From: "steve_cuccio" Date: Tue May 11, 2004 6:33 pm Subject: IC-151 failures Dave AK2A recommended I post some info about this to the group: For anyone interested in replacing a blown out uPC-1678 TX driver IC, make sure you change the value of the DC blocking cap C154 from 0.1uF to 0.01uF (and maybe the output cap C157 too). Two reasons: One is the 1678 amp is not unconditionally stable into all loads over it's broad frequency range, in particular around 1500MHz according to the datasheet S-parameters. The 0.1uF caps go self-resonant well below this frequency (probably below 100MHz) which means they will look like a lossy inductor at the higher frequencies and may contribute to circuit instability. I'm also sure keeping the IC powered up during RX while the amp circuit is unterminated does not help either (I recall some one did suggest running it's supply from switched TX DC bias which is a good idea). While a 0.01uF may still be self-resonant below a GHz, it is probably a better compromise since it's dielectric is typically a better performer at higher frequencies (X7R versus Y5V) and will still maintain a fairly low reactance (not mess up the driver VSWR) down at 1.8 MHz (160 meters). The second reason to reduce the input blocking cap is when the D151 PIN diode bias switches on during TX, the bias change causes one heck of a low frequency transient through the IC. It's 22uH RF output choke looks almost like a short circuit during this transient. NEC makes no claim on VSWR tolerance for this part which might explain why these things pop all of a sudden. I remember seeing some fellow change the 10 ohm DC bias resistor R157 to 47 ohms as a cure. This would only cause the amp's supply to sag both during the switch transient and actual TX drive which might result in other problems. Best thing is to reduce the transient before the amp's input. So far, I've done this fix on a few 746's and more than a few 746Pro's without any of them coming back (with the same problem). While replacing the 1678 with it's smaller footprint 1678GV is a pain, it does go in easy if you have a decent stereo microscope, a sharp X-acto knife to scrape solder mask and cut coppper track, and some patience. If anyone is interested, I could probably dig up some close-up pictures of the surgery. Just my two cents. 73, Steve NB3O DATE: Mon, 24 May 2004 08:12:50 EDT TO: SUBJECT: Fwd: 746Pro Yes, the older 746 has the same IC-151 (uPC1678) failures too, although not as common. Not sure why since they have similar circuitry, although maybe more heat in the Pro aggravates the condition or it possibly has a faster rise time for its T8V transmit supply line (see text). First, let's walk through the transmit DC bias path. In referring to the 746 (non-Pro) schematic, switching diode D151 is forward biased during transmit to complete the RF path to the input of IC-151 from one of the preselector filters (low-pass for broadcast band and 160 meters, band-pass for all of the higher bands). Its DC bias supply is derived from the T8V supply line (+8 volts during transmit) which is fed through R156 current limiting resistor, and the bottom end of the R154 "T" attenuator shunt resistor. The DC bias path to D151 is completed by one half of the "T" attenuator series resistor R153, through D151 itself with the other side exiting through L151 and going through R152 back to ground. The R151 resistor connected from R8V (+8 volts during receive) to L151 serves to provide back bias to D151 to turn it off during receive. While they saved an RF choke by re-using the "T" attenuator to pass the D151 DC bias, some one forgot that the other side of the "T" attenuator series resistor R155 will also have DC bias rising up during transmit. During the changeover from receive to transmit, the change in potential at R155 will go from 0 volts to 2.6 volts somewhere around 100 microseconds by virtue of the time constant formed by R156 and C153, given the assumption that T8V rises quickly during transmit. Voltage at R153,R154,R155 "T" attenuator junction during transmit can be calculated as follows (I'm picking the long hand version for simplicity): First determine the current through the DC bias path during transmit. (8 volts minus D151 0.7 volt diode drop) divided by the sum of (R156, R154, R153, R152) or 7.3/(330+270+4.7+220) = 8.85 mA Then calculate the voltage present across R152 and R153 and add the D151 diode drop back in. 8.85 times 220+4.7 = 1.99V plus 0.7V diode drop = 2.69V If the T8V supply was switched instantaneously during transmit, we could assume that C153 and R156 would define the time constant (maybe bad assumption, but I have run the scope there yet to see so I assume worst case). Lacking any rise-time measurements on T8V, multiply C153 by R156 to get the time constant of the DC bias rise time at D151. 