Received-SPF: neutral (gmail.com: 193.82.116.20 is neither permitted nor denied by best guess record for domain of owner-rsgb_lf_group@blacksheep.org)
DomainKey-Status: bad (test mode)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=btopenworld.com;
  h=Received:Received:Message-ID:From:To:References:Subject:Date:MIME-Version:Content-Type:Content-Transfer-Encoding:X-Priority:X-MSMail-Priority:X-Mailer:X-MIMEOLE;
  b=Q5u6hSDnHdZEMu2qxcTSJzuZTXurrWKA3sVwWavtbQSV1B8We3wOZONLibOLz/blth6zu9f0+HbVG3BokxRC8JDA7hSP6YKSbR9ycZ3tovDzX718m0PiJ+QKYSmhzgxvHmemLK1kjjzG8ta02KHfhg+rRAF9YyCa9J9H8p3Ts6o=  ;
Message-ID: <000d01c6ce6e$6b782c40$17637ad5@w4o8m9>
From: "James Moritz" <james.moritz@btopenworld.com>
To: <rsgb_lf_group@blacksheep.org>
References: <000001c6cdcc$c7d20490$e6a4c593@RD40002> <002701c6cdfd$2e48a2d0$6501a8c0@eagles>
Date: Sat, 2 Sep 2006 10:01:46 +0100
MIME-Version: 1.0
Subject: LF: Re: LPF inductor cores possible cause?
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Status: O

Dear J, LF Group,

I have been using air-cored coils in a 137kHz LPF for some time, with good
performance. The -2 mix iron dust toroids also work well. I have tried using
the ferrite TV LOPT cores (it is necessary to have an air gap between the
core halves), but while workable, they are generally less satisfactory and
more complicated to use.

It is certainly possible to get unwanted coupling between the air cored
coils, either to each other or to other parts of the circuit. But provided
the coils are arranged at right angles, are not very short with respect to
diameter (mine have windings abt 80mm long by 55mm diameter), and are kept a
couple of diameters from other parts of the circuit, coupling does not seem
to be a problem. So provided these precautions are observed, there shoudn't
be a problem with your air cored inductors. If you move  the filter
components around, the coupling will change, so if this leads to a big
change in the observed waveform, coupling could be the problem.

I suspect your spikes are reaching the oscilloscope via some other path
rather than through the low-pass filter. It is difficult to comment without
having the circuit to look at, but usually these things are due to grounding
in one way or another. What happens is there are multiple ground paths that
the switching transient currents flow through returning to their source. The
ground paths can have significant impedance, and if the same ground path is
shared by transient and wanted signal, a transient voltage ends up
superimposed on the wanted signal. One possibility is that some switching
currents are flowing via the output coax ground from the filter, through the
'scope ground via the "scopematch" circuit, and back through the mains earth
to the PA. Some things you could try are: -

If not already done, mount the PA and filter on the same metal base plate,
and make sure both are grounded directly to the base plate.

Eliminate the coax link, and connect the PA transformer secondary directly
to the filter input capacitor terminals. This provides the shortest path for
the circulating currents in the PA output. Remember that the filter is an
integral part of the class D PA - if you take the filter away, it isn't a
class D PA anymore. Direct connection will reduce unwanted stray inductance
in the PA circuit.

Try connecting the filter output to the scopematch input via a longer coax
cable (several metres at least, preferably coiled up). This will increase
the impedance for unwanted ground currents without affecting the RF load
impedance.

Cheers, Jim Moritz
73 de M0BMU


----- Original Message -----
From: J. Allen <vy1ja@northwestel.net>
To: <rsgb_lf_group@blacksheep.org>
Sent: Friday, September 01, 2006 8:30 PM
Subject: LF: LPF inductor cores possible cause?


> Jim and All,
>
> Thanks for the good explanation.  The spikes are displayed on the scope
> which is connected to the ScopeMatch and are on both the voltage and
current
> traces.
>
> The scopematch pickup has a half meter section of coax between it and the
> LPF, which in turn has an even shorter coax jumper connecting it to the
amp
> output. The LPF  uses two single-layer, air-core, solenoid inductors
mounted
> at right angles to each other.
>
> Your comments raised a set of questions:
>
> Could it be that the single 640 FET amp and this Push Pull amp with 2 640
> are such different designs that the air core inductor LPF which works well
> with one, does not work with the other?
>
> Is it possible that some odd inductive or capacitive coupling lets those
> high frequency spikes  be generated?
>
> If this is the cause of the spikes, will it require a pair of  toroid
cores,
> or will a pair of TV cores suffice?
>
> Order toroid cores,  yes or no?
>
> Thanks,
>
> J.
>
>
>
>
>