Posts by OE3GBB

    Hi all,

    Many thanks for all the contributions.


    Here are the results from my investigations:


    Yes, it is the 5th harmonic problem. I could tune my LNB to that frequency on 12.038 MHz using SDR-Console and could see some TV transponder signals at -115 dBm. But when I transmit at lower power I get the 5th harmonic of my transmission at -80 dBm. When I increase the power I get -45 dBm. So it seems that the activation of the doherty path of the amplifier is making this problem.
    I inserted now a 200W circulator and the situation improved a little bit. I can transmit now 333 kS at beacon level without increasing the noise level.
    I am using bandpass filters for 10W between driver and final amplifier, built for Wifi on 2410 MHz, 27 MHz bandwith, but 10W CW only.
    High power low loss low pass filters could be built according this: https://www.changpuak.ch/elect…mpedance_Lowpass_Coax.php. But I don't have the ability for that kind of high precession work.
    Alternativly those could be built from low loss PCBs with increased copper layer thickness.

    I also do have a WLAN duplexer, but only rated for 20W and I don't want to risc this eather.
    So next will work on optimizing the doherty amplifier to reduce the 5th harmonic.


    73 de Gerhard OE3GBB

    Hi Mike,
    that is a very good point. I do not have a SPA for this frequencies, but I will try to tune my LNB to 12025 and see what is there.
    If the 5th harmonic is the problem: This is the first PA I am using without a circulator. So, I also can try ma 200W circulator for 2,4 GHz and maybe this device is acting also as a low pass filter ? How to build a LPF for 200W on 2,4 GHz with minimum losses? I also will try to change the amplifier from Doherty to push-pull A-amp as a last try as maybe the problems arise when the C-part of the amp is activated.
    Will keep you informed.

    73 Gerhard

    Hello, I am using a 1,2m offset dish and a POTY feed. Never had that problem before with different amplifers from low power up to 70W. But now I have built an new amplifier using a Doherty transistor pallet module. I am getting increasing noise on the RX (MiniTioune Scan & Tune), if I increase the transmitting power. I don't have this with the diver stage alone, which is also good for 20W DATV. I do have a very sharp band filter (-70 dB) just before the final amplifier. And I cannot see any spurious signal up to 2,9 GHz on the HP spectrum analyser. Shoulders at 75 W output are at 30 dBc. The signal on the WebRX is very clean and I can produce a 16APSK signal at 15 dB without shoulders. I don't have an increase of noise on my nearby 85 cm dish, while I am transmitting with the bigger dish.
    So, what can it be? I don't have oscillations, I don't have a spurious emission on 2,6 GHz. The only difference to the other amplifiers is, that this one is a Doherty type. The operating point of the transistors are correct according to the data sheet and changed only a little bit to increased linearity, but without any change to this noise problem.

    Has anyone had this problem before or has some idea what I could check?

    73 de Gerhard OE3GBB

    Hi George, Sorry was my mistake. Somehow I thought you are using the LM35 to measure the voltage of some supply. That was a misunderstanding. I have chosen the Arduino Uno as it does have the Ethernet shield for easily sending data over IP. Never heard about the methode you are using by RS232 and a Mikrotic board. Can you send some info about this?

    Reagrds Gerhard

    Hi George, thank you for this hint. But how do you measure voltages with this LM35?
    I do have an Arduino Uno and a Ethernet Shield. So I will get the LM35s next.

