3.4 GHz Transverter Conversion (3.5 GHz WiFi transceiver)

[VK5ZT]

Hi all

And now for something completely different.....

I have a panel running with a 2 Metre IF. Its Rx only at present as I did it on my development system which has no Tx filters.
With rough testing it seems to have very little response on the image frequency. Rx sensitivity seems comparable with a 70 cm system.
The IF came from the output of the amplifier directly after the mixer. The following filter board and some other bits were removed from the main board. There is work needed on PLL1 to get the additional range from the VCO but it goes and locks fine.

The plan is to mod a 'virgin' panel completely for 2 metres. Of course the big issue will be how much image/LO gets through the transmitter and how easily can I reduce it. It would have been nice if they had used a proper mixer arrangement on TX....any balance to minimise LO output is probably destroyed by the balun!!

Cheers

Tim VK5ZT

[VK5ZT]

Hi all

Something else to consider....

At my Gippstech presentation I touched briefly on providing a lower frequency IF output for use with an SDR.

I have been working towards using a Flex 1500 with the transverters. Yeah, I know....you need to have a PC running all the time but I have PCs in the van and shack and can live with that for some of the benefits. We all know the features SDR software offers. In the case of PowerSDR, used by Flex, it has a transverter mode allowing you to set up a suite of transverters, factoring in frequency offsets and errors as well as transverter gain.

The transverter frequency is shown on the display, you can see a chunk of the band, waterfall etc etc.... One feature I like is the S meter. It can read in dBm and is accurate....by that I mean that it tracks my calibrated signal generator dB for dB from around 0 dBm down to below -100 dBm when the signal is competing with the noise! Of course, it can do the same for whatever transverter is connected. This makes gain comparison between antennas etc. much more meaningful than the "two blocks on the 817 s-meter" method!

What has all this to do with the 3.4GHz panels? Looking at the panel architecture I decided it was a prime candidate to run with a 28 MHz IF, making it usable with the FLEX, which stops at 54 MHz. You would not consider such a low IF normally as filtering out the LO and image would be almost impossible by simple means unless you used double conversion....which is exactly what the panels normally do on transmit. The baseband transmit signal must have been in the tens of megahertz when these were in use. A bit of maths showed I could place the Tx IF right in the middle of the passband of the SAW filter we normally remove (480 MHz). Then using the other mixer in the transmit chain and PLL2 to convert from 28 MHz up to that frequency (ie PLL2 = 452 MHz.) This is exactly how it normally operates so there is only one small mod (a wire link) to the transmit circuit to disable the Tx squelch giving total Tx control to the micro.

Next the small bit of circuitry on the first Tx mixer input is removed. The input of the mixer will be connected directly to the special transverter port on the FLEX....so 28 MHz from the FLEX will result in 3398 MHz output AND the image/LO signals are outside the passbands of all the filters in the transmitter so its output stays 'clean'! Transmitter sorted!

For receive, we need to do the filter mod where the original Tx diplexer filter is added to the receive chain in place of the removed ceramic device, fit a transmit relay and tune it all up. The Rx IF is now 480 MHz due to the modified Tx arrangement. This is likely almost where that filter board on the mixer output was tuned but I was unhappy with attempts to peak that board so I removed it! Next, using a short length of hardline removed from the filter assemblies (see previous post) I took an IF signal from the output of the amplifier on the Rx mixer output to the 'output' of the first transmit mixer.
(Ideally some sort of combiner resistor arrangement should go here but this was a rough 'proof of concept' mod.....)

This means I am using the first Tx mixer block (an ADE device) 'backwards' like we do in many of our other transverters as they are bidirectional. This results in a 28 MHz signal on the 'input' which goes to the FLEX. It works, apparently reasonably well too.... I need to add some sort of Tx switching arrangement, probably the usual DC up the coax and I will probably arrange to turn off the receiver mixer/amplifier during transmit to avoid complications! This whole exercise took less than two hours to do including all board work (reprogramming, removing cans, coaxs etc..)
(I needed some light relief after spending about four hours fixing one of my spectrum analysers!!!)

