Elcom Synthesizers

[9A4QV]

Hello Guys,

I am glad that I am not lonley spending time (and money) on this "bricks"
To be honest, I found not so much informations regarding the Elcom units and most of them were wrong, so decided to take the electric screwdriver in hands and start to explore and measure this units. Till you stay with the DFS units you can do it also by hands, but if you try to explore the DFSL units you know how many screws are on the RF side of the brick (88 pcs).

Just to let you know that from the 10 units I got (DFS1201, DFS1301, DFS1295, DFSL1201 and MTS1500) just three of them working properly, the others are money wasted...
Even among the same type DFS1201 there are differences, and not working with the same code. Elcom ceragon will work with the code published (if the bricks are not faulty), but the same code will not work with the DFS1201 Elcom Giganet units, even they are same inside. Then, there is a difference between the same units but the dash 0 and dash 1 type, the code is not the same...

So I am close to realise other idea. I am planing to use just the RF chain (VCO plus multipliers) and use my own PLL, or replace the the VCO and the PLL with my own and use just the multipliers. This will take less time, money and energy than finding out the scheme how this bricks are working.

Adam

[VK3HZ]

Hi Adam,

The units I have examined up close have used the ADF4252 PLL driving a discrete VCO, all on the "RF" side of the unit. I understand some others use a different PLL. The control lines (Data, Clock and Latch Enable) come through to the other "Control" side using wire "vias". It would be relatively easy to put your own PIC/ATtiny/PICAXE in the "Control" side driving those lines and programming the PLL to whatever you want.

BTW, it's worth noting that Doug had to repair two faults in a unit that looked factory fresh. Therefore, some of the units on eBay may be factory rejects or failed units swapped out of existing equipment - worth keeping in mind when that unit just will not program the way it should

Regards,
Dave
VK3HZ

[9A4QV]

Hello Dave,

I agree,
it is clearly stated by the vendor that he does not guaranty that the units are working properly and they are not tested, so it is on us to decide to buy or not.

One of mine DFS1201 lock fine, a diode blink and then lock, as it should be, then after some half and hour the unit is warm, the lock is OK, but if I switch it off it will not lock again till the unit is cooled down. Same thing cover on or off.
I think I have ti use some kind of FLIR camera to find the faulty part??

Most of the units are working off the frequency due to the crystal aging. Yes, this can be solved by tuning the small trimer capacitor but sometimes even this is not possible. This can be the indicator that unit is worn.

And, as the others noted, the programming frequency schemes are different, x2, x3, x3+1024, /5+1024

It seems like every unit is different.

Adam

[VK4OE]

Greetings to all again!

For some time I have been frustrated by the specification sheet for the likes of the Elcom 1201 synth (10-pin SIL connector type) which description does designate Pin 8 as the 'lock' pin, but that no significant change occurs on that pin (voltage or current) from switch-on to there being a frequency locked output. Having a LED that illuminates when the synth is frequency locked is what is needed by real people here on planet Earth. This LED should extinguish at any time the 'locked' state is lost but, as is, these Elcom synths don't ever change the electrical status of Pin 8 (except for a couple of very short bursts of data during the unit's boot-up sequence).

Well, Dave VK3HZ and I have put our heads together again and have come up with the following report and suggested modification.

The microprocessor on the control side of the unit 'talks' to the main PLL chip on the RF side by way of at least two of the five soldered-wire 'vias' that each synth has. These 'vias' go through the casing from one side to the other and they usually have yellow PTFE insulation as they pass through. They are distinctly different from the many DC links through the case by way of feed-through capacitors, and they are easily identified.

The 'via' that we're really interested in is the second one as viewed from the 'control' side of the unit. It's the 'DATA' pin and it does different things at different stages of the Elcom booting up and getting going. To possibly confirm which pin is being described, in both Dave's test unit and the unit I have modified, there is a 270 ohm resistor connected and located close to this second 'via'. We presume that many of the Elcom synth models are similarly configured in this area, but we obviously can't make a categorical statement - that many models do vary has already been demonstrated!

As another bit of preamble, in both 'editions' of the Elcom synth that Dave and I have been investigating, there is an SMD bipolar transistor the collector of which connects to Pin 8 of the unit's main input connector by way of a 47 ohm resistor. There is a 4K7 resistor from this collector to a +6V printed via that emerges through the board.

Dave and I have come up with the following modification sequence on the control side of the unit, as illustrated in the following image:

1. Cut two PCB tracks as visible in the picture (refer also to the second pic).
2. Solder a short piece of small-sized insulated wire from the DATA via to the 10K resistor that goes to the base of the SMD bipolar transistor.
3. Solder an SMD P-Channel MOSFET across the existing 4K7 resistor. (Many types will do the job here and you could even use a leaded one but SMD is more appropriate in the physical situation). The 'Source' goes to the +6V point and 'Gate' goes to the Collector of the existing bipolar transistor).
4. Solder a short piece of small-sized insulated wire from the 'Drain' of the P-Channel MOSFET to the bypass capacitor on the way to Pin 8 of the main connector.
5. Replace the 47 ohm resistor that goes to Pin 8 with a larger value - say 330 or 470 ohms.

Here is a close-up picture of the addition of the SMD MOSFET and other modifications in this area:

A LED can now be connected direct from Pin 8 to ground and it will illuminate soon after the unit is powered up, it will have a couple of almost imperceptible flickers during boot-up, and it will stay steadily 'on' unless there is a loss of 'lock'. This can be verified by simply disconnecting the 10 MHz reference (presuming that the synth has been modified for use with an external reference!). The locked output should recover itself when the 10 MHz reference is re-connected.

