any experiences with power ota's?

[raymond.shpeley]

Has anyone here had experiences with power operational transconductance amps for low and high frequency, high efficiency drivers?

I'm looking for a range of information from impedance and crossover effects to definition and clarity opinions. I have already researched the web on products like First Watt so I don't need that kind of info. What I'm looking for are experimental results.

Ray

[raymond.shpeley]

OK. Let's expand on this a little bit. A power OTA is a power amp with a high output impedance. It can be simulated to some extent by placing a resistor in series with the driver from the output of a standard power amp.

Normally this is how one generates an impedance curve for a driver. Altec had an article on damping and the effect on sound reproduction. I've noticed others here in the forum talking about inserting a resistor in series with a driver to modify damping characteristics.

I'm wondering if anyone here has used higher resistance values (active or passive) and what kind of results they've obtained.

[donaldpatten]

Shure Models VA300 & PM300 are current amplifiers. The Vocal Master is a 6 channel mic mixer,with eq and mono power amp. Intended as a portable volcal sound system .It was used by The 5th Dimension , Conway & Loretta, The Association ,...The Power Master is the mono power amp only. The power output is 100 watts to 8 Ohms with 1 volt input ; or 3.5 amperes out for 1 volt input. These products were very popular around 1970. The reasons for the current design (1 V in 3.5 amps out ) was that it provided equalization to the matching speakers and a extreemly reliable solid state , truly a first. Rolling Stone mag quoted that the Vocal Master system was used for the Sermon on the Mount.
The other amps in there hay day would destroy themselves when the output was shorted. Try to plugin a 1/4" phone jack at a speaker and not short it . The Voltage amps would see a short and over drive since they just lost their feedback siganl. In doing so, they would burn out their output transistors . And the SHOW dose not go on ! All VA & PM 300s were tested for one hour at 132 AC line with a dead short then a 12 hour "over driven" music power test with a simulated speaker load ( not just resistors). No DOAs when the customer opned the box.
Does a current amp sound different yes . It does not "Block" when over driven 10 to 20 dB ( 10 Volts in ). My HK Citation II tube power amp would shut down would when driven to clipping. On many occation I would put a Scope on a speaker line at a concert all I would see is square waves. As the man says "Loud is Beautiful" . Its frequency resp curve follows the speaker Z ( impedance ) curve. You could probalby find some out there.

[Richard C.]

Hello all,
I just went out to the Shure web site and downloaded the schematics in PDF, of both these amps. I see that the designers inserted a .12 ohm resistor in series with the output to the 4000uf output cap. They then pick off the voltage from the resistor, use an isolation transformer, and apply negative feedback (in series with the "normal" negative feedback), to the 2nd gain stage. Nice way to measure output current! I could see where output transistor heat dissipation could be limited in this way - even with a dead short!

Do you know of any more amps that may use this technique, or is this common on musicians amps?

It seems simple at first, but it might require a decent transformer to minimize phase problems with negative feedback across the passband.

Richard C.

[donaldpatten]

Richard
The heat sinks have thermal breakers to limit the max safe transistor temperature to 50 % of rated power. This shutdown the power supply and was tested in all units for one hour ; they would cycle off and on probably 20 times. I am not aware of any other production amplifier using current feed back; And yes, the feedback transformer is special.
Don

[raymond.shpeley]

According to this power/impedance diagram from the PM300 pdf there's only a small amount of current feedback used.

<img src=http://www3.telus.net/bayleafs/bayleafs/power-s.jpg width="400" height="358" />

I've added a red line to show constant voltage and a green line to show constant current curves. As you can see the PM300 is only slightly above the constant voltage slope. The emitter follower output stage impedance is already very low and apparently there's not enough loop gain for the current feedback to make a large difference.

A power OTA would be very close to the slope of the green line.