So there it is....use a 28 MHz IF if that suits your need! I will include more mod instructions, the software and results of testing in the next version of my TRI3400 Spreadsheet...out soon

Cheers Tim VK5ZT

[VK2KYP]

Hi All,
The power amp in the 3.4 GHz panel consists of 2 x HMC327 mmics. The drive is split for each device using a 90 degree splitter and the recombined after being amplified.

My question the driver has the marking "A62" is it known what this device is?

Thank you

73 Gary vk2kyp

[VK3PF]

Hi Gary,

I will take a stab in the dark: possibly a SGA-6289.
A Google search for "MMIC A62" returns lots of possible devices, but the Sirenza Microwave devices are at the top of the list. Given that the HMC327 is rated as 30 dBm saturated output each, or +27 dBm each at P1dB, the pair would give 1 Watt out at P1dB. Typical gain of the HMS327 is listed as 21 dB, so requiring a signal of around +9-10 dBm to drive the input splitter (allowing for a little loss).

The SGA-6289 seems to fit the bill: P1dB of 15.6 dBm at 3500 MHz at about 10 dB gain.

Others may have other suggestions..... Mine is simply an educated guess.

Good luck,

Peter VK3PF

[VK2KYP]

Thanks Peter, looks like this is the right part.

73 Gary

[VK4EA]

Tim,

You weren't kidding about these things being microphonic, wow!

[VK4EA]

The attachment IMG_1324.JPG is no longer available

First panel nearly finished, just in time for our 23cm on 23rd MAD this Saturday. Tim's statement re the green panels are better is quite true.

The attachment IMG_1329.JPG is no longer available

The attachment IMG_1327.JPG is no longer available

[VK3KGR]

Receiving my test beacon I get a S9 signal when I first power up the panel. Once I have transmitter once , even just a short PTT, the signal goes down to S8. If I cycle the power it goes back up to S9. Any ideas?

Geoff VK4KJJ

[VK4EA]

I thought I had finished a panel for our 23 on 23 day yesterday, all tested out fine on the bench with a signal source, output looked clean on the SA etc, etc.

Got to site and setup, never got a chance to fire up the panel before a gust of wind knocked it over, and now on the bench there appears to be a dead short on the 12v supply (across the electrolytic cap). I've disconnected the switcher PCB, inspected the board for anything obvious. So before apply 12v with no current limit to see where the smoke needs to come out, has anybody come across this? I recall Tim talking about a component letting go after a lot of power cycles, does anybody recall which bit fails?

[VK3ALB]

I thought I had finished a panel for our 23 on 23 day yesterday, all tested out fine on the bench with a signal source, output looked clean on the SA etc, etc.

Got to site and setup, never got a chance to fire up the panel before a gust of wind knocked it over, and now on the bench there appears to be a dead short on the 12v supply (across the electrolytic cap). I've disconnected the switcher PCB, inspected the board for anything obvious. So before apply 12v with no current limit to see where the smoke needs to come out, has anybody come across this? I recall Tim talking about a component letting go after a lot of power cycles, does anybody recall which bit fails?

OUCH Peter. Well you'd think it was a mechanical problem if it happened after a fall but I have had one dead LM2676 - you could try there.

[VK1JA]

Has anyone come across any panels with dead rx pre amps?

We have two panels showing the same fault. Feeding a 3.4Ghz signal into the antenna yields no signal, but feeding it straight into the beginning of the mixer stage (after the rx pre amp) the signal is present. Even waving a probe around the mixer stage with 3.4Ghz on yields a signal. By passing the filter that's been placed in line still has the same result.

Strange that both panels have the same fault. I'm yet to lift the lid on the pre amp stage but has anyone have any experience with the rx pre amp in these panels, any possible faults and ways to test?

