BAY96 70cm to 23cm trippler minutarization

Hello, this design ../imgqa/eboard/Antenna/rf-etivgpqhuie.gif ../imgqa/eboard/Antenna/rf-jgll0ab5dl0.jpg is a very well known to the hams before the microwave active devices were available at low cost. It can output quite a few Watts as well.

I wonder how can I minutarize it so that it can fit inside a rubber antenna.
I am planning to build a plug in antenna to use with my UHF handheld, so that it can TX/RX on 23cm.

Because this is a passive device it is suitable for fitting inside the antenna without the need for extra power.
My transceiver Kenwood TH-F7 has an FM deviation setting and I can additionally speak away from the microphone, if I find FM deviation is too high on 23cm.

The only thing I need is your help/Ideas on making the whole thing fit inside a rubber antenna, so some king of filters minutarization is needed.
For RX I think a simple capacitor to bypass the Trippler with some kind of pin diode to block feedback on TX, will do? (TH-F7 can receive 23cm).

Also what other diodes apart from BAY96 can I use? MAX power input is 5W, but probably I will feed it from 1W or so.

Any help is appreciated.

The problem with shriking that design is the box itself is part of the tuned circuit. Using lumped LC will probably not work well. Also bear in mind you will be dissipating ~0.75W of heat in the diode, enough to worry about in a very confined space. I don't know of any equivalent to the BAY96 which has been obsolete for many years and is very expensive from 'old stock' suppliers. You might be interested in this article which uses transistors instead of varactors though. I would up-mix if I were you!

Brian.

Thanks for the article Brian. The problem in these approaches is mainly the output power capability (which is too low) but mainly the need for a VCC. The BAY96 solution is a passive one, consumming energy only from the carrier (inefficiently). But since 1-5W are already available on the transceiver connector, that is not much of a problem.

I was thinking about the heating of the diode too. This diode can handle 15W of power so 5W or even 1W that I am planning to use won't heat it too much I think.

Obsolescence is a problem, aren't there any similar power diodes are produced today, even capable to handle only 1W or a bit less instead of 15W?

What worries me, is the schematic. It shows the diode connected with the anode to the enclosure (ground), but in the pictures I find, the diode is screwed to the enclosure, thus connected with the cathode to the Ground. Which way is correct?

As far as I see, the biggest space is consumed by the input uhf filter (LPF?). Can't I just use one of these mini circuits smd ceramic LPFs? They are rated to 5-10watts. The same applies at the output, can't I just use one of these mini circuits SMD BPFs?
All I see after all from the schematic, is a UHF LPF at the input and a 23cm filter (LPF?) at the output, and the diode between them, am I right?
In other words, what will happen if I try to feed RF directly to the diode and the parallel resistor without any filters?

I am sorry for asking too many things, but I am trying to find out how to do it, sonce I already have this diode and it will be a pitty not to "add another band" to this radio.

The filter is still primarily a lumped LC filter and could be replaced by a smaller SMD design, I believe. If you operate the tripler with 5W level transceiver output, you need to heat sink about 2.5W.

BAY96 availability is the major problem, and I won't be too sure that the devices offered here and there at Ebay are worth the money. It should be also mentioned that BAY96 isn't really optimal for 1.3 GHz, you would look for a smaller capacitance (and also smaller power) device.

Passive multipliers were useful in early days before faster devices were easily available but in terms of efficiency and cost they were never very good. You can use almost any diode as a multiplier, the BAY96 is just a high power rated varicap, optimised for multiplier use but any fast switching diode will work to some degree. You might find some newer devices work better than the BAY96 which is probably around 30 years old by now.

The schematic you posted is wrong, the diode is shown reversed. The cathode is the body and the anode is the tag. I've attached the data sheet for you.

The input filters are necessary I'm afraid. Their purpose is twofold, firstly to ensure only the TX frequency reaches the diode and secondly to ensure the multiplied frequency doesn't leak out back to the TX. In a conventional mixer or active multiplier there is an input port and an output port with circuitry between them. In a passive multiplier there is only one connection and the diode has all the frequencies across it simultaneously. The filters are there to isolate the frequencies and the values are picked to ensure the filters don't load each other as they are in parallel. You might be able to use minicircuits BPF modules but you will have to confirm their parameters (S22) at the multiplied frequency as well as the input frequency.

The diode will run hot without a heat sink but at ~0.75W it should survive.

Brian.

(Edaboard is allowing me to upload the data sheet - if it's missing I'll try again in a moment)

A link to the datasheet
http://www.mikrocontroller.net/attac...AY96_Valvo.pdf
As the diode is self biased, polarity doesn't actually matter in this case.

Thanks FvM. I've got internet problems again!

Brian.

