Single ended means that we can only plug the cans into the output - yes?
I would like to see something that has a front end less sensitive to cable length than a 12AU7. My headphone amp is next to the chair where I also listen to my speakers. There is about 15 feet of quality microphone cable needed to get there, it is audible. The MOSFET tail end seems to get along with my AKG 701s.
Perhaps the solution is a modular approach: a couple of driver stages and a couple of input stages so we can mix and match as desired, or build them all and ABX for the rest of our lives. I would like that! We call this type of thing a Eierlegendewollmilchsau in Germany - a pig that lays eggs, gives milk and has fur.
The true RG HA is a chameleon, just like our members!
Single ended means only 1 active element is providing the output signal So 1 transistor, FET, tube and to complement this 'active element' there is a passive element like a big resistor or a transformer. You could also see a current source as a passive element in this case. The opposite of single ended is an output configuration where there are 2 active elements providing the output signals as is done in most amps/op-amps.
If I were to design an amp in the future it would be modular (as to your suggestion)
There would be a 'motherboard' where you can mount an 'amplification section', an 'output section' and a 'power section' With possibilities of local and overall feedback and changing of power supplies. It would NOT become a portable amp though but something that can be changed/added e.t.c. easily. So you could configure it with cheap parts and expensive parts GD modules, Tubes, MOSFET's whatever you like.
But given the fact that MOST of us already have (several) good amps I don think it is worth the effort.
Lets just back up a bit, The Bravo/Indeed seems to have a kick ass driver section with Mosfets. 1 per channel(qualifying it as single ended) and no one has complained about the sound. The input valve section seems to be where we have several options.
Maybe my thoughts are a workable RG breadboard amp where we can all get first hand experience without lifting the traces on the chinese "motherboard" when resoldering. This would allow the less talented to listen to differences only otherwise accessable with investments in a lot of amps.
I started here with a BT928, then modded it as per Mikes article, Then came a CMOY and finally a modified Szekeres according to some tips that Rick gave me. I added a 12AU7 tube front end as buffer/gain stage and now am thinking about something SS discrete to compare with the tube. I do all of this stuff on Veroboard point to point and it is fun, even if it takes 6 months to do.
I would really enjoy a (cheap) development environment where I could try out/compare stuff. If you will a system in flux for the next 20 years or so. I think this is something that would be very good for RG. Then we do not tie ourselves down to a specific flavor of the week, month, year. Maybe we would also reduce the fear of getting started.
This development environment is my wish for a RG "product". Then we never have to agree on anything.
I have an idea about a variation with the emphasis on 'variation'.
Wondering if I should build an electrostatic HP and how to drive it (high voltage amplifier or transformers) I came to the following HP amplifier idea.
As an amplification part you generally use active elements like transistors, FET's, MOSFET's, opamps in all tastes and all sorts of tubes but you can also use a transformer for voltage amplification. In my younger years (no wife, kids and cats and better ears) I used to enjoy homemade electrostatic speakers with an active woofer (16Hz - 100Hz @ -3dB) the latter to 'unload' the electrostats because of the 'acoustical short'that troubles all narrow open bipolar speakers resulting in a dip around that frequency range. I (and all visitors I might add) was always amazed about the 'exact' sonics of it. Using a turntable with MC element (0.5mV) which had to be increased to a voltage of 1,500V meaning an amplification factor of 3,000,000 times (3 million !) a staggering 130dB and still the detail retrieval was phenominal. the MC was connected to a 10x transformer and the Electrostatic LS had a transformer amplification factor of 55x, so 550x times amplification just by passive transformers and still a respectable 5500 times (10 times more) with transistors.
So the idea is to build a 'passively amplifying' active class A single ended output (coupling capacitor free if possible) HP amplifier. Ofcoarse the input and output have to be matched very well and it has to have a lineair input resistance of at least 10kOhm and be able to drive 32 Headphones and 600 Ohms HP's. An amplification factor between 3 and 12 times (selectable in steps) would be enough and is easy to do with a wide enough frequency range (30Hz to 25kHz).
Given the fact that the output stage of the indeed thingies work perfectly and if driven correctly can easily reach 0Hz(DC) to 1MHz (1,000kHz) when there is no output cap. I intend to make an amp with this endstage (0dB amplification current loaded source follower) as an input stage to drive the transformer and a similar, but more powerfull, end-stage that can drive 32 Ohm HP's fully in class A. Obstacles have to be concurred. The output stage should be DC coupled (no output caps) and the transformer must not be fed with DC and loaded and driven properly, maybe some ringing/squarewave reproduction has to be addressed and possibly meaning DC servo has to be applied e.t.c.
Very curious to see and hear how this performs am gonna build it when there is time.
this is the idea (current status: completed project !)
that's an interesting idea I'm a bit new to solid state but I've used interstage transformers in my valve amps. just a thought but you could also make the driver stage a valve and put a choke on the anode in a para-feed stylee. I'm new to solid state and rusty with valves and for me it would be good to understand the a little of the circuit topology and a bit of the maths in a basic form. if thats gonna be possible thats just great. I'd be up for some of the hands on stuff if that helps
See it as 2 cathode followers (no voltage gain, approx 0.999x, only meant for impedance matching) and the transformer as the amplifying part. Most of the distortion introduced in amps comes from the amplifying parts due to non-linearities and transfer characteristics. source/cathode/emitter followers don't 'add' or screw up much in the audio signal. Ofcoarse transformers also add distortion and non-linearities as the core becomes more saturated when the voltage increases and there is also 'hysteresis' in the core material. But this is different distortion from those of electronic parts and thus has other sonic properties too.
the Source of the input stage only 'follows' the input voltage on the Gate, and is quite accurate at that I might add. The input is very high-ohmic and can be solely determined by the used input resistor and the output is very low-ohmic and thus can drive a 250 Ohms impedance transformer very well.
