE10-16-10k?
https://www.edcorusa.com/gxse10-16-10k
--
If you want better linearity use a current source for the anode load and capacitivly couple into the transformer. You need a very large cap (>40uF) to keep the LC resonance below audio so it is suppressed.
In addition to improved linearity, you will get very close to Mu gain witha current source load.
I'm designing a 71A headphone amp witha 8:1 (hand wound) transformer to drive a set of Sennheiser HD 600 headphones. Prototype sounded good, and the frequency response was 5Hz to 100KHz -3dB at 1Vrms into 300R. Distortion is predomonantly 2nd harmonic, 3rd is below -100dB in simulations.
The headphones are spec'd at 112dB at 1Vrms.
In addition to improved linearity, you will get very close to Mu gain witha current source load.
I'm designing a 71A headphone amp witha 8:1 (hand wound) transformer to drive a set of Sennheiser HD 600 headphones. Prototype sounded good, and the frequency response was 5Hz to 100KHz -3dB at 1Vrms into 300R. Distortion is predomonantly 2nd harmonic, 3rd is below -100dB in simulations.
The headphones are spec'd at 112dB at 1Vrms.
Last edited:
Sounds really nice, I'll bet. More high end than I was planning to do, though.
Do you know if the headphones' sensitivity spec is at 1V peak or 1V rms?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.