Thinking Outside the Box
No R, No C, No Batteries, No Hum
Steve Bench did some very interesting work outside the box.
See some of it here-
ARTICLE OVERVIEW
For those who want to avoid the PS cap in the audio cct here is an alternative developed by Steve Bench-
No R, No C, No Batteries, No Hum
Steve Bench did some very interesting work outside the box.
See some of it here-
ARTICLE OVERVIEW
If someone wants to cook an excellent dinner, why rob the chef of one of the strongest spices in the kitchen? If you want a clinical sounding, "neutral" amp, then it has nothing to do with music but a lot with those clinical emergency rooms. No music can survice in such an environment. Seems that you don't have realized what music and its reproduction is all about. For me, the complete wrong way. Done that, been there. It sounds horrible to me.
A direct coupled power amp with a coupling cap after the first stage. I cannot make head or tail of it.Here is the DC coupled amp described in the April 1951 edition of Audio Engineering. There are simpler ccts published more recently.
Fully agree. Capacitors are evil and even the most expensive types should be avoided if possible.
And this is why many of us enjoy inductive coupling between the stages. These days it is easy to get spoiled by all the exotic core materials available.
This said, if you really want an amp with no colourations, why are you even looking at valves?
Here is the DC coupled amp described in the April 1951 edition of Audio Engineering. There are simpler ccts published more recently.
Coupling capacitors are C1 and C2, in power supply. Instead of being coupled through the grid it is coupled through the cathode.
The only place where a coupling capacitor better to eliminate, is in the output stage, especially of a triode amp, where grid current on peaks is audibly non-linear. In pentode or ray tetrode amps with some power headroom it is not so critical. It is not so common to see dependence of control grid current on voltages, but some datsheets show it. 2 best ways to eliminate "capacitor sound" are either to use an interstage transformer, or a source/cathode follower.
Hi Guys
Though it's been a while since I build tube projects, I've always wondered why the focus has been on getting rid of C in signal path. From my experiments on dac , some SS amps, PS has greatest influence on the sound be it C in the way of signal or not. Even if you can get rid of C you'll still end up with R & C in the speaker XO unless your speakers are FR. Since C is such a great evil why not focus on getting rid of it on the whole chain, possible ????
Though it's been a while since I build tube projects, I've always wondered why the focus has been on getting rid of C in signal path. From my experiments on dac , some SS amps, PS has greatest influence on the sound be it C in the way of signal or not. Even if you can get rid of C you'll still end up with R & C in the speaker XO unless your speakers are FR. Since C is such a great evil why not focus on getting rid of it on the whole chain, possible ????
Do no remove capacitors, place more.
Caps in the input stage isolate the volume switch = it sounds better
Caps in the interstage = better than transformers
Caps in coupling help too with blocking low frequency, annoying DC elevation etc.
I thought I had built one solid state system with 0 caps, until i learned that all dacs had thousands if not more of capacitors in and out of the signal path
if you want to play God and reinvent the creation, remove all capacitors, use massive chokes , regulators and transformers, go bankrupt.
3 things are fundamentals of electronics: cap/res/coils if you remove one of those the world collapse.
Caps in the input stage isolate the volume switch = it sounds better
Caps in the interstage = better than transformers
Caps in coupling help too with blocking low frequency, annoying DC elevation etc.
I thought I had built one solid state system with 0 caps, until i learned that all dacs had thousands if not more of capacitors in and out of the signal path
if you want to play God and reinvent the creation, remove all capacitors, use massive chokes , regulators and transformers, go bankrupt.
3 things are fundamentals of electronics: cap/res/coils if you remove one of those the world collapse.
Last edited:
As I already explained, the problem is not in capacitors themselves, but in coupling of stages with non-linear resistances through capacitors.
I would get rid of them where possible and not all designs are equal, and/or allow this without penalty.
I've heard a marked difference and especially with low level stage coupling capacitors (types). Much greater than any power supply - assuming the PS was of reasonable standard to begin with. And best of all was without.
YMMV.
Getting rid of a capacitor can be done electrically, aswell as physically.
For example in a push-pull design, if the push side is biased (perhaps using a servo if necessary) to draw the same current as the pull side, then the power supply capacitor disappears from the circuit as only a constant current is supplied by it. By a similar token, the capacitors in a push-pull design also disappear if their "charge up" characteristic is mirrored by their discharge characteristic.
Finally, clever design can be used:- in Norman Coren's TENA amp, the operating points for the first stage and phase splitter were deliberately chosen so that as one valve drew more current, the other less, so overall, even the asymmetric bit before it became push-pull, drew constant current.
Ultrapath capacitors are probably another example where adding that second capacitor helps make the first one disappear, so to speak.
kind regards
Marek
For example in a push-pull design, if the push side is biased (perhaps using a servo if necessary) to draw the same current as the pull side, then the power supply capacitor disappears from the circuit as only a constant current is supplied by it. By a similar token, the capacitors in a push-pull design also disappear if their "charge up" characteristic is mirrored by their discharge characteristic.
Finally, clever design can be used:- in Norman Coren's TENA amp, the operating points for the first stage and phase splitter were deliberately chosen so that as one valve drew more current, the other less, so overall, even the asymmetric bit before it became push-pull, drew constant current.
Ultrapath capacitors are probably another example where adding that second capacitor helps make the first one disappear, so to speak.
kind regards
Marek
Note: This post has been edited by a Moderator to correctly identify the person being quoted. Please be more careful.
At least ponder about what resistors, capacitors and cables to use. These are not passive, ideal components. May components have iron
in the wires or cases.