The transformer has a ratio 1:6.3 (actually the one I am gonna use = 1+1:6.3+6.3) and you can serie/parallel windings and thus come up with different amplification factors) So the input voltage is amplified a factor 6.3 ( when used in 1:6.3 mode) and the impedance is transformed 6.3*6.3 and will be approx 10kOhm at it's output. With different windings selected you would have to match the drive and load resistors too.
The Source of the second stage, again, just 'follows' what is applied on it's Gate and being very high-ohmic means you can load the transformer 'ideal' with a resistor and/or maybe also a small 'compensating network' if needed. The Source output is near 0 Ohm and able to drive low-ohmic loads as long as the, by the load (read headphone) demanded, current doesn't exceed the class A current and is a near perfect copy of the output signal of the transformer.
I am just curious how it measures up to active amplifying parts and suspect it won't even sound bad at all, perhaps even surprisingly good. Given my experience with transformer driven electrostatic speakers and electret headphones it might even be excellent as long as the transfomer is too. Especially if you load and drive a transformer properly it 'conveys' signals very 'musically'.
It's a 'variation on a theme' that's the fun of this design and the thread is about variation . Not seriously meant to blow ones mind ... but then again... maybe it could. Won't know till it's actually built, which I intend to do just for the fun of it.
The DC servo needs to be worked out and if implemented properly won't affect the sound path.
Don't know why I would need a 6th HP amplifier though..
I'm new to headphones so still learning but in my experience with SE tube stuff the LF response can have as much to do with TX saturation. core construction can be an important factor and I've found that C-Cores are less prone to this.
of course C-Cores are comparatively expensive to construct which reflects in the purchase price. so i think your approach is spot on. see how the farnells go and if the circuit shows promise the better TX's become a more viable proposition.
it's an interesting idea that merits the good work you are doing to move it on.
I can't help noticing here and on other designs/posts, you do like your heatsinks
Have you compared it to anything else yet?
I like the amps to have a large voltage swing so need a high voltage on the parts. Also I like it to be class-A and needs to be able to deliver the needed currents for these high output voltages. To prevent (asymmetric) clipping in these cases the idle current must be higher then the maximum current that can ever be drawn. This is determined by the maximum output voltage swing and the output resistance + lowest impedance of the useable headphones (in this case 1 Watt in 32 Ohm) Therefore a high current is also needed. High voltages x high currents means high power demands (Watts) that are drawn constantly.
This needs to be dissipated as it does not go into the headphones it must go somewhere which is in the form of heat.
Since I don't like parts to heat up very much you need heatsinks that heat up only a little with a certain amount of power. In other words it must have a very low K/W value.
More surface and good thermal conduction and convection means BIG in size. hence the big heatsinks not because I like them big but because they must be.
Still this amp (it is not very big in size really) get's heated alright. Even with these heat sinks it still gets close to 50oC in a 18oC room with normal convection so in an open space on top of a table.
I did not yet compare it to my other amps but will attempt in the future. Going to make a special AB box where you can dial in output resistors in one path so you can do fair comparisons between amps. At least one amp needs to have a volpot. It will take me a while to finish this.
managed to give it a (subjective) listen with DT990, HD681(modded), SR850(modded) and HD650. Still 4 more cans to go. The 2 different output impedances have guite some audible effects dependant of the used can. No placebo and is easily verified, in the electrical output signal and thus also sonically, with an osciloscope and a frequency sweep or white noise and spectrum analyser (which I both have) .
DT990 prefers low Ohmic. At least I prefer that with the mentioned cans. A modded HD681 prefers high Ohmic but is at 'best' between 40 and 80 Ohms. HD650 is marginally better at higher impedances but has slightly different character on lower impedances and the preference for an impendance varies per music type. The SR850 performs about equal on both impedances, the lower Ohmic does make the bass tighter
I do not intend to make money on my designs.. so no patents.. If someone decides to use the transformer concept they can do so. It's all about maximum music enjoyment for everyone, not very commercial is it ? Not in my nature... don't need to be very rich either... I don't believe it will make you happier in the end. Having enough money does make your life easier... no doubt. What is enough ? If you feel you do not miss out on much and still have wishes left that will remain unfullfilled.
Have been thinking about shipping the amp to certain people to have it assessed 'subjectively'. Some names have popped up in my mind. Myself I will only to AB comparisons to stay true to my 'religion'
The weight included power supply is about 2.5kg meaning uninsured shipping will already be around E 20.- within Europe, not to mention the outside Europe rates. For now, however, it stays at home for some time.
You make music... without music there would be little use for headphone amplifiers... You can listen to music without amplifiers ...without amplifiers there can still be music... Now .. what was the more usefull job ?
Last Edit: Nov 27, 2010 14:20:10 GMT by solderdude
You make music... without music there would be little use for headphone amplifiers... To music you can listen without amplifiers ... Without amplifiers there can still be music... Now .. what was the more usefull job ?
Is that a chicken and egg joke?
The guy who makes my synthesisers?
Nah - it's a chicken and egg.
I still wish I had more practical skills. I used to put stuff together when I was younger till I got a nasty bolt and I think the shock rewired my brain.
I got 'bolted' a few times.... never could make music though. I did try and also 'helped' in 3 bands doing the sound. Maybe the rewiring is reveresed again at the second jolt and then stays that way...
Synthesisers do need some electronics... granted.. but I have a sneaky suspiscion you can handle a piano as well. Don't need no power (unless it's an electric one)