The AM Forum

I just zapped another MPF102 source follower for the zillion'th time that's mounted in the base of a modified D104 by changing mike plug from a cold (turned off) PA amp. (813 rig) to the cold 32V2.  Input impedance resistor in the gate lead is 16 megs with a 50 pf bypass. Any static could zap the gate I guess.  Sometimes it's a crap shoot. 

So I placed back to back 1000prv/2.5 a silicon diodes, gate to ground.
A. anyone done this?  what does this do to the 10 megs. gate impedance given that the diodes aren't supposed to conduct until 0.7 volts or so.
B. Any effect on the audio by some other means I haven't thought of?
The output looks reduced and 'highs only' on a scope but an air check's probably in order.
I've probably made a net few k ohms or worese out of what was once 16 megs.

modified post later.
Just zapped it again. diodes didn't help.

The diodes have some amount of capacitance, so depending on how much, this could reduce the high frequency response.

In addition to the capacitance that Steve pointed out, the diodes have a (non-linear) resistance at zero volts bias that is much lower than that of the FET.

I'm not sure why you have the a 16 megohm resistor between the microphone element and the FET. Separately, where is the 50pF bypass capacitor?

I would suggest the following:

The D-104 looks like an ideal voltage source in series with a 1000 pF capacitor (equivalent circuit).

Eliminate the 16 Megohm resistor.

Put a 1000pF capacitor from gate to ground on the source follower amplifier. This will form a 6dB voltage divider at the input of the FET... but it will not produce any frequency roll off or roll up. It will also protect the FET's gate from static discharges etc.


Best regards
Stu

The best solution would be to toss out altogether that shitty Astatic "power mike" module in the G-Stand base and use the mic as an original hi-Z output D-104,  with an outboard FET or tube-type microphone pre-amp.  The mic would also sound much better.  Doesn't the 32V use a high-Z microphone input?  What about the 813 rig?

Astatic recommends loading down the xtal element with five megohms, but I use 10 megohms with mine, which results in better low-end response.

To reduce hum and noise, Astatic used to recommend using the original microphone (that contained only the xtal element attached to a shielded cable) with a push-pull pre-amp, wiring the D-104 xtal element for balanced output.  A two-conductor shielded mic cord is used, with each terminal of the xtal element connected to one of the shielded leads, and the shield grounded to the body of the microphone case.  At the pre-amp input, a 5-meg resistor connects from each balanced wire  to ground, and two tubes or fet's at the input are fed with the hi-z balanced line, and the two tubes/fets are used in a push-pull circuit. 

I am considering building up a pushpull tube-type pre-amp for mine, and seeing if I can run the tube filaments on a.c. and still be hum-free.  Right now I use an unbalanced input tube pre-amp with a 12AX7, but I have to run the filaments on rectified and filtered 12 VDC to avoid hum.

The D-104 xtal element uses an ushielded plastic case, with neither terminal grounded to anything, making the balanced output circuit a feasible option.

1000 V PIV diodes are way too slow. I would use something like a 1N4148 or 1N914. They also have a lower C. Also put a transorb slightly above the battery voltage on the drain to ground.
Stu, I wouldn't unload the gate because it could build up a charge and hold the FET in saturation or off. The source has a resistor to ground so every leg has a path to limit high voltage

The Astatic preamps did not have any FETs in them. I'm assuming what Rick is using is homebrew.

Further thought

In addition to the changes I suggested in my earlier post... I wonder if the problem is really at the output.

The transistor is sensitive to the voltage between the gate and the source. Even if the gate is bypassed to ground with a 1000 pF capacitor (which is okay to do because the D-104 cartridge looks like a voltage source in series with a 1000 pf capacitor), you still have to worry about a transient across the source resistor (which would cause a transient between the source and the gate).

I suggest that you also place a capacitor directly across the source resistor... whose value produces a capacitive reactance that is approximately equal to the resistance of the source resistor at 10 kHz. I.e. if the source resistor is 1000 ohms, then bypass it with .015 uF.

