opamp driving tubes

[skyraider]

I know i will get bashed if 'opamp' is mentioned in here.. I know you guys never listen to those plastic chips...

Anyway, I want to open a discussion about the possibility of using modern opamps to drive output tubes, especially in SE design. The advantages of opamp in ultra low distortion and noise is a strong advantage from technical aspect. Also, if the opamp is powered with +/-rails, no coupling cap is required. This is useful for people who cant live with the idea of having caps in their signal path. Furthermore, 'capless' designs tend to be expensive because more iron is need.
I am pretty sure this idea has been tested so I want to ask for opinions from those who has tried it.

I am thinking of using OP627 to drive 2a3. One of my concern is the output current requirement to drive output tube. I am not EE, but I think this should be easilly solved with a BUF634 buffer. Both of these chips has been used succesfully on SS preamp designs.

Suggestions?

 

[EC8010]

Op-amps generally run off +/-15V rails. A 2A3 needs about 45Vpk to drive it, whereas your op-amp can probably swing to 13Vpk. It's not current that's the problem, it's voltage.

 

[pinkmouse]

Sounds like a job for a chipamp!

 

[Brett]

Sounds like a job for a chipamp!

So then, what's the point of the output tubes and transformer after it?

 

[analog_sa]

This is useful for people who cant live with the idea of having caps in their signal path.

Ah, yes, an opamp as a better alternative to caps, i like this

Once your design is complete you may further improve it by replacing the tubes with transistors and getting rid of the output iron.

 

[pinkmouse]

Once your design is complete you may further improve it by replacing the tubes with transistors and getting rid of the output iron.

Now that might just catch on.

 

[SY]

Or you could use one of those hideously expensive Apex high voltage opamps. Compared to the usual tube driver stage, you'll reduce the distortion of the complete amplifier from 6% to 5.999%. And expensive is always better.

Sorry, don't mean to be facetious, but there's not much logic to doing this. Compared to the worst coupling capacitor ever made since the days of Leyden, the distortion of the SET output tube/transformer is massively larger.

 

[Sch3mat1c]

So then, what's the point of the output tubes and transformer after it?

Placeholders on the chassis.

So yeah if you must use SS to drive tubes... use a discrete driver that can handle the voltage required.

Note that this will result in the exact SS analog of the tube driver stage you are replacing.

Tim

 

[alan brando]

adjust the feedback values to obtain the desired gain and bandwidth.

the power supply rails at this example are +/- 32 volts, but depends of the opamp, you could carry them to +/- 36 volts.

if you only need 45 volts peak to peak, a good start is a power supply of the +/- 24 volts with a 5534 op amp. I think that the 5534 will work fine up to +/- 32 volts, but you need to check the voltage at pin 4 & 7.

I chose a single opamp because they have low crosstalk, remember the output voltage 45 volts peak to peak, but you can use a dual opamp with +/- 18 volts maximum supply rating

the circuit has black background, if you can't see nothing, please contact me.

 

[skyraider]

Op-amps generally run off +/-15V rails. A 2A3 needs about 45Vpk to drive it, whereas your op-amp can probably swing to 13Vpk. It's not current that's the problem, it's voltage.

Yes, that is the first limitation. I was thinking of interstage trans to solve this.. but this will limit the design to a good interstage trans, thus offset the cost reason.

Sounds like a job for a chipamp!

Yes, can be another alternative.

Ah, yes, an opamp as a better alternative to caps, i like this.

Like I said, this thread is not for anyone who cant live with opamps. Personally I find nothing offending with opamps.

adjust the feedback values to obtain the desired gain and bandwidth.

the power supply rails at this example are +/- 32 volts, but depends of the opamp, you could carry them to +/- 36 volts.

if you only need 45 volts peak to peak, a good start is a power supply of the +/- 24 volts with a 5534 op amp. I think that the 5534 will work fine up to +/- 32 volts, but you need to check the voltage at pin 4 & 7.

Some opamps can go up to 34V p2p, but those opamps does not have the sonics I would like to integrate.

 

[skyraider]

The main intention is to make use of the low distortion and noise from modern opamp into a SE tube amp.

This has advantage by looking at it this way:
1. no need for good/exotic tube drivers
2. no coupling cap is required if the opamp is powered from +/- suplies.
3. Opamp input impedance can be varied to match cd players (20k-50k?)

