08/24/2026
At The Repair Bench - Hamplus MBD-8E Smart Antenna Switch Control - September 2026
By Chris Prioli AD2CS - [email protected] - www.ad2cs.com
One of our fellow GCARC members brought his Hamplus MBD8E Smart Antenna Switch Control (Figure 1) to me for repair, asking only that I replace the blown fuse. The factory-installed fuse is a two-ampere fast-blow size 1206 SMD fuse. I suggested replacing the single-use fuse with a self-resetting type, and the customer agreed. There was a method to my madness in making such a suggestion. You see, at that point we had no idea whatsoever what had caused the fuse to blow in the first place. I figured that by using a self-resetting fuse device, I would stand a better chance of determining the cause without the need to repeatedly replace a soldered-in surface-mount fuse.
When I opened up the unit, I quickly found several problems with the unit. Working without a schematic, as I had none for this unit at that point, I began to track down the problems.
This unit operates on a nominal 14VDC, meaning that it should operate properly from a standard 13.8V ham bench power supply. Coming in from the power inlet jack, the current flows through a pair of diodes and then into the fuse. From the fuse, the flow is to an LM7805 voltage regulator IC with a 100µF/25V aluminum electrolytic capacitor connected at the input to the voltage regulator. The output of the voltage regulator supplies the 5VDC supply to the rest of the board, and there is a 4700µF/25V aluminum electrolytic capacitor connected between the voltage regulator output and ground. These components cannot really be seen in the schematic excerpt at Figure 2, with the schematic having been obtained after the repairs were completed (more about that later). They cannot be seen because there is a difference between the schematic and the actual device as regards this portion of the circuit. The schematic shows a pair of capacitors at the voltage regulator input and another pair at the voltage regulator output.
In the schematic, we see capacitors C25 and C26 at the voltage regulator input side. While C25, at 0.1µF is probably a valid value and does exist on the actual PCB, the value of C26 shown is 10µF, while the actual installed value here is 100µF. Similarly, we see those same two values of 0.1µF and 10µF at the v/reg output, but in reality, only a single 4700µF capacitor is installed at this location. These differences point out the types of running changes that are often made as a product matures in the wild.
Let’s look at the 4700µF capacitor, as an example. The most likely reason that such a large value capacitor is installed here is to help to supply the current demanded by the switching appliances (which are external to this unit). This capacitor will charge to a certain level, and in so doing, it will store considerable current in its plates. In turn, when the downstream current demand is greater than the incoming source can supply, the capacitor can discharge some of its stored current into the circuit to meet that current demand. This is a much greater value capacitor than would ever normally be used at the output of a voltage regulator IC.
Further, a look at the related area of the PCB in Figure 3 shows that the apparent original intent was for capacitor C27 to be a smaller capacitor that would be installed to the component side of the PCB rather than hanging off the edge of the board. It is also apparent that although C25 and C26 are both marked on the board, while C27 is marked, C28 is not. This would seem to indicate that there were interim design versions of this unit.
After removal of the CPU board from the unit enclosure, the first thing that I noticed was a burned area in the paint on the inside of the enclosure floor, directly underneath the location of the voltage regulator pins on the printed circuit board (PCB). The IC itself extended off the edge of the PCB and was secured to the enclosure floor via a machine screw and nut at the tab of the TO-220 voltage regulator package.
A close look at the voltage regulator pins on the PCB showed that the trace from the center (ground) pin to the ground plane on the PCB had been burned away (Figure 4). Now… one might think that the voltage regulator IC would have been sufficiently grounded through its tab being fastened to the enclosure, but in reality, the enclosure itself is not at ground potential and the only real benefit to securing the voltage regulator IC to the enclosure floor is that it provides plenty of cooling for that IC.
Removal of the LM7805 voltage regulator IC from the circuit and basic resistance testing with an ohmmeter showed that the IC had failed. I began then to look more deeply for a cause of voltage regulator IC failure. I began by making a resistance measurement from the voltage regulator output pad on the PCB to the PCB ground plane. With that measurement, I found that there was a direct short circuit to ground. Going after the low-hanging fruit, so to speak, I desoldered the 4700µF/25V electrolytic capacitor from the PCB, and retested the resistance. This time, it measured about 6.5MΩ.
