Well, if you have a spare set of these highly inefficient but amazing sounding headphones, you can send me one and I can test it on my 300B amp to see if it would play loud enough, and if not, figure out how much more voltage swing would be needed. I've been on head-fi forums for a long time and never even heard of these headphones.
You could always run a balanced setup if you want more voltage swing.
You could always run a balanced setup if you want more voltage swing.
I agree 100%. For Grados which are highly sensitive my WE417 SET amp required a shunt regulated B+ and very well regulated DC heat. It has an absolutely dead silent background. But it is limited by the quality of the M19 SE OPT, it could be much better with a parafed permalloy core OPT but no one winds a core small enough to make "sense" for high efficient world class headphones.
But for modern orthos these are a different animal. Same impedance as the Grados, but require exponentially more voltage, these require a different design entirely.
There is a huge gap in the industry for high quality output transformers that work with headphones. I think with the orthos being very inefficient there is a better chance of finding a well designed OPT, but for low impedance high efficient headphones you can basically forget finding a well suited output transformer.
Many haven't heard how good a high end headphone can sound with an SET amp, I prefer them to speakers.
You have heard of them if you have been over there, Hifiman HE-5LE, very similar to HE-5 ad HE-500 and somewhat to LCD-2/3.
SET and balanced don't compute to me, you certainly can't get more voltage out of an SET with a center tapped secondary.
You would combine two SET amps into a push-pull configuration for more (double) voltage swing. The sonics should stay the same.
This would require 4 channels for stereo and an inverted signal.
Last edited:
Guess I would rather use a big power tube like a GM70 or 845 before I would build two set amps where the phases would never match, kind of defeats the concept of SET.
This is about the best I can come up with. The high order harmonics are dissapointing I don't know if it is the way I setup LTSWCAD or if this is typical for an SET. Also interesting is the 66p has much less distortion than a 5842 driving the 6c4c. The tubes are inexpensive may be worthwhile to try but probably better off with a hybrid for orthos.
An externally hosted image should be here but it was not working when we last tested it.
Well, if you really want to build a GM70, by all means, go for it. It is not any more expensive, cheaper than 300B actually. You would do D3A with CCS load -> GM70. A GM70 at full power would melt my full range speakers' voice coils, and those are far bigger than your headphones. Your headphone drivers would have to dissipate that power somehow, that same amount of power.
In a balanced amp, the + and - signals both come from your DAC. There is no phase inversion in the amp. They are two completely unmodified individual single-ended amps. The phase will add up perfectly. This is commonly done in the world of headphones, though not for power reasons.
In a balanced amp, the + and - signals both come from your DAC. There is no phase inversion in the amp. They are two completely unmodified individual single-ended amps. The phase will add up perfectly. This is commonly done in the world of headphones, though not for power reasons.
I really don't want to build a GM70, I'm looking for a practical solution to drive 2W into 38 ohms via an SET design that keeps the output DHT in a linear region swinging so many volts. But I do know that famed California headphone SET guru Frank Cooter chose the GM70 to power modern ortho headphones and really it makes a lot of sense when you do the math.
As far as the ideal of 4 SET channels for balanced operation the tubes would have to match so perfectly that I personally consider the idea to be really out there, it basically nulls any advantage of SET philosophy, meaning I would build a p-p amp should I want balanced output for voltage headroom/power before I would design a 4-channel SET to power two drivers. There are advantages to using a center tapped secondary on an SET for balanced out of a SET headamp but power isn't why, its a different subject entirely. But possibly and unfortunately P-P may really be the most practical option for these phones regardless. Thanks
Anyone have opinions on the spice model prediction I posted with the 6n6p driving a 6c4c (6V 2A3) ? Should I really expect that much high order distortion ?
Why not double the tubes in single-ended parallel, go with a 1.75K:40 transformer, and then you only need a 75V swing across the primary and have twice the power? Assuming that adding up the capacitances within the tubes in parallel doesn't start to limit the treble...
You say that wattage isn't the issue voltage is, but that's what a transformer is for, so voltage and current become interchangeable as long as there's enough wattage to do the job.
You say that wattage isn't the issue voltage is, but that's what a transformer is for, so voltage and current become interchangeable as long as there's enough wattage to do the job.
...and balance isn't as critical for a balanced amp as you might think. The two sides add up and if it's assymetrical in ampitude who cares when the sum is OK. In blanced it's not like one side is push and the other pull and you get inaccuracy due to waveform assymetry. In balanced both sides carry the full waveform and they add...if you watch one with the other as reference (like your load does) if one side is MUCH greater amplitude than the other there is NO signal fiedlity degradation, they still add to create an accurate waveform.
