Decreasing Intermod susceptibility of the Kenwood TM-733

Front panel of the Kenwood TM-733A
The front panel of the Kenwood TM-733A

The Kenwood TM-733A is an older dual-band mobile radio.  This radio is capable of 50 watts transmit on VHF, 35 watts on UHF, full-duplex crossband repeat, and simultaneous receive on two VHF or UHF frequencies.

It does have a problem, however:  It is terribly sensitive to being overloaded by other signals.  The effects of overload (or "intermod" - commonly referred to as IMD meaning "InterModulation Distortion") manifests itself as the reception of signals that aren't really on frequency.  During a 'bout of intermod, one will often hear QSOs of other repeaters (amateur and not) as well as odd bleeps and buzzes from paging and telemetry systems.  Because intermod is a mix of multiple signals, you cannot experience it from just one signal - it takes at least two signals!

Even before I got my own '733, I was aware of the problem:  Back in early 90's a friend of mine bought a TM-732 - the predecessor to the TM-733.  It had some pretty nice features and a reasonably intelligent (I thought) operator interface.  But it had terrible intermod problems:  It would seem to squeal and squawk at almost any time, anywhere due to front-end overload.  He asked me to look into it and see if there was something that I could do about it.

I took a look at the schematic and did some testing.  It became obvious that there was way too much RF amplifier gain ahead of the mixer.
What is "SINAD?"

For an FM receiver, the stronger the signal, the less noisy it is.  When quantifying sensitivity measurements, a common standard is to have a signal with a 1 KHz tone at a deviation of 3 KHz.  The "SINAD" is the ratio of that 1 KHz tone to the background noise.  The "12 dB SINAD" measurement is how much signal is required to get that tone 12dB above the "popcorn" noise one hears on a weak FM signal.

How noisy is a signal with 12dB of SINAD?  Pretty noisy - but if the other person is talking in a normal voice (i.e. not quietly or far from the microphone) 12dB of SINAD is good enough that, although noisy, one doesn't really have any problem understanding any of the words.

The fix was fairly simple:  Add some attenuation.

"What's that?  Attenuation?  How will that help?"

When it comes to intermod, it (generally) follows a simple rule:  Reduce the strength of the offending signal(s) by 1 dB, and the resulting intermod products will decrease by about 3 dB.  So, if you can reduce the amount of signal by, say, 6 dB, then the intermod will (theoretically) decrease by 18 dB. Note that if the interference problem is due to image response, then this rule does not hold - which makes sense, as an image problem is not the same is an IMD problem.

The trick is how to add sufficient attenuation without adversely impacting overall receive sensitivity.

As it turns out, there are multiple RF amplifier stages in the TM-732 (and the '733.)  As with any receiver, once one has enough amplification for the intrinsic RF amplifier noise to exceed the noise of the mixer, the further amplification does no good:  It simply amplifies the noise along with the original signal - it would be akin to turning up the volume on a noisy signal and not expecting the noise to get louder along with the desired signal!  Worse than this, each dB of additional unnecessary gain can also unnecessarily increase intermod susceptibility by 3dB - more or less.
12 dB (uV) SINAD 
S1 Signal (uV) 
Level (Before)
12 dB (uV) SINAD 
S1 Signal (uV) Level 
144 0.131 0.300 0.158 0.464
146 0.131 0.350 0.158 0.474
148 0.133 0.380 0.160 0.489
155 0.186 0.650 0.185 0.596
160 0.245 0.93 0.25 0.920
420 0.112 0.042 0.145 0.275
430 0.127 0.082 0.171 0.288
440 0.139 0.139 0.177 0.405
450 0.140 0.167 0.168 0.308
460 0.168 0.300 0.214 0.530
Table 1: This table shows the sensitivity of the "main" receiver portion of my TM-733 before and after modification.    These readings were taken with the "AIP" (attenuation) off.  With the AIP on, sensitivity on VHF and UHF was about 6db worse (i.e. about 0.35 microvolts for 12 db SINAD in the amateur bands - see text.)

The moral of the story is that with any receiver design - one should have just enough gain to obtain the desired sensitivity - and then no more.

