I was searching for a small driver for 2m in order to have enough power for my microwave transverters. I usually drive them with about 500mW. Since i have a transverter block with about 0dBm output it needs some gain. Again i found something from China. The advertisement told something about 1.5W, the pictures showed a PCB with a print 35dB/3.2W. However, i thought it could be good enough to get 0.5W out of it.
A small VHF/UHF amplifier
When the device arrived i noticed that the manufacturer milled the top of the two amplifier IC. Very funny that the guys there try to save their designs from unwanted copy ;)
At the input there is a 6dB attenuation pad which are followed by two amplifier stages of unknown type. The first stage is supplied via a 5V regulator, the second one is connected to the supply input directly. I used 12V to do some measurements.
The first action was to know the gain of the circuit. I wanted to see if it is really working from 1 to 700MHz.
In the picture you see the little noisy measurement. The drive from the analyzer was set very low.
Gain of small VHF/UHF PA
From the gain curve it can be seen that the device is probably not working that well at the lowest end of the shortwave. But at least from 30MHz on it should be usable. Towards upper frequencies the gain is decreasing a lot but still 27dB at 700MHz. For the 2m band roughly 43dB gain is achieved and about 33dB in the 70cm band.
Finally i wanted to know which output power can be achieved in the amateur radio bands.
VHF/UHF PA output power
At about 1W output the output always gets compressed. You might add about 0.4dB on 70cm and 0.3dB on 2m to my measurements since the graph does not include the attenuation of the output cable. For me it means that the amplifier should work well for my purpose.
Just for fun i tried to collect some components from the shelf and plugged together a simple monitoring receiver for 9cm.
I have a DF9NP PLL OCXO for 2592.2MHz and a pipe-cap filter tuned to 3400MHz. From China i ordered a wideband LNA some month ago. It has about 15dB gain on 9cm. The mixer is a ZEM-4300+ (+7dBm LO). The setup results in a IF of 807.8MHz which is received with an RTLSDR and the very nice GQRX software.
With a short wire attached to the input of the LNA i was able to receive the local 9cm beacon DM0TUD.
Since i use Quisk for my HiQSDR i found it interesting to experiment with FreeDV again. Quisk can directly support using the FreeDV API and work in this digital mode without building lots of cables. Since i wanted to try the latest mode 700C i had to compile the library from the scratch and add the mode to the configuration of quisk. Below you can find recordings of my CQ call in the modes 1600 and 700C which were made loopback.
Now i search for a QSO partner preferably on 2m somewhere in the area around JO61. I can work relatively well direction southeast. So east OK might be good.
Please remember the 700C mode uses only 1kHz of bandwidth.
If you are interested in some tests just send me a email to firstname.lastname@example.org
Some posts ago i wrote about my new 6cm transverter. I already did a few QSO and i got a 10W PA for a good price. Unfortunately this PA needs 1W of Input and the transverter gives maximum of 250mW output. So a small “driver” is needed. I was researching the options but found no easy solution that also fits my budget constraints. Then i saw some amplifiers sold as WLAN amp. Usually those are for 2.4GHz but i thought if there are some for the one band there are probably also some for the 5GHz band. During the search i found the nice Amp on the picture below that is sold as an amplifier for video transmissions for about 22Euro. It is specified with 3.5/4.5W. But whatever output power for the price i thought i cannot do wrong a lot.
A chinese 5.8GHz PA
When the device arrived i had to find out the supply voltage. In fact it works with 12V and a first test confirmed it has some gain. The bad thing was that the connectors were reverse SMA and for each connector it was quite some gap between the connector and the PCB. So i thought, “thats a strange matching method” ;)
After removing the soldered shield on top of the PCB one of the screws for mounting the heatsink also came out.
The semiconductors used in the PA are 2x Skyworks SE5004L with a P1dB of 30dBm typical and 34dBm max. So i thought 2W should be possible at least. The gain is specified with 32dB. At the input of the PA is a pi-attenuator.
Later i soldered some nice SMA jacks and then the PA went to storage for quite some while.
The “inner values” of the small amplifier
Now i found some minutes to remove the attenuator and do some raw measurement of the output power. The gain of the amplifier is about 28..29dB. I did not measure the cable attenuation exactly so do not nail me down on the last dB.
The input vs output figure you can find below. The P1dB is about 2.5W. So considering the samples to be typical there is quite some loss on the PCB. But even in that case for the given price its a nice thing.
Not all things work out… I was searching for some nice feed for a offset dish to be used for 6cm narrow band. In the internet i found a nice version from Russia.
It was announced for the 5GHz WLAN band without dedicated specification of the exact frequency.
