ad: k1jek

D-104 loading...

Discussion in 'Amplitude Modulation' started by W1BR, Jan 27, 2021.

ad: L-HROutlet
ad: l-rl
ad: QSOToday-1
ad: Left-2
ad: Subscribe
ad: abrind-2
ad: L-MFJ
ad: Left-3
  1. W1BR

    W1BR Premium Subscriber QRZ Page

    That is NOT what she said at all!!!!

    Reading 101: My thinking is that the electret mike element sounded too bassy without a VERY small coupling capacitor to treat the boominess.

    and she added further:

    The Kobitone crystal elements Mouser used to sell got good reports as a direct sub for the D104 crystal element. They are no longer available.

    Reread her above comments again regarding the Kobitone cartridges.


    I am using a Kobitone crystal mike element which is about as close to the original crystal mike element
    that was used in the early D-104 mikes. Later D-104 mikes went to a ceramic cartridge and added a
    transistorized amp in the mike base.

    Regardless, the problem was traced to overly aggressive action the SSM-2167 Analog Device chip.
    Reducing the compression range eliminated the distortion artifacts in my audio. That was
    a combination of the compression action and my close talking the mike. I'd suspect those
    artifacts would be a problem for any AM transmitter. Distortion is distortion. I'd also
    see some overshoot on the SM-2167 attack time; so I am not convinced that device
    is the best approach for an audio processor without some sort of look ahead audio delay.

    Why use the board?? It provides all phase correction and a high audio level to my transmitter
    distribution switch box, where the level is dropped to an appropriate level for any of the
    transmitters that are under its control. It also provides a six megohm load for the
    Kobitone cartridge.

    :)

    -W1BR
     
    Last edited: Feb 4, 2021
    AC0OB likes this.
  2. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    As far as I know, the D-104 stayed with the crystal element to the end as default, with the ceramic version available as an option. The ceramic element has less output than the crystal, and I believe, poorer high frequency response. But it is less susceptible to self-destruction in hot humid environments.

    The amplified version was produced in response to the low-z input impedance of more modern ham rigs, and of course, CB transceivers. The amplifier is primarily designed to be an impedance transformer, not to increase mic gain, converting the high impedance of the crystal element to the more useful 150 to 500 ohms input impedance, standard in most modern rigs. The stock in-base pre-amp suffers from a noticeable degradation of the audio quality, although circuits for much improved homebrew substitutes using FETs can be found on line.

    According to a pre-WWII Astatic data sheet regarding the D-104, the output impedance of the crystal element is 80,000 ohms at 60 cps. No data is given for higher frequencies.

    A consideration when a triode pre-amp is used with a high source impedance such as a crystal microphone, is the Miller Effect, which effectively increases the input capacitance by a factor equal to the grid-plate capacitance multiplied by the gain of the amplifier. This may limit high-frequency response in any amplifier, solid state as well as vacuum tube. A good explanation of the Miller Effect is found here:

    https://www.pearl-hifi.com/06_Lit_Archive/02_PEARL_Arch/Vol_08/Sec_31/2071_The_Miller_Effect.pdf
     
    1 person likes this.
  3. W1BR

    W1BR Premium Subscriber QRZ Page

    For the heck of it, I bypassed around the 2167. Getting good audio reports without asking. That leaves the J-fet high Z input and the all phase filter in the mike. No more compression, etc.
     
  4. WD5JKO

    WD5JKO Ham Member QRZ Page

    I had a similar issue with the 2nd generation K7DYY processor in the UG8 base of my D-104. The JFET output would "shoot through" the 2167 on voice peaks. I replaced the JFET source resistor with a POT, and set the level to
    something that the 2167 could handle. This was better, but during the 2167 attack time, the 2167 would occasionally clip. I found no easy fix other than an external clipper that only went in limiter mode during a loud voice after being quiet
    for a while. The audio was much better with the clipper IN because the big occasional spike was eliminated.

    AFTER THE CHANGES:

    Post_Mod_No_Clip.jpg

    RESULT WHEN THE PCB WAS STOCK:
    Pre_Mod_Clip.jpg

    SKETCH OF UG8 BASE:
    UG8_Base_Wiring.jpg

    Jim
    Wd5JKO
     
    Last edited: Feb 11, 2021
    AC0OB likes this.
  5. W0RIO

    W0RIO Ham Member QRZ Page

    So, if you plug the hole, does that turn your D-104 into a barometer? ;-)

    It might actually work if you ping the crystal and look at the echo, not that one should bother.
     
  6. KF4ZGZ

    KF4ZGZ Ham Member QRZ Page

    In stock form the D-104 ( which you don't really have ) is actually a quite trebbly mic.
    Thats why they put the hole in the diaphragm .... it NEEDED more bass response.
    That's also why they "extra hole hack came about" people wanting more bass response .... and it does work btw.

    Have you simply tried to borrow a real un-molested stock D-104 just to see what it sounds like?
    I'm almost willing to bet this would work ..... maybe even seal up the hole in the diaphragm.
     
  7. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    Maybe if one carefully added another hole or two. The one I have that was molested by a CBer sounds worse than a tin-can telephone. The diaphragm looks like someone stabbed it multiple times with an ice pick. The same alleged effect should result from drilling a hole in the side or back of the Bakelite case of the mic element, which could easily be plugged up with a drop of epoxy if it didn't achieve the desired outcome. I don't understand why Astatic didn't take that route (if that's really what the hole is for). Just make sure none of the dust or debris from the drilling process gets inside the cavity.

    From my experience the best way to gain additional bass response (in the unamplified version) is to increase the load resistance on the xtal to between 5 and 10 megohms. This is done in the pre-amp stage, not inside the mic head.

    I wouldn't go punching or drilling holes anywhere in a good D-104 element unless I had an ample supply of extras on hand. Since Astatic doesn't make them any more, they are rapidly turning into unobtanium. Every stock element that is destroyed, means one fewer left on the face of the planet.

    None of the "substitute" elements circulating in ham (or CB) circles are anywhere near the same as the original. All the ones I have heard over the air sound worse or much worse, except in a few cases where someone gutted the entire D-104 head and replaced the element with a known high-quality mic like a large diaphragm electret condenser. But then you no longer have a D-104, just some other microphone dressed to look like one.
     
  8. W5JAY

    W5JAY Ham Member QRZ Page

    HI Jim,
    One could experiment with R17 which is what they call Rcomp. That sets the gain and compression of the IC 2167. The K7DYY board sets R17 to infinity! I would think dropping R17 resistor down to 35K ohms for a 3 to 1 ratio or 47 K ohms for a 4 to 1 ratio would be better than set for infinity? Also most folks didn't like the gate as it slams shut way to fast, but it is easily defeated by removing a jumper. Also like I mentioned before, this board runs at -10dB. That will probably make the audio on the hot side for a stock Ranger input? Might be better to use the second stage at the cathode or the phone patch input if one is available? Just some more random thoughts, mileage will vary, hi. 73 de w5jay/jay..
     
    AC0OB likes this.
  9. W1BR

    W1BR Premium Subscriber QRZ Page


    I bridged R17 with a 37k resistor and it helped a lot. Regardless, for my application I finally bypassed around the Analog Device chip and will use external processing if needed.
     

Share This Page