Quad Eight 712

[Gold]

I've had these raw QE712 modules for a while. I haven't been able to get flat frequency response through them. I never had a test I thought was good so I was unsure of my results. I now have a good test and something is wrong.

http://img.photobucket.com/albums/v357/plushmore/QE712.jpg

I have the 5k trimmers set to 2k5 and have 25 turn 10k trimmers set to approx 5k. This seems like a good place to start for a "0" setting. I am using an HP oscillator and an HP 400E for measurement. Measuring at the frequency points: 68Hz=0dB, 160Hz=0dB, 400Hz= -3dB, 1K= -8dB, 2k2= -11dB, 4k5= -12.5dB, 8k2= -13dB

Playing with the 5K trimmers doesn't help much.

Any ideas on what is going on?

 

[PRR]

The schematic is missing a ground at the bottom of the left capacitor group. I don't think that explains your result.

The circuit is two parts. What oddity do you get at the top of R16, where they meet?

Disconnect all the 10K pot wipers. Short one wiper to ground with a 300Ω resistor. That should give a flat response (if not, the amps are sick). Turning the 10K pot should raise/lower wide-band gain. Trimming the 5K pots should set the max/min gain: about 7-8dB with 2.5K trim, about 20dB with trim shorted. Try all the 10K pots this way to be sure you wired them right.

Offhand, I'd bet a penny that there is a 0.1uFd cap from the right end of R3 or R15 to ground, instead of to a pot. i.e.: check all wiring, and have a chum triple-check your work.

 

[bovox]

If I see it right there is a Lowpass filter ( odd) from output series resistor
27 ohms via 0,015 uF + 10 ohms to ground. That explains your decreasing
Hf responce. Cheers Bo

 

[Gold]

Thanks for the help. This board is is a life saver. I'll do my homework and report back to the class.

 

[PRR]

> 27 ohms via 0,015 uF + 10 ohms to ground.

That's down at 400KHz; not Gold's problem (unless he mis-read a part).

It is a moderately routine compensation technique; you see it on most loudspeaker power amp chips, often as 0.1uFd and 10Ω. It is ugly (it forces the amp to strain) but sometimes necessary, and does swamp any resonance in the load.

 

[Gold]

Umm. Err. Uhh. I figured it out. The new oscillator isn't so good. It's an HP465A with selectable output impedance. Even with it terminated I'm not getting flat frequency resopnse. At least it was free. I do appreciate the help and I have had similar problems with the QE in the past under different conditions. I try to do my homework before I ask questions.

 

[PRR]

> The new oscillator isn't so good. It's an HP465A

HP465A is an amplifier, not an oscillator.....

And never trust your oscillator or your voltmeter until you test them face-to-face and get "flat" or near-flat. (Yes, it would be possible for them to have complementary error, but unlikely.)

For non-radical EQ like this, use a cheap function generator. The sine isn't really sine, but the error in EQ response curves are tenth-dB, and a function generator is a lot cheaper and more stable than a wein-bridge.

I'll be danged. The Global Specialties 2001A is still around. New face, price risen 50% in 25 years, but clearly the same machine. Problem is: Mouser is showing no stock and 5 week lead time. Has something replaced it as the basic audio signal source?

There are the 40-step testers like BK 3001, $80.

Then again you can pick up HP 200 weins for $5-$50 "as-is", they look real cool, and they usually work. Your worst fear is that some jerk has been inside "calibrating" and screwed things up. The usual mis-calibration (assuming no real jerk) is a droop/rise at the top of each band. The safest thing to do is measure it and subtract it from readings.