The service manual is a must for such a project and I had no luck finding an original for sale. VA7OJ recommended the manual download from the Icom site and a family member with high-speed Internet access downloaded it for me. The manual is in .PDF format and was of much better quality than I expected. The schematic diagram quality is marginal in some places, but with the help of magnification, you can usually decipher the lettering on component labels.
The regulator unit itself is housed in a cast aluminum compartment on the right rear of the 781. It's directly next to the final amplifier unit. Unfortunately, the housing is poorly ventilated and the heat builds up owing to the presence of several hot-running regulator devices. In order to check out and repair the unit, you must remove many screws and the service manual diagram is very helpful here in identifying the right items to remove. One of the regulator modules that's fastened to the housing must be insulated from ground and there's a mica spacer used with that one. Pay close attention to which module uses the spacer since there are three other regulators that do not use one.
When the board is loose, remove the high voltage (30VDC) wires (one red, one black) that supply the PA stage. These are secured to terminals in the PA compartment using screws and ring terminators and so are easily removed. At this point, there are still several wires connected to the regulator board, and you must unsolder these from the board in order to remove it from the 781. Thankfully, the wires are terminated in very small metal tips that keep the wire strands under control and allow the de-soldering operation to be done pretty easily. After removing the solder around the wire tips, the tips can easily be pushed through the board with a small pointed awl.
After years of struggling with cheap soldering pencils, I finally bit the bullet last year and purchased a Weller WTCPT temperature controlled iron. There's no substitute for an iron of this type when de-soldering or soldering circuit boards. The very small tip can deliver large amounts of heat without overheating components or the board. For de-soldering, I strongly recommend a good de-soldering braid that has plenty of flux embedded in it. There's a vast difference between the braid that Radio Shack sells, and the Chemtronics "Chem-Wik" brand that I bought from Digikey. I used the .030 size, and it really sucks the solder off connections quickly.
Upon inspection of the regulator board, I could tell that someone had been in there before me as clear RTV compound was smeared between the four electrolytics and the L7 coil. Icom recommended the RTV to repair centers as a means of minimizing movement of the L7 coil which isn't secured to the circuit board except via its two heavy leads. Not a good design, as the coil is a bit heavy and will rock back and forth, cracking the circuit board traces to which its leads are soldered. It makes no sense to secure the coil to the caps, as they themselves will move a bit due to their mounting method. Clearly, the RTV should be applied between L11 and T2 which is secured to the board via many soldered pins and will make a much more substantial anchor point.
The first thing I inspected was the circuit board land patterns for L7, and just as Icom said, one of them was badly cracked. There was still continuity, but the remaining copper bridge was very narrow indeed. (See the photo below.)
There were a couple of scorch marks on the circuit board and I had intended to replace only the 4 or 5 components the Icom literature recommended (L11, C4, C28, R35). However, after pricing out the electrolytics, I decided to replace all of them before reinstalling the board. (Removal of the regulator board was something I didn't want to do again!) The total price for replacing all the electrolytics (from Mouser) as well as the L11 coil and R35 (from Icom) was less than $25. I learned something here - the electrolytics were all high temperature rated (to 105 degrees C) while normal electrolytics are rated at 85 degrees C. That increases their price only slightly, but makes it just a bit harder to find all the values you'll need.
Another key item is to re-solder the ends of the small metal bar that's used in the over-current sensing circuit. This bar is a resistor of only .001 ohms, so it's very important that the bar-to-land pattern junctions be as low resistance as possible. It's on the under-side of the board, about an inch long.
I removed the large L7 coil from the board completely, and decided it needed something more than its leads to secure it to the board. After cleaning the board where L7 mounts, as well as the bottom of L7 with alcohol, I put a thin coating of RTV on the bottom of L7 and pressed it firmly onto the board and allowed it to set up overnight.
is a view of the REG Unit board showing the positions of L7 and T2, with the electrolytic capacitors to the left of L7 removed. After securing L7 to the board with RTV, the gap between L7 and T2 was filled with RTV to further minimize movement of L7. (This picture was taken prior to the RTV application). In stock form, L7 was held in position via one heavy lead on the left side of L7, and the other heavy lead which can be seen soldered to the lead from T2. Movement of L7 was transmitted to the circuit board trace through the left lead, causing the trace to flex and break.
