This new HPA from Schitt touts an output stage that has no output transformer and no output capacitor between your headphones and the tubes. How can this be?
https://audioxpress.com/news/schiit-introduces-limited-edition-10-tube-dc-coupled-otl-tube-amplifier
https://audioxpress.com/news/schiit-introduces-limited-edition-10-tube-dc-coupled-otl-tube-amplifier
Use bipolar supplies, and some kind of trim or correction (servo. etc.) for drift and the remaining offset.
Here's the 1952 first commercial op amp as an example (tube of course).
Here's the 1952 first commercial op amp as an example (tube of course).
Last edited:
Headphones have relatively high impedance and they need low power, that is low current. Parallel connected cathode followers in bridge arrangement can easily provide the current and voltage swing.
I would set the idle current higher than the peak output cirrent need. 100mA shared among 5 tubes? Doesn't seem too difficult.
I would set the idle current higher than the peak output cirrent need. 100mA shared among 5 tubes? Doesn't seem too difficult.
https://www.diyaudio.com/community/...id-tube-circlotron.386403/page-3#post-7066642 It's possible to make a circlotron with tube output - the chassis I have as space for 4 ecc99s per channel. I decided to switch to BJTs simply because I'm running low impedance. I may at a later stage switch out the low voltage and make a all tube headphone version.
You can then run 4 ecc99 triodes at a bias of 24mA. The difficult part for a multi-triode output stage is getting the triodes to balance without current hogging causing tubes to run hotter than the others. Broskie got around this for the flutterman style by having each top and bottom pair connect through a cap allowing the pair to be balanced w.r.t to the others.
You could do the same with the circlotron but.. one grouphave to collectively balance the other group to prevent the same issue or distortion occurring.
Not having a hybrid means they can run a simple four power supplies with one extra for a common heater supply (the signal to the headphones never threatens the cathode-heater limit).
EDIT: I just noted their headphone jack is a 3 pin unbalanced. Option there is to use a single-ended circlotron, or, use a futterman style with a servo and hope the tubes are closely matched.
You can then run 4 ecc99 triodes at a bias of 24mA. The difficult part for a multi-triode output stage is getting the triodes to balance without current hogging causing tubes to run hotter than the others. Broskie got around this for the flutterman style by having each top and bottom pair connect through a cap allowing the pair to be balanced w.r.t to the others.
You could do the same with the circlotron but.. one grouphave to collectively balance the other group to prevent the same issue or distortion occurring.
Not having a hybrid means they can run a simple four power supplies with one extra for a common heater supply (the signal to the headphones never threatens the cathode-heater limit).
EDIT: I just noted their headphone jack is a 3 pin unbalanced. Option there is to use a single-ended circlotron, or, use a futterman style with a servo and hope the tubes are closely matched.
Interesting bit from 2009 tube cad about impedance multiplying: https://www.tubecad.com/2009/09/blog0171.htm
The OP's link gives a hint along with each pair of tubes having a DC servo:
So basically it sounds like each tube is biased by servo, the common output is then linked via a solid state impedance multiplier and out to the headphones with an 'analogue' safeguard ensuring a mute is performed if the tubes misbehave.
The OP's link gives a hint along with each pair of tubes having a DC servo:
So basically it sounds like each tube is biased by servo, the common output is then linked via a solid state impedance multiplier and out to the headphones with an 'analogue' safeguard ensuring a mute is performed if the tubes misbehave.
Last edited:
Why are you interested in a vacuum tube headphone amplifier that does not have an output capacitor, and that does not have an output transformer?
I do not want to hear that all capacitors and all transformers interfere with your hearing a very pleasing sound.
As soon as all the recordings you listen too do not use any signal transformers, any coupling capacitors, and not even any bypass or filter capacitors, then I might consider the concept of such headphone amplifier(s).
A car that is designed to run on 4 tires can run with only 3 tires. Remove one tire, and do not move the position of the other 3 tires, just add a proper counterbalance. (Engineering can make almost anything that is not practical run relatively smoothly).
All you give up is: Safety, cornering capability, mileage (added counterweight), maximum load rating and where the load has to be distributed, and the increased cost, etc.
As always, just my opinions.
Your mileage may vary.
I do not want to hear that all capacitors and all transformers interfere with your hearing a very pleasing sound.
As soon as all the recordings you listen too do not use any signal transformers, any coupling capacitors, and not even any bypass or filter capacitors, then I might consider the concept of such headphone amplifier(s).
A car that is designed to run on 4 tires can run with only 3 tires. Remove one tire, and do not move the position of the other 3 tires, just add a proper counterbalance. (Engineering can make almost anything that is not practical run relatively smoothly).
All you give up is: Safety, cornering capability, mileage (added counterweight), maximum load rating and where the load has to be distributed, and the increased cost, etc.
