ad: ProAudio-1

A more elagant SB-220 keying interface?

Discussion in 'Amateur Radio Amplifiers' started by W1BR, Jan 6, 2011.

Thread Status:
Not open for further replies.
ad: L-HROutlet
ad: l-rl
ad: Left-3
ad: Subscribe
ad: L-MFJ
ad: l-BCInc
ad: Left-2
  1. WA7PRC

    WA7PRC Ham Member QRZ Page

    The terms "reverse biased" and "conducting" are mutually exclusive. A series diode does nothing to prevent back EMF from reaching the switching device. Just to prove it, I modeled it in LTSPICE. We know the SB220 uses a 110VDC/10KΩ relay, and I guesstimated 200mH inductance. Without a diode across the relay coil, I see ≈ 10µs spikes of around 300 volts across the switching device when it opens. With a diode in series with the switch, I see the same thing:
     
  2. W1BR

    W1BR Premium Subscriber QRZ Page

    That Spice model doesn't show the .02 bypass capacitor, which is effectively across the coil. It will dampen the spike to some degree.

    The back EMF voltage is reverse polarity, so I guess I am failing to understand why a series diode would conduct in both directions? The only concern I could is the junction capacitances forming an unwanted voltage divider when the sharp reverse EMF voltage pulse occurs, but a small cap from the collector (emitter for the PNP circuit) to ground should correct that issue.

    Where did that 10K resistance value for the coil come from? I recall Measure's stating 0.25amps current draw for the relay?

    Pete
     
    Last edited: Jan 7, 2011
  3. W1BR

    W1BR Premium Subscriber QRZ Page

    I believe the polarity is incorrect as shown for the Spice simulation. The counter EMF spike has the incorrect polarity.

    Pete
     
  4. WA7PRC

    WA7PRC Ham Member QRZ Page

    At first, I didn't add it, in order to see the spike. With it across the switch, it "sucks up" the spike. Of course then, you're back to having to discharge the capacitor when pulling the ANT RELAY jack low unless you add a small R in series with the jack as AG6K recommends.

    The polarity of the pulse is correct (LTSPICE doesn't lie). Rectifiers conduct in only one direction. If the series diode is reversed, it's cut off, and no current flows (the relay won't pull in).
    It would. Your switching device needs to handle the peak current (with a capacitor) or handle the back EMF voltage without it. Without it, the relay will drop out a tiny bit faster. It's a non-issue becaue, if timing is important (for full break in, etc), you wouldn't be using the OEM relay anyway.
    For my simulation, I used a relay from Tyco/Midtex (Potter & Brumfield), similar in appearance to the one that Harbach supplies as a replacement:
    [​IMG]
    Using Tyco's parametric search engine, I found pn KA-14DG-110. This is specified as being 3 Form C (3PDT), 110VDC/10KΩ/125mW (the datasheet says 1.2W).

    0.25A at 110V is 27.5W! :eek: You must have the decimal point off by one place. After I did the LTSPICE simulation, I found references on the web that say the coil measured 4600 ohms, which would jive with 25mA. It appears Measures was correct but, the precise values aren't terribly important - I was just trying to demonstrate the (non) effect of a series rectifier.
     
  5. W1BR

    W1BR Premium Subscriber QRZ Page

    Wow Bryan, now you're waffling. I stated a diode in series with circuit would prevent the back EMF voltage from exceeding the breakdown voltage of the keying transistor. You stated that the diode would conduct when the voltage was reversed. Quote "A series diode does nothing to prevent back EMF from reaching the switching device." You can't have it both ways. Either the diode conducts in both direections, or it doesn't. I asserted it would provide reverse emf protection, and then you used flawed spice analysis to try and prove otherwise. Now you're doing a 180 and trying and to save face.

    It is quite obvious that a series diode will prevent the reverse EMF voltage from reaching the device, and it will not the clamp the coil voltage adversely affecting the dropout time either. Mirage and others have been doing this for decades in their designs.

    Eeeek my butt. You provided your hero's webpages to support your claims, took unwarranted pot shots at W8JI who isn't even involved in this thread. Measures claimed .25 amp, and even states the voltage drop across the suggested resistor. if you have a problem with the data, remember you supplied the source.

    Your spice model is flawed since it doesn't take into account the pulse width, amplitude, and the capacitance of the reverse biased diode junction. That would be useful information in determing whether a small shunt cap between the diode and keying transistor to ground would be beneficial in swamping that capacitance and forcing the back EMF across the reverse biased junction.

    Thanks for the input.

