Can a chorus effect be built using only discrete components?

[kimelopidaer]

I was also wondering about other examples. For example, has someone
demonstrated a 'discrete vibrato' or phaser?



It would be rad if there was a section of the forum that dealt just with discrete component designs
or is there?

thanks
K

[R.G.]

First, anything that can be built with an IC could, in theory, be built in discrete components

=> if you're willing to use enough of them and pour in enough money, time, and equipment <=

[/b]
to make it work.

I have a couple of times did a rough design of an all-tube digital delay. It requires several thousand tubes and megawatts of power.  Needless to say, I didn't build one.

Some fast logic chips cannot be built from discrete devices and work at full speed because the speed of an electric field in copper ( about 1nS per foot ) is too slow to keep all the timing coherent across circuits more than a few millimeters across. They worry about picoseconds of clock skew on fast micro processors these days, and these a significant fraction of a billion MOS transistors. As a practical matter, no, you can't build a discrete Pentium.

So, for the actual questions:
Can you build a discrete chorus? Yes. However, it's impractical in terms of size and complexity because of the need for sampled signals and time delays.
Can you build a discrete phaser? Yes. The Univibe is one of these.
Can you build a discrete vibrato? Yes. The Univibe is one of these.

What others?

[PRR]

> an all-tube digital delay

Stretching "all tube".... there were some. Though not for audio.

You have a stream of a few hundred bits you want again later, and maybe over and over. (Inputs to a gunnery table maybe.) Fill a bucket with mercury. Mount a crystal to beam "sound" into the bucket and bounce it off the bottom. Pick up the return, amplify and clean it. Ta-da! You got your hundreds of bits back with only a few tubes.

What is a chorus? A wobbly delay? Then I guess you mount the crystal on a cam-lever on a 300RPM motor.

I know the speed of sound in Mercury is far too high to get audio-useful delay with any amount of Mercury which should be legal in one county. Air-delay would be better. Yes there were echo/reverbs made from a garden hose. If you want chorus, add a trombone slide.

[iccaros]

to be Fair R.G.
Creating a TUBE CPU was done, and the Russians had flight control like the SAU-155m, Which when we pulled apart the first An-124 Condor Cargo Plane the US got a hold of had a variant of that system, that was all tube, that could beet the pants off a Pentium at flight control.. but it was limited to just that role.....


but there was the EDVAC
only 6,000 tubes, 12,000 diodes and used 56KW of power.. But was a complete computer system as powerful as a $1 calculator now..  Well maybe it was not that fast or could do the higher math a $1 calculator can do now..

[Mark Hammer]

First, anything that can be built with an IC could, in theory, be built in discrete components

=> if you're willing to use enough of them and pour in enough money, time, and equipment <=

[/b]
to make it work.

I have a couple of times did a rough design of an all-tube digital delay. It requires several thousand tubes and megawatts of power.  Needless to say, I didn't build one.

Some fast logic chips cannot be built from discrete devices and work at full speed because the speed of an electric field in copper ( about 1nS per foot ) is too slow to keep all the timing coherent across circuits more than a few millimeters across. They worry about picoseconds of clock skew on fast micro processors these days, and these a significant fraction of a billion MOS transistors. As a practical matter, no, you can't build a discrete Pentium.

So, for the actual questions:
Can you build a discrete chorus? Yes. However, it's impractical in terms of size and complexity because of the need for sampled signals and time delays.
Can you build a discrete phaser? Yes. The Univibe is one of these.
Can you build a discrete vibrato? Yes. The Univibe is one of these.

What others?

And just so everyone is clear, when you see the switch marked "chorus/vibrato" on a UNivibe or workalike, the "chorus' setting is NOT what frequently/generally gets called a "chorus".  It is the swishy phaser-type effect we associate with a Univibe.

[george]

First, anything that can be built with an IC could, in theory, be built in discrete components

=> if you're willing to use enough of them and pour in enough money, time, and equipment <=

[/b]
to make it work.

I have a couple of times did a rough design of an all-tube digital delay. It requires several thousand tubes and megawatts of power.  Needless to say, I didn't build one.

Some fast logic chips cannot be built from discrete devices and work at full speed because the speed of an electric field in copper ( about 1nS per foot ) is too slow to keep all the timing coherent across circuits more than a few millimeters across. They worry about picoseconds of clock skew on fast micro processors these days, and these a significant fraction of a billion MOS transistors. As a practical matter, no, you can't build a discrete Pentium.

So, for the actual questions:
Can you build a discrete chorus? Yes. However, it's impractical in terms of size and complexity because of the need for sampled signals and time delays.
Can you build a discrete phaser? Yes. The Univibe is one of these.
Can you build a discrete vibrato? Yes. The Univibe is one of these.

What others?

And just so everyone is clear, when you see the switch marked "chorus/vibrato" on a UNivibe or workalike, the "chorus' setting is NOT what frequently/generally gets called a "chorus".  It is the swishy phaser-type effect we associate with a Univibe.

oh dammit Mark I was going to say "AND a unvibe has a chorus switch, so there ya go!" .... and you got in first.