The trick was to put the attenuation after the first RF amplifier stage:  Signal levels to the mixer could be decreased without reduction of overall sensitivity if one didn't go overboard.  Also, I knew that the actual sensitivity of the radio was significantly higher than the published ratings - so I was willing to decrease sensitivity just to the point of being within the original specifications of the radio.

It is worth mentioning that any stage of amplification in a receiver can contribute to intermod problems - but it is most likely to be the mixer stage that contributes the lion's share of the problem:  The amplifier stages are intended to be linear - but the mixer, by its very nature, is not!  The "ideal" mixer would only have the expected mixing products (the sum and difference of the local oscillator and the input signals) but most mixers are quite far from ideal - especially when signals are strong.

After modification, sensitivity was (pretty much) just within the published specifications of the radio and the susceptibility to intermod was  considerably reduced.  Alas, my friend's '732 was stolen not too long after that - and I had lost my notes on the modification...

Modifying the TM-733:

Not too long after that (in about 1995) I bought a new TM-733.  Unmodified, the intermod wasn't quite as bad as that of the unmodified '732 - and it had an "AIP" circuit (read more about this circuit below...):  This circuit simply reduced the gain of the first RF amplifier stage in order to reduce the amount of signal getting to the mixer - and therefore the susceptibility of the receiver to intermod.
Schematic showing modification to the VHF receive portion of the TM-733
Figure 1:
Schematic showing the modification of the VHF portion of the TM-733
Click on image for a larger version

Even with the AIP turned on, the performance of the '733 (in terms of intermod) was still pretty bad in metro areas - but owing to prior experience with the (very similar) '732 I knew that there was something that I could do about it.

This time, I documented things.  Table 1 shows the "before" and "after" performance of the main receiver section.  Keep in mind that the actual receive sensitivity cannot necessarily be used to determine the change in gain of the RF amplifier sections:  The readings on the low end are often limited by the intrinsic noise of the amplifier stage(s).  Rather, the amount of signal required for an S1 S-meter reading is more useful for that - as long as the noise doesn't contribute to the meter reading.

A note:  The TM-733 was adjusted in accordance to the service manual - before and after modification.  From the readings it is apparent that, at least on VHF, even with the added attenuation, the receiver sensitivity does not suffer on frequencies in the 160 MHz area - more on this in a moment.  However, in the frequency range of interest (the 2 meter band) the modification resulted in an approximate 1.5-2dB reduction in signal level to the mixer.  Theoretically, this means that, according to the 1-3dB rule, the IMD is to be reduced by 4-6 dB or so.
Schematic showing modification to the UHF receive portion of the TM-733
Figure 2:
Schematic showing the modification of the UHF portion of the TM-733
Click on image for a larger version

In looking at the schematic in Figure 1, one notices something else:  R10 - originally 47 ohms, changed to 150 ohms, also affects the loaded Q of the bandpass filter, consisting in this case of L4, D17, and D5.  The higher Q also reduced the bandwidth - further reducing the number of off-frequency signals getting to Q3 - and, by extension, the mixer.  An additional modification - R9 and the added component R9A reduce the amount of signal fed to the "Sub" receiver section.  This is important, as both the VHF and UHF "sub" receiver aren't particularly intended to have either the sensitivity or the intermod resistance of the main receiver: Refer to Table 2 for "before" and "after" sensitivity performance of the "sub" receiver sections.

