DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=btopenworld.com;
  h=DKIM-Signature:Received:X-Yahoo-SMTP:X-YMail-OSG:X-Yahoo-Newman-Property:Message-ID:From:To:References:In-Reply-To:Subject:Date:MIME-Version:Content-Type:Content-Transfer-Encoding:X-Priority:X-MSMail-Priority:X-Mailer:X-MimeOLE;
  b=sR6rdS97MXRspvByPExtFUKWUOEqtOyciqx2M549PO/aEhpzJ5+7KA+IFh13dT+QZV9H1p7y18IMPUQNxFqhCb8tpnZJcNwBVxihlFuety0/oV0PxKoPMimxrY9omTJYJidaFD+uhmH+lXrcSnBvskNW0dpFCncolhBv2h0yPFw=  ;
Message-ID: <6F17BBDDB2624CB4B81DF1B446E928F2@JimPC>
From: "James Moritz" <james.moritz@btopenworld.com>
To: <rsgb_lf_group@blacksheep.org>
References: <4E2410BC.7070405@iup.uni-heidelberg.de> <CAA8k23TQMhV_Ew-WMA8i1P5H94ty5bLaksWqhxsjwDxZKsEbEQ@mail.gmail.com> <4E2452BC.5050408@iup.uni-heidelberg.de> <D81113E5E794498285F29E6EEED8C2DC@JimPC> <4E255A9B.8090502@iup.uni-heidelberg.de>
In-Reply-To: <4E255A9B.8090502@iup.uni-heidelberg.de>
Date: Tue, 19 Jul 2011 22:11:52 +0100
MIME-Version: 1.0
DomainKey-Status: good (testing) 
Subject: Re: LF: TX converter for 137 kHz, help needed
Content-Type: text/plain;
	format=flowed;
	charset="iso-8859-1";
	reply-type=response
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
X-AOL-SCOLL-URL_COUNT: 0  
X-AOL-SCOLL-AUTHENTICATION: mail_rly_antispam_dkim-m217.2 ; domain : btopenworld.com DKIM : pass  
x-aol-sid: 3039ac1d40ce4e25f3731357
X-AOL-IP: 195.171.43.25
X-AOL-SPF: domain : blacksheep.org SPF : none



Dear Stefan, LF Group,

>>> So what is missed in my circuit is the allpass network that makes a 90
>>> deg delay of the input signal.
>>
>> ...Unless the software generates I/Q outputs from the sound card L/R 
>> stereo outputs.
> Why? The design Andy shows has a mono audio input. This is what i am 
> searching for :-)
>
That was just an incidental remark - it is possible to generate high quality 
quadrature waveforms in software, which avoids the need for hardware phase 
shift networks and their limitations.

>> Scale the values of either R or C proportional to the inverse of the 
>> centre frequency (or both - it is 1/RC time constant you need to scale). 
>> Yo might find faster op-amps neccessary at high audio frequencies.
> OK, i did that for a center frequency of 12 kHz but the result was poor. 
> Then is systematically played with the values and found two pairs of 
> 1n/5k8 and 15n/2k27 which showed at least 47 dB image rejection in the LF 
> band, peaking at -76 dB at 12.55 kHz, which will be a theoretical value of 
> course. I'm going to try these values in a practical circuit and see what 
> comes out.

Well, it just depends what you are trying to optimise - do you want a wide 
frequency range with moderate unwanted sideband supression (this is what 
Andy's initial values on the spreadsheet seem to be aimed at), or a higher 
supression over a narrow frequency range - you might like to try C1 1n, R1 
5470, C2 4.7n, R2 6790. These are feasible values - but to get the response 
shown on the spreadsheet, you might need a screwdriver with reduction gears 
for adjusting the preset resistors ;-)

BTW, the LM358 is primarily a low-frequency device - it is getting a bit 
short of gain and slew rate at 12kHz, it might be OK but you might want 
op-amps with a bit more bandwidth - e.g. TLC272 should work fine in Andy's 
converter circuit

Cheers, Jim Moritz
73 de M0BMU