Date: Thu, 23 Aug 2018 21:04:20 +0100
From: Chris Wilson <dead.fets@gmail.com>
Message-ID: <539641811.20180823210420@gmail.com>
To: N1BUG <rsgb_lf_group@blacksheep.org>
In-Reply-To: <88af04ed-ef1c-03e3-464f-fc5e4155d0eb@n1bug.com>
References: <1613303827.20180823120801@gmail.com> <88af04ed-ef1c-03e3-464f-fc5e4155d0eb@n1bug.com>
MIME-Version: 1.0
Subject: Re: LF: Anyone good with Elsie?
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


Hello Paul,

Thanks  for  the reply. I added a 470pF cheap disc ceramic across each
of  the 2000pF polystyrene caps, and indeed the filter moved lower, to
pretty  much the optimum spot (thank you Stefan). But I see a lot more
attenuation with these two caps added. I presume disc ceramics are not
ideal?   The  inductors  are  cheapo  pre  made  ones (look like 1/4 W
resistors), again off Ebay so probably far from optimal. I may need to
look  at  raising  the  Q of this thing! Are there high Q commercially
made  inductors in 680 and 2.2 uH or do I need to wind my own? How sub
optimal is using disc ceramics for parallelling the branded poly caps?

Here  are screen shots of the same level noise signal with and without
the  modded  BPF in between the amplified noise source and the pre amp
for the Red Pitaya SDR.

http://www.chriswilson.tv/with-modded-bpf.jpg

http://www.chriswilson.tv/without-modded-bpf.jpg


the modded BPF drops the noise signal from -78 dB to -98 dB

I am running the RP tonight with the modded BPF in place to see how it
goes but i fear it's a bit of a mess now!

Thursday, August 23, 2018, 7:26:10 PM, you wrote:

> Hi Chris,

> I sent an earlier reply which seems to have gone into the bit bucket
> somewhere, possibly to due to server settings in my mail client
> which I've just now changed back to what was working previously.
> That first reply was mostly off topic so if it does show up, please
> ignore it.

> I've been following the comments on your topic and just wanted to
> toss something in the mix. It appears the response of your actual
> filter is more rounded and the skirts less steep than the model. I
> leave it to those with more knowledge to correct me, but I believe
> this suggests Q of components in your filter is significantly lower
> than Q assumed by the model. Using higher Q components, if possible,
> may give you more attenuation of unwanted out-of-band signals. I
> built one of these filters and my response appears to be in closer
> agreement with the model.

> Not speaking to a solution but a possible cause of the unexpected
> filter response: If you used capacitors with reasonably good
> tolerance, perhaps the inductors came out a bit low in value. I've
> forgotten whether you used pre-made inductors or wound your own
> using cores of some type. If the latter and the cores are ferrite,
> the specific batch of cores might be enough to throw it off. I
> believe ferrite materials typically don't have tight tolerance.

> Paul


-- 
Best regards,
 Chris                            mailto:dead.fets@gmail.com