I?m developing a simple, five-frequency superhetrodyne receiver to operat e between 2.5 MHz to 20 MHz in which the local oscillator crystals and whip antenna tuning circuits are selected (switched) using diodes. Here?s the problem:
The whip tuning circuit is comprised of a single variable capacitor with fi ve fixed-value inductors, each in series with a BA277 "bandswtiching" diode . A single diode/inductor combination draws about 6 mA when ON (switched in ). This works fine, but at and above 15 MHz, the BA277 diode significantly contributes to the series resistance of the inductor, inducing a 3dB "inser tion loss" (as compared to the same circuit with the diode short circuited) .
The BA277 datasheet indicates that at 6 mA and 100 MHz, the series resistan ce of the diode should be much less than 1 Ohm. No performance data is give n for any other frequencies. My bench measurements indicate that at 15 MHz, the diode presents much more resistance than that (probably tens of Ohms). No surprise there.
QUESTION: The BA277 works OK for what I?m doing, but is there a bandswitc hing or PIN diode out there better suited for operation below 30 MHz?
ate between 2.5 MHz to 20 MHz in which the local oscillator crystals and wh ip antenna tuning circuits are selected (switched) using diodes. Here?s t he problem:
five fixed-value inductors, each in series with a BA277 "bandswtiching" dio de. A single diode/inductor combination draws about 6 mA when ON (switched in). This works fine, but at and above 15 MHz, the BA277 diode significantl y contributes to the series resistance of the inductor, inducing a 3dB "ins ertion loss" (as compared to the same circuit with the diode short circuite d).
ance of the diode should be much less than 1 Ohm. No performance data is gi ven for any other frequencies. My bench measurements indicate that at 15 MH z, the diode presents much more resistance than that (probably tens of Ohms ). No surprise there.
tching or PIN diode out there better suited for operation below 30 MHz?
Dave, I don't know much about what you are doing. But isn't the dynamcial resistance of the diode the thermal voltage divided by the current? so ~25mV/6ma ~ 4 ohms At 20 MHz you could use a swtich or a relay.
The problem is to find a PIN diode with a low on resistance and a long carrier lifetime, so it keeps conducting for a long time when current is flowing in the back direction. If the signal is slow compared to lifetime, it acts like an ordinary diode.
There's no carrier lifetime specified for the BA277. It's probably pretty short.
A little telecom type relay is a good idea.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
Motorola, GE & RCA used DC switching for crystals in business radios for decades. You have to turn the diode on hard, or it causes all kinds of problems. A relay is unreliable, if you need good accuracy. The only manuals I have left are for synthesized business radios.
You've got an 8:1 range in frequency, which, working against your constant capacitance, means an 8:1 range of impedances in your inductor. That's just going to make life difficult for you.
If you could find a two-section capacitor you could ease up on your variation in impedance.
You're probably also getting distortion from the PIN diode, too -- at low frequencies it'll distort like a diode, instead of acting like a resistor.
Suggestions:
I've seen power diodes like the 1N400x series suggested as HF PIN diodes. You're using them "off label", so you can't be sure that you'll get what you want in every one (at some point someone will decide that last year's fast recovery diode exceeds all the specs and will start printing "1N4005" on it -- then you'll be screwed). But it may be worth a try.
You may consider a different kind of switch. Diode switching works great in places where the dynamic range is known and limited. That's not the situation in a radio front end. If you can find any that make sense from a capacitance point of view, you may be able to use JFET or MOSFET switches.
(I'd be inclined to use a 5-position two-bank rotary switch, myself. Or relays).
--
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com
Ok, that's the receive frequency range. What the crystal oscillator frequency range?
Below about 10 MHz, a cheap 1N4148 diode is good enough. You'll get about 0.6 ohms series resistance. If you're using a parallel resonant crystal, the usual worst case crystal Rs is about 50 ohms. A commodity PIN diode might get that down to 0.3 ohms at 14ma. A few tenths of an ohm isn't going to make much difference.
Here's a schematic of a direction finder (Homer) I helped design for the USCG in about 1977. The crystal switching is about as simple as it can be: About 14ma forward current. However, this design used an HP 5082-3168 PIN diode because the USCG wanted to use over the counter crystals from multiple sources (and other radios). A good substitute for that PIN diode is a 1N5719. The commercial version of the Homer used
1N4148 or Motorola MPN3401 PIN diodes:
How are you matching the whip antenna to this front end filter (presumably an LP or BP filter) or is the "tuning circuit" just a matching circuit? The short whip antenna is going to need a large series inductor, which will tend to have low Q and high losses.
