On 4/9/07, Ken Arck <[EMAIL PROTECTED]> wrote: > At 09:27 AM 4/9/2007, you wrote: > > >Hello, for DTMF IC you can find Clare M-8870 or Teltone M-8870.
I believe the original company to make all of these was Mitel. Clare and Teltone were repackagers, I believe. Thus the "M". > <----The 8870 (and its derivitives) are decoders ONLY. Sure you're > not thinking of the 8880? And I don't think Teltone is even around > anymore but the Zarlink MT8880 may be what he's looking for: > > http://products.zarlink.com/product_profiles/MT8880C.htm Perhaps. I think Zarlink bought up Mitel and/or the rights to produce their stuff, years ago, didn't they? California Microdevices also made their own flavor (or licensed it from Mitel?) of the 8870 and 8880, I believe. CM-8870 and CM-8880. They (if I remember correctly) had slightly better specs (we're talking VERY slight) than the original Mitel's, and different packaging available. The original Mitel's are damn good at what they do for such a cheap device. They detect the CRAPPIEST DTMF you could ever look at on a scope... with amazing accuracy. Their switched-capacitor filter design is nifty stuff to look over for those of us without a BSEE or any of the resulting math background, just to learn a little something about filter control. (It's widely rumored that that TS-32 and TS-64 from CommSpec for CTCSS encode/decode use similar "ideas" for CTCSS detection as the Mitel chipset does at audio/DTMF frequencies, just lower. I've not seen any documentation to that effect though, and I doubt CS is going to share how their stuff works while they're still around and selling [nice] products.) Some fancy DSP algorithms can beat the Mitel's for sure, but for the Amateur -- the Mitel style chips work great. I recently dealt with a DSP algo at work that someone wrote that has a hard mathematical cut-off at -28 dBmV on a T1 carrier circuit, which was WAY above where the customer sometimes had DTMF coming in from a foreign country... I bet the Mitel would have at least TRIED to detect those weak signals, but the DSP math had been told by some programmer somewhere... "Just ignore anything below -28 dBmV"... and it did... they had to drop a new version of the DSP code in to go down to -31 dBmv or -32, I forget. The customer [thought they] wanted -40 dBmV! But when you get down there, echos off of walls from speakerphones are enough to double and triple DTMF digits... The only place a DTMF detector should go THAT low is at the switch, perhaps... where there's nothing going on but dialing a number or receiving ANI, or similar... a detector always in-line shouldn't go that low -- in telco anyway! Nate WY0X
