I never said capacitors don't block DC.

 

A DC signal as high as a thousand volts can be blocked with nothing more than a $1 capacitor, for decades, while allowing the AC components through

Anyone who's ever build or modded an amplifier will disagree with you. Including myself.

I never said capacitors don't block DC.

 

I was implying that a 1$ capacitor is going to sound horrible, but anyway this is getting off topic.

After doing a little more research I realized that the biasing effect of an amplified signal on a tube is much greater than I initially thought.
So I'm thinking that as long as the music is playing there shouldn't be any audio signal problems with the lane switching.

I was thinking of using something like this http://www.tubecad.com/2007/11/blog0126.htm along with added design choices in order to reduce tube strain as much as possible.
The major unknown factor is, how long does it take for a tube to reach its meltdown point? From the reading I've done so far most of them seem to be pretty robust, maybe a few seconds to many seconds before bad things happen.

So in other words this could still work, the bias-on switch timing may not be all too important but the bias-off switch timing would still be very important.

 

Also I am wondering, you mentioned that I cannot use my dac as an on/off indicator because of the way it works.
Why not make an alternative usb dac specifically for this purpose? Both signals will be the same coming out of the PC.

 

I was implying that a 1$ capacitor is going to sound horrible...

No, it isn't.

And neither will a $0.10 capacitor. First, you are equating measuring differently, sounding differently, and sounding horribly. In fact, these are three very different things. Second, you are proclaiming that the difference is perceptible, when in fact it usually is not.

Now, I know that the one thing golden-eared audio crazies hate is research, but it turns out that while the average person-on-the-street cannot hear the difference between a DC-servoed power amplifier and one with an aluminum electrolytic output capacitor, he (she) *can* hear the difference between 0.1% THD and 1% THD. 1% is a rarely achieved goal for an all bottle class A output stage, while 0.1% is almost trivially simple with solid state gear.

But more to the point, claims that the difference between a preamp with all polypropylene capacitors and one with all ceramic capacitors is *perceptible* (not measurable) are wrong. Properly controlled double-blind studies are rare, but the ones that are credible also are consistent. Also, that $50 per foot speaker cable is bogus.

I know golden-eared audio crazies. Golden-eared audio crazies are friends of mine. One of them taught me many things while I helped install the studios and master control room for a classical music FM station. But later I worked for a psycho-acoustics research lab and a speech and hearing research lab. Turns out that not only can you *not* hear most of the differences you claim to exist, but also - we know why.

ak

 

From your first post I got the impression you wanted to switch the grid bias in order to protect the output amplifier tube. That's why I asked about the reference information, because I can't find anything that says the tube needs that kind of protection. Even so, I can see how it could help by preventing electrons from hitting the plate when they're not needed. Now I think you may be thinking that cutting off the electron flow will eliminate some random noise. Is that true?

 

Also I am wondering, you mentioned that I cannot use my dac as an on/off indicator because of the way it works.
Why not make an alternative usb dac specifically for this purpose? Both signals will be the same coming out of the PC.

Yeah, well, Let's check my assumptions. It's my understanding that the USB signal is a complex bit stream. The complexity means that it needs to be constantly analyzed to determine what the original analog signal was. That analysis is done in a CPU, which ultimately keeps a 32 bit number internally representing the next analog value to output. Is that number externally accessible on some output pins anywhere inside the DAC? I assumed not, so let's check that. Do you have a schematic for the DAC?

 

No, it isn't

I'm not going to have this conversation, it's been done a million times before elsewhere. The difference in sound between a crappy capacitor and a high end capacitor in the interstage position is nearly 100%, nothing subtle about it. It's like having two very different amplifiers. No I am not quoting someone else, I only believe my own experience. Also in my experience the only people that claim that things don't make a difference are people that have never personally tried it themselves, also people that think THD (emphasis on T)is the only thing that matters in the signal. But anyway please, don't bring this up again, it's turning the thread off topic.

From your first post I got the impression you wanted to switch the grid bias in order to protect the output amplifier tube. That's why I asked about the reference information, because I can't find anything that says the tube needs that kind of protection. Even so, I can see how it could help by preventing electrons from hitting the plate when they're not needed. Now I think you may be thinking that cutting off the electron flow will eliminate some random noise. Is that true?

No no no, for example, in normal operation the cathode is heated and electrons are "boiled" off of it which then fly toward the plate at high speed because the plate is positively charged, if this happens unhindered the plate will melt and the tube will be destroyed.
This is why we have a grid, the grid applies voltage fluctuations (AC signal) between the plate and the cathode, when the grid becomes negative the electrons are repelled, when the grid goes positive, the electrons are sling shotted toward the plate.
What happens if there is no ac signal? This is why you have to bias the tube so the grid is always x amount of voltage below the cathode so if the signal is not applied then the electrons do not flow to the plate and cause catastrophic failure.