Figure 2 shows the modifications done to the UHF section.  For the main receiver, R206 is changed and R206A is added.  IC203 is actually an additional converter section that allows the '733 to receive in the 300 MHz area - and, in fact, is used to route 70cm receive signal to the remainder of the 70cm receiver section.  Note:  IC202 is present on those (non-USA) units that can receive in the 800-900 MHz range and simply adding it to a USA-version of '733 will not enable 800-900 MHz reception owing to different CPU code.  R205 and the added R205A reduce the signal levels to the "sub" UHF receiver, located on the VHF board.
12 dB (uV) SINAD 
S1 Signal (uV) 
Level (Before)
12 dB (uV) SINAD 
S1 Signal (uV) Level 
144 0.123 0.180 0.159 0.420
146 0.123 0.190 0.155 0.470
148 0.115 0.205 0.157 0.530
155 0.125 0.231 0.167 0.560
160 0.153 0.218 0.180 0.510
420 no data taken - 0.147 0.291
430 no data taken - 0.147 0.292
440 no data taken - 0.151 0.275
450 no data taken - 0.157 0.296
460 no data taken - 0.176 0.361
Table 2:
This table shows the sensitivity of the "sub" receivers of my TM-733 before and after modification.  These readings were taken with the "AIP" (attenuation) off.  With the AIP on, sensitivity on VHF and UHF was about 6db worse (i.e. about 0.35 microvolts SINAD in the amateur bands - see text.)

Looking again at Table 1 one can see that the actual added attenuation (in the UHF receive signal path) amounts to 8-10 dB - resulting in a theoretical IMD reduction of 48-60 dB.  In actuality, the IMD reduction is probably nowhere near this amount - but in real-world use, the radio's susceptibility to IMD is greatly reduced.

Considering that the rated sensitivity of the TM-733 for a 12 dB SINAD signal is 0.16 microvolts on both 2 meters and 70cm (in the amateur bands) the target sensitivity was pretty closely reached.  One will notice that the sensitivity of the main receiver on the 70cm amateur band is closer to 0.18 microvolts than 0.16 - but that is only a difference of only about 0.5 dB - a (practically) insignificant amount.

Table 2 shows the effects of the modification on the sensitivity of the "sub" receiver.  Unfortunately, I seem to have forgotten to record the original sensitivity of the UHF "sub" receiver.  It is worth mentioning that the rated sensitivity of the "sub" receiver is 0.25 microvolts - so even with the added attenuation the sensitivity of this receiver exceeds the official specifications of the receiver.  With the "AIP" turned on, the sensitivity decreases to approximately 0.35 microvolts in the amateur bands.

Analysis of the results:

If you look at the data in Tables 1 and 2 you might notice an interesting thing:  Being that the S-meter reading is representative of the actual amount of signal power getting to the demodulator (and therefore the mixer) you'll notice that, in many cases, the sensitivity decreased by an amount less than the signal level did.

A case in point:  Look at the readings for 146 MHz in Table 2:  You'll notice that the receiver sensitivity was reduced from 0.123 microvolts to 0.155 microvolts - about 2 dB - while the S-1 sensitivity was reduced from 0.19 microvolts to 0.47 microvolts - nearly 8 dB.

How can that be?  This is a classic case of excessive gain.  Before modification the receiver was hearing quite a bit of the noise intrinsic to the RF amplifier stages themselves.  In fact, one could have reduced the gain by 3-4 dB without having any significant effect on sensitivity at all!  What we know now is that every dB of extra gain in our receiver can increase our susceptibility to intermod by 3 dB, so it wasn't unreasonable to decrease the gain (and thus the sensitivity) by as much as was done.  If we wanted to be sticklers, we could have knocked the gain down by another 4 dB or so (for a total of 12dB) and still would have come out about at Kenwood's sensitivity specifications for the "sub" receiver (e.g. 0.25 microvolts.)

The "AIP" circuit:
Closeup of the circuit board showing modification to the VHF receive portion of the TM-733
Figure 3:
Schematic showing the circuit board modification of the VHF portion of the TM-733.  These components are mounted on the component side of the board.
Click on image for a larger version

The '733 has a "feature" that has become pretty common on mobile rigs these days.  Kenwood calls it "AIP" (tm) which stands for "Advanced Intercept Point" (tm).  All this does is to decrease the gain of the first RF stage of the receiver - reducing the total amount of signal energy that the mixer section gets hit with, therefore reducing the susceptibility to intermod as well.  As it turns out, the AIP reduces the gain of the '733 by about 7dB on VHF and just under 6 dB on UHF - theoretically reducing the susceptibility to intermod by approximately 21 and 18 dB respectively.