0.1E-6 times 330 = 33 microseconds. Using the natural log function, 33 microseconds will charge up the circuit to 63%, 66 microseconds will charge up the circuit to 86%, 99 microseconds will charge up the circuit to 95%, etc. Within 100 microseconds, R155 will be at 2.55 volts, again assuming that T8V has an instantaneous rise time and lacking any effect of IC-151 loading the circuit. It really would be nice to get the scope out and look at this, but I ain't pulling the covers off until it breaks again. The time constant on the other side of R155 through C154 (our cap in question) is a little more complex, since the input impedance of IC-151 is not listed below 50 MHz. I'm not sure that its relevant anyway since the data sheet says the absolute max input RF level is +10 dBm. How does this compare with the sudden 2.6V transient in approx 100 microseconds on one side of C154? +10 dBm into 50 ohms is 0.7 volts or 2 volts peak to peak or six tenths of a volt less than the transient. Does any of this voltage get differentiated (high-passed) by the time constant of C154 and the input impedance of IC-151? Maybe. But every time I read the following in its data sheet, I figure it just ain't worth the gamble and would rather reduce the cap value to be sure: "PIN Input Power dBm +10" "Notes: 1. Operation in excess of any one of these parameters may result in permanent damage." (Bam! DOA!) Could you fit a "T" high-pass in front of IC-151 that will pass 1.8 MHz? Yes, with a little hand carving. Like I said in an earlier email, if IC-151 craps out again after initially changing C154 to 0.01uF, then I also change R155 to 0.01uF and add an inductor to form a high-pass "T" filter. I would probably "Tee-Pee" a 0.01uF chip cap with a 4.7 ohm resistor instead of yanking out the resistor entirely to help maintain the integrity of the 1.5 dB "T" attenuator for stability's sake. A 22 microhenry chip inductor (X=247ohms at 1.8MHz) is attached from the two series caps to ground. Coilcraft makes the 0805LS-223XJBC which is in a tiny 0805 package (they supply samples, too), the next package size up from the 0603 size R155 and should squeeze in there just fine if you can clear a section of ground to solder its other end. Icom uses a 22nH inductor on the output of IC-151 to feed the DC to its output pin, so when in Rome, etc, etc. Why keep the 4.7 ohm resistor in the circuit? I had second thoughts after my previous email and took another look at the data sheet. It should be noted that R155 at 4.7 ohms does serve as an element in the 1.5 dB "T" attenuator. Given the gain of IC-151, I don't think 1.5 dB plus or minus means a hill of beans, however the original intent of the designer was to help normalize the input match of IC-151 to keep everything (amp and filters) in a better termination impedance for stable operation. Placing the 1.5 dB pad in front of IC-151 will provide an additional 3 dB of return loss. An example would be the difference between a match of 2:1 VSWR and 1.6:1 or in the out-of-band case, from infinite VSWR to 6:1 (the difference between an amp and an oscillator). This can prevent IC-151 from running unstable (oscillating) above 600 MHz since the data sheet shows a dependency on keeping a good match throughout the part's range of gain where its stability factor (K) is less than 1. This goes past 2GHz! The preselector filters look pretty awful above a few hundred MHz, so it might be wise to tread lightly if you consider yanking out R155 to put only a cap in there. If I did decide to do this mod without Tee-Peeing the 4.7 ohm resistor with the cap, I would use a vector network analyzer to confirm stable operation with all combinations of preselector filters, or at least, confirm using a spectrum analyzer that IC-151 is not oscillating all the way up past 2 GHz. To be honest, I'd rather be on the air instead. This is why I am hesitant (lazy) to do anything more than swap out C154 unless my hand is forced. Hope this helps. 73, Steve, NB3O DATE: Mon, 24 May 2004 08:13:04 EDT TO: SUBJECT: Fwd: 746Pro I recently purchased a 746 (non-Pro) from Dave AK2A. This is my fourth 746. We use all of them for Emcomm, contests, Field Day, etc because they have low phase noise and are easy to operate. Two out of my previous three had the sudden NEC uPC-1678 driver amp (IC-151) failures. I mentioned to Dave that I replaced them with the smaller 1678GV after carving up the PCB, but also lowered the DC blocking cap at the input and have not seen them crap out yet (after about two years). He asked me to sign up and post to the 746 group at yahoo.com to share the info. My experience with the uPC1678 goes a few years back before the 746 used them. I designed them into local oscillator amp sections in cellular base stations because they have decent gain and can drive the high level mixers we were using. It was at this time we noticed that hitting the input of these things with a DC spike, especially when its output had a small value RF choke blew them up easily. I also noticed the same thing with EIC Corp's ECG-003 and other similar high-gain, high-freq MMIC's. When the devices are powered down, they do not blow up with the same input spike, so my guess is the last stage in the amp is failing. I suspect that they just cannot tolerate a brief short at their