Trying to make noise reduction circuit with LM1894

[bancika]

Hi
I just finnished prototype of noise reductor based on LM1894 according to N.S. data sheet.
Since it's intended for stereo application I just ignored one channel, left pins open. All caps are as suggested, and for R1 and R2 I put 1K trimmer (R1+R2 should be 1K). The thing is, I can't tell the difference in noise level with and without circuit. Can someone with more expirience than me (I'm a noob  ) take a look and suggest some mods so it can be more suitable for guitar generated noise.
By the way, it's plugged into regulated 9V PS (8V is suggested, but it can work with up to 18V according to sheet)
Thanks in advance!

[bancika]

One more interesting, circuit sounds same with or without power 

[Mark Hammer]

1) If it sounds the same with and without power, my guess is that it isn't working properly.

2) Just because it is a stereo chip being used for a single channel doesn't mean you can't use both sides.  I've seen a number of instances where the two channels of such a chip were cascaded to provide double the noise reduction.  It is also worth considering use of both channels for noise reduction at different points in the signal path, such as earlier and later.

[bancika]

Thanks for reply Mark,
Good idea about two pass filtering. I will consider it, just to make single pass version work
Idea is to put in the samo box as tonemender EQ, I was planning to use that after my line out from Firefly amp.
Do you know maybe what are desired voltages so we can compare with mine?
Thanks again.

[Mark Hammer]

Unfortunately, I don't know the voltages.  I do have a project article for this chip in an old issue of Electronics Today International in my basement somewhere.  I see that the datasheets from National include a PCB layout.  Is that what you used?

[bancika]

Nope, I just followed schematic on first page of data sheet and put it on vero board. I just messed around it little more and it pisses me off. Noise level is less with DC power off  Moving trimmer affects tone just a bit, and when it's all way down sound gets muddy. Strange

[GFR]

What kind of noise are you expecting to reduce with this circuit?

It is intended to reduce high frequency hiss, like you hear from cassete tape recordings.

If you're using the standard values from the datasheet and feeding it to a guitar amp, you may be reducing noise that the amp is not reproducing anyway

Getting muddy when the sensitivity is maximum is normal - you're not only reducing noise but also signal.

[Paul Perry (Frostwave)]

The chip is designed to remove noise from hissy audio tapes, by using a couple of psychoacoustic tricks (details on the data sheet). The main thing is that it has a low pass filter, the cutoff of which is cranked according to the music spectrum. In particular, when there is no high freq music present, the high freq noise is cut.
So I wouldn't expect it to be much use for guitar.

It is possible that a Thomas Henry could pervert the device into something useful......

[Mark Hammer]

Don't be too harsh about the chip just yet.  It functions in a similar fashion to other noise-filter pedals that use dynamic filtering rather than gating.  This includes the Rocktron HUSH system, which many people like.  Of course, whether the chip itself is suitable is separate from whether the overall circuit built around it meets the needs of guitar.  Given that it is probably easier for most folks to get than an SSM2000 HUSH hip, not to mention simpler to use, I think itis worth exploring as a potential solution to some noise-reduction scenarios.

Keep in mind a couple of things:
1) The audibility of a lot of hiss will be greater with full-range multi-driver hi-fi speakers than with your typical 6khz bandwidth rock and roll speaker without any added piezo or other tweeter.  If the speaker can't reproduce it, then removing it from the signal before the signal hits the amp won't really be detectable.
2) The amount of hiss, or spectrum to be covered by the sliding filter of the 1894 may need to be different in the guitar context.  For instance, it may be perfectly fine to have a filter swing down from 20khz to, say, 4khz during quieter sections, because you'll still have plenty of bandwidth left.  However swinging down by a proportional amount in a lower range (e.g., from 6khz to 1200hz) may well muffle the tone more than you'd like.

This is all the long way of saying, once again, that the circuit may need to be altered to bring out the best in guitar.  I think it is possible, but not with the stock circuit in the appnotes, which is oriented towards full-bandwidth program material.

[bancika]

Can you suggest some mods?
There's cap 3300pF which sets bandwidth, I tried everything between 1000pf and 68000pf and no major difference
If anyone has some advice please help, I'd really like to make this thing work
Tnx

[Paul Perry (Frostwave)]

Bancika, what are you hoping the chip will do? What kind of 'noise' do you want to reduce?

[bancika]

Well mostly hum going on at higher gain

[Paul Perry (Frostwave)]

I'm sorry to say, it isn't going to help with the hum.
The first thing to do with a box that hums, is to make a dummy input (a plug with a 220K or so resistor soldered across the terminals). Then plug that in, instead of the guitar, and see if htere is any hum... that will tell you whether the noise is from the guitar, or the box.

[bancika]

Bummer...it's not some enormeous hum, just normal hum you get when you crank all pots all way up

[pedalgrinder]

if you check out over on freestompboxes they have the ibanez nb10 schematic that has the lm1894 quite nicely done. Check it out.

[deadastronaut]

  2006....thats gotta be a record thread revival surely!....