[R.G.]

> an all-tube digital delay
Stretching "all tube".... there were some. Though not for audio.

Yeah. for fun, though, I did the by-the-book version. One duotriode per bit, sixteen bits per sample. 44k samples per second for CD quality (or, pick your quality here), so a one second (or pick your amount of delay here) gives 1*16*44,000 = 704,000 duotriodes. This is so big that you can either add the surrounding logic or just ignore it as trivial.

A duotriode used for logic usually has about 2ma flowing when it's on, none when it's off, and on average half the tubes will be on at any one time. So you need 352,000 times 2ma, or 704 amps (1) of plate supply. We could drop that down to maybe 150V for logic, so the plate power is 704*150V = 105.6kW. The heaters all run all the time, and a duotriode as was actually used for logic uses 150ma at 12V, or 1.8W/tube, times 704,000 = 1.267MW. Ooops, I just said the "M" word. 

It gets worse if you really try to build one. All that power going in comes out as heat. You have to keep them cool, so you have to take the heat out by letting the tubes heat air, then cooling the air. At best, one could build this outdoors and let the sun/planet power the cooling. But sometimes it rains. Us humans would probably put it indoors to keep the bugs and dust out of it, so we'd use blowers and probably air conditioning to pump the heat out. About the best AC ratio is 20:1, so add another 5% power to do the heat pumping.

It's probably best to locate this thing near a hydroelectric power station... 

[Fender3D]

lol
wouldn't be easier put the amp in a room then running with wireless guitar from that room to another and back.
Usain Bolt would play his guitar with "natural" chorus in less than 1 sec. each 10m travel.  

[garcho]

to be Fair R.G.
Creating a TUBE CPU was done, and the Russians had flight control like the SAU-155m, Which when we pulled apart the first An-124 Condor Cargo Plane the US got a hold of had a variant of that system, that was all tube, that could beet the pants off a Pentium at flight control.. Smiley but it was limited to just that role.....

I heard that same story (from an audio electrical engineer, of course - who else would be talking about tubes in 2011?), but I heard that the use of tubes had something to do with the ability to fly post-nuclear attack. Chips don't like fallout? Are tubes in use at nuclear power stations? Some 12AX7's hooked up to a flashlight? 

[Earthscum]

As I understood it, the tubes were being used not for post nuke, but because tubes can withstand the EMP a lot better than silicon components. EMPs will fry a modern computer, but tubes are a whole lot tougher in that area (at least thats what I've been told from the old dogs in the military).

[Mark Hammer]

but there was the EDVAC
only 6,000 tubes, 12,000 diodes and used 56KW of power.. But was a complete computer system as powerful as a $1 calculator now..  Well maybe it was not that fast or could do the higher math a $1 calculator can do now..

When I think about that, I can't help but think about this: http://www.youtube.com/watch?v=mDaPHVcUdS8&feature=related
The good part is around the 2-minute mark.  Know what I mean?

[iccaros]

but there was the EDVAC
only 6,000 tubes, 12,000 diodes and used 56KW of power.. But was a complete computer system as powerful as a $1 calculator now..  Well maybe it was not that fast or could do the higher math a $1 calculator can do now..

When I think about that, I can't help but think about this: http://www.youtube.com/watch?v=mDaPHVcUdS8&feature=related
The good part is around the 2-minute mark.  Know what I mean?

Like starter turnover sound track for a "Electric car"...

[iccaros]

As I understood it, the tubes were being used not for post nuke, but because tubes can withstand the EMP a lot better than silicon components. EMPs will fry a modern computer, but tubes are a whole lot tougher in that area (at least thats what I've been told from the old dogs in the military).

AS tubes are Thematic devices, they do not rely on junction connections, they use heat over a vacuum to allow electrons to move, like how heat in a Jacobs ladder allow the arc to rise.
"Many people have bought or kept old vacuum tube radios for use after an EMP attack.  Although vacuum tubes are thousands of times more resistant to EMP than transistors (and discrete transistors are much more resistant than integrated circuits), other components of vacuum tubes radios can be damaged by EMP" source --> http://www.futurescience.com/emp/emp-protection.html

To be fair the Condor had up to 32 different computer systems and the SAU-155m was the backup, and came from the MIG-31. I am not sure they ever used it. I got the change to see one, we conducted Joint Airborne operations over FT Brag, Great jump from that plane, a little different than a C5, but the Russians were nuts.. Safety standards were no where near the same as they are for US airborne operations. We dropped 1,500 paratroopers on 5 drop zones.  We were not awarded Russian Jump wings, but I was giving a pair by my Russian jump buddy, in exchange for US wings..

but I remember looking at the computer configuration and noticing that the Concord was designed to be functional over a range of emergency situations.  But at the time I had never seen CAT Eye tubes, but they had them in the radio systems to detect radio waves and give the navigator the ability to triangulate based on known radio transmissions. Now that was something interesting to a Radio Repair tech.