Does this actually work?  Yes - it does.  The sensitivity of the main receiver on both VHF and UHF is in the area of 0.35 microvolts or so when the AIP is turned on - about the same sensitivity of a typical "commercial" (e.g. non-ham) receiver.  In metro areas, one can expect nearby repeaters to have signal levels in the area of 1 to 100 microvolts - so even with the "AIP" mode turned on, one doesn't miss anything - and the intermod susceptibility is reduced even more (that is, the "AIP" does an even better job.)  In more rural areas, there aren't usually the numbers of strong signals to cause intermod, so you can operate without the AIP active.
Closeup of the circuit board showing modification to the UHF receive portion of the TM-733
Figure 4:
Schematic showing the circuit board modification of the UHF portion of the TM-733.  These components are on the foil side of the board.
Click on image for a larger version
Note that the AIP setting also affects the "sub" receiver for that band, as both "receivers" share the same 1st RF amplifier stage, and thus, the AIP affects both receivers equally.

Doing the modifications:

It should go without saying that one should not even attempt these modifications without having the appropriate skill and familiarity with surface-mount electronics - and before you even ask:  No, I will not do the modifications for your radio, nor can I provide copies of the service manual!

Note:  These modifications were performed in 1995 or 1996 and based on my notes, I took the attached pictures to properly document these modifications.  Hopefully, I have covered everything - but it is possible that, over time, I have forgotten some important detail.

Figures 3 and 4 show close-ups of the TM-733's circuit board, showing the added and changed components.  Note that these pictures are intended to provide a rough guide to the modifications - but one should have access to the entire schematic of the radio and, preferably, the service manual as well.

Ideally, one would use the same size of surface-mount components as on the radio, but I had only a larger size, necessitating that some of them be mounted on edge.  Some of the added resistors (the ones with the "A" suffix) had to have one end connected to ground, so a suitable ground location was found, scraped, and soldered to.  It is not recommended that ordinary leaded resistors be used - especially for the UHF modifications - as clearance to the chassis (which is grounded) needs to be maintained.

Consult the service manual for additional information on removing the circuit boards - a required step for the UHF-band modifications.  The following steps are for removing the two circuit boards - but remember to consult your service manual.  Note that the task is fairly difficult, requiring a desoldering tool at the very least - and one should consult the service manual before doing this:  First, the RF amplifier modules should be carefully removed.  Next, the RF coaxial cables should be removed - carefully avoiding overheating of the coax to avoid melting.  Next, the DC power leads should be disconnected.  Next, the front circuit board should be removed (there are screws that hold it into place) along with the fragile ribbon cable.  Finally, there are some heavy metal posts that provide DC connections.  After removing the screws, the boards will lift out (albeit awkwardly) and you'll notice that there are several pins that pass signals to-and-from the two boards. Note that the old heat-sink compound (if there was any - there was a period in the mid '90s during which Kenwood didn't seem to be putting enough compound on the power modules!)  should be cleaned from the power module and the TM-733's chassis and fresh heat sink compound should be applied prior to installation!

After modification, one should perform an entire receiver alignment to maximize performance as tuning (especially on VHF) may have changed - although the tuning difference will slight.

The "Fine Print":

Note:  Terms such as "Kenwood," "TM-733," "AIP," "Advanced Intercept Point" and probably a few others are trademarks of Kenwood Communications Corp.  The portions of schematics are shown only to provide an accurate description of the modifications.  I neither can or will be able to provide complete schematics or copies of the service manual.  The schematic originally came with the radio and copies of the service manual can be obtained via Kenwood spare parts.

These modifications are without any sort of warrantee - either expressed or implied.  While reasonable care was taken in the preparation of this document, I cannot be responsible for the outcome of these modifications.  It is strongly recommended that you do not do these modifications unless you have the schematics of the radio, a service manual, appropriate skill, and test equipment to perform and validate them.

Your mileage may vary.  Do not taunt Happy Fun Ball.

Go to The KA7OEI FT-817 "Front Page" 

Any comments or questions?  Send an email

This page maintained by Clint Turner, KA7OEI and was last updated on 20060328.  (Copyright 2002-2006 by Clint Turner)

Since 12/2010: