817 servicing, don't do it this way.

This was inspired by Ft 818 discontinued

I know from experience Yaesu UK parts service is pretty damned good. However, with the 817/818 being discontinued I didn’t want to find that I should have bought some 817/8 specific bits when I had the chance and not waiting till I was desperate. Parts were ordered last week and arrived 2 days later.

OK, what’s the biggest 817 failing? Yes, that’s right the case screws fall out! Well my main 817 (used since 2009) is missing 4. I could have stolen them from the other 817 but I ordered 15 screws (£1.01 +VAT). At the same time I ordered a new SEL encoder and knob as the #2 817 was iffy, it was fine turned CW but was rubbish turning CCW. You need a knob because the shaft is now different to my old one. 15 screws, encoder, knob, VAT and P&P £9.22, bargain.

OK, I read up on the SEL encoder change… “Don’t lift the tracks on the PCB”. Fair enough. No pictures or videos. First take case off 817 and don’t loose screws! I looked at the bottom and thought “something goes wrong here” and that is the battery hatch spring flies out and gets lost. This time it didn’t but it’s happened before :frowning: Unplug speaker, remove battery. All parts safe. The front panel is held on with 4 plastic clips which are fragile…don’t break them or your wallet will be unhappy. Then lift the locking tab on the flexible PCB connector on the main unit, the cable lifts out and you can separate the two items.

Looking at new encoder, the knob pushed on so I carefully removed the old one. There’s a self-tapper that holds the front panel PCB in the plastic case, removed, but the panel is fast. Ah, remove tuning knob. The correct way is to slip a thin screwdriver between the rubber grip and the knob body, slip off the rubber grip and loosen the Allen grub screw and the knob pops off. My way was to think it was a spring loaded push fit… well it came off along with the brass innards of the encoder. Bu**ger. Well we may need to order a new one of those but that’s a detail. Undo retaining nut, tuning encoder slips out, unplug and it’s free. I checked the connector was polarised, it is. OK, remove the board.

Yes, remove the board. SEL end is loose but rest moves a few mm but no more. Jiggle it. Shake it. Poke it. Jiggle it. It cannot be this difficult. Resort to HM’s Naval grade swearing and a bit of gentle levering and it came out in a rush and something vanished under the bench. That was the up/down mode and band buttons. 5 mins of groping about and I recovered that. Hrrmph, this is being tricksy. DO NOT USE A METAL LEVER, I scratched the very soft older resist in a few places. Use a plastic spudger and some swearing.

Next, I checked the new encoder was the right size and pinout. It was so onto removing the old one. OK, don’t damage the tracks. I’ve been soldering for 51 years so I did a quick test on one pin of the encoder. The verdict: I’m not de-soldering easily. I have a mechanical solder sucker and a decent iron so i removed as much solder as I could. Last year I treated my self to a new pair of side-cutters, about £45 worth and they’re the mutz nutz. Without a good de-soldering station the best way is to cut the pins on the encoder then clean the holes. 5x snip and the encoder was not moving. Look at new one, aha, there’s 2 mounting tabs! Snip them and old one is out. Fine tweezers onto old pins and each came out easily. De-soldering braid did wonders and all 7 holes looked good and clean. New encoder into position, fits well. We’re now 90mins into the job, go and prepare something to eat and leaving it cooking then solder up new encoder.

It soldered up easily. There are some springy metal tabs that connect the earth plane on the front PCB to the chassis, I bent them back to better contact. Then I reassembled the PCB and button rubbers into the front panel. After 20 mins I was getting annoyed. It was a complete bitch to get out and it did not want to go back in. Finally after another 15mins it was in situ and the band buttons and A/B/C buttons where in place. Mechanically all felt OK apart from the power button… OK fit main tuning encoder and tighten nut, that brought the PCB into correct place and power button felt good. Time for a rest and some fine home made vegetable soup.

Fed and relaxed it was time to check the 817 worked. Reconnect flexible PCB ribbon, apply some volts to the power socket, press power button, backlight on, display appeared but no beep… no speaker. OK ignore that and onto the SEL knob. twiddle it and the frequency jumped in 2.5kHz steps and it was fine either direction. Check all buttons, think it was working… lets try the tuning knob, will it fit back in or is it gubbed? It went in with a plop and tuning it the frequency moved. Hurrah! Hurrah! And thrice hurrah!

Well no. Turning the SEL or tuning knob and the PTT light flashed red and the relay clicked over. It was PTTing for a moment when the knobs were turned. OK, connect up power meter and mic. Double bu**er, no PTT on mic and no power on meter. Andy: stop, remove power and have a cup of tea. Scream and swear and remove PCB. This time the tuning knob was removed correctly and it only took a few minutes to get everything apart. Careful examination of all the soldering with a jewellers loupe looked good. I checked for trapped wires etc. No all was good. Maybe you need to do a full reset when you have had the front off because everything seemed to working and it looked like it was the software screwed up. (40 years of professional embedded programming experience talking here.)