A test of the removed capacitor with my capacitor leakage tester showed that it was indeed shorted and therefore merited replacement. I decided to remove and test the remaining electrolytic capacitors on the CPU PCB. The included the 100µF/25V capacitor at the voltage regulator input, and two 10µF/25V capacitors at other locations on the PCB. Of these, the 100µF type showed excessive current leakage, while the two 10µF types were intact. However, due to the fact that they were all presumably of the same age, I opted to replace all four electrolytic capacitors.
While examining the foil side of the CPU PCB, I noted that three of the BC517 transistors used as the antenna circuit switches had evidently been replaced at some point in time (Figure 5). This was indicated by the appearance of the solder joints for those transistors. I filed that information away in my mind, and proceeded to finish the assembly of the unit.
With the damage that had occurred to the PCB at the voltage regulator IC ground trace, the board of course needed some repair work. I accomplished this by scraping away some of the solder mask off the ground plane in the vicinity of the voltage regulator IC. I then fashioned and installed a wire loop with a tail, soldering the tail to the PCB in such a manner that it held the loop directly in line with the center pin position for the voltage regulator IC. However, I raised the loop up over the PCB surface by about a thirty-second of an inch so as to prevent short circuits between the IC pins.
I installed the new LM7805 voltage regulator IC at its proper location, except that I repositioned it so that its pins came into their PCB holes from above the board rather than from underneath the board as it was originally installed. This was done to take best advantage of the board repair that I had made. After the IC was installed, I applied some conformal coating to the PCB over the repaired area.
With the new electrolytic capacitors and the new voltage regulator IC installed, I next installed the self-resetting fuse in place of the original single-use fuse. I then applied power and tested the board operation. At first, it seemed to operate properly, but after about fifteen seconds, the power went out as the fuse tripped. Apparently, I still had some detective work to do.
I began the next phase of the diagnosis by isolating the CPU PCB from the second board in the unit to verify that the cause of the tripped fuse was on the CPU board, which it was. Next, I began taking operating voltage measurements for each of the thirteen transistors used on the CPU PCB. It did not take very long to determine that five of the thirteen transistors had failed! Three of the failed transistors were in the bank of eight BC517 NPN Darlington devices that are used as the basic switches for the antenna circuits. The other failures included one each of the BC327 PNP and BC547 NPN general-purpose transistors used in the various control locations in the unit. Needless to say, I replaced the failed transistors… and then I decided to also replace all of the transistors, failed or not, as a preventive measure. This was spurred by the fact that among the non-failed transistors, there was a wide range of hFE values reported during operational testing of the transistors. For example, when testing the BC517 types, I saw hFE values as low as 16 and as high as 248. Similar results were reported when testing the BC327 and BC547 transistors, though here I was comparing the results more to datasheet values than to the other transistors as I had with the BC517’s.
Moving on, with all of the new transistors installed, I once again applied power to the board, and once again, the fuse tripped out after about twenty seconds. The left me digging for more clues as to where the problem was hiding.
This CPU PCB uses a total of sixteen diodes, four of which are of the two-ampere surface-mount ES2A type, and the remainder being 1N4148 through-hole devices. All of these tested OK. There are many resistors and ceramic capacitors, all surface-mount, and every one of them tested OK. Finally, there are eight additional integrated circuits (IC’s) beyond the voltage-regulator IC that we already discussed. All eight of these other IC’s are surface mount devices.
Unfortunately, the only sure method of eliminating any one of these devices as the cause of the tripped fuse is to remove the device from the PCB, and then to apply power to the PCB and see if the fuse trips. As it turned out, the causal IC in this case was the most difficult to replace.
When I speak of difficulty, I am not referring to the actual act of removing and installing the IC. That is a pretty straight-forward task accomplished with some hot air for removal, and a fine soldering iron and thin solder for installation. No… the difficulty lies in the fact that the problem IC was the microcontroller, a type of chip called a PSoC (Programmable System on Chip) device. Of course, as luck would have it, the one chip with proprietary programming installed is the one that needed to be replaced.
The chip itself is readily available at either Mouser or Digikey. It is a 28-pin SOIC type of chip, requiring fine soldering work to install properly, but the physical installation is only half of the process. Once the chip has been soldered in place, it must be programmed.
