[hpsdr] [Fwd: Re: Tiny Demeter]

[Jeroen Bastemeijer]

Dear Ahti,

Thank you for your input! For Tiny Demeter the idea is simplicity. In this 
case the simplicity will generate some extra heat ;-)

Only for the -12V we need a switcher. I think, this should be the only 
switcher which will be initially in the Tiny Demeter design. However, in 
the block-diagram you can see an optional preregulator for the +5V. In the 
diagram this is drawn as an optional linear regulator. But by providing 
some sort of screw-connections to the outside world, one could easily add 
a switching regulator instead. This will give each one the opportunity to 
make the trade-off for himself between simplicity and/or higher 
efficiency.

Maybe a part of the -12V switcher could be used to drive an external 
switching transistor (FET or bipolar) to generate the pre-regulated +5V?!

In general, I think the power dissipation in Tiny Demeter is determined by 
the external voltage applied. So we can/should design the linear 
regulators to 
work with minimal voltage drop, but the final dissipation will the depend 
on the input voltage.

For the GPS-power (see far below ;-) ) I think an extra regulator 
generating the necessary voltage (this could be 3.3.V) is the solution. Up 
till now, each card generates its own 3.3V, 1.8V etc. with small 
regulators from the +5V. For simplicity and flexibiltity point of view 
this is the best solution. Specificially for the GPS, I have some modules 
around which will work on 5V, others work on 3.3V..... which one should I 
pick?! ;-)  Maybe it is best to leave the GPS power supply voltage to the 
final user. But: For Tiny Demeter provisions could be made to have a 
"standby 
power circuit".

Thank you all for getting the discussion going!

73 Jeroen PE1RGE

On Sun, 15 Jun 2008, Ahti Aintila wrote:

> Jeroen and Greg,
>
> It may be a good idea to use discrete linear regulators together with
> switch mode preregulators assuming that you can prevent the conduction
> and radiation of the switching transients.
> In order to keep the power dissipation low in the series pass
> transistors, use power MOSFETs with the drain on the output side. That
> way you can use minimal preregulator output voltage and still provide
> good gate voltage level to the MOSFET. Naturally you have to use
> P-channel devices for the positive voltages and N-channel for the
> negative voltages.
>
> 73, Ahti OH2RZ
>
>
> On 15/06/2008, Greg - ZL3IX <zl3ix at inet.net.nz> wrote:
>> ***** High Performance Software Defined Radio Discussion List *****
>>
>> Hi folks,
>>
>> Phil VK6APH asked me to comment on the proposal for Tiny Demeter.  My
>> comments and his reply copied below.
>>
>> 73, Greg, ZL3IX
>>
>> Hi Greg,
>>
>> Agree with your comments - we may want to have the GPS power on at all times
>> to lock the associated xtal.  Suggest we have this discussion on the forum
>> so that others can join in on the suggestions. Would you like to move you
>> comments overthere.
>>
>> Thanks
>>
>> Phil...
>> ----- Original Message -----
>> From: "Greg - ZL3IX" <zl3ix at inet.net.nz>
>> To: "Phil Harman" <phil at pharman.org>
>> Sent: Sunday, June 15, 2008 3:29 AM
>> Subject: Tiny Demeter
>>
>>
>>> Hi Phil,
>>>
>>> I had a look at this as you asked.  Just two comments.
>>>
>>> 1)  I would use the DC/DC converter to pre-reg the 5V supply to, say
>>> 6V.  As it is, that series pass transistor is dissipating over 12 W at
>>> 1.5A output.  The -12V can then be taken from a flyback winding or
>>> something like that.
>>> 2) The original Demeter spec talks about a supply for a GPS module.
>>> Many of these require 3V3 today.  Not sure if Tiny is also supposed to
>>> power a GPS module, but if so, this should be taken into account.  Could
>>> be a linear reg from the 6V, as these modules don't draw a lot of current.
>>>
>>> My 2 cents.
>>>
>>> 73, Greg
>> _______________________________________________
>> 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/
>>
>

 1213523118.0