Thanks Brian and FvM for your replies, all being very helpful!
From the datasheet it seems this diode is purposively designed as a multiplier.
If I use this I may make the cylindrical enclosure of the circuit from machined aluminum and attach the diode to it as a small heatsink. I won't use more than 1W input anyway.

Can you think of any idea of how to bypass the multiplier circuit on RX passively?
My transceiver can receive FM on 23cm.
I was thinking a small value capacitor at the output of the trippler, that is connected to it's input. The UHF LPF at the input should prevent 23cm feedback to the diode, and since the transceiver will be in TX mode, there would be only VSWR to it.
On RX, this capacitor will couple the signal to the transceiver.
Maybe I should connect an anti-parallel diode pair somewhere.

Any ideas?

At those frequencies, even a small capacitor would wreck the filter characteristics.
I've used these at 1GHz before with good results: http://www.te.com/usa-en/product-1462052-3.html
They will give better isolation and less 'on'attenuation than PIN switches. The pins are arranged to allow direct soldering to 50 Ohm tracks.

Brian.

I think that I found a simpler solution.
TH-F7 transceiver, like most commercial ones, have a low pass or band pass filter just before the antenna. I checked this on the service manual. Thus no input LPF is needed.
Based on this, I guess I can directly connect the diode and the parallel resistor at the output of the transceiver?
After the diode and resistor, I could connect just a 23cm BPF, either plig in or coaxial type, and finally the antenna.

So the whole configuration will be: UHF fIlter inside the transceiver -> Diode/resistor -> 23cm BPF -> Antenna.

On RX, no change need to be done, since RF will come through the 23cm filter (with some attenuation, based on the filter loss) to the receiver. Since RX signal is of low level, it won't be able to make the diode conduct it to ground.

On TX, the multiplied signal will try to go to the antenna and to the LPF of the transceiver, which can't pass through.

How do you find this simple setup?

The TH-F7 has a complicated output network that works at 2m and 70cm band frequencies without any switching. You can of course try without any filters but I suspect the results may be poor. The important factor isn't how well the 70cm gets out of the antenna socket - the manufacturer has taken care of that, it's how high it's impedance is at 1.2GHz when looking back into the socket. I would guess the 2m and 70cm output network are not going to isolate 1.3GHz very well.

The only way to find out is to try of course.

Brian.

What does it actually mean? My guess is:

1.3GHz will be driven not only at the antenna but also back to the transmitter, leading to:
1. Less radiated power (how much less?)
2. Heating of the final amplifier due to vswr (even if the reflected power is at a frequency other than the amplifier was designed for?)

I said 1W max input power but I see the transmitter in the low setting is 0.5W (battery operation), so 0.5W will be the maximum input power to the trippler. Do you think that in the case of high vswr on 1.3GHz, this would fry the final V/U amplifier?

My personal guess is that if the 5W capable transmitter is operated without an antenna at 0.5W, it will not be damaged, what do you think?

PS. I am only thinking of this solution (no LPF at the input) so that it can be easily used on RX as well, without any switching. Since 0.5W is the max input power, I would love to find any other diodes that can be used for the purpose, they do not need to be that powerful as the bay96?

VSWR isn't the problem and I agree that such low power is unlikely to damage the PA transistor whatever the load is. The problem as I see it is you want maximum power at 1.3GHZ but you would be connecting a tuned circuit at 145/435MHz directly across it. Such a detuned circuit, especially one designed to eliminate harmonics, would almost certainly 'short out' most of the 1.3GHz signal. The original input filter isn't just to prevent frequencies outside of 70cm reaching the tripler, it is to block the 1.3GHz from being heavily loaded by the TX output circuits. L1, L2, C1,C2 and C3 are to block 1.3GHz, not to accept 70cm.

You can try other diode types, as I understand the varactor operation in this circuit, it releases energy at the higher frequency because it's charge changes as it's capacitance varies in sympathy with the voltage applied across it. The same will happen with almost any diode but to allow it to work at microwave frequencies it need to have low initial capacitance but a relatively high V/C ratio. You would have to experiment to find which diode worked best but I would start with silicon signal switching diodes with fast recovery time.

Brian.

Thanks Brian.
I am looking about the output filter now. Obviously it has to be small in order to fit in the rubber duck antenna of the handheld transceiver.
After a lot of search I have found something that looks promising and can be made easily at home http://dl4xav.sysve.de/0.5lambda.fil...da.filter.html
I do not expect a very sharp responce, but this is my estimation, I have not measured anything.

I am thinking of using a semi-rigid coaxial and solder it's outer conductor to the inner of the tube. Then bend it's inner conductor at 90 degrees and solder it on the resonator line. I am thinking this, because it would be convenient to take the I/O from the sides of the tube. Ideally the semi-rigid should be soldered outside the tube though and I do not know how will this affect the responce/resonance.

