D-104

[kb3rdt]

Hello I found this when I was Looking for a D-104 circut and found it good info to pass on!

http://www.qsl.net/wa2mzf/d104.html

[k4kyv]

Many hams love their Astatic D-104 microphones, both for their audio qualities and for their traditional
appearance.  The problem is that they do not work well with most of the newer rigs which mostly
require a low (500-600 ohm) impedance, while the D-104 has a high (50K ohm) impedance.

And here is the reason why so many of them sound crappy.  The manufacturer's recommended load impedance is 5 megohms.  I have a 20 meg load on mine.  A crystal microphone element is roughly equivalent to an ideal a.c. generator with about a 500 pf capacitor in series. In order to get full low frequency response through a 500 pf capacitor, a very high impedance load is needed.  50K just doesn't cut it, but I have been told that certain "vintage" AM rigs use a 50K grid resistor in the mic preamp stage.

Testing and Adjusting

...If the final output seems too tinny (not enough base in the audio), you may be able to experiment with different capacitor values added in-line with the audio wire. I actually discovered a different method for lowering the frequency response of the microphone. I had a couple of spare D104 element heads. I took one of these spares, opened it up and proceeded to punch a series of pin holes in the element foil. This seems to relieve some of the tension on the foil and results in a much warmer response. I’m sure there is a way to do this electrically (so as not to destroy a perfectly good mic element for future use), but I was unsuccessful in my attempts to do this. Since I had a number of spare elements to play with, I wasn’t inhibited in punching holes in the one I used.

So at last we see an explanation for the mysterious but lame-brain destruction of perfectly good D-104 crystal elements that has apparently been popular in some CB circles!

[ke7trp]

The later D104s had the holes from the factory. That is where he likely got the idea. I too saw this years ago.  Placing one pin hole mid way from center to outside edge does bring in the bass.  Why this works is simple. It gives a vent hole for and relieves pressure behind the diaphram.

C

[k4kyv]

But one hole should be enough.  The first time I replaced a D-104 element, many years ago, it came with what looked like a hole punched in the  cardboard box.  Then I opened it, and removed what looked  like packing material in front of the diaphragm.  I assumed that's what it was, and noticed a hole in the diaphragm.  Again I ASSUMED it had got damaged in shipping, and something had accidentally rammed a hole through the box and into the diaphragm.  So I stopped up the hole with a drop of epoxy.

It worked OK, but at the time I was playing around with high level speech clipping, ultramodulation, etc, so slopbucket quality audio was pretty much the norm.  Later on, after I became aware that the  padding material was supposed to be left in place, I fabricated another pad to go over the top of the diaphragm out of fibreglass insulation, and carefully drilled another hole, using a tiny electric drill.  Don't recall if that improved the audio quality any, though.  The problem with the original hole was that it looked like someone just shoved a large pin or small nail into the diaphragm, leaving a dent and ragged edge.  That's why I thought it was accidental damage.  As for the fibreglass, it was missing from my defunct original when I removed it, so I thought it was just protective material in the new one to protect the diaphragm before installation.

Astatic should have included instructions with the replacement element indicating the purpose of the padding and that the hole was supposed to be there.  I'd bet I'm not the only person who has ever made those mistakes. 

I picked up a mic at a hamfest, and when I tried it out, the sound was atrocious.  Worse than a tin-can telephone made with sardine cans.  When I opened it, I found that someone had punched numerous holes in the diaphragm to make it look like a salt shaker, and dented the foil in several places to boot.  That's when I was told this was a common "treatment" in good-buddy circles. I doubt that more than one hole would make it sound any better or different than the one punched at the factory.  It would also have caused less confusion if they had drilled a clean round hole with a tiny drill bit,  rather than just pierce it with a sharp obeject.

[NR5P]

Many hams love their Astatic D-104 microphones, both for their audio qualities and for their traditional
appearance.  The problem is that they do not work well with most of the newer rigs which mostly
require a low (500-600 ohm) impedance, while the D-104 has a high (50K ohm) impedance.

And here is the reason why so many of them sound crappy.  The manufacturer's recommended load impedance is 5 megohms.  I have a 20 meg load on mine.  A crystal microphone element is roughly equivalent to an ideal a.c. generator with about a 500 pf capacitor in series. In order to get full low frequency response through a 500 pf capacitor, a very high impedance load is needed.  50K just doesn't cut it, but I have been told that certain "vintage" AM rigs use a 50K grid resistor in the mic preamp stage.