However, from the feedback I gathered, the opamp is limited by output voltage. Other than this, any other technical challenges?

 

[skyraider]

What I have in mind is something like this:

 

[EC8010]

By the way, when I said 45Vpk, I was using exact terminology, that's 90Vpk-pk. A 5534 can stand +/-18V rails, that's not nearly enough.

 

[audiousername]

IOne of my concern is the output current requirement to drive output tube.

It shouldn't be a problem if you're running class *₁ since the only current flowing is that charging the Miller capacitance of the output valve(s). If the opamp can output a few mA, maybe class *₂ operation may be a possibility.

This has advantage by looking at it this way:

1. no need for good/exotic tube drivers

The same can be achieved (as already suggested) by using a discrete transistor based driver stage, which then also eliminates the problem of insufficient voltage swing capability to drive the output valves

2. no coupling cap is required if the opamp is powered from +/- suplies.

This can be achieved with a valve based stage as well, it's just that the negative supply will have to be deeper.

3. Opamp input impedance can be varied to match cd players (20k-50k?)

The input impedance of a valve input stage is essentially arbitary, with the maximum being the maximum grid leak resistance allowable by the valve's datasheet, in the megohm range for some valves.

However, from the feedback I gathered, the opamp is limited by output voltage. Other than this, any other technical challenges?

I completely fail to see the logic of this approach, but if you must pursue it, consider using valves that will require less voltage swing on the grids to get to full power. If triode output is a must, triode strapped EL84 may be up your alley since they require very little voltage to get to full output power. If the power is insufficient, parallel the output valves, or go for PP operation.

OT: SY, have you been watching Iron Chef? (your sig...)

 

[Brett]

This has advantage by looking at it this way:
1. no need for good/exotic tube drivers

An EL84 in pentode will drive a 2A3 superbly (to clipping), with low distortion (look at pete Millett's tests) is common, cheap as dirt and you already have the PS for it. If you're not so purist as to contemplate an opamp, then a pentode shouldn't be outside your religious leanings

 

[SY]

OT: Jason, note the avatar. That's Dr. Hattori himself. My 4 year old knows all the chefs' names and specialties, asks me every night when I put him to bed, "Where's Kaga?" and pumps his little fist in the air and yells, "Arrrgh Kizzim!"

BOT: It's not a technical challenge, it's more a matter of "why bother?" Use a high voltage op amp; there are plenty of those around.

 

[boholm]

FYI:

In Electronics World October 1995 issue under the title "Hot audio power" Jeff Macaulay designs a 32W power amplifier with a full power bandwidth of 5 Hz to 55 kHz and 0.07% distorsion at 200W - taken from the intro of the article.

So it has already been done. His trick was to put a transistor on the output of the opamp to drive the valves.

The article can be purchased at www.electronicsworld.co.uk for £4.00.

 

[darkfenriz]

HEY GUYS

how about floating supply design????
the opamp could easily drive the cathode follower and anything.
right?

 

[Sheldon]

BOT: It's not a technical challenge, it's more a matter of "why bother?" Use a high voltage op amp; there are plenty of those around.

Sy, you have said a number of times that there is little (no) rational basis for doing much of what you do with tubes. Why a different set of rules for skyraider?

Sheldon

 

[skyraider]

Audiousername,

It shouldn't be a problem if you're running class *1 since the only current flowing is that charging the Miller capacitance of the output valve(s). If the opamp can output a few mA, maybe class *2 operation may be a possibility.

Checking the datashett for opa627+buf634, this should be able to go up to 250mA, which is more than enough to drive the interstage trans.
But may I ask what is *2 operation? I am no EE, need to learn more.

The same can be achieved (as already suggested) by using a discrete transistor based driver stage, which then also eliminates the problem of insufficient voltage swing capability to drive the output valves

Another possibility. But using transistor will have us to end up with a design similiar to normal valve driver, only that valve is replaced with transistor (as already suggested). Given the same topology, I would rather go valve instead of transistors (obviously
My idea is to combine the sonics of modern opamp.

I completely fail to see the logic of this approach, but if you must pursue it, consider using valves that will require less voltage swing on the grids to get to full power. If triode output is a must, triode strapped EL84 may be up your alley since they require very little voltage to get to full output power. If the power is insufficient, parallel the output valves, or go for PP operation.

I may have to switch to different valve which is easier to drive, but EL83 is not my cup of tea..