So basically instead of having a permanent bias, I want to switch the bias in and out when needed.

Do you have a schematic for the DAC?

Anything I have on the dac is here http://www.audio-gd.com/Pro/dac/NFB12015/NFB12015EN.htm
But I'm wondering, if the dac is able to "find" this signal, why can't we? I'm assuming it would take a software solution.

 

But I'm wondering, if the dac is able to "find" this signal, why can't we? I'm assuming it would take a software solution.

Yes. You can "tap" the serial bitstream that is entering the DAC, pick off the frame numbers, and know the exact start and end points of any track. That's what a CD player does and how the track select operation works.

BUT - I don't understand the primary question. By definition the average value of audio, whether it be compression/rarefaction of air or the direction of electron flow, is zero. Thus, the average value of the grid voltage during cuts and between cuts is exactly the same. Same for the plate current. So I don't see why a tube would overheat while sitting at its operating point.

To prevent some language problems, two questions. How old are you and where are you?

ak

 

So I don't see why a tube would overheat while sitting at its operating point.

Its operating point is whatever bias you set it to i.e. the difference in voltage you set between the cathode and grid.
There no default operating point other than 100% . If there was no difference in voltage between the cathode and the grid at idle then the current inside the tube would destroy it because there would be nothing preventing electrons from bombarding the plate.
For a 6080 tube If the grid is -30v DC and the cathode is 0v DC then at idle the -30v charge would prevent electrons from flowing to the plate when no AC is present.
The bias is the default operating point.

To prevent some language problems, two questions. How old are you and where are you?

25 and USA.

 

I'm not going to have this conversation, it's been done a million times before elsewhere. The difference in sound between a crappy capacitor and a high end capacitor in the interstage position is nearly 100%, nothing subtle about it. It's like having two very different amplifiers. No I am not quoting someone else, I only believe my own experience. Also in my experience the only people that claim that things don't make a difference are people that have never personally tried it themselves, also people that think THD (emphasis on T)is the only thing that matters in the signal. But anyway please, don't bring this up again, it's turning the thread off topic.


No no no, for example, in normal operation the cathode is heated and electrons are "boiled" off of it which then fly toward the plate at high speed because the plate is positively charged, if this happens unhindered the plate will melt and the tube will be destroyed.
This is why we have a grid, the grid applies voltage fluctuations (AC signal) between the plate and the cathode, when the grid becomes negative the electrons are repelled, when the grid goes positive, the electrons are sling shotted toward the plate.
What happens if there is no ac signal? This is why you have to bias the tube so the grid is always x amount of voltage below the cathode so if the signal is not applied then the electrons do not flow to the plate and cause catastrophic failure.

So basically instead of having a permanent bias, I want to switch the bias in and out when needed.


Anything I have on the dac is here http://www.audio-gd.com/Pro/dac/NFB12015/NFB12015EN.htm
But I'm wondering, if the dac is able to "find" this signal, why can't we? I'm assuming it would take a software solution.

OK, thanks, that's consistent with what you said in your first post, so you're keeping me on track. In my experience there are two ways to control electron flow to the plate, not counting the grid. If you reduce the cathode heating, you reduce the electron flow. And if you reduce the potential of the plate with respect to the cathode, you reduce electron flow. In fact, if you leave the plate unconnected, virtually no electrons leave the cathode and hit the plate, they just cluster in a cloud around the cathode.
Here's where we have a difference of opinion, perhaps based on personal experience. I think the tube is designed to operate at some normal electron flow in a steady state condition for as long as the cathode can function, the heater doesn't open circuit, and the grid isn't contaminated, I.e., years to decades. The grid allows us to modulate that quiescent state to vary the output current based on an input signal.
Now, if I choose to run the tube at such a high plate-cathode voltage that enough electrons flow to burn up the plate, then I'm likely operating outside the design range of the tube. OK, sometimes that's the only way to solve a problem, like overclocking a CPU. But even if I did that, once I put the grid into the equation I have to set the grid bias so that the tube is operating around midpoint between zero and full electron flow, which should be a safe value to run at for the tube. Do you agree? You can see this in both tube and solid state design graphs based on the load line. You're familiar with the load line, yes? Is your understanding in line with mine at this point, or are we diverging?

As far as the DAC goes, I couldn't find enough information on the web site to hack into it and use internal signals. But here's a thought. Why not switch the bias when the USB signal starts and stops? Even if the USB audio signal is basically quiet, it's still the incoming signal, so it should be amplified. But if the incoming USB signal stops, then no audio needs to be transmitted. Do you have an oscilloscope?

Also found this site with amplifier designs that avoid the use of resistors and capacitors.
http://diyaudioprojects.com/mirror/members.aol.com/sbench/norc_pa.html

 

I think the tube is designed to operate at some normal electron flow in a steady state condition for as long as the cathode can function, the heater doesn't open circuit, and the grid isn't contaminated, I.e., years to decades.