Frank:

I agree with your point about having a path to leak off any charge that accumulates on the gate of the FET. I wonder if the small amount of input conductance of the FET is sufficient. I.e., I don't know whether having a 10M ohm or a 15 M ohm resistor from gate to ground is really necessary. The source follower is not too sensitive to a voltage offset between gate and ground.

Stu

Stu,
I've seen FETs of many die sizes hold a charge enough to make you think there was a gate source short. We made this cool high power pulse driver Customer asked us to run some transient tests on the 28 volts. Good thing an 80 volt positive transient came darn close to turning the driver on. We had a relay in series with the output  so nothing would have happened but we did go back and change the RC slope control on the gate to less R and more C to bleed off the transient coming through the drain-gate C

Frank

Based on your comments... a 10 or 15 megohm resistor between gate and ground is a necessary addition.

With 1000pF from gate to ground and 10 megohms in parallel, the time constant will be .01 seconds... which should be small enough to leak off any charge accumulation.

Happy New Year
Stu

I should have said drain source short. I would think so on the rc compared to the human body model for esd. (not milm  std 331 though). Aiso with protection on the drain

My 2 cents:

http://members.cox.net/n4jk/d104.htm

Seen this page for a few years.  I admit that the 2.7 meg ohm is probably a bit low

Al VTP

Thanks for all the thoughts. Yes, I've been using the MPF102 homebrew source follower (or switchable phase using the drain connection) in the base of the D-104 on the 813 rig for many years.  You did hear it the other night Don.  The original crappy Astatic 'power mike' 2 transistor amp was removed but the battery clip and long terminal lug were retained for the MPF102 single stage.   

 It's just when I unplug it from, say the 813 and reconnect it to some other rig that something zaps the MPF102. Sometimes it's a year or so before I zap it and sometimes, like today, twice.  I try to remember to turn off the rigs; if I don't it definitely will pop.  Sometimes even with the rigs off.  So I've been trying to sink extraneous voltages such as T/r residuals from the 32V2's PTT line (you know there's a lot of voltage there.) that stray into the audio without sucking down fidelity in the audio in/out lines.

One sure cure is too leave this mike plugged into the 813 rig all the time.  It's sounds fine with it's high resistance load as you've all heard. Thence into a PA amp, w/ 5 ch. EQ,, thence into the 811A or 572B's modulator (through a backwards ex-output transformer driving the grids.) 

I removed the diodes ; they did suck down the highs and knocked audio down overall. Was worth a try.   You know, a 6C4 or half 12aX7 cathode follower might be the way to go.  ... a lot more EMP proof  ;D  ..just a little hard to fit in the base.

Al

The circuit in that reference has a few shortcomings... but they are easily fixed.

After correcting for the typo (the electrolytic capacitor at the output probably was supposed to be labeled 1uF) the circuit cannot drive an input having an impedance lower than about 5k ohms at low frequencies. A 1uF capacitor has a reactance of 5305 ohms at 30 Hz.

If you replace the 1uF capacitor with a 10uF capacitor (to be able to drive a 600 ohm transceiver input), or if you are sure that this preamp will be driving an input whose impedance is 5,000 ohms or greater... then that will fix the low end roll off at the output.

Likewise, as we know, a 5 or 10 Megohm resistor from gate to ground would take care of the roll off at the input. The 1000pF series capacitor in the equivalent circuit of the D-104's crystal element has a reactance of 5,305,000 ohms at 30 Hz. Thus a 5 Megohm resistor from gate to ground will produce a 3dB rolloff at 30 Hz.

This still leaves open the original question regarding protection against static discharges and/or other transients.

Perhaps the rf filters at the input and the output of this circuit would be sufficient to protect against transients.

Happy New Year!

Stu

 

This is an interesting thread. I like some of the ideas presented by Stu with modeling the crystal element as a capacitor.