Testing and Adjusting

...If the final output seems too tinny (not enough base in the audio), you may be able to experiment with different capacitor values added in-line with the audio wire. I actually discovered a different method for lowering the frequency response of the microphone. I had a couple of spare D104 element heads. I took one of these spares, opened it up and proceeded to punch a series of pin holes in the element foil. This seems to relieve some of the tension on the foil and results in a much warmer response. I’m sure there is a way to do this electrically (so as not to destroy a perfectly good mic element for future use), but I was unsuccessful in my attempts to do this. Since I had a number of spare elements to play with, I wasn’t inhibited in punching holes in the one I used.

So at last we see an explanation for the mysterious but lame-brain destruction of perfectly good D-104 crystal elements that has apparently been popular in some CB circles!

I put a 3Mohm resistor across the grid on my apache.  It seemed to help out a lot.  I was afraid to go higher for fear there might be instability problems.  I've thought about increasing it more but everyone tells me to leave it alone because it sounds great so maybe I shouldn't.  I think originally it had 50k across the grid.  I can't believe they would have put such a low resistance across the grid like that.

[k4kyv]

That's one reason the stock Apache has such crappy sounding audio, and has the  reputation "scratchy Apache".

[flintstone mop]

So many holes will get the low end down to 30hz??Then Jay N3WWL, would be FB OM using a holy D-104.
When I read some of the chatter about modifying the microphone, I was getting ready to for some oxygenated wire to amplify the high end.

Fred

[NR5P]

It's not hard to replace the diaphragm though it just take some patience.  If somebody got a hold of a diaphragm that was beyond repair would there be any benefit to using something thinner like tin foil?  I guess maybe if it isn't broke don't fix it maybe the thickness of the factory diaphragm is just right.

[kb3rdt]

I like D104 they are hard to beat mic I use a MPF 102 10 meg ohm on the gate.  Don't know why they don't use any resistors on the gate! I think why a 4.7k to the input not needed want bass play with caps on the input to ground if I am wrong help me out!

[k4kyv]

It's not hard to replace the diaphragm though it just take some patience.  If somebody got a hold of a diaphragm that was beyond repair would there be any benefit to using something thinner like tin foil?  I guess maybe if it isn't broke don't fix it maybe the thickness of the factory diaphragm is just right.

I wouldn't think it would be easy to stamp out a new diaphragm of the proper shape and end up with something with the same acoustical resonance characteristics, even with the proper foil stock. OTOH, I have a couple of defunct elements but haven't thrown them away.  Some day I might try my hand at salvaging  the diaphragms from the bad ones and try to use them to replace the damaged ones.  I have one that has been mangled by having multiple holes punched through the diaphragm (it sounds like total crap), and another with practically no low frequency response apparently as the result of someone either inadvertently or intentionally denting up the diaphragm.

I'm surprised that some manufacturer hasn't started producing an after-market replica of the original crystal element since Astatic discontinued them.  They aren't exactly rocket science and it should be easy to produce something identical to the original.  Like automobile parts suppliers, they could offer a discount for "core replacement" and re-use the bakelite cases and possibly even undamaged diaphragms taken from the defunct ones traded in. Maybe there are still too many working D-104s in circulation for there to be enough demand to make an after-market product sustainable. I must have seen at least a dozen D-104s at the Dalton, GA hamfest this past weekend, and it wasn't all that big a flea market even though it was decent enough to be worth the trip.

[Pete, WA2CWA]

That's one reason the stock Apache has such crappy sounding audio, and has the  reputation "scratchy Apache".

I suspect the real reason was that users interpreted the front panel "Gain" control as the actual audio gain control when, in fact, it was used to set the clipping level. The actual audio level control was located behind the key jack.

[kb3rdt]

yep why radio has an audio gain or mic gain but some don't know how to use them I talk close to my D104 cause if not you lose the bass responce of your voice!

[W9BHI]

Here I go again mentioning the RB Micro replacement element kit at the risk of being flamed again.
This time I'm wearing my fireproof suit.
I would like to invite K4KYV to try one of these kits and give everyone an unbiased opinion.
I'm sure your words would hold more water than mine.
The website (RB Micro.com) states that you will get your money refunded if you are not satisfied.
$28.00 to get your D-104 sounding like new.(that is if you like the way a D-104 sounds to begin with).
It sure beats screwing around with an old damaged stock element.
BTW I have NO affiliation with RB Mico whatsoever.

W9BHI

[k4kyv]

I think I have seen information about these kits somewhere on the internet.  If it is the same thing I remember, there is one problem: the case is metal and one terminal is grounded to the case.  The original D-104 element is non-polarised.  The case is bakelite with two terminals, and no polarity is assigned to the two terminals.  The ambiguous polarity allows the mic element to be wired to provide balanced output. 