output without some sort of current limiting because the manufacturers are trying to squeeze all they can to attain high output drive with a pretty small silicon die. The current spike probably punches through them, maybe more frequently when they are warm and exhibit slightly higher DC gain. Our fix in previous days was to limit any DC or low frequency blast at their input by using a "T" type high pass filter consisting of small DC blocking caps and a shunt inductor designed to pass the lowest frequency of interest unattenuated. This seemed to keep them happy and reliable. The other mods like the one mentioned somewhere for the 746 that increase the DC supply 10 ohm resistor (R157) to 47 ohms will cause the +5V supply to drop from 4.5V at the VCC pin to 2.65V which is out of spec for this part (based on its nominal current draw of 50mA). The lower voltage may create instability, but definitely reduces the part's dynamic range which will either limit the available RF drive to the PA or cause more IMD distortion. I guess I should have done the "T" highpass trick given the time it takes to replace these things in the 746, but I just got lazy and figured that changing the 0.1uF DC blocking cap to a 0.01uF at the input would reduce the spike from the D151 PIN diode bias when the rig is switched back and forth from RX to TX. So far, neither of the two I repaired have given up despite being whacked by a switching transient at the antenna from an unstable Kenwood TL-922A amp during several long weekend contests (I've since fixed the amp, but that's another story). Typically the 160 meter and broadcast bands will give more problems with antenna ESD since the selectable input filters for these bands are lowpass as opposed to bandpass filters on the other bands. If they ever fail again, I'll probably change R155 at the input from 4.7 ohms to a 0.01uF cap and bridge a 22uH inductor between it and C154 to keep energy below 160 meters from sneaking in. The input VSWR should not degrade much unless I planned to operate the rig bootleg in the AM broadcast band. R155 is used as part of the 1.5 dB "T" attenuator (along with R153 and R156) to help normalize the low input impedance of the 1678 closer to 50 ohms for the sake of the band pass filter matching. Removing it will probably not introduce any noticeable filter ripple, and R153 and R156 can be adjusted of needed. Not sure I would choose to use back-to-back Shottky diodes at the input to clamp the voltage since their 0.6V drop may still be too high to prevent damage. 73, Steve NB3O DATE: Sat, 22 May 2004 11:20:17 EDT TO: SUBJECT: Re: 746Pro Hi Mark: Like you, I have been trying to get to the bottom of this problem and provide some understanding of the issue(s).........I am the person who first discovered the ESD problem. http://www.mods.dk/view.php?ArticleId=2725 http://www.mods.dk/forum.php?action=ViewArticleThread&Thread=1815 I wrote a letter to Icom Japan, to the manager in charge of the U.S. FCC Certification (his name was on the original FCC application). I pointed out my less than 30-day old 746Pro failed immediately AFTER I CHANGED antenna's, and I noted a distinct ESD when I touched the ANT connector. My investigation of the schematics found not only a lack of back to back PIN diode protection on the HRX line (as in ALL previous 746 abd 756/pro designs) but that the PC Bd was already laid out for the parts, and they had been depopulated, possibly by accident in manufacturing. Approx 6-weeks later, I received a very vaguely worded letter (actually a C.C.) from the manager to the Icom U.S. cust serv manager, stating that as a result of my letter, I should send my rig in for internal service and to reference memo 920, which the service manager explained to me, and I performed myself, and then documented on the URL, as dated above. This created a real hornets nest for ICOM. Icom denied the existence of the 920 memo, until it was recently leaked on the web, and now is officially posted on their home URL...... You have done a great job on your heating analysis, although the actual calculations for thermal resistance from the silicon junction to the pins to the etch are a little vague. My problem/question with your analysis, as well as that of the fellow in Europe who first mentioned and documented quite thoroughly the improper heat sinking issue (same URL as above after my posting), is that I have noticed that BY FAR (I would guess a factor of 10-to1) most 746Pros seem to fail instantly, with no apparent loss of power (i.e., with heating). One day the rig is fine, next power up it's dead.....this leads me to believe that there are MORE THAN one issue with the UC151, since some people with the ESD fix have reported the same UC151 failure.....however, a review of dealers and service techniccians all seem to agree the incidence has dropped dramatically to 746PROs that DO HAVE the pin diodes added. While I agree that: 1. The part should be OFF during receive; (transmit mode in RTTY or AM or FM must actually exceed the pwr disipation rating, so the OFF