As always, just my opinions.
Your mileage may vary.
Saving up for these electrostatic headphones! https://www.stereophile.com/content/jecklin-float-electrostatic-headphones
Amp is ..... "doable" .........
5 such triodes in parallel can be biased at 112mA which is enough for 1W into 80 ohm
No need for bridging, outputs can be connected in totem pole fashion.
Complex independent supplies needed, probably signal and DC voltage shifting to couple positive and negative swinging halves, but hey, guess hype sells and makes this kludge posible.
No complete Silicon avoidance though, does a voltage shifting Zener count as such?
And the
He claims it´s not because it´s in parallel with (meager) tube current. so not "really" a buffer.
let´s be polite and say it´s a "grey area".
Is it a good design, from the Engineering point of view? .... not so sure about that, but since he plainlybrecognizes it, he´s cheating nobody, is he?
But at least I like he´s candid about it, so ......
5 such triodes in parallel can be biased at 112mA which is enough for 1W into 80 ohm
No need for bridging, outputs can be connected in totem pole fashion.
Complex independent supplies needed, probably signal and DC voltage shifting to couple positive and negative swinging halves, but hey, guess hype sells and makes this kludge posible.
No complete Silicon avoidance though, does a voltage shifting Zener count as such?
And the
stinks of a SS buffer.
He claims it´s not because it´s in parallel with (meager) tube current. so not "really" a buffer.
let´s be polite and say it´s a "grey area".
Is it a good design, from the Engineering point of view? .... not so sure about that, but since he plainlybrecognizes it, he´s cheating nobody, is he?
But at least I like he´s candid about it, so ......
This is the reason I switched for low impedance to a BJT hybrid. Not that it's simpler with nine floating power supplies. It was either that or spend on two lundahl headphone transformers (maybe in the future once the OTL craze has worn off).
The original poster may have missed that unless his valves push and pull with complimentary operating points and thus only recirculate current in a loop as the programme material varies, then the power supply capacitor is part of the signal path.
By a great stroke of irony, an electrostatic speaker is nothing more than a huge flat capacitor.
kind regards
Marek
Newby here, so my original thought was about safely coupling headphones to the output without electrocuting a user that unplugs their phones at the phones end. Since then I've searched out various push pull circuits that use a split supply which would be safe (I guess). Turns out John Broskie has published a lot of such circuits.
You do need to check the behaviour in scenarios such as tube short anode to cathode, open tube or headphone short.
The safest tube headphone amplifiers would include an output transformer (single ended or push pull),
It has to have one end of the secondary grounded to the ground of a proper 3-wire power mains system.
But, now the problem is if you grab another equipment or device that is not grounded via the same 3-wire power mains system.
That could be the stereo, hi fi, fan, light, refrigerator, etc. That is true, especially if the device fails, or is just poorly designed and has some leakage current.
Your grounded headphones might provide a good path from your head to your hand that contacts an ungrounded or faulty equipment.
Safety First!
Prevent "The surviving spouse syndrome"
It has to have one end of the secondary grounded to the ground of a proper 3-wire power mains system.
But, now the problem is if you grab another equipment or device that is not grounded via the same 3-wire power mains system.
That could be the stereo, hi fi, fan, light, refrigerator, etc. That is true, especially if the device fails, or is just poorly designed and has some leakage current.
Your grounded headphones might provide a good path from your head to your hand that contacts an ungrounded or faulty equipment.
Safety First!
Prevent "The surviving spouse syndrome"
And start-up with cold tubes. No they will not all warm-up in perfect sync.
I do not know what's in that circuit, but it is likely a simple current mirror with dissimilar sized mosfets. These are commonly used inside IC chips where all the mosfets are identical except for size. Doubling the size of the driven fet creates a mirrored current that is twice the sense fets current.
Many years ago I wired a simple mosfet current mirror into the cathode circuit of a 45 tube. There were 5 identical mosfets wired as a 4 X current multiplier. I manually plotted curves of this thing and they did indeed look like a 45 on steroids. I never actually built an amp and listened to it though, so the idea fell through the cracks long ago. There were some experiments done on an SSE board using 600 ohm OPTs, mosfets, BJT's and 6W6 tubes since they work well at low voltages, they are cheap, and I have plenty of them. My old paper notes indicated some promising results, but the comments from the "change your name to Transistorlab" vacuum tube purist crowd hadn't died off from the TSE design yet, so I shelved the idea along with the dsPIC powered high efficiency SE amp that's now being reinvestigated.
Remember that many of his circuits only exist in the mind of a simulator and may not work in the real world. Do your own breadboard testing before launching into a full build.
I'm thinking a White CF as the output stage with stabilized +ve & -ve supplies would be a good start.
Driven by a differential stage front end. Something to look at when time allows.
Driven by a differential stage front end. Something to look at when time allows.