This is achieved by using a two-conductor shielded mic cord.  Each of the two conductors is wired to one of the terminals on the element, and the shield is grounded to the case.  This allows the use of a push-pull preamp.  Each tube in the preamp has a grid resistor to ground, and the two grid resistors  in series, with the common point to ground, act as a voltage divider to provide the 180° out-of-phase signal required for the grids in the push-pull circuit.

The push-pull pre-amp has two advantages over the more normal single ended circuit: (1) There is substantially less rf and 60~ a.c. pick-up, and (2) the two grid resistors in series allow the crystal to be loaded down with twice the resistance as with the single ended circuit; there is a limit to how much grid resistance can be used in a tube type amplifier before the bias voltage becomes unstable.  Astatic recommends 5 megohms.  In the RCA Tube Manual, the recommended maximum for most tubes is only 0.5 megohms!  Up to 10 megohms is usable, if tubes are carefully selected for stability.  With the push-pull circuit, the total load resistance can easily be set to 10 megohms, and by carefully selecting and matching tubes in the first stage, up to 20 megohms load can be used.

Using the highest load possible allows the maximum low frequency response of the microphone.  A crystal element can  be thought of as an ideal a.c. generator with approximately 500 pf of capacitance wired in series.  In order to pass lower audio frequencies through 500 pf, an extremely high load resistance is needed.

The push-pull circuit is not my creation.  The original data sheet with the D-104, dating back to the early 30s, describes this circuit and recommends it for runs of mic cord over a few feet, to reduce hum pickup.  I have observed that it also reduces stray rf pickup from the transmitter.

This might be a moot point if a FET pre-amp is built  right into the microphone or microphone base. Some circuits inherently have a very high input resistance, and the short distance from mic to pre-amp would eliminate the hum and rf pickup problem.  Of course,  the pre-amp could include pre-emphasis to adjust the response curve to be equivalent to the stock element.

This replacement would not be practical in my case, since my pre-amp uses the push-pull balanced circuit with a pair of high-mu triodes (octals with nearly identical characteristics to one section of a 12AX7).

[DMOD]

I can't believe they would have put such a low resistance across the grid like that.

In the earlier days they wanted the higher frequency audio to be transmitted so the audio could get through the noise. A 50k ohm load acted as a differentiator with high frequency components, whereas a 1-5 MOhm acts more like an integrator which rolls off the highs.

Phil - AC0OB
 

[W9BHI]

FB Don,
I see your point on the balanced inputs.
I use the FET mod on the mic I use on my FT-950.
I use an unamplified mic on the Valiant, it has a 10 meg grid resistor installed.

W9BHI

[KA2QFX]

The purpose of the pinhole in the diaphragm was to equalize pressures that were damaging elements when subjected to cold temperatures, as in shipping.
Apparently, a change in manufacturing technique of the rear casing created an unusually tight chamber when the diaphragm was glued on. The resultant air pressure was sufficient to crack crystal elements or break the armature from the diaphragm.
 
It had nothing to do with changing the frequency response, at least not intentionally.  Nor do I believe it actually did change the response.

The JFET mod, which is applicable to both amplified and un-amplified stands, due to gate leakage current has limits on the maximum resistive load that can be applied to the element. This is much the same limit as Don mentioned regarding tubes with too much a grid resistance.  Some gate return resistance MUST be utilized to hold the gate voltage at ground and enable self bias from the source current.  I found 10Mohm a reasonable value using 2N3823 and MPF-102 as shown below.
http://home.comcast.net/~msed01/images/D-104%20Mods.JPG

FWIW: The crystal element has a bypass capacitance across it formed by a foil overwrap. The value and affect of this has yet to be determined.  
http://home.comcast.net/~msed01/images/MC320.jpg

Both of the images above are part of a larger work in progress.  Eventually, we intend to make the crystal elements themselves.  Making large Rochelle Salt crystals isn't too difficult, slicing thin slabs and handling them is another matter.

Hope this is informative.

[k4kyv]

IMO, there is no reason to use a blocking capacitor between the mic element and the grid/gate.  The DC voltage at that point should be very close to zero, and the crystal element itself functions as a capacitor, with the two sides of foil wrap serving as plates and the crystal as dielectric.  In my push-pull preamp, the two-conductor mic cord wires are connected directly to the tube grids, with nothing else but the grid resistors to ground. I don't use any kind of bypass capacitors or other rf filtering.  With the single-ended preamp, I had to use a small by-pass capacitor combined with a series resistor to keep rf out of the audio.

What best method would you suggest for non-destructively unglueing the diaphragm from the rim of the case and detaching it from the armature, if/when I try to replace the damaged diaphragms with good ones salvaged from defunct elements with bad crystals, and what kind of glue or wax should I use to re-attach the new diaphragm?