during RX suggestion is really a moot point......since TX requires a heatsink any way......) 2) The GND pins should go to a buried GND or VOLTAGE plane for adequate square area to provide proper heatsinking; 3) Some parts MAY have failed due to heat..... I think there is an OTHER problem, which is pointed out by: http://groups.yahoo.com/group/IC746/message/7930 This seems to me to be a much more likely candidate for the problem than over-heating, which when it does occcur seems to display itself by a degradation of output power first, as pointed out by: http://www.kb2ljj.com/data/icom/ic-746pro.htm (scroll down to Kark, OK1DNH, who reported this same observation as you on DK.MODs on 05 November, 2003) .......and NOT an instantaneous failure. Also, see: http://www.qsl.net/oz2m/ham/ic746pro/ So, all of that having been said, please consider all of the above with an open mind and let me know what your thoughts are. I would like to talk to the person (AK2A) who thinks there is a transient with the C154 value....that seems like a more probable root cause to me, because most (granted not all) UC151 chips do seems to fail instantly, and not degrade in output power first-although the heatsinking problem should indeed be addressed with a re-layout of the board by ICOM. Vince, K1VF --------------------------------------------------------------------------------------------- To: IC746@yahoogroups.com From: "steve_cuccio" Date: Sun, 20 Jun 2004 20:48:48 -0000 Subject: IC746 ESD Protection, I stand corrected A while back I mentioned lowering the cap value in front of the dreaded IC-151 (uPC1678) to reduce the transient caused by the PIN diode RF switch bias. Well, after having two more 746's with this change come back with more blown-out IC-151's, I figured it was time to do some elaborate testing. In both cases, the 746's were switched to dipole antennas that did not have a DC path to ground and had a direct DC path from the rig to the antennas. In both cases, wind static in excess of 50 volts open circuit was measured on the antennas during similar weather conditions on the days of failure. For those that have not experienced this situation, try grabbing the open end of your dipole's coax when the wind is stiffly blowing. On second thought, please don't do that. Under the right conditions, the voltage can become great enough to arc from the center pin to ground. This indicated a definite ESD problem, but those two pairs of back-to-back diodes (D1 and D2, 1SS302) should have clamped the volage, especially since the one 746PRO had them installed as a factory update. Or so I thought. Both sets of diodes in the two rigs were blown out too. And they were blow open, not shorted. Both antennas were fed with about 100 feet of coax. The one fed with RG-58 has about 2900pF worth of coax (28.5pF per foot) charged up to greater than 50 volts on the day the operator lost his IC-151. The other station is fed with LMR-400 which is about 2400pF total feedline capacitance. Given that either of these coax lengths make great low-loss capacitors with low series resistance tells me that when the antennas were switched to the rig and into those two tiny silicon diode clamps, the instantaneous current flow melted their internal bond wires. I suppose this explains why the older rigs generally used a big RF choke to ground at the antenna connector instead of diodes. With the failure of the clamp diodes, the wind static voltage injected into the rig can become high enough to do a lot more damage. Where is this threshold of pain? A destructive test was run to measure the equivalent effect of switching an antenna with wind static into the 746 antenna input during receive in hopes to understand why IC-151 always blows out during these conditions. This was done on my own personal rig. A square wave generator supplied a 1 kHz square wave to the antennna input of the rig. An oscilloscope was connected to the intput of IC-151 while the rig was in receive mode. A volt meter was connected across the 10 ohm supply resistor (R157) to measure its voltage drop and calculate the average current of IC-151. As the 1 kHz square wave voltage of the signal generator was increased above 5 volts, a 100 microsecond wide pulse in excess of a half a volt started to show up at the input of IC-151. As the voltage was increased to 10 volts, the pulse widths became wider. The current of IC-151 increased from approx 50 mA to 55 mA. Then IC-151 blew out (again). What happened? The sharp square wave pulse edges get through the first blocking cap (C2) on the HRX line. The bottom diode in D3 is forward biased during Band 1 through Band 7 selection which feeds the pulse edges to C152, then to D151, the input PIN diode switch to IC-151. Unfortunately, D151 is back biased during receive with only a measely 1.44 volts because the R8V (8 volts on receive) is divided by R151 (1k) and R152 (220). An impulse arriving at D151 which is a diode drop greater than its back bias (1.44 + 0.6 = 2.04) will throw it into forward conduction and allow the transient to pass