An SE amplifier should drive these phones just fine, we're talking 2W maximum here, and a balanced output which just requires an output transformer with a pair of 10 ohm secondaries wired in series and grounded at the electrical center. (Balance doesn't need to be perfect here since CMRR is not really a consideration)
A single 300B per channel should give you excellent linearity at 2W starting with a 5K primary and 400V fixed bias operation at 60 - 70mA. Jack Elliano can set you up with exactly what you need for an output transformer.
You can drive the 300B with a single triode D3A using a plate choke and capacitor, bias from a red led and running on a supply of 200V or so.. Use a CCS to set the plate supply voltage to the choke with a high quality cap for local supply decoupling and to set the LF corner > 10uF or more.. Operating point will be dictated by the CCS, AC behavior by the choke and cap, or you could just go with a CCS plate load and run a higher supply voltage to the CCS.. You can also use a 5842/417A with an IT having a step up ratio of 1:2.25 or so, using the same technique or something similar.. Like battery grid bias have a look at lithium iron AA/AAA batteries...
It is not that difficult, no need to over complicate or go negative about its feasibility. I have a 700mW device sitting in my basement that does essentially the same thing for 32 ohm headphones.
A single 300B per channel should give you excellent linearity at 2W starting with a 5K primary and 400V fixed bias operation at 60 - 70mA. Jack Elliano can set you up with exactly what you need for an output transformer.
You can drive the 300B with a single triode D3A using a plate choke and capacitor, bias from a red led and running on a supply of 200V or so.. Use a CCS to set the plate supply voltage to the choke with a high quality cap for local supply decoupling and to set the LF corner > 10uF or more.. Operating point will be dictated by the CCS, AC behavior by the choke and cap, or you could just go with a CCS plate load and run a higher supply voltage to the CCS.. You can also use a 5842/417A with an IT having a step up ratio of 1:2.25 or so, using the same technique or something similar.. Like battery grid bias have a look at lithium iron AA/AAA batteries...
It is not that difficult, no need to over complicate or go negative about its feasibility. I have a 700mW device sitting in my basement that does essentially the same thing for 32 ohm headphones.
Nah, many triodes can do that or at least come pretty close. Not nearly impossible by any stretch.
71A for example can swing around 230 volts peak. The 1626 is in the same category; tubes like 2A3 can swing a lot more (around 360 volts) not to mention the 300B.
Depends on the load impedance of course. I run parafeed with gyrator loads, for maximum output signal swing voltage. With a 2.5k transformer primary you'll get less, but still for example the 2A3 will swing 250+ volts easily.
This is the design keeps pointing to a 300B but hoping to find a cheaper DHT than the 300B. Also I've always liked the idea of parafeed for headphones, it seems a good fit. Thankyou for the input.
A 71A may be able to swing 230V peak but the trick is finding a tube that can do it with linearity. The best way to think of this if you're a speaker guy is I need a 9W SET.
But the 9W figure really is only 2W
Its Ohm's law IR=V, my headphones are 4 times the impedance so need 4x the voltage. That's asking a lot from a DHT .Thanks for all the input, as with anything new like 2w watt headphones
things can be a bit controversial. But its nice to step out of the box.300B is cheap! Maybe $30. Gm70 same price. 300B can deliver the 2W using standard 8ohm speaker transformer, but with such a high damping factor it may sound drier like an SS amp. My Hashimoto transformers have a 16ohm tap which is useful for headphones, which you could consider. My EML 300B XLS can run on 600V. Also you can consider.
You can also build a 6C33C OTL tube amp. This tube is cheap. It is not as linear as a 300B (which has a parallel plate design), and does not sound as good, but you won't need a transformer. Saves you a lot of money right there.
You can also build a 6C33C OTL tube amp. This tube is cheap. It is not as linear as a 300B (which has a parallel plate design), and does not sound as good, but you won't need a transformer. Saves you a lot of money right there.
http://www.mif.pg.gda.pl/homepages/frank/sheets/021/7/71A.pdf
Well, if those curves aren't evenly spaced enough, you really are going to be in trouble!
My advice would be to get some 4p1l (real cheap but linearity doesn't really get any better), run them balanced with a cascode CCS (I have a darlington pair and a MOSFET there) in the tail, single MOSFET gyrators on the plates. Put the OT between the gyrator MOSFET sources, thru a 2µF cap (go soviet for best quality and price).
Transparent as anything, and with plenty of power.
I have a quad of 4p1l's, but am not at all familiar with the topology you are describing. Do you mean run them parallel para-feed? Where could I find a similar schematic ? The problem I see with the 4p1l is the Vg runs only around -18V, not leaving much headroom.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.