I never really thought the MPF-102 was particularly static sensitive like IG-FETs (the ones that come with all leads tied together with a wrap around shorting wire). I typically use the 2N5458 or 2N5459 when I need a N channel J-Fet. I still got some left over from a poly-paks assortment. Remember poly-paks?

I once built a source follower using a GaAs FET (MRF966) and I protected the gate with two hot carrier diodes where each was biased off 2 volts to limit the gate swing to +/- 2v Peak. This was used on a military active antenna Direction finder array for 20-80 Mhz. I never lost a FET even with a 10w 225 Mhz transmitter antenna about 1 meter away.

The original post brought up using power supply diodes as a clamp for the MPF-102 gate. These diodes can have a lot of junction capacitance. See the following link here:

http://www.hanssummers.com/radio/varicap/diode/index.htm

I have a tube VFO with varactor diode tuning using dual 1N4007's. This works extremely well. If I were to clamp the fet input, I'd use something more like a signal diode such as the 1N4148 or 1N914. These have < 1 pf junction capacitance and this does not vary much with changing the reverse bias voltage.

HNY,
Jim
WD5JKO

Hi Rick,

 Plenty good circuit theory here OM FB, Good Thread... I was never much on Amplified Mics either like Don Old school all the way but, in the situation of multi xmitters why not establish your audio at one location an then apply it via a mixer then on out to the xmitters..just flick a switch or knob at the IP...I'm Hip on the EMP thought though...

73 OM Happy New Year..

Jack.

See attached diagrams

HNY!

Something you could also try is an ESD supressor. The place I work for makes them:

http://www.st.com/stonline/products/families/protection/protection.htm

there are many configurations including diode arrays, diodes arrays with zeners, etc, and they are very fast. Some with capacitances below 1 pF, others with higher cap as you may prefer. They clamp at 5V and up so should protect the FET. Any of them designed for USB protection would likely do well. I'm not an expert but some big names use them by the millions. The only difficulty might be they you'd have to deal with a surface mount package. The DVIULC6-4SC6 has a SOT-223 package which is not too hard to use and it can protect 4 lines.

Stu,
The crystal element, as you described, is essentially a capacitor
in series with a voltage source.
(you might also want to insert a small ~2K series resistor)

The addition of the shunt capacitor to ground across the
5Meg gate resistor has another interesting effect at low freqs.
In addition to the 6db voltage drop caused by the capacitor voltage
divider action (assuming Cshunt = Celement), the effective series
cap will be doubled (impedance lowered) and low-frequency response
will be extended by a factor of two.

This effectively is a Thevenin capacitor equivalent to a resistive
circuit where, for example, a 1volt source with series 1k, when shunted
to ground with a 1k has an output of 1/2 volt (-6 dbV) with a source
resistance that has been lowered to 500 ohms.

This is the reason that a crystal mike can drive a long length of
shielded cable (shunt cap to ground) with few ill-effects,
as long as shielding is not compromised by ground loops.

I would go with Franks suggestion of using in4148's back-to-back
at the output with the addition of a 1k series resistor after
the diodes to protect them current-wise. The 1k resistor will
not affect frequency response or gain if the following stage
is 25k or so,
Regards,
BobbyT

Bob,
Those 1N4148s will surprise you in transient current rating and blow short protecting the FET.  I test a lot of transient protection devices. Some of these larger transorbs will surprise you when they handle over twice their rating at 85 degrees C and not break a sweat. The only problem is junction c so you need a good diode ahead of them to isolate the junction. We even bias them sometimes. Manyof these little RS485/422 transceivers go boom when they see more than 15 volts.......When you induce 100 volts on the twisted pair life can be short.

Bobby T.

Agreed on the effect of the parallel capacitor. I have sometimes used a 5000pF capacitor from grid to ground (given the high output of a D-104) so that I could use the D-104 with a preamp having an existing 1M ohm grid-to-ground resistor. Per your comments: with a 5000pF from grid to ground, one obtains 1/6th the output... but the roll off frequency (for a given resistance from grid to ground) is 6 times lower.