I bought those nice (Chinese) metal film resistors. Held a magnet besides them: all the wires are iron meshed with a bit of maybe copper and some white outside metal. The old cabon resistors I buy from Element14 do have copper wires. And then the band that sticks the wire to the substrate is most often iron too (do the test with the magnet. There are resistors that have no magnetic material. Those have the best frequency behaviour.
I inadvertently used 'super-duper' silver coated shielded interlinks and also used these for connecting components. Turned out these were cables for HF use (radio) and the core is iron, so any low frequency signal will get a counteraction (slew rate). Replaced then, thin sound became rich again.
Some buy PIO capacitors only to find out that they are in a metal case. So they take off the case. An improvement is claimed from that treatment. Saying that I had true Western Electric caps in a metal case that sounded great and dynamic.
At least ponder about what resistors, capacitors and cables to use. These are not passive, ideal components. May components have iron
in the wires or cases.I bought those nice (Chinese) metal film resistors. Held a magnet besides them: all the wires are iron meshed with a bit of maybe copper and some white outside metal. The old cabon resistors I buy from Element14 do have copper wires. And then the band that sticks the wire to the substrate is most often iron too (do the test with the magnet. There are resistors that have no magnetic material. Those have the best frequency behaviour.
I inadvertently used 'super-duper' silver coated shielded interlinks and also used these for connecting components. Turned out these were cables for HF use (radio) and the core is iron, so any low frequency signal will get a counteraction (slew rate). Replaced then, thin sound became rich again.
Some buy PIO capacitors only to find out that they are in a metal case. So they take off the case. An improvement is claimed from that treatment. Saying that I had true Western Electric caps in a metal case that sounded great and dynamic.
Last edited:
Nickel is also ferromagnetic and frequently used in metal film resistors, a magnet does not prove it is iron ...
nickel chromium (NiCr) is a popular film material
You are right: white gold contains a lot of nickel and hence gold is sometimes magnetic
In my case the central conductor (silver plated, teflon isolated) is very magnetic and holds a multiple of its weight while the shield (copper/silver and teflon outer isolator) is non-magnetic. It probably is a >75 ohms cable for radio. Should I make a picture? I try to make the point that in my case using such cables 'squeezes' out the sound, it becomes thin, it pinches dynamics. [But hey I am only a bricoleur.] And if that is in the feedback loop it will be really weird.
Note: This post has been edited by a Moderator to correctly reflect the person quoted, hence Tubelab_com's response.
My latest UNSET amp design had TWO coupling caps in the signal path....and I don't use pricey caps. The design is new and only one prototype exists currently, but as of now I'm not sure which is the better sounding amp, the TSE-II with 45's, 2A3's or 300B's (tried them all) or the UNSET with a total of $3 worth of tubes in it.
The TSE_II uses good Transcendar OPT's and the UNSET is currently running 13 year old Hammond 1628SEA's and has tamed the "Hammond notch" by virtue of it's low driving impedance (fat TV sweep tube in triode mode) although I have stepped up to $10 tubes in my latest testing.
UNSET is coming?
Wavebourn stated the true problem. If the coupling cap sees a non constant time varying load when driven with real music, it can and will misbehave. This is why different caps will react differently.
In the TSE-II and the UNSET designs the coupling caps see a constant load of 1 MEG OHM in parallel with a few pF of gate capacitance. The fets that follow the coupling cap are chosen and biased such that that capacitance remains constant through the range of signal seen in the amplifier's use.
Frequency response of either design stretches from about 10 Hz to hundreds of KHz except for the OPT which is by far the most dominant pole in the amp, and there is no feedback applied from the OPT's output.
The TSE-II, like the original TSE is a no-feedback design while the UNSET uses local feedback loops around each pentode to emulate triodes. No global feedback is applied.
Enclosing a capacitor inside a feedback loop is another way to bring out it's worst qualities, since correction signals through the cap can be nonlinear.
It wasn't me who said that. I use plenty of wire, resistors and caps. My speaker cable comes from Walmart, and my interconnects were grabbed up when a local Radio Shack store crashed and burned.
My latest UNSET amp design had TWO coupling caps in the signal path....and I don't use pricey caps. The design is new and only one prototype exists currently, but as of now I'm not sure which is the better sounding amp, the TSE-II with 45's, 2A3's or 300B's (tried them all) or the UNSET with a total of $3 worth of tubes in it.
The TSE_II uses good Transcendar OPT's and the UNSET is currently running 13 year old Hammond 1628SEA's and has tamed the "Hammond notch" by virtue of it's low driving impedance (fat TV sweep tube in triode mode) although I have stepped up to $10 tubes in my latest testing.
UNSET is coming?
Wavebourn stated the true problem. If the coupling cap sees a non constant time varying load when driven with real music, it can and will misbehave. This is why different caps will react differently.
In the TSE-II and the UNSET designs the coupling caps see a constant load of 1 MEG OHM in parallel with a few pF of gate capacitance. The fets that follow the coupling cap are chosen and biased such that that capacitance remains constant through the range of signal seen in the amplifier's use.
Frequency response of either design stretches from about 10 Hz to hundreds of KHz except for the OPT which is by far the most dominant pole in the amp, and there is no feedback applied from the OPT's output.
The TSE-II, like the original TSE is a no-feedback design while the UNSET uses local feedback loops around each pentode to emulate triodes. No global feedback is applied.
Enclosing a capacitor inside a feedback loop is another way to bring out it's worst qualities, since correction signals through the cap can be nonlinear.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.