Regulator Unit Adjustments:When the regulator is repaired, there are two adjustments that you must make in the regulator compartment: Setting the 15V regulator output level, and setting the 30V regulator over-current protection level. This is more complicated than you might think because you must supply an appropriate load to those supplies. For the 15V supply, you need only a 40 watt resistor in the range of 5.6 to 6.8 ohms. The 30V supply is another matter, as you need to be able to accurately measure the current draw as well as be able to dissipate the heat produced. That load must be 1.66 ohms at 540 watts (!). Don't even think about using the Ic position of the 781 meter to monitor the current. That, as well as the Vc metering function is probably more than a little bit out of adjustment.
Unless you have access to a high current DC meter, you'll need to devise a way to measure 18 amps, and the easiest way to accomplish that is with a meter shunt. Those are available surplus and typically produce 50 millivolts across them at the rated current. I found a 20 amp shunt at Nebraska Surplus Sales that produces 100 millivolts at that current level. All you need then is an accurate milli-voltmeter to measure the drop across the shunt. The 6.8 ohm resistors I obtained were just a bit on the low side and the resulting resistance measured 1.67 ohms. With the 30 volt supply, that gave me 17.96 amps - close enough to the 18 amp target.
Since I had to wait for parts from Icom, I also ordered the Sherwood Engineering cooling fan for the 781. The fan assembly includes a new outer black metal housing which replaces the original regulator cover, and even has the rear connections screened on top as the original had. It also includes the inner metal shield, which is drilled with ventilation holes directly in front of the fan position. This is one quality product and should prevent future regulator unit failures. I found the fan produces a bit more background noise than I'd like, so I may increase the value of the resistor in series with the fan to reduce the speed a bit. A word about Icom parts: I've never run across a manufacturer's parts department that offered more pleasant and quality service than Icom's does. They quickly ran down replacement part numbers for superceded parts and they shipped very quickly. If you ask nicely, they'll even send small parts via USPS Priority Mail for no extra charge. When you order from them, be ready with the Icom part numbers shown in the Service Manual.
While I had everything apart, I decided to make the latest modification to the 781 ALC circuit to eliminate the well-known spike. I needed to change a resistor on the IF board, and one on the PI board. I did the IF board first, thinking that would be the toughest, since there were many cables to disconnect before the board could be removed. It turned out that the PI board change was much worse, as the unit is buried under the scope module, and even getting all four mounting screws out was a challenge. It turns out that was good experience, as you need to check the regulator output voltages at connectors on the PI board.
When I set the regulator over-current protection level, I found that the meter Ic indication was off by a couple of amps and that the Vc reading was a volt low. After reading through the Adjustment section of the Service Manual, I decided to run through almost all of the receiver and transmitter adjustments. That proved to be a long drawn-out process, but I was able to find and correct several adjustment problems that otherwise might have gone un-noticed for a while. For example, the actual power output was low, while the 781 meter was reading about 30 watts high. The combination meant my 781 power meter reading was off by 30% on some bands. I also adjusted the scope center frequency on the three ranges. That can be done by observing a marker signal on the scope while making the adjustments.
I was able to do just about all of the alignment without any special test gear. Here's what I used:
- Ancient HP 200CD audio oscillator
- Old Beckman bench DVM (good millivolt measurement capability)
- B&K Precision 5380 DVM
- Bird 43 wattmeter with 250 watt slug
- Oil-filled paint can dummy load
At the suggestion of Adam, VA7OJ, I used the 781 Service Manual available via download in PDF format from the Icom site. With a decent inkjet printer, just about everything (including the schematics) is readable. On the schematics that include notations of voltages and waveforms, much of that information is difficult or impossible to read because the original was printed in a color that doesn't copy very well. Icom no longer carries originals of the 781 Service Manual.
The repairs took quite a bit of bench time, but are certainly not impossible to do if you're technically adept and want to save some money. I'm very glad that Icom took the time to publish their list of common service questions as I could never have figured out the solutions on my own.
If you have a 781, don't have the Sherwood fan and have never had the regulator board serviced, you can probably count on experiencing regulator-related problems in the future. As was pointed out to me, the 781 was designed to run on a line voltage of 100 V and uses a shunt regulator which means that it's pouring out lots of heat even when the 781 is in receive mode. When you power it with 120 volts (my line voltage is a high 124.5 most of the time), the heat generated will be much more than can be safely dissipated from the sealed regulator compartment. You can probably forestall the inevitable by reducing the line voltage to 100 volts by using either a separate low-voltage transformer with its secondary connected so as to buck the line voltage, or by using a Variac adjusted to 100 volts. You can often find nice 10 amp Variacs in an enclosure for around $100 on eBay.
Copyright 2007 - K8AC - all rights reserved