Frank et. al.

Yes, on the output side of the preamp (much lower impedance than the input side), a diode-type protection circuit would provide a higher degree of protection than just the capacitor shown in the diagram I posted. On the other hand, the capacitor might be sufficient for the application at hand.

Best regards

Stu

all depends on the width of the pulse....

Rick,

I have a homebrew preamp in the base of my D-104 using an MPF102 and have had no static zap problems. I utilized the switch in the desk stand to remove power from the preamp when not transmitting and to also short the output of the preamp to ground when not transmittng (unkeyed). If you are not doing this it may be a good idea to provide additional protection.

Mark WA1QHQ

Thanks Mark.
If I leave the mike plugged in, T/R operations pose no problem.
Somehow switching the mike to other rigs sooner or later poses the problem.
My internal battery is also switched by the stock column T/R switch set of spare contacts in D-104.

There are IEC standards and GR1089 standards, and also human body and machine models for ESD. If a few people would like a couple samples of ESD or transient (line crossing) protection devices, PM me with the PNs from that site. But is it really ESD or is something else going on? Could there be a voltage on that connector's pins?

Someone needs to make kits to convert the D104 to a FET amp, instead of the stock amp. I took the stock amp out of mine, but rewired it to run straight from the mike element to the radio. But since there is an impedence mismatch between the radio and the D104, the audio doesn't have any lows, unless you are 1 inch from the microphone. I figure that is because the low frequencies are loud enough to basically force their way through the mike. Am I on the right track, or am I way off?
So I tried to fix the problem with a small transformer, but that doesn't quite work the way I thought it might. Where can I get the parts to build one of those amps, anyway?
Shelby KB3OUK

Here ya go. Preamp and a compressor - all-in-one.

http://k7dyy.com/products.htm

I would seem to think that maybe you could put together another mic plug and tie all the pins, or just the problem ones, then connect it to ground, and stick that in there before you plug the mic in to short the pins to ground, so that it won't discharge through the mic. However that may not be a good idea, depending on how the  mic jacks on the two rigs are wired, you may be shorting some voltage to the ground.
Shelby KB3OUK

Howdy lollypop microphiles,

I used to work for ADI and those designers put a lot of work into protecting the chips from ESD. In fact finding clever circuits that protect Hi-Z inputs without destroying frequency response was guarded information. So as radical as it may sound, how about designing up a simple Hi-Z Op-Amp mic preamplifier? The thing can even be fully differential on the element side with your choice of single or balanced Low-Z on the output.

In a D104, which needs little actual gain, I would use the classic 2 op-amp INAMP circuit for simplicity. Vo = (Vin+ - Vin-)( 1+ R1/R2) where R1 = R4 and R2 = R3
Ground the minus rail and set VREF to 1/2 of the 9V rail for single supply operation.

FET Cascode inputs allow input Z's of 10 to the 12 or 13th power with capacitances of less than 2 pf.

There are plenty of dual FET op-amps you can get on Digi-Key or EPay for less than 5 bucks which are granchildren of the TL082 which would run on a single 9V battery for a long time. http://parts.digikey.co.uk/1/1/94083-ic-dual-fet-input-op-amp-8-dip-opa2137p.html

73's Mike - on the road again!

U2d, interesting op amp design.  "lollypop microphiles," - good one.
Back to strays: The 32V is polarity checked for proper ground. I've marked the old two wire AC line with a black mark on the plug.  Yeah, thought of stray 115 field somehow being present at places it shouldn't.  Come to think of it both mike cable wires are inside the shield, both audio and PTT lines.  On the 813 rig the PTT line carries about 6 to 8 volts, on the 32V2 it's 115.  So there's an obvious solution.

I've probably been too lazy to do anything other than change out MPF102's, they're so cheap and the fix lasts up to a year when I forget not to wildly change out mikes without turning off and upplugging everything.
 
Been an interesting discussion. Thanks all for the neato input. A lot to think about...

Steve, I thought about DYY's preamp but have heard some stories about it too. Maybe I don't want to be compressed  ;D

I think someone already mentioned something similar to this already, but what if you used the spare contacts on the PTT switch to short the output of the amp  to ground, that way the amp is disconnected from the connector when it isn't keyed up. Then the only way the MPF102 would get zapped would be for you to be holding the PTT switch down as you plugged the mic in to either rig. Also, you may want to check to see if there is any stray voltages on that connector that could cause the MPF102 to get zapped.

Yea, the DYY thing is just another option. One could always bypass the compressor.

The circuit below has been used successfully by many. TNX to K1DEU.


http://www.amwindow.org/misc/gif/micamp.gif

Joining in on this a bit late, but here goes...

To help with JFET static issues, I generally use a 5-megohm resistor ahead of the FET's gate.  Also, a bifilar wound CM choke at the gate will assist in suppressing transients, while leaving the differential mode audio waveform intact. 

I see far too audio many designs that use brute-force differential-mode filtering (e.g., individual inductors on separate cores).  Even with proper termination, the result almost always manifests with overshoot and ringing, owing to the passing of complex audio waveforms through inductance.  Absent winding leakages, the CM choke completely cancels differential-mode inductance, while raising common-mode inductance by a factor of 2L.  A well-designed CM choke with a very large L value is capable of preserving the integrity of complex non-sinusoidal audio waveforms.

On my D104, the crystal cartridge (+) and (-) leads are twisted from the mic head connection,  then pass through the 5M resistor and CM choke, both components located near the FET's gate.  No other components are used at the gate.

Although R and C values between the FET gate and ground may be used to help with static, I have always been concerned with degrading the input Z of the FET buffer circuit.  If you conduct a web search on "D104 FET" you'll get many FET buffer circuit designs that IMHO, are either deficient, or incorrectly designed. 

For example, on the source-follower designs, I've seen 100K resistors brought from the gate to ground.  Why?  For gate input protection?  For "gate leak" biasing as if the gate is identical to the grid of a vacuum tube were a "grid leak" may be required?  In any event, that's far too extreme and any value below the meg-ohm range will degrade low-end performance from the crystal cartridge.  A "gate leak" resistor is not required in the source-follower.  The FET as a source-follower is 100% self-biased.  The V drop across the Rs value produces and sets the Vgs bias. 

To set mid-point bias on the source-follower, here's a good approximation:  Rs = 1/Gmo; where Rs is the source resistor, and Gmo is the transconductance value from the device's data sheet.  Many FET datasheets will show large sample disparities between Gmo values.  Without manually setting the optimum point, I take the geometric mean between the lower and upper limits from the datasheet.  Moreover, self-bias stabilizes the Q point against changes in FET parameters (e.g., Gmo, Idss, etc.).

Whatever components are used at the gate, it's important to keep the loading of the crystal cartridge at a minimum in order to preserve low-frequency response.

I used that classic twin opamp design 25 years ago using LM308 opamps.    I used it as a substitute for the "instrumentation" balanced input amplifiers for thermocouple amplifiers and other balanced inputs.   The LM108 worked even better.   The common mode noise rejection was pretty good as long as you used 1% resistors.    I would add a 10Meg resistor from each input to ground to provide a current return.   The gain is adjusted by the ratio of the resistors.

Avoid bipolar op amps around RF. da suck

    Yes I've use clamps on the output to bullet prof the pre-amp since 1987.

  Here is a link to my shaky drawing

  http://hamelectronics.com/k1deu/pages/ham/audio/mikes/jfet_mike_preamp_equalizer.htm


   JFET Mike Preamp/Equalizer

Some typical Voltages are;

Battery 9 VDC Velcro-ded underneath

Drain  7.1 VDC

Source 1.8 VDC

Gate= near Zero

The 1 Mfd. Output Cap is non-polar.  To RF proof the input I put a miniature 470 micro henry choke in series between the gate input and the 10-20 megohm resistor along with a 18 pf cap from gate  to ground at JFET. All short leads please !  You may build the circuit without pots. Full bass boost would be up in the drain near +9 VDC the .1 across a fixed 10K Resistor, which is good feeding many transmitters. In most cases I would use a pot for treble/sibilance boost control running more boost when my signal was weak at the other end, less when strong. 
   The bass boost magic is from a triode Vacccccum Tube ckt in the 1938 Radiotron Designers Handbook ! Aparently  a triode is a triode sometimes.
Sorry I'm so late to reply.  Regards  John

Interesting circuit.  I have used the  same basic idea for treble boost for many years, in  cascaded 12AX7 stages.  The low value of cathode (or source) bypass capacitor makes the cathode (or source) resistor appear bypassed at higher audio frequencies, while at lower frequencies it appears to be unbypassed, resulting in degenerative feedbark at the lower frequencies.

It had never occurred to me to do the same thing with the plate or drain resistor to boost the bass.

Does the bass boost shift the phase at low frequencies?  I have an Altec tube-type mixer board, which is equipped with a set of consumer-audio-like bass and treble controls.  The bass boost works well, but it reverses the phase of the lower frequencies.  I could phase my audio for upward positive peaks, but with the bass boosted, some of the higher amplitude peaks would go in the negative direction while others went in the positive.  To  maintain my original voice asymmetry, I had to run the controls at flat response.

Actually, that circuit is not a  treble boost.  It is a bass attenuator.  And the bass boost is actually a treble attenuator, since they depend on the degenerative feedback of an unbypassed cathode/source resistor, and the gain reduction of a lowered plate/drain resistance.

My treble boost is not variable.  I use the fixed value that gives me the desired pre-emphasis curve.  Two 12AX7 stages gives about 9 dB of boost (or attenuation, if you prefer).

Ok,
In the K1DEU/W1ECO circuit I see back to back 5v zeners. Why are they in series and not paralled from signal to ground? They are already on the AC side of the 1uf capacitor.  How can they act at 0.7 volt threshold (+/-) when they counteract each other?  That might have been my problem with the hv silicon diodes.
Somedays the fuzzy logic portion of my head just can't seem to compute stuff so I'm perplexed. 

The back to back zeners act to clip both the positive going and negative going peaks to about the zener rating.

Rich

W7SOE

Yes your saying what it does but how does it work?

Why can't I get the same effect with the diodes paralleling each other but flipped in polarity?

so your saying that the 'top' diode will pass half cycles less than 5 volts to the bottom diode (but block those same phased half cycles over 5 vac).  But then doesn't the bottom diode and pass the other half cycle (inverse) to the bottom diode which will then block ..   uh, guess I better sketch out my thinking.

Also if I inadvertantly put 115 vac across the output from the T/R line of, say, a 32V2 then I'm going to pop the 5 volt zeners unless there is a dropping resistor, say 20k or so between the jct. of the top of the zeners and the output.  20k ought to be fairly impervious to audio at the impedances involved.

....I must be missing something  ;D

The top diode will pass half cycles MORE than 5V to the bottom diode, you then have to add the forward diode drop of the bottom diode.  That way the top diode clips half cycles more than 5V (ish).

On the negative cycle the bottom diode conducts in reverse when the voltage is below -5V (ish) plus the (now forward biased) voltage drop of the top diode.

Rich

ok I got it.
well you know..... brain cells are the second thing to go.  8)

Don; The Bass boost circuit changing the plate load resistor alters the GM (Transconductance) around 77+- (broad(low q)) Cycles per second. Yes, many more hams in the mid 30's studied and knew vaccccccum tube theory.

http://en.wikipedia.org/wiki/Transconductance

Regards  John, K1DEU


Please click on http://www.tvfools.com  , once its a bookmark you will use again in our future even if we have cable or satellite TV. Do not be mislead by the url its serious terrain path software !

http://amfone.net/Amforum/index.php?topic=18294.0