Reassemble again. We had left Naval strength swearing and were on to Special Forces strength swearing. It was a complete and utter ********** to get back together. Then a full reset (oh boy just how much data to enter again?). Looks good but no, SEL and tune caused a blip of PTT and sometimes the set would stay PTT’d with a confusing display. Power off. Stay calm.

I know, on Who Wants To Be A Millionaire they ring a friend for advice? Who should I ring? Well obviously a one time Analogue RF designer and man who ran a ham radio repair outfit. He used to fix all the things were the big 3 repair centres wouldn’t touch. Yes, Mr. Barry GM4TOE.

I called him described the fault and what I’d done. He agreed it was a total pain working on the front panel. He hated fixing 817 because they were tiny and fiddly. So SEL and tuning caused momentary PTT, Menu A button not working always, no PTT and permanent S3 on S meter. With over 110 years of analogue and digital work between us we talked over the faults and the work done. I just happened to press F and got into the menu and then noticed SEL worked fine, so did tune, no phantom PTT. Out of that mode and just a quick F and in that menu again SEL worked and so did A. Barry’s considered response “are you sure the ribbon cable is in correctly? It looks like everything is working, software screwup or shorted wire”. I removed it from the chassis and then thought well is it OK in the front panel as it’s be jiggled a lot, has it been pulled out of alignment? I removed it from both connectors and why, I don’t know, but I looked at both ends up the jewellers loupe. It’s 0.05mm pitch, so at my age I need plenty of help seeing it.

And there, as clear as day, on the front panel end of the cable where 3 tiny balls of solder about 0.05mm. They were perfect spheres. They wiped off with rub of my finger. “Solder splashes” I shouted. Quickly refitted the cable, powered up and normal service had been removed. Some more expletives and everything seemed fine and dandy. Now thinking about what was happening, the shorts could have been jumping one of the lines from the encoder with PTT. The pulses on the encoder would drive the PTT for a moment. Once the encode stopped the output would be fixed and the when I PTT’d the line would be held by the encoder. Likewise scanning signals from the front panel CPU could have been shorted causing corruption on the S-meter signal.

How did those splashes get so far into the connector. Maybe when the solder was being applied to the joints? Or when I was using the sucker. But perfect spheres suggests the were molten in the air and surface tension caused them to form a sphere. The connector is near where I was soldering. Still after all my work experience I was very surprised to see those contaminants where they were. I’d not think it possible for them to get in there. The important point was when things were wrong was to stop so I didn’t make them worse and talk over the issues with someone similarly experienced. I could have zapped the CPU with static but considering the mechanical issues and soldering, the most likely reason would be damaged cables or shorts. OK it was shorts on the cable from solder.

Barry was thanked for just helping me talk through the issue. I told him he could treat himself to a drink for his efforts!

I started programming the memories settings etc. and thought “sod this for a game of marbles.” FT817 Commander installed and a serial cable connected. After much option clicking I got it to connect. Refamiliarising with FT817 Commander I tried to reprogramme all the memories. Last time was using Windows 7. On this Windows 10 machine FT-817 Commander is not reliable and tends to lock up when downloading to the 817. Uploading is fine. So I tried maybe 7 or 8 times before I got the memories set.

What should have been a 1hr job at most took about 4 hours in the end including phone calls, swearing and thinking I’d have to buy some costly parts. I also aged a lot more than 4 hours.

Lesson learned… you probably never need to do a full reset on an 817, FT-817 Commander doesn’t run smoothly on W10 Pro 64 any more, get a tin of dry compressed air for blowing detritus away from where you’ve been soldering, drink more tea when working, check for Youtube videos about jobs as some poor mug may have been through this already, if you want to remove the tune knob you need a 1.5mm Allen key.

So I now have screws in both radios and the new SEL feels fabulous. I’m so versed in this repair I might just buy another encodef for the other 817 as it feels sloppy compared to the new one. But I’m not going to fit it till I absolutely have to! And no, I don’t want to start a business replacing these for people with worn 817s :wink:



Yes resumed. :slight_smile:

… and someone asked me why I sent mine to MLS for repair. :hushed:

1 Like

If you save your memory settings to a file, you can retrieve and replace them in the radio after such surgery.

I know this is being wise after the event, but I actually have used the saved memories to set two 817s to have common memory channels, (mostly one CW and SSB freq per HF band, then some ssb frqeuencies and beacons on vhf, then some repeaters, rarely used in my case).

It’s also worth saving the IF gain and power settings in the extended menu functions. The big daddy reset sets them all to 100 which is too high on any band.

But well done getting the encoder swapped. I did that on an IC703 and it wasn’t a trivial exercise. The new one works much better.

73 Andrew VK1DA/VK2DA

1 Like

I can understand your emotion very well. Also had the dubious pleasure of swapping two 817 encoders a few years ago.I was very lucky that my solder spatter only ended up permanently in my fleece jacket.