    Pete
     
    Last edited: Jan 8, 2011
  6. W1BR

    W1BR Premium Subscriber QRZ Page

    I'll retract the comments about the amperage. It was stated as .025 amps on the website. 25mA, not 250mA.

    Pete
     
  7. WA7PRC

    WA7PRC Ham Member QRZ Page

    –30–

    Good grief. I don't believe I've been self-contradictory. NEVER in my life have I intended to state (save for a typo) that a rectifier conducts (except for miniscule leakage current) with reverse voltage.

    You stated a series diode prevents back EMF. You are incorrect, and LTSPICE proved it. If you have an issue with LTSPICE being in error, you can take it up with Linear Technology. They wrote the software and models. All I did was use them in the configuration that Heath did, and you intended.

    Richard Measures isn't my "hero". I agree with most (but not all) of what he says. Someone needs to reread his work but that someone isn't me (I've already read it many times). And, I took no potshots at Tom/W8JI. I have the utmost respect for his knowledge and skill.

    Since you found my input so flawed, I don't understnad why are you thanking me for it. Lastly, public ridicule of my honorable intentions earns you a permanent plonk from me.

    –30–
     
  8. W8JI

    W8JI Ham Member QRZ Page

    Right, and that is what slows the relay down. Pay attention to the current flow direction. When the relay coil field collapses the voltage at the relay jack does NOT reverse. It goes higher in the same polarity.

    It is electrically impossible to quench counter EMF with a series diode. This is because the voltage increases with the SAME current flow direction.

    The diodes ICOM puts across the transistors do nothing at all unless the keying line is accidentally reversed. To be effective as back-pulse clamps, they would have to be zeners or placed across the amplifier relay coil.

    Another flaw in the ICOM circuit is voltage is not reduced. Generally there are bypass capacitors in the radio. Let's say you connect a SB220 to the radio. The capacitors in the radio will charge to 130 volts or so. When you close the relay contacts, the bypass capacitor discharge into the contacts. That is never good long term.

    If it is negative, like an SB200, the same thing occurs. The only difference is voltage is negative.

    If you have a voltage arcing problem in the contacts eating them up, the ICOM circuit is not a good fix.

    An additional reed relay is not so hot either if switching time is not an issue, and it requires the reed relay handle switching voltage and current. Better to just use transistors, but takes more than one to do it right if you need to limit voltage and current.
     
    Last edited: Jan 9, 2011
  9. W1BR

    W1BR Premium Subscriber QRZ Page

    It is electrically impossible to quench counter EMF with a series diode. This is because the voltage increases with the SAME current flow direction.

    The diodes ICOM puts across the transistors do nothing at all unless the keying line is accidentally reversed. To be effective as back-pulse clamps, they would have to be zeners or placed across the amplifier relay coil.

    Another flaw in the ICOM circuit is voltage is not reduced. Generally there are bypass capacitors in the radio. Let's say you connect a SB220 to the radio. The capacitors in the radio will charge to 130 volts or so. When you close the relay contacts, the bypass capacitor discharge into the contacts. That is never good long term.


    And the counter EMF would be in series with the supply voltage, compounding the problem. Guess using Mirage as a design example was a mistake.

    Yeah, the emitter keying does put full negative voltage on the radio. I was considering adding a second PNP transistor to alleviate that problem, although it isn't an issue with the JRC relay keying. The series resistance is beginning to look simpler.

    Pete
     
  10. W8JI

    W8JI Ham Member QRZ Page

    Mirage does that to protect the amplifier, not to protect the radio.

    I have seen Internet and print articles where people claim an external diode can quench backpulse. That is not true though, unless it is a properly sized zener.


    Most radios have a bypass cap inside the radio. Even if you decided to ruin the RF bypassing in the amplifier by adding a series resistance (you might as well cut the bypass out as to add a resistor), the radio still generally has a cap.

    Also any backpulse that causes an opening contact arc will still occur, it is just current limited by the base resistor.

    I have no doubt the transistor will extend relay life. My only concern is it still does not eliminate contact spark on closing because the "unimportant" capacitors that bypass RF discharge into the relay.... and the single transistor system passes backpulse through (but it will be current limited).

    To tell you the truth I think you would be better off with a single transistor by holding the base on a fixed low voltage supply and opening and closing the emitter to key. My bet is most relays would like the <100 mA a whole lot better than the 120 volts. It is the arc that normally kills a relay contact, not the steady current. Most relays will hold a hell of a lot more current than rated at if the contacts don't arc.

    By the way, who cares about a few milliseconds more on opening?? Closing speed is what messes things up and makes nasty signal.


    73 Tom
     
Thread Status:
Not open for further replies.

Share This Page