For 40 Euro including shipment it was ordered and after only 2 weeks in transit (russian saying tells: For bad children Santa Claus sends the gifts by russian post… ;) ) and 1 week in customs i finally got my russian present.
5GHz horn from Russia
Measuring the input matching shows that its well in the middle of the lowest 5GHz WLAN band.
WLAN horn matching measurement
Far away from 5760MHz. But ok, lets try. First i removed the N connector.
5GHz horn from Russia
The connector with the radiator seems to be a robust construction. I tried to shorten it step by step. Unfortunately it happened what could be expected. The feed can be de-tuned to about 5.5GHz but above that the matching vanishes. Probably the distance to the backside of the horn is too big. So this one was a fail. Fortunately i can still use it for WLAN if i make the radiator a bit longer again. If you want to use the lowest of the 5GHz bands it is really a nice and very solid construction.
I have a Tytera MD380 DMR transceiver now. Flashed latest MD380tools and database yesterday and hope to make a first qso soon. A mmdvm hotspot is in the pipeline but i still need to connect a transceiver.
Now finally i have my 4m rig ready. It consists of a DF2FQ transverter with some modification to power module mounting and RX amplifier stage.
Currently i use my FT-817 to drive it. I also added a separate drive input which i need for the SDR. With 1mW input on 29MHz it can achieve about 25W output on 70MHz.
So far i did not benefit from ES conditions. Seems it missed all the nice propagation from last weeks. Hopefully there will be another opening until end of August, when the permission ends here.
As antenna i use a Hentenna Quad with direct 50Ohm assymetric feed point. The antenna has some significant Null in the radiation pattern but is a good compomise for the fibre mast at the balcony.
My fist contact was with DK2EA in JO50UF in CW. Meanwhile i also worked DD1VD and DL2VPO locally.
4m band rig on the balcony
4m Hentenna Quad
[Edit 2017/07/23] meanwhile i worked EA1BFZ in IN81SS and today SV2JAO in KN10DN. ES propagation seems to be very short and spotty. You really need to look to the cluster and search the frequencies all the time. 5 minutes later all could be gone. 100km away the conditions can be significantly different.
Just before the May DUR competition i got my new 6cm Transverter working. Before I faced some problems with the multiplier chain.
Murphy made that saturday afternoon i burned the control circuitry of the SMA bistable relay – poff. So i had to exchange it with my own sequencer circuitry that needs some highside FET switch to do that job.
I used a dual-band ring feed i made for tests and connected it to a old 60cm dish. The result you can see on the photo below. Very beta version ;) The PA is still missing and therefore i had only about 250mW.
I also wanted to take 3cm with me. In order to limit the amount of equipment i decided to minimize effort for 23cm. 13cm is out of order at the moment anyway. That means only using the former Wimax patch antenna for 9cm.
The following result i can claim:
23cm 12QSO 772P
9cm 4QSO 481P
6cm 2QSO 351P (x2)
3cm 4QSO 489P
Next steps for 6cm will be adding driver (ordered from China) and the PA (10W, already here) as well as creating some better feed suitable for the mesh dish.
My goal is to operate 23/13/9/6 with one mesh dish.
This time i had a neighbour in JO61XA. DM4SWL was operating few meters away on 23cm FM only. Thanks for taking the foto !
Attiny2313 + Uln2003 + highside FET switch for PA + lowside FET switch for RX-TX relais and a few bipolar Low Power stages on a two-layer board. The PTT input can be either high-side or low-side active. It also has lock input to connect several transverter to one antenna system. Via the driver IC a pulse relais and a fan can be controlled. There is also a Tx inhibit output that can be used with Yaesu tranceivers.
Schematic: Attiny Sequencer Rev2 schematic
The circuit has two PTT inputs and two Error inputs (decoupled by diodes). A active high PTT input is available as well. There is a highside FET switch that can be used to drive a PA and a lowside FET switch that can be used for an RX/TX relais. The TX inhibit signal is available to drive the inhibit input of Yaesu transceivers. The active out is used to signal a sequencer in TX mode. This can be used to lock other sequencers via the error input. The ULN2003 driver is currently configured for another RX/TX relais, a fan which is running until 30s after releasing the PTT, another low active signal parallel to TX-inhibit (as PTT for TRX without TX inhibit) and output signals for a pulse relais (bi-stable RF relais). The serial interface (also used for ISP) is unused in the application. It might be possible to connect a temperature sensor for example.
The missing connection from Q4 to R7 is a print error.
The design was done with KiCad which is a open source PCB layout system. The code for the microcontroller was written in C. The 2k program memory are almost full now but one might be able to optimize. If you are interested in the design data let me know. I also have some spare PCB left over.
I wonder what else could be implemented with this PCB with the small program memory size. Feel free to give some ideas.