My curses were heard when I realized that the pads were significantly smaller than the pins on the new encoder. So sharpen the pins with the nail scissors.

73 Chris

1 Like

All modern electronics (except military, avionics & medical - ask yourself why) are glued together with Devil’s Spawn (Lead free solder). Before attempting to desolder, especially with desoldering braid, resolder the connection with real solder (Lead/Tin) to flood the connection with some Lead.
Much easier to get the residual solder off (and the melting point is lower too so less risk to those delicate tracks).

I took Andy’s advice and treated myself to two bottles of fine ale - obviously worked as it resulted in the correct outcome at Twickenham :laughing:


I haven’t considered that! Does the service manual give a calibration process to get the IF gain back to the “right” value. I can monitor the output powers and set them back easily.

1 Like

Professional repair shops used a special low melt solder to bring melting points even further down:
Low-melt desolder alloy 1x16.5cm

This will drop the melting point to below 100 °C. So with a hot-air station one will not desolder any other parts in the process.
But it needs to be cleaned of with desolder wick properly.

Repair videos on YouTube are helpful to learn such tricks.

73 Joe


I don’t have the service manual, but since those gain and power limit settings are provided for several sections of the HF, 6m, 2m and 70cm, there are a lot. I suspect device variations mean that ideally you do each one separately. However it does not matter all that much if your lowest power setting results in 550 mw or 450mw instead of the nominal 500. More important that the highest level is not too high (it can be brought up to the 6w without any danger, but why would you?) and that ALC action is only slight at the peak of the transmitted signal.

For transverter use, you do need to know that the level you use is not far from the right power level.

If I was doing it from scratch, I’d set the FM or CW output power limit higher than the end point, then adjust the IF gain so that it makes that power level, then bring the output level control down so it “just” affects the output power, ie, it would then be generating ALC. doing it all via ALC means the radio is always producing adjacent channel crud as the ALC is only compression in the RF domain and distortion must result from sudden gain changes. These changes are all soft changes so no trimpots are injured in the process. And it all gets done at each power level, and in each frequency band (HF being divided into several bands only for this purpose). You need a power meter of appropriate ranges to do this.

Making the settings equal to another radio of the same model would be a reasonable starting point. The settings for this are only accessed with the appropriate combination of keys held down when powering up. The middle level of Reset just opens up the menus without resetting anything but the top level of reset puts all these gain and power level settings to 100, ie max.

Fun with 817s, Same process is applied to the 857 and 897 I believe.

73 Andrew VK1DA/VK2DA

1 Like

Yes - TX Gain adjustment and RF Gain Adjustment.
All done in software

1 Like

I have problem with rear antenna - front antenna relay. MM0FMF, where did you order your encoder? Is there any Yaesu part dealer in EU? I ask csy&son and waiting their answer

Well I used Yaesu UK ( Welcome to Yaesu.co.uk ) who have a simply brilliant team handling parts etc. It helps if you have the Yaesu part number from the service manual as Yaesu have a proper follow up system working for when parts get replaced so that if you need the replacement part they also know if you need other parts to go with it. In my case I needed a new knob for the new style of encoder.

Sadly the UK is no longer in the EU so I don’t know if this will affect your ability to purchase from the UK. I know it is very difficult for me to buy from small traders in the EU now. Coming soon: having to pay for a ETIAS “visa” to enter the EU. Ah yes, Brexit, the gift that never stops giving :wink:

You can try them but they may direct you to a Yaesu office in the EU nearer to Croatia.

1 Like

Thank you for your answer. I am aware of problems with brexit but need to start somewhere.

Problem with ft818 is that I do not have schematic or service manual to find Yaesu part number. I have only part of ft817 schematic. Are the relays in ft817 and ft818 the same? I do not know

1 Like

RL3016 is manufacturer’s part number ARK115DC9V and Yaesu P/N M1190094

Thank you for your answer. Sorry for my bad english.

This is ft817 shematic. I do not know if ft818 is the same. This was just to ilustrate the problem.

By the way, ft817 is using two different relais for rl3016. ARK115 and G6Y-1

1 Like

I did not repear my 818 yet. Still waiting. But made a test with another 818 and strenght of receiving signal differs between pl and bnc. It is stronger on pl. Test was on 2m with beacon 150 km away.

Is there anybody willing to test his 818? Is there any difference between pl and bnc?

To use PL and BNC you need different cables or an adapter. Maybe one is wrong?

By the circuit diagram, the relay’s normal position (not energised) is to use the front panel socket (bnc). If the rear socket is working better, the relay is the faulty item. This assumes the 818 is wired as per the 817nd in this part of the circuit.

1 Like

Thank you for your hint.

Our schemes are drawn little bit different. I assumed that not energized position is rear connector.

Now, I am sure that relay is a problemm

1 Like