For this, I reached out to a firm in Texas that is the USA service representative for Hamplus, a Brazilian company. The principal at that Texas firm is a true gentleman by the name of Eduardo Cunha, who was extremely helpful to me. Eduardo provided me with the proprietary firmware file for programming into the chip. Now all that I needed was an appropriate programming tool. This tool turned out to be a MiniProg3 programmer, a tool that is made available (at substantial cost) by the chip manufacturer. I ordered in the chip and the programmer kit, and then waited for them to arrive.
Three business days later, I had the chips (I had ordered two of them) and the programmer kit. Soldering the chip into place on the PCB was no problem. The chip is a 28-pin SOIC type of SMT device, so to solder it into place was quite easy. I used my FNIRSI 100W iron with a knife tip, and the solder that I used was the 0.012” (0.3mm) rosin core solder about which I have written before.
My technique involves a thorough cleaning of the board and pads, followed by a light application of solder to the four corner pads for the chip (Figure 6). Then, I placed the chip on its proper footprint on the PCB, aligning the four corner pins with their assigned pre-tinned pads. While applying light downward pressure on the chip, I heated each of those corner leads in turn to cause the solder on the pads to reflow and then to bond the chip in place. I then went down each line of pins, soldering them as I go along. Finally, I went back to the corner pins and apply the proper amount of solder there to provide a proper solder joint at each corner. It is a quick and easy way to install these high pin count SOIC chips (Figure 7).
The programming was not quite so smooth and simple. Never having installed or programmed one of these Cypress chips before, I was not familiar with their procedures and software. It seems that the software that is called for by the Quick Start Guide for the programmer kit that I purchased does not actually support the programmer, nor does it support the chip family of which the CY8C29466-24SXI is a member!
It took a great deal of frustration and detective work to determine that I would need to use a legacy programming software with the programmer kit that I purchased, and then I would need to install an “unconfigured” device driver for the actual programmer in order to get the software to communicate properly with the programmer. Friendly warning : nothing is simple in the Infineon/Cypress world when it comes to hardware, software, or the documentation for either one of those.
Suffice it to say that I eventually figured out how to get the programming software to talk to the programmer, and then to get that combination to work with the target chip. It turns out that the chip must be externally powered for programming with this programmer, and that the programmer must be placed in its “reset chip” mode in order for it all to work, none of which is clearly explained anywhere that I could find in the Cypress documentation.
Ultimately, I got the chip programmed using the hex file that Eduardo had generously sent to me, and when the unit was tested afterwards, it worked flawlessly. Success! I went ahead and reassembled the rest of the unit and set it aside for pickup by its owner.
The story is not yet complete, however. Almost as an afterthought, I dropped another email to Eduardo and asked him if it was possible to obtain a copy of the unit schematic diagram, which after a short delay, he provided to me. I decided to see how the unit that I repaired matched up to the schematic, and what I found was a bit of a surprise.
The bank of eight BC517 Darlington pair transistors that do the main antenna selection switching are indicated on the schematic (Figure 8) to be the less-capable BC547 NPN Silicon general-purpose transistors. While both the BC547 and the BC517 are rated at 625mW total power dissipation, the BC517 is rated for a full ampere of current, while the BC547 is rated for only 100mA, or one-tenth of the current that the Darlington transistor can handle. This is an important point, because as I said earlier, there was evidence of some of the BC517’s having failed previously, and I had found several of them failed when I started this job. It is my bet that even at one ampere, the BC517’s are not right for the task that they are being called upon to perform, and that the earlier design level that used the BC547’s there must have had numerous repetitive failures due to current flow through these transistors. The BC517 would appear to be a compromise “drop-in” solution that could be implemented with no other circuit design changes being necessary. Unfortunately, it would appear that this solution really does not go far enough.
It is my belief that this unit will continue to have problems in use, as its design does not appear to be robust enough to handle the current levels required to do the job that this device is designed to do. In my opinion, the switching transistors used should have been higher-current devices in the TO-220 package. Substituting the BC517 devices for the BC547 originals provided the power advantage of the Darlington pair transistors there, but it really is not enough to solve the problem completely. The advantage is that they fit directly in place of the original BC547 transistors. Sometimes the manufacturer just gets it wrong… and this is one of those times.
See You Next Month!