Yep. That's what happens when you switch at low impedances. The Rs of the diode switch is a substantial percentage of the input impedance, resulting in plenty of loss. If your BA277 is 1 ohm, and your loss is 3dB, then the matching circuit must have an input impedance of 1 ohm. With small coils, you could easily have 3dB of loss in the matching network due to low coil Q, even without the diode. I'm not going to try and figure out where the loss is hiding without a schematic.
The way I see it, you have 2 choices:
Redesign your input matching network or filter network so that it runs at a much higher impedance making the diode losses less important. That usually means a broadband transformer at both ends of the matching network or filter network.
Use magnetic latching relays like most of the HF radio vendors have been doing for years. Lots of benefits such as lower average power consumption. PIN diodes draw power all the time. Magnetic latch relays only draw power when changing state. Photo of Elecraft mini antenna tuner: Those are magnetic latching relays.
Yep. You're probably measuring the diode lead inductance.
I don't know. Going from a junk diode with maybe 1 ohm of series resistance, to a fancy PIN diode with maybe 0.3 ohms, isn't going to do much if the filter impedances are all 50 or 75 ohms. However, it will have a big effect if the filter impedance is very low in order to match your short whip antenna. Personally, I would recommend trying a
1N4148 as a sanity check. If it works the same as the PIN diode, you have a problem elsewhere in the design.
Also, at HF frequencies, atmospheric noise dominates the receiver sensitivity. You could have really great receiver sensitivity, and as soon as you plug in an antenna, all you hear is amplified noise. You can blame all the lightning in Florida for that. So, it doesn't pay to squeeze too much sensitivity out of an HF receiver below about 20 MHz. Note the atmospheric noise, both for day and night.
Also, pay attention to how much reverse voltage you might be applying to the diode or PIN diode via the antenna. The upper end of the dynamic range is set by when the band switching diode becomes reversed biases, and turns the received modulation envelope into garbage. There are plenty of things that might overload further downstream, but you don't want it to be the input switching diodes.
Good luck.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
It doesn't take much wire to screw things up. I'll make it easy and pretend the diode is connected directly to the measurement bridge(?) or RF analyzer, and that there's only one clip lead involved. The measured 1 ohm series resistance seems about 0.7 ohms higher than I would expect for the PIN diode. If all of it were produced by lead inductance, and ignoring any contact resistance, then: L = Xl / (2 * Pi * freq) L = 0.7 ohms / (2 * 3.14 * 15*10^6 Hz) = 0.0074 * 10^-6 H L = 7.4 nH Finding a handy straight wire inductance calculator: For a 1mm dia wire, it would require only 12mm of wire to produce the required 7.4 nH inductance.
However, he said "tens of ohms" which implies that he's not getting a stable value. That suggests clip leads. If the clip leads really did produce more than one "tens of ohms", which would be 20 ohms minimum, doing the same arithmetic produces 212 nH or a 183 mm clip lead. That's still within reason for a crude test fixture.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
True. I discovered that the hard way. We were using 1N4007 diodes in the T/R switch of a 150 watt PEP SSB marine radio. Someone (it might have been me) substituted some lower voltage diodes in production. They blew up if the RF voltage went high enough to reverse bias the TX diode, and reverse recovery time was too short. I think the 1N4006 has the same structure as the 1N4007, but the 1N4006 was rejected by one of the other engineers for some long forgotten reason.
However, that was in the 1970's and 80's. Companies have changed hands, processes have changed, and far too much junk is in distribution. I'm not sure I would use a 1N4007 these days for fear of dealing with a moving target:
More on recovery time of various devices: with curve tracer photos and comments on using SPICE with PIN diodes.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
How long is the whip ? It is going to have a very high capacitive reactance if it is less than 3 m (1/4 wave) at 25 MHz and much worse at lower frequencies.
Is there a risk that a real outdoor antenna is to be connected to the receiver or is it just whip only ?
There are several ways to connect an electrically short antenna into the receiver.
One way is to put a equally large inductive reactance (loading coil) in series with the actual whip and there is a pure resistance between the antenna and the receiver input. However, the radiation resistance for such short antenna might only be a few ohms, so still some matching might be needed.
Considering the atmospheric noise levels, it might be a good idea to have a resonant antenna (with loading coils) at 20 MHz, but at lower frequencies, there are still going to be sufficient signals available despite mismatch.
An other approach is to use some kind of source follower as a voltage probe. The antenna capacitance of the whip (perhaps 10-100 pF) form a voltage divider with the gate/source capacitance of the FET.
The voltage at the gate is then transferred to the resistor (or emitter follower) at the source electrode, which can be loaded with a typical 50 ohm receiver input impedance.
Of course, the source follower should operate with a sufficient operating voltage (5-24 V), so that combined waveform from the antenna does not clip.
