Page 5 of 8

Re: Signal generator output ranges

Posted: Sat Mar 16, 2013 7:10 pm
by juniorhifikit
I breadboarded it using NTE375 & NTE398 transistors (all I had available). With +/-35 on the transistors, and an equal 10K/10K voltage divider, I had around +/-16.5V on the op amp rails. I could only get about 26.7dBu output before it does this:

[img]deleted[/img]

And the rails do this:

[img]deleted[/img]

I tried several variations of gain and input level, but always got the same results. Not close enough to 30dBu like I need. Haven't had time to look into it any further, but plan to as soon as busy work dies down.

Re: Signal generator output ranges

Posted: Sat Mar 16, 2013 7:14 pm
by mediatechnology
With +/-35 on the transistors, and an equal 10K/10K voltage divider, I had around +/-16.5V on the op amp rails.
Looks like approx. (35/2)-Vbe.
Very similar to what I found.

I think you'll need +/-45V to get to +30 dBu.

I'd modify the reistor values to get +/-20V on the 5534.

Re: Signal generator output ranges

Posted: Sat Mar 16, 2013 10:27 pm
by juniorhifikit
mediatechnology wrote:
I think you'll need +/-45V to get to +30 dBu.
Haven't got that much available. It's kind of a one-off test situation, so I'm not going to build a separate power supply just for this. I did order some samples of the LTC6090 though...

Re: Signal generator output ranges

Posted: Sun Mar 17, 2013 7:29 pm
by ricardo
30dBu is 34.64Vp.

34.64Vp is also 75W rms into 8R .. ie not a small power amp but a fairly big one.

Re: Signal generator output ranges

Posted: Sun Mar 17, 2013 11:05 pm
by JR.
ricardo wrote:30dBu is 34.64Vp.

34.64Vp is also 75W rms into 8R .. ie not a small power amp but a fairly big one.
While this may all be relative... You can make a class AB 75W audio amp with a single pair of TO-3 power transistors (on a real heat sink). So not a "big" amp in sound reinforcement terms, perhaps getting up there for a tube guitar amp.

FWIW modern SR amps are routinely outrunning the mains cord's ability to supply enough power, even with class D efficiency so several kW,

JR

PS: Hi rico is summer fading yet?

Re: Signal generator output ranges

Posted: Mon Mar 18, 2013 12:00 am
by juniorhifikit
Well, I'm not going to be driving a speaker, just a high impedance op amp input.

Re: Signal generator output ranges

Posted: Mon Mar 18, 2013 4:57 am
by mediatechnology
30dBu is 34.64Vp.
And you'll need a Vcc/Vee that's at least 3-5V higher on each polarity to get 35Vp output.
With a bootstrap transistor maybe 6-10V per polarity.
The 40-45V rails I mentioned earlier aren't that unrealistic.

The LTC6090 doesn't have that great of distortion vs. frequency performance.
A power amp IC would likely provide better performance than the LTC6090.
Or, for lower output currents reconsider the bootstrapped op amp.
It should perfrom as good as an unbootstrapped 5534.

Re: Signal generator output ranges

Posted: Mon Mar 18, 2013 10:17 am
by juniorhifikit
mediatechnology wrote:
30dBu is 34.64Vp.
And you'll need a Vcc/Vee that's at least 3-5V higher on each polarity to get 35Vp output.
With a bootstrap transistor maybe 6-10V per polarity.
The 40-45V rails I mentioned earlier aren't that unrealistic.

The LTC6090 doesn't have that great of distortion vs. frequency performance.
A power amp IC would likely provide better performance than the LTC6090.
Or, for lower output currents reconsider the bootstrapped op amp.
It should perfrom as good as an unbootstrapped 5534.
I see that you're right. I was hoping to get the output closer to the +/- 35V rails I have available, so I wouldn't have to build a new power supply. Can't change the laws of physics...

Re: Signal generator output ranges

Posted: Mon Mar 18, 2013 10:54 am
by JR.
juniorhifikit wrote:
mediatechnology wrote:
30dBu is 34.64Vp.
And you'll need a Vcc/Vee that's at least 3-5V higher on each polarity to get 35Vp output.
With a bootstrap transistor maybe 6-10V per polarity.
The 40-45V rails I mentioned earlier aren't that unrealistic.

The LTC6090 doesn't have that great of distortion vs. frequency performance.
A power amp IC would likely provide better performance than the LTC6090.
Or, for lower output currents reconsider the bootstrapped op amp.
It should perfrom as good as an unbootstrapped 5534.
I see that you're right. I was hoping to get the output closer to the +/- 35V rails I have available, so I wouldn't have to build a new power supply. Can't change the laws of physics...
The laws of physics don't change but different design topologies behave differently. The common emitter topology like used in the chicken scratch schematic I posted will pull pretty close to the rails, limited by the saturation voltage of the output devices and the emitter resistors.

The common emitter topology is a little harder to stabilize because the driver stage has voltage gain, but some voltage gain is useful for that application.

Decades ago I made a hifi amp for my younger sister using a variant on feeding the opamp PS rails into the bases of common emitter stages. Simply imagine zener diodes between the power device bases and the opamp power pins. This way the opamp voltage is happy and outputs swing near the rails (opamp output drives into ground to send current out the ps pins.So input polarity is reversed for negative feedback connections).

There are many ways to skin this cat and get close to the rails. Some are harder than others to stabilize and set nominal class A current in. For a low distortion buffer I'd bias it to run a little hot so there is no crossover distortion to speaks of (near class A).

JR

Re: Signal generator output ranges

Posted: Fri Mar 29, 2013 8:02 pm
by juniorhifikit
Just wanted to report back on where I ended up on this topic. My maximum supply voltage for my test was +/- 35V, and the swinging op amp couldn't get my output close enough to the rails to approach +30dBu. I ended up breadboarding a booster circuit using the LTC6090 on an adaptor board (it's SOIC), which got me right to +30dBu (24.5Vrms) before clipping - exactly where I wanted to be. Driving the super high impedance input of the OPA134 circuit I was testing, the LTC6090 didn't heat up at all.