[donaldpatten]

Hi
This curve shows the max power out at 5 % THD.
The max Voltage out is 30 rms the red line shows this.
At 8 Ohms : 30 V =112 watts
16 Ohms 33V = 68 watts
These are limited by the power supply Voltage ( 94 Vdc )

At 7 Ohms and below the limit is output transistor current
As shown by the green line . The max transistor current is 3 Amperes
Since we have parallel pairs. The max amplifier rms output
Current is : 2*3*0.707= 4.24 amps

6 Ohms 4.24*4.24*6 = 108 watts

4 Ohms : 4.24*4.24*4 = 72 watts

The above has no effect on current or voltage feedback.

To see the effect run a frequency response with an 8 ohm dummy resistive load and then switch to 16 or 4 . The amp Voltage gain will change. Or run a response curve with real speakers . The low end will show an uphill bump between 50 Hz to 200 Hz. Also there will be a slow rise from 300 Hz out to 10kHz.
Don

[donaldpatten]

Hi
This curve shows the max power out at 5 % THD.
The max Voltage out is 30 rms the red line shows this.
At 8 Ohms : 30 V =112 watts
16 Ohms 33V = 68 watts
These are limited by the power supply Voltage ( 94 Vdc )

At 7 Ohms and below the limit is output transistor current
As shown by the green line . The max transistor current is 3 Amperes
Since we have parallel pairs. The max amplifier rms output
Current is : 2*3*0.707= 4.24 amps

6 Ohms 4.24*4.24*6 = 108 watts

4 Ohms : 4.24*4.24*4 = 72 watts

The above has no effect on current or voltage feedback.

To see the effect run a frequency response with an 8 ohm dummy resistive load and then switch to 16 or 4 . The amp Voltage gain will change. Or run a response curve with real speakers . The low end will show an uphill bump between 50 Hz to 200 Hz. Also there will be a slow rise from 300 Hz out to 10kHz.
Don

[raymond.shpeley]

I agree that this kind of curve makes it pretty difficult to predict amp behaviour below clipping, but it does give some hints as to what may be happening.

Sure I think it would be a good idea to set up dummy loads and see what power is delivered for each load. I presume you would prefer to do this at less than max power to get an idea of actual amp output impedance without voltage and current limiting coming into play. That is of course what the 5% distortion curve shows.

Your calculations show a 3v rms increase in voltage at 16 ohms (33v) from 8 ohms (30v). Interpolating this back to zero load gives 36.6v rms or a power supply voltage of 103.5v without accounting for voltage drops in the emitter resistors or output transistors. Let's add about 1.5v (I'm being gracious here) to give 105v for power supply voltage. Not quite the 94v mentioned. The bias current in the output stage would have to be pretty high to load the supply down to 94v from 105v.

So if I can assume that not all of the output resistance shown in the curve is due to power supply resistance it would be safe to say that some of it is due to current feedback. And if some is due to current feedback how is it limited if not limited by the power supply resistance? Perhaps by the amount of loop gain available in the amp?

I wrote my previous post basically falling asleep at the keyboard so I kept it short. What I didn't mention was the patent I had read on this type of circuit a couple of months back. I don't have it available right now but I do remember some of the claims made and why it was an improvement on earlier designs. Basically for this kind of amp to be stable with transformer-coupled current feedback the loop gain must be kept low and consequently the output impedance of the amp is limited. Transients in particular will suffer.

Compared to class A direct coupled OTA designs this amp design has many limitations. One of these is the problem of the power/resistance curve from my previous post. At mid-power levels you can attain a semblance of constant current operation which disappears as you approach full power.

You mentioned, "To see the effect run a frequency response with an 8 ohm dummy resistive load and then switch to 16 or 4 . The amp Voltage gain will change. Or run a response curve with real speakers . The low end will show an uphill bump between 50 Hz to 200 Hz. Also there will be a slow rise from 300 Hz out to 10kHz."

This is what is to be expected from a high output resistance amplifier given either resistive or reactive loading conditions. Are you saying you measured this on the Shure?

An additional problem I have with believing the Shure PM300 was a full-blown current output amp is that their literature didn't mention the changes needed to crossovers. Basically a crossover designed for a voltage regulated amp will not work properly with a current regulated amp.

Tell you what. Run some curves and prove me wrong.

Ray

[Richard C.]

I don't have any experience with power OTA's, and the way this thread is going; I don't think I want to know any more than I have seen here.

Richard C.