[eagle] Re: Preliminary CAN-Do EMI Findings and Suggestions to Mitigate in the 70 cm Receiver

Chuck Green greencl at mindspring.com
Fri Jun 15 08:59:44 PDT 2007


Hi Juan,

I don't know of any specifications but the idea that the power will be 
noisy has been the assumption all along.  The CAN-Do! has some filtering 
at it's power input with this in mind.  It starts out with an inductor 
(10 uH) followed by two capacitors (.1 uF and 33 uF).  Unfortunately, 
the fuse precedes this filter so if you end up bypassing the fuse for 
your module power you won't be able to benefit from this filter.

We had a bad experience with using capacitors for the first element of 
power filters on P3D.  So I'd suggest your first element be an inductor 
if you are going to bypass the CAN-Do! filter.

Chuck

Juan Rivera wrote:
> Chuck,
>
> I actually think that the noisy power coming out of the CAN-Do module is
> perfect for these tests.  The more I think about it the more delighted I am
> that its there.  I have no other way to generate noisy power.  This is
> great!  I'll attempt to capture this noise with my oscilloscope and post it
> in my log if I can.  My scope is not feeling well these days so I may need
> to dig up another one.
>
> Even if the receiver eventually receives its power from another source, we
> need to assume that it will have switching noise on it from other payloads.
> That's the only reasonable assumption as far as I'm concerned.  The receiver
> must be able to function in the presence of noisy power and the CAN-Do
> module is a perfect noise source.
>
> Based on the above I think I'll leave the CAN-Do module alone for now and
> see if I can add some simple filtering to kill the spurs in the IF while
> running on the switched CAN-Do power.  If it can do that then I think we
> will be on the right track.
>
> Every other payload designer should take heed and plan to operate with noise
> on their DC inputs too.  This is probably a topic that needs some high-level
> attention.  Are there any power specs that describe maximum expected noise
> levels?  Perhaps a standard DC input filer should be designed and
> incorporated in every power supply input.  Why leave this up to each
> designer?  Come up with a standard filter using SMD components.  I would
> think that this filter would also work to reduce switching noise from the
> payloads and help prevent that noise from propagating back out to other
> payloads on the power bus.  All this looks like an area that needs
> attention.
>
> I'm open to advice here.  All I was thinking of doing was adding about 10 uF
> across pins 39/40 to ground, an RF choke in series, and then another 10 uF
> cap and maybe a second smaller value to ground at the load end of that RF
> choke.  I can cobble something like that together and stick it on the PCB in
> dead bug style.  I can keep adding stages until I get rid of the noise.
>
> 73,
>
> Juan
> WA6HTP
>
> -----Original Message-----
> From: Chuck Green [mailto:greencl at mindspring.com] 
> Sent: Friday, June 15, 2007 7:53 AM
> To: juan-rivera at sbcglobal.net
> Cc: eagle at amsat.org; David Smith; Dave Black (Work); Dave Black (Home);
> Samsonoff at Mac. Com; Juan.Rivera (Work)
> Subject: Re: [eagle] Preliminary CAN-Do EMI Findings and Suggestions to
> Mitigate in the 70 cm Receiver
>
> Juan,
>
> I have a couple of suggestions.
>
> Based on previous messages I presume you will not want to use the 
> switched power from the CAN-Do! regardless of the noise issue. I also 
> see that the switch transistor in the CAN-Do! is in the same area of the 
> PCB as the CAN-Do! power supply.
>
> 1) Put a jumper between pins 2 and 3 of the switch transistor (TO-220). 
> This will bypass the switch and any noise it is picking up from the 
> power supply. This will keep the 2 amp fuse in your power source and the 
> current measurement circuit.
>
> 2) You might also eliminate the 2 amp fuse from your power source by 
> placing a jumper from pins 1,9 of the DA15P to pin 2 of the switch 
> transistor after removing the transistor (or just cutting it's leads). 
> This will preserve the current measurement of the power going to your 
> module.
>
> 3) You might also bypass everything by putting a jumper from pins 1,9 of 
> the DA15P to any unused user pins on that same connector (pins 
> 2,3,5,10,11,12) then picking up the power from the corresponding user 
> pins on the 40 pin connector. If you do this, you should provide a 
> current measurement circuit on your module and attach it's output to one 
> of the analog input pins of the 40 pin connector.
>
> The above may eliminate or reduce the noise in your power source while 
> accomplishing other power source objectives.
>
> Chuck
>
> Juan Rivera wrote:
>   
>> All,
>>
>> I've completed a preliminary evaluation of the CAN-Do RFI problem. 
>> Please see my log for details by clicking here 
>>
>>     
> <http://www.juanr.com/pages/hobbies/ham_radio/Eagle/CAN-Do_Noise_Troubleshoo
> ting.htm>.
>   
>> Preliminary Findings...
>>
>> 1. The spurs seen in the receiver IF enter the receiver in two ways
>>
>> o As radiation from the CAN-Do inductor
>>
>> o Via CAN-Do module switched +14 VDC power on pins 39 and 40
>>
>> 2. I/O line filtering appears to be unnecessary
>>
>> 3. The switched power from the CAN-Do module pins 39 and 40 must be 
>> filtered
>>
>> 4. The CAN-Do module will either have to be kept physically separated 
>> from any sensitive components, redesigned, shielded or a combination 
>> of the three.
>>
>> 5. Moving the Can-Do module 2-1/4" forward of its normal location on 
>> the receiver PCB lowers the EMI level below the noise floor.
>>
>> Suggestions...
>>
>> . Create a full size PCB for the next generation receiver and shift 
>> the RF components to the back of the board (away from the CAN-Do module.)
>>
>> . Add DC input filtering to the receiver
>>
>> 73,
>>
>> Juan
>>
>> WA6HTP
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Via the Eagle mailing list courtesy of AMSAT-NA
>> Eagle at amsat.org
>> http://amsat.org/mailman/listinfo/eagle
>>   
>>     
>
>
>
>   


More information about the Eagle mailing list