[w9ac]

...I found 10Mohm a reasonable value using 2N3823 and MPF-102 as shown below.
http://home.comcast.net/~msed01/images/D-104%20Mods.JPG

In that diagram, the 10K source resistor (Rs) is way too large for self-biasing.  The correct value is found by taking the reciprocal of the FET's transconductance (gm) value.  With effort, it can be measured.  Or, take the geometric mean of the gm range found on the FET's data sheet.  Depending on the FET, Rs should only be a couple hundred ohms, typically 200-300 ohms. 

Referring to the diagram, the FET's 10M gate bypass resistor is optional.  If used, keep in mind that it's in parallel with the FET's input Z, thus lowering an 11 meg input Z to 5 meg.   Mostly a non-issue, but it's still a design consideration.  The resistor is a good idea for use in low humidity environments for input gate protection.  The output of the crystal element is already a capacitor, and the input to the gate has no DC voltage present.  The input cap is not needed unless some other form of element is used with DC biasing.   

There's no PS de-coupling shown.  Astatic didn't need de-coupling when using a 9V battery.  However, if this circuit powered from the mic jack (typically +8V from a modern transceiver), then de-coupling is needed.  I would de-couple the supply line with 100 ohms, then filter at the FET drain lead with a combination of electrolytic and film bypass caps. 

The second stage uses collector feedback bias and the Q point varies with output.  Arguably, it's a sub-standard design.  If gain is needed since the FET source-follower (the first stage) only provides unity gain, then I would consider a single common-source FET.  It retains the Hi-Z input advantage, yet provides good gain and works well into moderate input Z.  So, why bother with two stages?  Cut the parts count roughly in half, and get better operating performance.

Paul, W9AC

[KA2QFX]

Paul,
    You are quite right on all. However, I chose the 10K to minimize battery drain. The Vs will find it's equilibrium point with a variety of FETs that a modifier might have on hand. The increased output Z to the next stage seems to be a non-issue. Eliminating the second follower would definitely require a much lower source resistor.
I leave the input cap in place in the TUG amplified mic because some folks like to use the dynamic head with the stand.   I don't agree that the gate resistor is totally optional. I have found some FETs do exhibit enough reverse gate leakage to require some ground path. 20M would be better, but this is intended to be a more universal mod using junk box parts. This circuit therefore uses very forgiving values with predictable results that most folks find very pleasing. 
Let's face it, you AM aficionados are a whole other class of operators with much more discriminating tastes and technical ability.  

[KA2QFX]

Oh, forgot Don's question.  Removing the diaphragm: I often find the glue fairly brittle if not detached altogether on the old stuff I've seen.  I melt the armature adhesive with a soldering iron. I'm told later units used epoxy, but I haven't come across any of those yet. 

For tough diaphragm glue I've used a combination of heat gun, or soldering iron on low, and methyl-ethyl-keytone. Not at the same time!  Fortunately the base material can tolerate guite a bit of heat and solvents. Phenolic maybe?  Of course, you must be careful to confine the heat to the edge to avoid damage to the crystal. 

I don't think there's any exact method but this has worked for me.

[k4kyv]

Oh, forgot Don's question.  Removing the diaphragm: I often find the glue fairly brittle if not detached altogether on the old stuff I've seen.  I melt the armature adhesive with a soldering iron. I'm told later units used epoxy, but I haven't come across any of those yet. 

For tough diaphragm glue I've used a combination of heat gun, or soldering iron on low, and methyl-ethyl-keytone. Not at the same time!  Fortunately the base material can tolerate guite a bit of heat and solvents. Phenolic maybe?  Of course, you must be careful to confine the heat to the edge to avoid damage to the crystal. 

I don't think there's any exact method but this has worked for me.

Have you tried to re-assemble any of them?  If so, what kind of glue did you use?  Epoxy would work, but you have to be 100% sure everything is correct with that stuff, since there is no turning back.  Heat will soften cured epoxy, but it would probably require enough to ruin everything else, too.

[KM1H]

Its pretty hard to get beyond a 4.7 meg grid resistor without instability in a typical factory ham rig without a bit to a lot of extra work.

IMO its often better to terminate the D-104 externally in a real preamp and then drive the rig at the correct level.

[KA2QFX]

Hi Don,

I heard you and Al VTP last night. Sounding good as usual...

I reassemble with Weldwood's or Pliobond Contact cement.  I thin it with toluene (acetone or MEK work also) and apply several coats with a small brush trying to keep it to a 1/16" band around diaphragm and capsule edges.  I machined a small piece of PVC with a ring sized to contact just the glue area but I've also done it without the "caul" with acceptable results.  

I'm scattered today.  

[kb3rdt]

D104 amp can make them sound how ever you like!