[hpsdr] ANAN-100D Angelia_test_v3.9 code

[Brian Lloyd]

On Sun, Feb 9, 2014 at 3:41 PM, Joe Martin K5SO <k5so at valornet.com> wrote:

> Hi Brian,
>
> You're not asking for much, are ya!  From what I can see it would involve
> a complete disassembly of the ANAN-100D in order to gain access to the
> relays themselves on the PA board, and I've had the thing for less than a
> day, hihi.
>
> Yeah, sure, I'll look it over and see how much trouble I can get into
> doing it to check out the possibility for you; if you're sure you don't
> want to dissassemble yours instead, that is!!
>

Well remember, you don't own it until you take it apart! I make a point of
dismantling nearly everything I get new. My ANAN-x radios were no
exception. I am quite comfortable with my ANAN-10 and ANAN-100D in pieces
on the bench. (I have made board-level SMD changes on the Hermes board in
my ANAN-10 to fix the external 10MHz reference input. I don't have a
hot-air rework station for nothin'!)

But in this case the only disassembly required would be for the code going
into the FPGA. The chips that drive the relays in the ANAN-100D PA BPF
(turns out they are bandpass filters, not lowpass filters) are a pair
of cascaded TPIC6B595 8-bit, serial-in, parallel-out shift registers with
high-power open-drain outputs. The way you select a filter (or an antenna)
is to put a one on the SPI_SPL_DATA bus at the input and then clock the
shift register until the one appears at the appropriate point in the shift
register and pulls that output low to activate the correct pair of relays
to select the desired BPF. (Actually, you use that to select a number of
things including antenna relay, the bypass relay, and the bandpass filters
so it would be normal to clock more than one '1' bit into the shift
register.) But if somehow you clocked in TWO one bits and shifted them into
the BPF0-BPF5 outputs you would end up with TWO BPFs selected. The
interesting thing about this bug is that the PA would still work but you
would probably get strange gain changes at different frequencies due to the
paralleled bandpass filters. And this is something that is a
software/firmware problem rather than a hardware issue. SPI_SPL_CLK and
SPI_SPL_DATA both come in from the Angelia on the 10-pin ALEX umbilical.
Just look at the schematic for the ANAN-100[D] PA board to see what I am
talking about.

Hmm, this is where a bunch of little LEDs showing the status of these
open-drain lines would be nice. Better still, bring them out to the front
panel of the ANAN-100D so we have something to look at. :-) But I digress
...


> 73,  Joe K5SO
>
>
> On Feb 9, 2014, at 1:42 PM, Brian Lloyd wrote:
>
> On Sun, Feb 9, 2014 at 1:37 PM, Joe Martin K5SO <k5so at valornet.com> wrote:
>
>> ***** High Performance Software Defined Radio Discussion List *****
>>
>> Thanks, Bruce.  Great to hear it!
>>
>
> When doing the calibration one thing I noticed that surprised me is that
> gain is not consistent. Used to be on my ANAN-100D and ANAN-10 that gain
> was steadily lower going from 160m to 6m. I now have some bumps where not
> expected, e.g. the PA gain is higher for 15m than for 17m.
>
> Hmm, the relay drivers for the PA BPF are a pair of TPIC6B595 shift
> registers. That means that it would be possible to shift in two bits and
> select two of the BPF in parallel. That might have a significant effect on
> the passband ripple. Can you confirm that only one bit gets loaded into the
> shift register?
>
> --
> Brian Lloyd, WB6RQN/J79BPL
> 706 Flightline Drive
> Spring Branch, TX 78070
> brian at lloyd.com
> +1.916.877.5067
>
>
>


-- 
Brian Lloyd, WB6RQN/J79BPL
706 Flightline Drive
Spring Branch, TX 78070
brian at lloyd.com
+1.916.877.5067
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20140209/875c8344/attachment-0003.htm>