through C154 and into the input of IC-151. Changing C154 from 0.1 to 0.01 uF as stated in my earlier post reduced low frequency transeints to IC-151 but did not eliminate them. With an impulse duration in excess of a 100 microseconds into IC-151 at over a half a volt, the L152 (22uH) inductor on the output stage of IC-151 looks like a dead short. The part blows up. Why doen't this happen on the IC-756 series? The power supply for IC-151 is switched on only during transmit through a transistor. Could this fix be implemented on the 746 series? Yes, but with some minor pain to fly-wire a PNP transistor and two resistors in place of R157. I'll let you be the judge of whether this fix or another idea is easier for you. Would that fix solve the input transient damgage to other components? No. The clamp diodes will still fail when hit with a large dump from a charged up length of coax (or a nearby strike). D3 or any of the other PIN switching diodes in the preselector circuit can also be damaged. What's the best fix? If you are using a DC grounded antenna like a loop or using a DC blocked lightning arrestor like a Polyphaser, then you probably don't have to do anything. For the rest of us, I suggest a solution like the old timers did in the boat anchor rigs, and put an RF choke directly across the HRX line (maybe in place of D1 and D2 clamps). I put a small axial leaded molded ferrite choke on the bottom of the PC board right at the center pin of the HRX connector and ran it to one of the available ground pads nearby. The beauty of this is the choke does not have to handle the transmitter power. It just needs to fit under the PC board and has to be invisible from the lowest frequency of interest to the highest at the HF connector, or in most cases from 100 kHz to 60 MHz. Chokes can be obtained from Digikey or Mouser electronics. You can also install a choke to ground by breaking into the center conductor somewhere on the HRX cable however this fix would require a mounting method. And yes, given enough energy, the choke will blow open like the diodes. However the difference is the choke can withstand higher current for a longer period of time which might be all you need to discharge any potential at the antenna connector. How do we determine the proper value? Since the HRX line is 50 ohms, a shunt inductive reactance of 500 ohms would go unnoticed. Divide 500 by Two times Pi times lowest frequency of interest. I get about 800 microhenries for 100 kHz. I measured a 820 microhenry molded axial leaded choke in the junkbox from 100 kHz to 60 MHz. Although its self resonance is around 2 MHz (the place where the inductor looks like a capacitor), it still leaves the VSWR very low. Its resistance is less than 10 ohms, about the best you can do for something this small. This is still a lot better solution than a couple of small diodes. You can do a similar choke performance test using an antenna analyzer. Place a 50 ohm termination on the analyzer in parallel with the inductor in question and observe the VSWR. It should remain below 1.2:1 out to 60 MHz for most small ferrite molded inductors in the 100 to 1000 microhenry range. If it rises above 1.5:1 somewhere, try another value choke. A lot of the chokes' behavior above its self-resonance is defined by how tightly spaced its wire is wound over the ferrite core. We are fortunate that the 746 does not use the HRX line for 2 meter operation. For those wondering about my previous post concerning the transient at IC-151 created by its D151 PIN diode bias; I measured a very slow rise time of 400 microseconds from 0 volts to 2.4 volts on one side of C154 and very little on its other side facing IC-151. Since C154 and the input impedance to IC-151 form a fairly fast time constant, the amplitude of the transient hammering the IC may not be of great issue. However, I'm still inclined to change C154 to 0.01uF given the Achille's heal of D151 back bias and the extra minute or two for the job once the board is out of the rig. 73 To: IC746@yahoogroups.com From: "steve_cuccio" Date: Mon, 21 Jun 2004 06:17:34 -0000 Subject: Re: IC746 ESD Protection, I stand corrected After seeing what happens to silicon diodes when a static charge from a coaxial capacitor hits them, it might be a good idea for any modern rig that does not have an RF choke across its antenna input. I have not seen a complete 756 schematic to identify a potential ESD path of destruction, but I gotta believe the 756 and 746 may be similar with the exception of turning off the uPC1678 transmit driver during receive. The first receive mixer and receive RF amps generally are immune to small static discharges because they are coupled through an RF input transformer to block the DC, but not so in the preselector section (and the 2nd RX preamp in the 746). Most of our newer rigs have the receive antenna routed directly into PIN diodes to switch preselector filters. When relays or band switches were used in the older rigs for this, ESD was less of a concern. By the way, one of the 746's also had D22 in the control unit shorted. This diode and its associated transistor swtich provides an RF shunt path to ground on the HRX line during transmit to prevent any RF from feeding back. It drove me crazy because the symptom was about a 6 dB loss of receive sensitivity, not enough to find it easily. All because of a dipole without a DC ground. If I only knew then what I know now........ To: IC746@yahoogroups.com From: "steve_cuccio" Date: Mon, 21 Jun 2004 06:42:38 -0000 Subject: Re: IC746 ESD Protection, I stand corrected When the D1 and D2 back-to-back receive diode clamps were left out of the 746PRO, it made any small transient (about 5 volts or greater) more prone to override the much-too-little back bias on D151 and bang IC-151's input. While these diodes won't provide a robust path to ground like a choke will, they still had some utility in this regard. I think this is why Icom installed them as a retrofit in the earlier 746PRO's that were sent back for warranty repair. See http://www.icomamerica.com/amateur/hf/ic746prodiodefix.asp IC-151 cannot stand a long duration transient (greater than a couple of microseconds) because the RF choke on its output is only 22 microhenries. This inductance looks like a short at lower frequencies which cause the second internal stage of the IC to blow or weaken if driven hard enough. I have also read about the heat issue, but after today's destructive testing on the bench using a 10 volt square wave, I really believe this to be the main failure mode in these rigs. While I agree with Icom that some form of protection to suppress ESD is necessary to save IC-151, I just don't agree that a pair of silicon diodes is the best answer given that the diodes blow open too easily. I guess this is why Icom mentions the Polyphaser and antenna grounding in their technical brief. Some one recommended elevating the back bias to D151, however this only pushes the problem to what ever the available voltage source is (which is still less than wind static potentials). The real answer is to clamp low frequency transients before they reach any piece of silicon. 73 To: IC746@yahoogroups.com From: "Mark Brueggemann" Date: Wed, 23 Jun 2004 09:07:33 -0700 (PDT) Subject: Re: IC746 ESD Protection, I stand corrected Thank you for your informative post about the IC151 clamping diodes. It's refreshing to see someone willing to take the covers off their radio and do some work to resolve the issue instead of whine about it in the forums. I'd like to make a couple of points. I don't want to dispute the data you presented, but add that I don't think this is the only failure mode. My antennas are all terminated with 10K resistors to dissipate static, I have an ICE surge arrestor for each transceiver connected at all times, and the antennas are automatically disconnected outside the shack when the rig's supply is shut off. Yet my IC151 still failed, and when tested the HRX line diodes were still good. My static bench testing of the SSOP version of the part showed that the bias current is enough to heat the unsinked part to the upper extreme of the specification range. You've suspected a potential voltage spike due to overcoupling. So it seems if one failure mode doesn't take out the part, another one will. You're right, a back to back pair of diodes isn't a cure-all for this problem. There is probably no single best solution, but a combination of fixes. Compared to some of the "mods" to fix other radio's problems, adding a heatsink tab, a shunt choke, changing the coupling caps and modifying it for switched TX Vcc are all pretty easily and inexpensively done. Better yet as a preventative before the IC fails. I fixed mine even though it was still under warranty, $30 in shipping and a 12-week (at the time) wait wasn't worth it. Steve, would you mind if I quoted some of your messages to add to my 746Pro webpage, and summarize a list of fixes? Mark K5LXP Albuquerque, NM To: IC746@yahoogroups.com From: "steve_cuccio" View Contact Details Date: Wed, 30 Jun 2004 03:07:29 -0000 Subject: Re: IC746 ESD Protection, I stand corrected Hi Mark, no problem quoting the messages. Maybe it will save some one the trouble we've gone through. I agree that a combination of all fixes would be the best solution. I always liked the idea of shutting the IC off during receive as in the IC-756. I never heard of a similar blow-up on those. The 10 ohm R157 supply resistor can be replaced with a PNP or other switch of preference. I like the Micrel MIC2514 switch because it is a fully integrated high side switch with low on-resistance in a SOT- 23-5 package (Newark usually stocks these). This would be a one-chip- one wire-and-no resistor solution. Its input control is CMOS compatible which will allow non-invasive control from the junction of R154 and R156 (approx 5 volts on TX, 0V on RX). Wish I had the time to stick one in the previous repair, but the regular day job is now day/night. Can supply more info, however I get to check this site only occaisionally, so drop me an email to nb3o!arrl.net if needed, just replace the ! with @. 73