This kind of construction is common in semiportable receivers with antenna input for a real outdoor antenna as well as equipped with a whip for portable operation. The outdoor antenna is connected directly to 50/75 ohm antenna input, but when whip antenna is selected, the whip and source follower drive the receiver input, instead of the outdoor antenna.
What is the capacitance range of this capacitor ? If it is wide, then at Cmax and large inductor at 2.5 MHz or alternatively Cmin and smallest inductor at 20 MHz, the required inductance range is reasonable. However, with a narrow capacitance range, a 1:8 frequency range requires nearly 64:1 inductance range, which can cause problems with strays and impedance levels.
Is this a mechanical variable capacitor or some varicap ?
If this is a mechanical capacitor, I can understand using only one, but still I would consider switching in serial/parallel capacitors (which of course reduces tuning range), instead of simply inductor switching.
With varicaps, why not simply use 5 varicaps and have 5 independent resonant circuits and select between these resonant circuits with PIN diodes. In this way, there are no lossy switching element _within_ the resonant circuit, which would dissipate the resonant current and hence lower the LC circuit Q and hence increase bandwidth.
With separate resonant circuits, you can also tailor the impedance values as needed for each band, considering also the (short) antenna reactance.
Thank you for all the constructive suggestions. A few comments on your rep lies:
Regarding the BA277, the manufacturer (NXP Semiconductors) bills the device as suitable for ?Low loss band switching in VHF television tuners? and doesn?t explicitly call it a PIN diode. If it is one, then compared to most PIN diodes, its carrier lifetime must be fairly long if it can be used at VHF frequencies (30 to 300 MHz).
I substituted the BA277 with 1N914 diodes in my prototype circuits and they ?re definitely more lossy in the frequency range I?m working with (2.5 to 20 MHz). Therefore, the 1N914 and its equivalent, the 1N4148, are unsuit able.
In my application, the capacitance is constant at all frequencies. Consequ ently, as someone correctly pointed out, the reactance of the tuned circuit is low at 20 MHz and so the BA277 diode in series with the inductor notice ably contributed to the loss (lowers the quality factor Q).
Based on your posts, here are the two options I?ll consider:
1) Keep the whip antenna tuning circuit as is (single tuning capacitor, mu ltiple inductors switched with BA277 diodes) and live with the loss. This keeps the circuit as simple as possible.
2) Redesign the tuning network to minimize the BA277?s contribution to t he loss. This can be done with a capacitor (varactor) paired with each ind uctor, and switching these L/C pairs with BA277?s. This keeps the diode out of the L/C parallel combination.
Indeed. In a test lab we had no problem making 4-digit accurate and repeatable (4-terminal) measurements multiplexing scores of devices in thermal chambers with to-5 class latching relays and careful layouts, wiring and fixturing at 10 MHz. Did have to be careful to use consistent cable lengths (80 coaxial conductors and 196 pin connectors).
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510 USA
+1 845 480 2058
hobbs at electrooptical dot net
http://electrooptical.net
plies: Regarding the BA277, the manufacturer (NXP Semiconductors) bills the device as suitable for ?Low loss band switching in VHF television tuners? and doesn?t explicitly call it a PIN diode. If it i s one, then compared to most PIN diodes, its carrier lifetime must be fairl y long if it can be used at VHF frequencies (30 to 300 MHz). I substituted the BA277 with 1N914 diodes in my prototype circuits and they?re de finitely more lossy in the frequency range I?m working with (2.5 to 20 MHz). Therefore, the 1N914 and its equivalent, the 1N4148, are unsuitab le. In my application, the capacitance is constant at all frequencies. Cons equently, as someone correctly pointed out, the reactance of the tuned circ uit is low at 20 MHz and so the BA277 diode in series with the inductor not iceably contributed to the loss (lowers the quality factor Q). Based on you r posts, here are the two options I?ll consider: 1) Keep the whip a ntenna tuning circuit as is (single tuning capacitor, multiple inductors sw itched with BA277 diodes) and live with the loss. This keeps the circuit as simple as possible. 2) Redesign the tuning network to minimize the BA277 ?s contribution to the loss. This can be done with a capacitor (var actor) paired with each inductor, and switching these L/C pairs with BA277 ?s. This keeps the diode out of the L/C parallel combination. Thank s again! -Dave
--------------------
You're asking for very low resistance.
I've use the SMP1352 (Skyworks) before, but it only approaches 1 ohm. (1.5 ?s CL, 50 ?m I-region)
You can arrange things at those frequencies such that by paralleling them ( as far as AC goes), you get the benefit of "resistors in parallel."
I/O O | --- --- | +----+ | | D^ Dv | | Ib->+ +->Ib | | --- --- --- --- | | +----+ | O I/O
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.