Other than that, the dimensions of the filter are ideal, because it fits perfectly in a rubber antenna.

Construction would be more complicated than the diagram suggests. To solder the semi-rigid co-ax to the inner of the tube (or outer for that matter) requires a lot of heat so you would have to be very careful not to damage anything. Even just soldering through a side hole in 15mm pipe would be hard enough! It might be easier to very slightly insert the tube of the co-ax through a tight fitting hole and solder the outside to the pipe instead.

I made an SWR bridge using similar methods many years ago but made the stripline on a PCB that slid in one end of a 15mm pipe, you might be able to adapt my method, it makes construction far easier because the capacitor can be mounted on the PCB and adjusted through a small hole. Similarly the co-ax inner could pass through a small hole in the PCB and be soldered through small holes in the 15mm pipe.

Brian.

Thanks!
I have tried to solder stuff on copper pipe before and I know how hard it is. It needs hot air preheating if you want to do it with the soldering iron.
On 1296MHz the tube is quire small though.

The attachment show a similar filter made of square tubing instead. The interesting thing is the tuning capacitor.

Since the capacitor is shunt to ground and since 1-10pF is needed on 1296MHz, a brass screw can be used, forming an air dielectric capacitor. A little bit of experimentation is needed though to get the desired capacitance, maybe a small nut at the end of the tube, or a copper penny, is needed to achieve the desired capacitance.
But it is much easier to make homebrew and no expensive and hard to find capacitors. Not to mention that the mechanical stability is better.

Have you got any idea of the formula used to calculate the filter dimensions in different frequencies?

What if I increase the length of the cavity a bit, will this decrease the frequency at all?

I am also not sure if the tap points need to be perpendicular wires to the center conductor, or if they can leave the conductor at an angle.

I've used the brass screw trick many times, not only to make capacitors but as the inductors in interdigital filters too. If you make a small disc of brass and solder it to a threaded brass rod it makes a good capacitor plate. I suggest threaded rod rather than a screw because it's easier to fit afterwards, a screw has to be fitted through the pipe and have it's securing (soldered) and locking nuts in place before you fit the disc to it. It's essential you solder the disc at 90 degrees to the shaft so it turns parallel to the tuning wire or it will be difficult to adjust, the capacitance will go up and down as you rotate it!

I would keep the connecting wires perperdicular if possible. It doesn't make much difference to the signal but the more they are in line with the tuning wire, the more they increase it's effective inductance. It makes the lengths more difficult to calculate. I would suggest using a CAD package for more exotic designs. The same LC laws apply to a great extent but the effects of nearby conductors, particularly the enclosure/pipe become more important when their distance from the tuned circuit is small, particularly when less than 1 wavelength away.

Brian.

I was thinking of a coaxial design. As the output of the transmitter is likely to be capacitive at the higher operating frequency, you need a shunt L to resonate it at the output frequency, so a half wave line (at the output frequency)should then provide reasonable rejection. So you need to split a bit of 22mm copper pipe and assemble all your circuit in one half of it with the other half re-bent to overlap the bottom bit. Perhaps with some nuts soldered on the inside of the bottom bit with the top cover held on with screws through them.
Frank

Here is the mechanical construction of the tripler antenna I am thinking of.
The relative dimensions of the components of the antenna are more or less scaled ok, just to get a feeling of the relative size of the antenna.
The top antenna is a sleeve dipole that is used in many Wi-Fi rubber antennas. The diagram is not quite right, the velocity factor of the cable has to be calculated for the ground element.
I believe, in addition to the filter, the dipole, being resonant, will help as well in radiating the frequency of resonance and reject (vswr) the other varactor byproducts?

you really need to use that "sleeve of copper braid" to be your heat sink. otherwise the diode will quickly fail. It is just a silicon diode, and the gold metallization will diffuse into the silicon body if the temperature gets too hot, and the diode will eventually short out.

I would mount the diode right at the junction of the shield and radiator element (a virtual ground point). I would put the matching network as close to the diode as possible if it is a broadband use, but if narrow band the matching network can be further away, like at the bnc connector to the antenna. If there is line length between the diode and the matching network, obviously the matching network design may have to change to take into account the transmission line effect.

and as far as the suggestions to find a lower capacitance diode...well that would also mean a smaller diode diameter, and hence a much higher thermal resistance, and it would probably overheat again! so maybe stick to this diode if you are planning on hitting it with watts of energy. They do make (or at least they used to) diodes with small capacitance and good thermal resistance, but they were pretty exotic (used diamond heat sinks, etc) and would be hard for a ham to find.

Also, at those power levels, you would need a high breakdown voltage in the diode to keep it from clipping.