It depends entirely on the tube, most aren't designed to withstand no bias.

Now, if I choose to run the tube at such a high plate-cathode voltage that enough electrons flow to burn up the plate, then I'm likely operating outside the design range of the tube

Not at all, I've never seen a tube like this, Tubes are designed from the beginning to run on crazy voltages, except maybe some low voltage radio tubes or something. It is not like overclocking a CPU, overclocking a tube would be doing pushing voltages above the rated specs, this can be done and tubes are usually pretty robust and can handle it, but it reduces the lifespan of the tube by a lot.
The datasheets usually give instructions that the tube must be biased, the grid voltages on the load lines are the typical bias points for any given plate voltage/current.
If you are going to use a tube for audio you have to look at the load lines on the datasheet to find the most linear area of operation with enough headroom for linear voltage swing on either side which all involve high plate voltages and currents depending on the tube, sometimes over 1KV and amps of current, you can't just pick a lower voltage because you don't want to deal with biasing.
At any practical operating point for audio it is going to be a bad idea to have a tube unbiased, unless we are talking about certain pentodes and power triodes that are designed to have a positive grid bias and have massive dissipation capabilities.

Do you have an oscilloscope?

No but I just ordered one.

 

But here's a thought. Why not switch the bias when the USB signal starts and stops? Even if the USB audio signal is basically quiet, it's still the incoming signal, so it should be amplified

Isn't that basically the same as my binary interception idea? It's all binary before it hits the dac.

 

Isn't that basically the same as my binary interception idea? It's all binary before it hits the dac.

Yes, it's a binary bit stream coming in over the USB port. But I understood that you wanted to switch bias voltages whenever the audio went below some threshold value. Now I understand that's not accurate. You want to switch off the bias when there is no input stream. Very different requirement.

Ok, we do have a common understanding of load lines, so that's good. Somehow, I'm not sure we agree on tube design yet. In my experience, a tube that is biased per its operating specs will not self destruct. Is that not your experience? I've worked with tubes off and on since 1968, and I've never come across a circuit to switch grid voltages like you propose. Which doesn't mean it isn't a good idea, but understanding the need often helps with the solution.

At this point, I propose the following for a solution, no matter if it's a good idea or not, it's your idea to pursue.
You need two things. First, a detector to know if there is incoming audio to the DAC. Easy. Second, a multiplexer that can deal with higher voltages. A little trickier, but doable. Agree?

 

. You want to switch off the bias when there is no input stream.

I want to switch on the the bias when there is not input. So like if a track ends and no signal is going to grid, the tube wont die on me.

In my experience, a tube that is biased per its operating specs will not self destruct.

It really just depends on the tube, for example the 6080 tubes in my current amplifier are only dissipating about 7.5 watts of their 11 watt maximum even with no bias, but some tubes will glow red as soon as there is no bias applied which is known as the "death glow".

 

I've never come across a circuit to switch grid voltages like you propose.

I have found that tube innovation has remained stagnant since the 60s.
I read that it has something to do with designers being afraid to change from tried and tested designs because buyers were afraid to buy new designs for some reason. Tubes nowadays are a lost memory so not many people are left to innovate to begin with.

 

At this point, I propose the following for a solution, no matter if it's a good idea or not, it's your idea to pursue.
You need two things. First, a detector to know if there is incoming audio to the DAC. Easy. Second, a multiplexer that can deal with higher voltages. A little trickier, but doable. Agree?

Ummm sure, I have no idea what a multiplexer is, my experience is limited to analog electronics at the moment which my brain is already overloaded trying to figure out

 

Okay, a multiplexer is just a digital switch, that seems alright for now. I question whether it affects the sound like an opamp or transistor but if it does I can always use it to switch the plate on and off rather than switch between the signal.

 

Ok, let's start with the easy part, the detector. Do you know how to use a 555 timer chip? It's mostly an analog circuit so I'm starting there first. And it's a good fit for what we want to do here. Google "555 timer circuits". We need a timer circuit that can detect a change on the D+ line of the USB four wire connector. You can tap the Cable, the DAC, or the computer to get the three wires you need, which of course are Vcc, GND, and D+. When data starts flowing, the change from logic low to high on the D+ line will trigger the 555 circuit output to go from low to high. You'll set the timing of the output pulse to whatever amount of time you want that will protect the tube, but constrained by the longest gap that might occur in the incoming bit stream. For example, let's assume that the bit stream will have gaps of no more than 0.5 msec in a continuous piece of music. Maybe it's actually zero, I don't know, but that's what the oscilloscope can tell you. So you decide that if there's no digital signal for a full millisecond, you want to switch the bias voltage. Can you handle this thus far?

 

Yeah, give me a bit of time to research it and get my oscilloscope.