LM1894 DNR peak detector - Full-wave or not?

Hi all,

Has anybody around here used this chip enough to know if the peak detector is full wave or half-wave?

(Yes I know this is an old part. And I have read the datasheet.)

Here's a link to the data sheet. It's the pictoral of figure 11 that made me wonder if if was single ended, although that would seem kind of slack for a device that is supposed to have less than a 1 ms. attack time.

http://www.ti.com/lit/ds/symlink/lm1894.pdf

Anybody know???

Thanks

mr coffee

Its shown in the data sheet as half wave.

It only has to detect high-frequency program.

And it is no-way a precision device.

So I take  it the answer is

no, not full wave.

Thanks

mr coffee

They specify the attack time at 500 uSec as measured with a 10kHz tone burst, missing one half cycle at 10kHz is not significant wrt 500 uSec .

=====

Back in the 90s I designed a high performance version (Peavey AMR SENR) using a DBX VCA but in subtractive mode... (the hiss filter was subtracting out low level hiss). This meant that HF transients would not have to wait for the filter to open up, but would be passed mostly intact since the subtraction circuit couldn't subtract out a large HF transient.

These do not make very natural sounding single ended noise reduction, but using them in combination with downward expansion helps...  Combining gentle sliding LPF with modest downward expansion, sounded better IMO (these combo approaches were made popular by Dynafex and other old studio products). 

JR

PS: Burwen pioneered single ended NR in the 70s.

Thanks all. Good ideas, too.

mr coffee