[hpsdr] Power supply filtering

[Andreas Troschka]

Getting a look at the dimensions of the power supply lines/planes and contacts of connectors! on the Atlas Backplane, common sense suggests to forget PAs over say 6-8W rf, continuous wave.

I would charge no more than 36W on the 12V line and less on the over voltage-lines (don't forget we've one 0V return line only).

If you really need more power from the PA, then separate supply has to be provided to that board even if we are used to have external PAs so I don't see any requirement for rf power-blasters on the HPSDR.

Apart from this considerations, if memory helps, the use of FPGAs, quick ADCs and the like is announcing a consumption of a lot of power.

As project manager of the ATLAS board Phil N8VB should give some indications on the power limits of the various power-lines of the backplane.

A study should be made and maintained up-to-date to verify what board combinations fullfill the ATLAS power maximum ratings.

This study should also provide a table containing the current drains (or power requests) of every HPSDR board in its various function modes (or at least the max power consumptions).

I hope we aren't running low from the beginning. In that case solutions can be studied as modifications to the existing ATLAS backplanes and to its layout (and alternate power lines included) for future production lots.


vy 73s de IK2WQI - Andreas


Jeroen Bastemeijer wrote:
> ***** High Performance Software Defined Radio Discussion List *****
> 
> Dear All,
> 
> Thanks for the link, N0UU. I downloaded the document. Allthough general 
> principles still apply, I noticed the AN is quite old (1986!).
> 
> Michael Ardai gave a few good suggestions for the supporting electronics 
> of Demeter (like the ADC etc.).
> 
> Andreas Troschka and Tony Langdon send mails with regard to the EMI of 
> switchers. On this point I would like to add the following:
> 
> Designing a "quiet" switcher is not as easy as building a "quiet" linear 
> supply. However, it is my strong belief that it can be done. By using 
> proper grounding, effective filtering and a proper lay-out of the PCB. 
> The main advantage of a switcher will be the lower power consumption and 
> less heat.
> Some other words about EMI: We shouldn't panic too much about the noise 
> which is generated in Demeter. The whole HPSDR is filled with all kind 
> of sources for EMI (e.g. clock-oscillators and data-busses). One might 
> argue that the power in Demeter is higher, but this doesn't mean 
> necesseraly that the noise induced in the SDR will be bigger. Some times 
> there is a stronger electric coupling (E-field) instead of a magnetic 
> coupling. So a small signal clock line close to a sensitive input could 
> cause more noise than a "heavy" switcher at couple of inches away....
> 
> I'm certainly willing to make some experiments with switchers and see 
> how quiet they can get.
> 
> Tony wrote: " I would consider a low current linear supply for 
> everything except the PA. "
> 
> Thsi brings me to the point: How much power do we need? I made an 
> assumtion, based on some measurments (from the Jnaus/Ozy combo), but the 
> rest of the power consumption is still a mystery....
> 
> Next step has to be to mak an inventory of all the "power-hungry" cards 
> which will populate the HPSDR. Is this a nice task for Phil? Are you 
> willing to do that? Otherwise, I will contact each project-leader myself 
> and ask the about real life numbers or good approximations of the power 
> needed.
> 
> Thank you,  73 Jeroen PE1RGE
> 
> stoskopf at tri.net wrote:
> 
>> ***** High Performance Software Defined Radio Discussion List *****
>>
>> For a good discussion of output filtering on switcher supplies see:
>>
>> http://www.linear.com/pc/downloadDocument.do?id=4176
>>
>> starting at p. 59 of AN19.  Be sure to read the final comments at the end.
>> Great sense of humor.
>>
>> I still haven't figured out the proper source/loads for modeling in say,
>> SwitcherCad.
>>
>> N0UU
>>
>> _______________________________________________
>> HPSDR Discussion List
>> To post msg: hpsdr at hpsdr.org
>> Subscription help: http://lists.hpsdr.org/listinfo.cgi/hpsdr-hpsdr.org
>> HPSDR web page: http://hpsdr.org
>> Archives: http://lists.hpsdr.org/pipermail/hpsdr-hpsdr.org/
>>
>>
>>
>>  
>>
> 

 1183072300.0