[teemuk]

Is it considered discrete if you replace the solid-state delay element with some of those electromechanic delay devices, e.g. tape loop or an oil can delay?

[R.G.]

Problem is, chorus in particular is a short delay, and must be varied a bit.

As to discrete with an electromechanical delay? Depends entirely on what you think. In the words of the immortal William Jefferson Clinton, testifying under oath, "it depends on what your definition of "is" is."

[teemuk]

Chorus and flange effects require very short delay times (1 to 30 milliseconds) that you can't normally do with a tape machine. Luckily, a multi-track tape machine can use its record head for playback (essential during overdubbing to keep the tracks in sync). To create a chorus or flange effect, the play head was routed to the mixing console, and the record head (which plays the track early) was routed to a second tape machine running at the same speed. By slightly varying the speed of the second machine, chorus, flange, or vibrato effects were created. All the effects on Beatles records were done this way.

[Perrow]

The heaters all run all the time, and a duotriode as was actually used for logic uses 150ma at 12V, or 1.8W/tube, times 704,000 = 1.267MW. Ooops, I just said the "M" word. 

It gets worse if you really try to build one. All that power going in comes out as heat. You have to keep them cool, so you have to take the heat out by letting the tubes heat air, then cooling the air. At best, one could build this outdoors and let the sun/planet power the cooling. But sometimes it rains. Us humans would probably put it indoors to keep the bugs and dust out of it, so we'd use blowers and probably air conditioning to pump the heat out. About the best AC ratio is 20:1, so add another 5% power to do the heat pumping.

It's probably best to locate this thing near a hydroelectric power station... 

But wouldn't the heat mean that you could use less energy to heat the tubes. I understand we need to keep certain things cool but we do actually want the tubes to be warm. I admit I know little enough about tubes to know if the heating inside the tube serves some other purpose than just keeping the tube warm.

[R.G.]

Chorus and flange effects require very short delay times (1 to 30 milliseconds) that you can't normally do with a tape machine. Luckily, a multi-track tape machine can use its record head for playback (essential during overdubbing to keep the tracks in sync). To create a chorus or flange effect, the play head was routed to the mixing console, and the record head (which plays the track early) was routed to a second tape machine running at the same speed. By slightly varying the speed of the second machine, chorus, flange, or vibrato effects were created. All the effects on Beatles records were done this way.

Yep. However, the LFO for the chorus was someone's thumb on the supply reel, slowing down or letting the reel catch up to speed.

A person's thumb is kind of an ultimate "discrete" LFO. 

And again, in the words of the immortal William Jefferson Clinton, testifying under oath, "it depends on what your definition of "is" is."

But wouldn't the heat mean that you could use less energy to heat the tubes. I understand we need to keep certain things cool but we do actually want the tubes to be warm. I admit I know little enough about tubes to know if the heating inside the tube serves some other purpose than just keeping the tube warm.

The heater is not just to keep the tubes warm. Tubes are called "thermionic" because the heat from the heater is what excites electrons in the cathode to jump into the vacuum so the tube can work at all. No heater, no electron emission, and no electronic action.

Note that there do exist some highly specialized tubes which can run with cold cathodes. These are very much the exception; all normal amplifying tubes use heated cathodes to get electrons to work with.

[PRR]

> we do actually want the tubes to be warm.

We want a very small area (cathode) to be red hot. We need concentrated heat just in that area. All else as cool as possible.

No, you can't reasonably recycle the overall heat back to the cathodes. That would mean pumping heat uphill. That's not impossible, we pump heat into houses from cool air. But it is still significant work, And concentrating that heat at red-hot to the myriad of cathodes would be a plumber's nightmare.

I could suggest some slight reductions in R.G.'s estimates. Plate current could be lower. However in most small-signal tubes the plate power is already much smaller than heater power. There are tubes which use heater power better, however the experts (tube-computer builders) didn't use them. Probably a reliability issue.

The Russians continued developing vacuum tubes after the West went into their transistor orgy. Filimentary rod-grid tubes could compete with early Germanium transistors for overall power consumption.

Reliability is really the larger problem. One tube may run 10,000 hours. With 10,000 tubes, the long-term average failure is one per hour. Indeed the early VT computers tended to run about 40 minutes on average. Women with shopping-carts traced and replaced the dead bottles and re-started the computation. Relay computers would run for months non-stop. However the VT machine got more done in 40 minutes than the relay machines did in a month.

"Flanging" is specifically a tape-trick.... run dubbed tapes on two machines mixed together then drag your finger on one machine's reel _flange_ to slip the sound one tape against the other. Any other "flanger" is a cheap (but more practical) fake.

Reverb is a large empty room with speaker and mike. Very low-tech (but high-rent). Echo can be a long room with hard ends. Or of course another head on the tape machine. (In Wagner, echo is a horn far off-stage repeating the horn in the pit.)

Chorus is many singers, or one Mary Ford singing the same part over and overdub.

[petemoore]

  ..Gas' powered tubes.
  We got those too, even a candle can become a triode.