[VK2JDS]

Hi Jayson,
closely examine the way you have soldered in your rx filter, there is a chip component nearby and if you accidentaly removed it when getting rid of the old rx filter then that would give the symptom you have.
i think the track goes inside the pcb and across into the preamp.
If you cant find a suitable picture from the ones posted in this thread of the rx mixer area i can post one up for you, otherwise pop off the preamp can and see if you can inject some 3.4g into that area. the dc may be missing , or the track has gone open from the output of the lna into the mixer block. Something to think about if its an orange pcb
73 Dave

[VK2JDS]

Just clarifying the rx fiter removal method i use.
A small oxy/gas CIG torch on a low gas fine point blue flame aimed right in the middle of the filter will quickly release it from the pcb. It provides sufficient heat to release The main earth substrate join, while allowing the pins to soften nicely without track damage. It will easily flick away carefully from its position using this method.

73 Dave

[VK1JA]

Hi Dave,

Thanks for the info.

I'd had a look at the pictures in the how to instructions, and it appears that all components are still intact. Although I've uploaded a picture of the panel in question - if you can see anything glaringly wrong please let me know! I've marked in red where I have soldered the in/out of the filter.

Also, would you know what spot in the preamp block is the final output that leads into the mixer block - and what spot is the input (from the preamp block) in the mixer block?

Regards

Jayson

[VK5ZT]

Hi Jayson

Expect the unexpected....where these boards are concerned!

I have marked up your photo with where I think the track goes (from memory, I am at work!)

Also, check DC on the big SM resistor. One side will have 5 or 6 volts on it...just to make sure the preamp is powered.
I have not had any issues with the preamp section but with these boards anything can happen!!

I can do a few typical voltage measurement easily tonight if you need any help...

Cheers

Tim VK5ZT

[VK1JA]

Thanks Tim, this will be a big help.

Tonight I'll make sure there's power getting to the preamp stage and test for conductivity between the two spots you've marked, as it 'may' have gone open when removing the can.

If you could make a few typical voltage measurements it would be great help for everyone, especially me

Regards

Jayson

[VK4EA]

Tim, while you are at it, I have one panel that is not TX switching, the signal is getting to the comparator, but nothing changing on the output pin. I cannot see any 5 volts on the comparator's supply pin, does it come from the 5v reg near the TX PA?

[VK5ZT]

Hi all

Here's a few voltages around the RF/Mixer stage of the board.
I also confirmed the connection between the RF and mixer sections (thin black line).

As for the Tx switching.....

The low dropout 5V regulator supplies power to the comparator IC. Note that +5 is on Pin 7, output is pin 6 which is LOW for TX.
Pin 3 should be around 2.5 V and pin 2 around zero until DC is applied via the resistor. When pin 2 volts exceeds pin 3 the output goes low.

(Note that in my mods where the micro controls the TX, I bias the comparator into TX mode then rely on the other control input to that circuitry from the micro....)

Any questions...just ask.

Tim VK5ZT

[VK1JA]

Thanks Tim, this will help a lot!

Also, for the -50dBm input (@3400Mhz) when testing/peaking the rx, what sort of output is everyone getting at the 444Mhz side to the transceiver?

Or, with -50dBm input what rough S reading are people getting on their radios - noting that S meters on radios can be vastly different (lets not get into THAT debate ).

I suppose what I'm trying to get at is it would be good to know if the rx is on par with other panels.

[VK5ZT]

Jayson

-50dBm is a HUGE signal! Your guess meter should be full scale

You should be able to see signals down below -100dBm when all is tuned and working correctly.

I can't easily generate an accurate level at 3400 but I can measure levels (on a spec an) there. As such I know the second harmonic of my sig gen at 1700 MHz is about 40 dB below the fundamental so I can get a reasonable idea and I know the transverters can see signals below -100dBm

Cheers

Tim VK5ZT