Executive Summary
Thanks to the internet and a variety of mailing lists, it's no secret there's a problem with the 746Pro transceiver.  My own 746Pro crapped out in less than 6 months of operation, so this problem was involuntarily foisted upon me. Rather than go through the Icom warranty mill I did my own investigation into, and subsequent repairs of the problem. Since that time I have discovered others who have done a similar effort and devised their own solutions as well.  It only made sense to incorporate all of the known fixes into the radio at one time and take care of the problem, no matter which specific issue is the root cause of the failure.

There are several failure modes suspected as the cause of the "sudden death" 746 transmit problem. They are:

• ESD event
• Heat
• Input switching transient
• [Reserved] (for Future Failure Discovery)

IC151 undergoing thermal testing using a mockup heatsink.	A view of the competed SSOIC installation on the SOIC pads, along with the homebrew heatsink.

The UPG1678 from Digi-Key is an SS08 package, not an S08 as originally used in the radio.  If you're careful, you can bend the leads out and get them to fit on the S08 pads.  Alternatively you can get the exact replacement from Icom.  If you're reading this before your radio died and are doing the fix as a preventative measure, you can leave it alone.

I couldn't find a retail (or even OEM) source for the 1SV252 diodes, so the most convenient source becomes Icom for those. If you don't care about using "genuine" Icom replacement parts, the  BAT54 diodes from Digi-Key are a form/fit/function replacement and I think a better substitute than the original.  While it may seem superfluous, I would order at least two sets of everything.  This will cover at least one botched attempt or lost part.  Having spares on hand will save you much time and effort than if you'd only ordered one, had a problem, and had to repeat the order and wait.  It's only a few extra bucks.

Something you may consider doing while having your 746Pro on the operating table is to install the CR-282 High Stability Oscillator module.  If you operate much on 2 or 6M weak signal modes, the 7ppm frequency (in)stability of the 746Pro is pretty noticeable.  You'll have the radio apart to that point anyway, and you can fix two problems at once (drift and sudden TX death).  Just a thought.

The 1SV252 SOT diodes are the ESD protection diodes.  If they aren't present in your radio, it would be a good idea to install them at this point.  Even if your radio already has these installed from the factory, it might not be a bad idea to replace them.  If your IC151 was blown by ESD, it's a given that one or both of the diodes are fried or weakened and should be replaced.  Even though their overall effectiveness is questionable, it doesn't hurt to put them in.   You will have to remove the RF Unit PCB to get to the backside of the board to install/replace the diodes.  It will help to remove the small shield that covers one of them.  Details on how to remove the RF Unit PCB can be found on page 91 in the 746Pro Instruction Manual, in the section that describes how to install the high stability oscillator module.

After the diodes are in and the shield soldered back in place, trim the leads of the 1mH choke to about 1/8" and solder it on the bottom of the PC board connected between the pad for the center pin of the HRX connector (J1) and the PCB groundplane.  You will have to scrape away some solder mask on the groundplane to expose the copper.  That should shunt any errant DC and static that should come down the HRX pipe. This is the intended function of the Icom diode fix, but a choke is a lot less likely to get blown out as the diodes are.  This completes work on the bottom side of the PCB.

This diagram shows the location of the "protection diodes" and the soldering points for the choke.	Here is a view of the PCB with the choke soldered in place.  There is plenty of room for the choke even standing up like it is.

 
On the top side of the PCB, R&R C154 with the .01 cap. That takes care of the RX-TX switching transient issue.

If your IC151 is dead you will have to remove and replace it. If you order it from Icom you'll get an exact form/fit/function S08 replacement. If you order it from Digi-Key or Mouser you'll get one in a different package, as the original SO8 version was obsoleted by NEC/CEL some time ago. The two versions are functionally and electrically interchangeable.  It is possible to fit the SS08 version on the S08 pads, but will take some lead bending/extending and careful soldering.   If anyone has a source for real S08 UPG1678's other than Icom I'd love to hear about it.  The only sources I've found are obsolete parts brokers that will only sell them by the reel.

You may consider adding a heatsink at this point as described on OZ2M's webpage, though fabricating one to work with the tiny SS08 version can be a challenge.  I made a couple attempts at making a heatsink before I finally came up with one that was good both electrically and mechanically. Unfortunately it's not something most hams would probably be comfortable in dealing with. This IC is barely 1/8" square, including the pins.  It fits comfortably on the end of a pencil eraser.  I used some 18ga copper sheet from a hobby store for heatsink material.   It had to be made out of several small pieces, otherwise it worked too well as a heatsink and I couldn't get it hot enough to melt solder on it.  I made the base by filing away the areas where the SO8's input, output and Vcc pads were on the PCB. Then I bent the SSO8's input, output and Vcc leads up so they wouldn't touch the copper ground/heatsink.  Kapton tape went over the PCB where IC151 resides to insulate a hot via there. I tinned everything, fluxed it and in one step using an 800F iron I flowed the copper heatsink base to the PCB ground pads, and the SSO8 IC to the copper heatsink base all at once.  Then in a second step, I soldered two square copper fins to the tabs on the base.  Now, since the group of fixes detailed here includes switched TX VCC just like the 756Pro, which doesn't have a heatsink and doesn't suffer the sudden TX death problem, it may be overkill to heatsink the part.  It's up to you.

The last step is to install the zener which will switch power to IC151 only during transmit. Remove R157 and solder it standing up on it's pad closest to IC151. Solder the anode end of the zener to the top end of R157. Using some 30ga wire wrap wire, connect the cathode end of the zener to the end of R156 closest to IC151 (the TX switched 8V point) as shown in the PCB diagram.  Between the voltage drop of the zener diode and R157, IC151 sees about 4.7VDC during transmit, which is perfect.

This diagram shows where the various affected components are located on the top side of the PCB.  The red dot indicates the TX 8V connection point.	This is a view of the competed mods done to my own 746Pro..

Smoke Test
That completes all the mods that are known to date.  Button it back up, put the fire to it, and if everything went right your 746Pro is once again a happy camper.  It shouldn't require any alignment or other adjustments.  Any slight differences between the old and new amplifier IC is automatically compensated for since this stage is within the ALC loop.  A while back, Allen AE0Z posted a question on the 746Pro Yahoo Group page wondering if the switched TX VCC modification contributed to any QSK CW bit truncation.  To verify this I "unmodified" mine, tested the CW keyup delay, then tested it again with the switched TX voltage in place. No measureable difference was noted between the two (about 12mS).  Now, rather than push your luck again, this would be a good time to install some quality DC shunt type arrestors on the feedlines coming into the shack, such as the ones available from Array Solutions.

Back to the Main Project Page	Back to the K5LXP Home Page
© 2004 Pane Relief Computer Services