Delta Audio Modifications

By Jeff Depolo, WN3A

Edited By Dave Karr, KA9FUR


The modifications presented here were specifically developed for utilizing Canadian split, UHF Delta-S radios for Auxiliary Link service use in the Amateur service.  We have seen at least two different versions of TRS boards which contain minor board layout differences.  The pictures presented here should generally be correct for all Delta-S and -SX radios but I would encourage you to cross check component locations with the appropriate GE/Ericsson manual for your paticular radio.



First of all, most of these Canadian radios have pin 26 on the front connector wired up to DATA IN.  This was a custom mod; most radios from other sources probably wouldn't be configured like this.  DATA IN is a connection that you can put audio into which bypasses the mic processing circuitry (i.e. no preemphasis, no limiting).  IMPORTANT: this point is DC-coupled!  You MUST AC-couple it with a cap.  The DC bias is about 4.5V at this point, input Z 10K.  A non-polarized electrolytic would be a good choice (math in my head says 2.2 uF should be plenty to pass PL without attenuation).

However, even using DATA IN and bypassing the mic processing, the high frequency response has some roll-off.  The DATA IN connection is summed with the limited mic audio and channel guard tone, and then goes into a second-order low-pass filter.  So, if you need flat response up beyond 2.5kHz, you will still need to do some mods to flatten it out (either that or boost the high end yourself externally).  I've come up with two versions of the mods, one shifts the filter up to around 4.5 kHz, and another that makes it a "barn door" transmitter that is flat up into the ultrasonic range.  If you chose the "barn door" mod, you damned well better do your own audio processing outside of the radio.

If you want to use the mic limiter, I have also provided the mod to eliminate the preemphasis.  Keep in mind that the gain of the mic stage is pretty high (from roughly 6 dB to 30 dB depending on frequency due to the preemphasis), so either you have to feed it low-level audio, or modify the gain of the stage.  I'm assuming just about all of us will want to run flat audio through these transmitters, so the deemphasis is going to get axed anyway.  With the deemphasis gone, the stage gain is only about 6 dB so no sense in modifying it further.

So, there are a couple of combinations of mods you can do depending on what you need/want.  If you're going to use pin 26 (DATA IN), then there's no need to modify the preemphasis in the mic limiter.  If you're going to use the mic limiter for deviation limiting, you will still need to do one or more of the other mods.

Also remember that there is DC bias on the mic line, so AC couple accordingly, or perform modification #5 below.

You don't have to remove any shields to do these mods.

  1.  Remove C311, a .047 uF leaded cap.  Pictures show where it is.
  2. That's all.
With the deviation limit set to 5kHz, the microphone input sensitivity will be about 100mV per kHz of deviation.

  1. Replace C305 (.068 uF) with .047 uF.  Regular leaded component-see pics.
  2. Replace C309 (.001 uF) with 470pF.  This is a surface-mount device. If you want, you can remove the surface-mount C309, and then just tack-solder a 470pF leaded cap between pins 1 and 2 of U301 (4558 dual op-amp, 8 pin).
The frequency response will now be flat (within .1dB) up through 3500Hz and the -3dB point will be about 4.6kHz.


This mod is mutually exclusive with Mod #2 (if it wasn't already obvious to you).

  1. Replace C305 with .0056 uF or remove completely.
  2. Remove C313.
  3. Remove C309.
  4. Don't answer the phone if "HOLLINGSWORTH, RILEY" shows up on your caller ID.


  1. Replace C307 (.022uF) with 0.22uF.  This will extend the low frequency response.  C307 is a 1206 sized surface mount part.
  2. Replace C310 (10uF) with 100uF.  This dampens a positive feedback path around U301-B eliminating oscillation of this stage after the low frequency gain is extended by changing C307.  Stability confirmed in both real life and with pSpice Models.
With this modification the low end -3dB point will be about 35Hz.


DC Bias can be removed by either removing R651 or R652.  If R652 is removed, the microphone input impedance will be approximately 27k ohms (R307).  If instead R651 is removed, the input impedance will be dominated by the series combination of R652 (649 ohms) and C656 (15uF) which are series connected to ground.



Decreasing Carrier Squelch response time

When used as auxiliary link receivers it may be desireable to speed up the squelch action.
  1. Change C605 (.1uF) to 0.047uF, or adjust to taste.
  2. That is all


All radios include the provisions for an internal squelch control.  However not all radios are equipped with them.  The picture below shows the location of the fixed squelch control.  In order to enable the internal squelch control the shunt on J605 needs to be removed.  In the picture, the shunt is shown in a stored position with one half of it on pin one (the far left pin).   Note that if the shunt is moved between pins 1 and 2, that Pin 15 on the radio connector becomes a squelch disable line.

For radio's that are not equipped with the control, it appears that a Bourns 3309 Series trim pot will fit the board.  Digikey part number 3309P-103-ND.

No modifications are neccessary for the receiver audio, but here is some information on the receivers audio performance characteristics.

Volume Squelch High (VSH) is the best point to obtain audio.  VSH can be optionally passed through the CTCSS filter (if installed) by applying a shunt to either J608 (unfiltered/bypassed) or J609 (filtered).  See picture above for shunt locations.

The swept frequency response of the receiver is shown here and is quite flat, within 1dB from 60Hz through 3.5kHz.  Flatter than most other radios I've put under the microscope.

For repeater use you may or may not want to use the the CTCSS High Pass filter.  Its swept frequency response is shown here and is "fair".  The attenuation in the CTCSS region isn't quite as good as some other filters (like the Micor LC filter).  This filter was obviously designed to only reject tones less than 210.7Hz, i.e. the "M" tones wouldn't be filtered well.  As you can see by the plots, there is about a 3dB peak in the response around 320Hz.  The filter is underdampened to steepen the slope of the filter below the cut-off frequency.  If I were going to use this filter for tones in the 100-150Hz range I'd probably move the cut-off down a little lower in frequency and make it critcally-damped to eliminate the overshoot.

Distortion numbers were obtained using a Potomac analyzer.

I put a Delta, Micor, and Mastr II on the bench to test distortion.  The numbers I got are a bit odd (at least I think).  I tried various ways of tuning the IF filtering in the Mastr II and the Delta.  Optimizing the distortion using 1kHz tone at 5kHz deviation seems to yield about the best average numbers.  The Micor doesn't have anything to adjust other than the discriminator.

I ran one set using the internal 400Hz and 1000Hz tone generators in the Fluke 6060A, and another using the Fluke modulated by the HP 3311A function generator.  The HP read 1.62% at both 400Hz and 1000Hz on the Potomac when measured directly, so it's hard to say for sure what I'm seeing here.

On the Micor, readings were taken off the discriminator buffer amp (emitter-follower).  The Delta and Mastr II were taken off Volume Squelch High.

The other thing I determined is that the tone generators in the IFR1500 suck, and that it is useless to try to use the IFR as a receiver for distortion testing of transmitters.

Distortion Comparison.  Results expressed in percent
Audio Source Delta Mastr 2 Micor
Fluke internal, 400Hz tone @ 3kHz Deviation 1.0 2.6 1.3
Fluke internal, 400Hz tone @ 5kHz Deviation 2.0 3.3 2.9
Fluke internal, 1kHz tone @ 3kHz Deviation 2.3 3.4 1.8
Fluke internal, 1kHz tone @ 3kHz Deviation 4.6 4.9 3.7
HP 3311A, 400Hz tone @ 3kHz Deviation 1.5 3.3 1.4
HP 3311A, 400Hz tone @ 5kHz Deviation 2.1 3.8 2.7
HP 3311A, 1kHz tone @ 3kHz Deviation 1.6 3.8 2.0
HP 3311A, 1kHz tone @ 5kHz Deviation 4.4 5.1 4.1
Simple Average 2.44 3.78 2.49