    73 Gerhard

    Hallo,
    I have also finshed first tests last week and got about same results. 60W DATV at 30V/10A. Now waiting for a bigger power supply. There is no tuning necessary. About 15 dB of gain at 2406 MHz. Maximum gain is around 2350 MHz, but only 1 -2 dB more. These pallets do have a nice PTT design. There seems to be a temperature switch on the pallet also, but I have not tested it yet. I have used a 4 mm copper plate for heat distribution and a heatsink with blower.
    My pallets seem to be B-ware, but obvisiously have never been used. There are markers on the back pointing maybe to a point with lack of soldering to the copper plate. So far no problem.
    I am now going to construct a PA with two of these pallets to get about 120W DATV out for my small 1,2m dish.
    There was another sale on eBay, but with a pallet for 2500 - 2600 GHz. Also ordered one to see the difference.
    73 de Gerhard OE3GBB

    George, that's great news. To understand it right: You are transmitting using a Pluto with F5OEO software. The repeater is consisting of another Pluto running Gnuradio on a Linux PC. You are receiving with miniTiouner-Express.
    Would it be possible to use a RaspberryPi for the Gnuradio repeater?
    Please keep me informed about your further developments.
    Regards and 73 de Gerhard

    Hello Thomas, Yes I have seen this. But I am confident, that the pallets are OK. They are removed from some amplifiers and so they are used and without any warranty. I have bought 2,4 GHz PA PCBs advertized as scrap in the past and they all work without problems. Will keep you informed, but delivery is proposed at the end of December.

    Great! I have taken down our DATV repeater OE3XNK a year ago because of technical problems with DVB-T. I will try the software and maybe install the repeater again.

    Hi George,
    This amp seems to be quite new, so I don't have any experience. From the data sheet it seems to be quite linear up to 25dBm output in Class A. It's high gain and relatively low price is interesting of SDRs.
    73 Gerhard

    Thanks to all for the responses. We are in an early stage of the project and your part is only to set up the antenna and RX/TX system. The communication part is done by someone else and we just know that part of it will be a high speed downlink on 13cm. To our calculations we are on the limit with a 2 m dish. So at the moment we are investigating the optimum for this setup regarding dish-size / rotator accuracy and speed / system costs. We also have to account for future developments as this institut will not be able to make a similar investment within the next 10 years.
    Anyway, further informations and reports are welcome.

    Our Amateur Radio Club OE3XNK has been asked by a local Fachhochschule to help setting up their satellite station. It is planned to install a 4 m dish for LEO cube satellites from 2m - 13cm. Prosistel is advertizing the PST71D-PRO + PST71E COMBO for this size of antenna. Is there any experience with this setup? What kind of dish will be appropriate, segments or one piece?
    73's Gerhard OE3GBB

    Hello dg0opk,

    TNX, but I did read this before. M0VKK did use the pallet with the smaller transistors 2x MRF7S21110HS, or? I am having the bigger transistors MRF7S21170HS on my pallets. I think the behaviour will be similar, but would like any comment before I start the assembly.

    73's Gerhard OE3GBB

    Hello All,

    I have two of these Andrew pallets, which I want to use in push-pull with 90° phasing. These amplifiers are each with two MRF7S21170HS transistors and Xinger XC2110E-03S. I have found some reports from Andrew pallets with other transistors, but not with these ones. From the Xinger parts data sheet I would assume, that each pallet could deliver 100W DATV on 2,3 GHz, from the transistors I would think of 2x170W-6 dB or about 75W in DATV.
    Has anyone used this pallet for DATV on QO-100 yet? I would need some experience reports and results in DATV, and also modifications, if necessary.
    Thanks in advance.

    Gerhard OE3GBB

    Hi George,

    Sorry, but I am not the right person to discuss that topic, I am just a tester.
    As far I have understood, the sampling of the voice in codec2 is 20ms and 40ms for lower bitrate modes. So 50 Baud is a pretty good match for any type of modem. To get the whole coded data over the channel, David has used several carrieres in parallel and used OFDM to get them as close as possible. Any other modems would be maybe possible and there might be some information about the theory on freedv.org or http://www.rowetel.com.
    In my view the easiest approach to get a best result regarding audio quality without neural network is to fill the 2700 kHz bandwith with as many carriers of 50 Baud. Using differential coding helps to reduce the BER.

    73's

    Gerhard