This indication is not everything that everyone would want - it doesn't tell us the moment when the PLL locks up on the programmed output frequency, but it is better than no indication at all!

Short of reading the data in the two brief data bursts during boot-up, this modification should be useful to many users of these interesting microwave sources!

Very best 73,

--Doug Friend, VK4OE.

p.s. The pics were taken of a model ILCDFSL1201 synth. There may be variations from other '1201 synths, but probably not so much in the area of the 'control' side that relates to these suggested modifications. It is very likely that the mod's will remain applicable to other '1201 units.

[9A4QV]

Hello guys,

I went quickly through the standard 12F675 controller that most of the guys are using and the PIN 8 was used exactly in the same way, just having the NPN transistor with the LED on the same PCB where is the 12F675.
Just to add that the CDFSL-1201 is different then the DFS-1201 unit. The latest Giganet 1201 units that can be found on the e-bay are the same as the Ceragon DFS-1201. I have a couple of them not running at all and some of them not locking or losing the lock when running for half an hour. Found the cause, there is the MMIC amplifier in the PLL chain just after the lowpass filter before the PLL chip. This mmic is problematic, oscilating, or not working. When replaced with the simple MAR-6 and changed the bias reistor the units were back to proper operation. Found such a problem both on Ceragon and Giganet units.

What I found out, the sign that something may be odd with the unit is how fast the LED blink and return to ON state after the unit has been powered. If this blink is realy short then the unit is OK, but if the blink is a bit longer than you may have some problems with the PLL and mmic in question.
I have checked some 20 Ceragon/Giganet different units, and this is the clue that I pull out.

Have to say also that the Art in Part guy send me a replacment for the one unit I was not able to repair, free of charge. He also looked to help about the MTS-1500 YIG unit that is not working properly here but at the end we found out that this one was purchased through the other supplier from the same country, so not his fault.

I have now DFS-1201 running also on my 47GHz transverter, quite handy. On the other side the 3.33MHz step does not allow the in band 47GHz beacon frequency. For that, the MTS-1500-01 should be just perfect...

Adam

[VK4OE]

Greetings to all!

It seems to me as if the available supply of useful Elcom synths is gradually drying up (or, for some models, has already dried up), but I am guessing that quite a few are still stashed away in drawers waiting for project impetus to happen.....[Hey, if anyone has a '1295 model that they don't want any more, please get in touch by direct means!]

Whereas most of my previous postings were about the Elcom model 1201 and its many derivatives, this posting is about Model 1301, which is designed to cover 12.65- 13.35 GHz. And, yes, this model is still sometimes available. And I have to admit that this posting is a gathering together of bits of information that I have gleaned from various sources, notably the practical experiences of Peter VK3APW for whom I have programmed a couple of PICs to work with a couple of models of Elcom synth.

The short story is that, due to its internal RF architecture, model '1301 doesn't have the 3.33333' MHz minimum frequency step that the '1201 model has, but that its minimum frequency step is a neat 5.0 MHz. This means that the .ASM file used to start the programming process needs to be different in format from G4FRE's original one which has been successfully used many times by many folk around the planet.

An internet search revealed that an acquaintance of mine, Barry Chambers G8AGN in the UK, had successfully got his '1301 synth operating so I asked him and he has graciously provided his edition of the modified 'ASM file to be used with model '1301.

[I tried to add this particular .ASM file to this posting but it was not allowed as an attachment to this posting. So, if anyone wants a copy of it please just ask by one or other direct contact method.]

Of course, in exactly the same way that we adjust .ASM files for programming the '1201 synth, this alternative .ASM file needs to be adjusted for your chosen frequencies in the '1301 case. That's an easy exercise after you recognise that you're dealing with a number that is the desired output frequency x2 and not x3. And this is the reason that the minimum frequency step is 5.0 MHz and not 3.333 MHz.

Another interesting point that came up via Peter VK3APW's experimentation is that, whereas the '1201 (and other models as well) work happily with a +8V lower voltage rail, the lower input voltage rail fed to the '1301 needs to be only +6V. It hardly ever obtained 'lock' with a +8V input voltage but, with +6V supplied, 'lock' was reliably obtained. Interestingly, it was radio amateurs who have recommended an increase of this rail voltage over what's clearly written in the original Elcom data sheets and, that has been OK for many models. But the experiences obtained with Peter's particular unit suggest that, for the '1301 model, the manufacturer's recommendations on this point should be adhered to.....

Continuing with the results of Peter's experiments with his '1301, I had added 12600 MHz to his desired 12885 MHz when programming the PIC for it. Peter reported back that the 12885 MHz output was good, strong and stable but that, whilst it still 'locked' at 12600 MHz (which is 50 MHz below the designed lower frequency limit for the '1301), the output spectrum had noticeably worse phase noise (assessed by the noise pedestal around the output frequency as viewed by a close-in look with his spectrum analyser, of which some noise contribution may have been due to the SA itself). Furthermore, the 12600 MHz output had weak but noticeable close-in sidebands (~ +/- 1.5 KHz) that were not present when viewing the 12885 MHz output under the same conditions. Peter further observed that those sidebands could be minimised by adjusting the main +12V rail voltage to +10.5V. There's further room for experimentation there!

I trust that these reported results and comments may be useful to other experimentally-minded microwave radio amateurs, wherever they may be!

Cheers and 73,

--Doug Friend, VK4OE.