PCB v-1.1 test

[ik1xpv]

Hello List,

hereafter some report of  PCB v-1.1 Hermes Lite

( https://groups.google.com/d/msg/hermes-lite/HfMIURmR_pA/Wf7A0S00FrwJ )

using the Hermes_Lite_Test firmware, modified with assignment of ad9866_rxclk to PIN_P14 instead of PIN_P11, to use with pcb 1.1

( https://groups.google.com/d/msg/hermes-lite/lFu49OLCt-s/2n_J-M82TIQJ )

The PCB is Rob's v-1.1 equipped with the 73.728 MHz osc. Thanks to Steve and Rob advices

( https://groups.google.com/d/msg/hermes-lite/lFu49OLCt-s/z3c17JshmFAJ ).

Settings:

Hermes-Lite running Hermes_Lite_Test Steve's firmware with default gain, +20 dB.

Window 7 64 bit running PowerSDR^TM OpenHPSDR mRX PS v3.2.19 (8/29/14)

no NB or ANF

AGC slow

mode: USB

filter: 5 kHz

50 Ohm load connected.

The readings were taken from the the dBm meter, I report the mean value, the reading was +1, 0, -1 dB.

-131 dBm at 4MHz

-131 dBm at 7MHz

-130 dBm at 10MHz

-129 dBm at 14MHz

-129 dBm at 21MHz

-127 dBm at 28MHz

The result is very similar to Steve's one:

https://groups.google.com/d/msg/hermes-lite/0CMOG6NKilI/RAnCiE_DJF8J

The BeMicroSDK and the HermesLite are powered via a linear power supply 7805.

I placed an aluminum thermal radiator on top of the AD9866 with a rubber thermal pad.

During RX it reduced the IC temperature of about 20°C from 60°C to 40°C .

I measured the receiver noise floor with and without the thermal radiator and I got no difference at all, following the AD9866 datasheet pagg 12-13. I keep the radiator in case it helps to increase the life of the IC, maybe in TX condition ( to be tested ).

Thanks to Steve, Rob and all the List for the support.

73',

Oscar

IK1XPV

[Steve Haynal]

Hi Oscar,

Thanks for the data. It is interesting that you do not see the rise in noise on 10M that Rob saw. Also, it is interesting that you came up with similar results with PowerSDR. How is PowerSDR working? Are you able to adjust the LNA gain via changing dither?

Did you have to add a capacitor for reset to function correctly?

What balun (core, windings) did you use for TX and RX?

I made the same measurments with my v1.2 last night and was surprised to see the noise values were 8-10 dBm lower. So many things have changed that I don't have confidence yet in those numbers. I will post hard numbers once I make sure of a few things. 

I am very impressed with your build, enclosure and heat sink. It looks much more professional than the sprawl I have on my workbench. I've been thinking a bit about how to make the whole setup more mechanically stable. Rev 1.2 has mounting holes. I also want to post some OpenSCAD custom plastic standoffs designs for the BeMicro SDK that people can have 3D printed. (I have a buddy with a 3D printer.)

Any pointers you can share about your build process would be helpful to others.

73,

Steve

KF7O

  

[ik1xpv]

Hi Steve,

PowerSDR is fine on Win7, as you wrote, it is possible to use VAC audio path to listen via PC speakers.I used IP direct Ethernet connection to Hermes Lite (APIPA). No server in between.

I understood that in Hermes Lite Verilog the AD9866 gain setting is controlled using the dither flag as most significant bit and then the other bits came from the attenuator setting. In PowerSDR the dither check button is in the HPSDR setup tab, it controls the AD9866 gain msbit. Under setting option tab you can select the attenuator steps and then use the GUI's combo box for attenuator selection. It modifies the gain setting in real time, but I didn't  check the real gain levels yet.

For the noise level testing I used your Hermes_Lite_Test configuration with fixed 20 dB gain setting and just readings of the PowerSDR dBm meter.

<< Did you have to add a capacitor for reset to function correctly? >>

At the beginning I used the 100n capacitor, then I made a cut on the pcb and connected the AD9866 reset with a wire patch to RESET_EXP BeMicro SDK connector pin3. See following image. It seems to run ok.

<< What balun (core, windings) did you use for TX and RX? >>

In RX i used a 1: 9 balun. I used an Amidon FB-43-2702 binocular core.

2 turns primary and 3+3 turns secondary as described in http://f1frv.free.fr/main3c_Baluns.html

It would be 50:450 Ohm. The differential input impedance of AD9866 is 400 Ohm.

I have not jet made any serious test of the wound transformer. I operates but I did not tested its loss.

I have still to test the receiver with a calibrated generator and to verify the LNA gain setting and I made no test at all in TX mode. I verified that with the Hermes_Lite firmware (no Test mode) there are 2 RX instances running

The box comes from local flea-market. Possibly It was a Motorola Mobile ETACS cellular (I'm in Italy). It has an H internal shape and offer good heat-sink. I plan to use the other side to house the radio front end, amplifier and filter.

Audio?

Do you think possible in some way to reintroduce the Hermes audio output path. I imagine it could have a shorter latency than the PC path via VAC ?  What is your thinking about?

73,

Oscar

IK1XPV

[Steve Haynal]

Hi Oscar,

Thanks for the information. Please share your front-end, amplifier and filter with the list when ready.

Did you modify the RTL to generate a reset sequence on the RESET_EXP pin? As it is, this pin connects to a FPGA pin that is currently not programmed or used by the firmware. There is a pull-up resistor on the BeMicro. Perhaps connecting to this pin adds enough capacitance or your build is just not too sensitive to the reset issue Rob had. This is the pin that I am using in the upcoming firmware for reset so you are in good shape. 

Regarding a dedicated audio DAC, I don't want to include one on the main board to keep things simple and expenses down. However, you should be able to add a small I2S card to your system which requires only 4 signal connections plus power and ground. The original Hermes code uses a TLV320 ADC via I2S so it shouldn't be too difficult to reenable that code for a different I2S device. I removed that code but it can be put back. I don't know how many FPGA resources it takes, so you may not be able to get 2 receivers to fit if it is included. 

Here are some I2S boards:

http://www.wvshare.com/product/UDA1380-Board.htm

http://www.hifiberry.com/dac/

http://www.adafruit.com/products/1780

Adding sound via I2S to the Raspberry PI was a popular hack:

http://volumio.org/forum/list-i2s-dacs-for-raspberry-t1103.html

Google raspberry pi arduino i2s

73,

Steve

KF7O

[ik1xpv]

Hi Steve,

I made no rtl modification to drive the RESET_EXT at all. I'm waiting for your solution. :-)

Thanks for the audio information. I'm looking at the rtl I2S code and codecs. 

I will keep the List informed.

73,

Oscar

IK1XPV