Last week I was imaging for a new astrophotography project and my Losmandy G11 stopped responding correctly to NINA and PHD2. It was already midnight, but I spent a few more hours trying to solve the problem. Some commands worked intermittently, but it correctly executed commands so infrequently that there was no way to image with it. I swapped USB cables, USB-to-serial converters, and computers. No luck.

After some sleep, I brought the Gemini 1 computer into my shop. With just the Gemini and a power supply, it connected to my laptop without a problem. I didn’t have motors hooked up at that point, so I couldn’t try slews. I left it connected to my laptop with the ASCOM driver and NINA to see if the problem would happen again. Sure enough, after a few hours the ASCOM driver started complaining about serial port problems. After another few minutes, I couldn’t get the computer to connect to the Gemini at all.

I dug through the Gemini Users group and the Gemini 1 FAQ on Gemini-2.com. Through those, I learned that a fairly common failure in the Gemini-based mounts is the serial/RS-232 interface integrated circuit (IC). The Gemini 1 FAQ outlines that the typical failure mode is the negative voltage side of the RS-232 IC and also describes how to test it. Sure enough, mine was dead. I started hunting for the right parts to order and posted on the Gemini Users group about it.

At the great suggestion from Michael Herman, I also checked the capacitors around the RS-232 chip and found several of them were bad too.

The Gemini 1 FAQ lists the RS-232 IC as a MAX3221EEAE. I researched adding more protection to the chip since this is a common problem, but decided to use a different part number instead and also add an external DB9 serial opto-isolator (Amazon links are affiliate links). The current version of the MAX3221EEAE appears to be the MAX3221EEAE+, but M. Collins recently recommended the MAX3226EEAE+ in a different thread. So, I went with that. (Note: aside from reading Collins’ post and verifying the datasheet for compatibility, I did zero real research in why I should choose the MAX3226EEAE+ over the MAX3221EEAE+.)

After looking at pictures of my Gemini 1 PCB, Brendan Smith also recommended I look at my RV-1 through RV-4 components. I tested those and they seemed fine for now. I bought spares just in case though. I found capacitors in the right specs and size (0805) and ordered extras of everything.

Since 0805 capacitors are fairly small, they seem to fly away on their own when you remove them from the packaging. I bought 10 of each capacitor I needed. Thankfully, they’re cheap.

I’m ok at hot air soldering & reflow work, so I wasn’t too concerned about swapping out the serial IC. 0805 capacitors are easy to do with a fine-tipped soldering iron. Still, my Gemini 1 is 17 years old and it can be pretty easy to tear off solder pads if you’re not careful. I did have a backup plan in case I ruined the Gemini completely, but certainly didn’t want to ruin anything.

I tried to clean up the pins on the RJ22 port, but there was still some leftover corrosion and scoring. So, I chose to replace that too.

Here are the parts from Mouser that I ordered for my repair:

• MAX3226EEAE+ (I needed 1, but I ordered 2) MAX3221EEAE+ is closer to the original part. I’m not convinced the MAX3226 is better than the MAX3221 – it might not be. See my connection sequence section toward the end of this post for more info.
• Yageo CC0805JRX7R9BB473 0.047μF 5% 0805 capacitor (I didn’t need any, but I ordered 10)
• Yageo CC0805JKX7R9BB334 0.33μF 5% 0805 capacitor (I needed 2, I ordered 10)
• Yageo CC0805JPX7R9BB104 0.1μF 5% 0805 capacitor (I needed 1, I ordered 10)
• TE 5520257-2 RJ22 connector (I needed 1, I ordered 2)
• Littelfuse V5.5MLA0805LNH 5.5V varistors (these are for RV1 through RV4. I needed 0, but ordered 10)

Depending on your shop supplies, you may need to pick up a few other things. Here are some other things I used for this project:

• 858D hot air reflow station
• Solder paste
• Plastic syringe tips. I think these are a lot easier to use for laying down solder paste than the flat-tipped metal needles. If you don’t get the flow you need, you can trim the end with an X-acto knife or scissors.
• Solder. For surface mounted components, I think this 0.020″ no-clean solder is easy to use.
• No-clean flux
• Solder wick
• Soldering mat

If you have this problem and you’re not already good at soldering small components, I recommend you contact one of the people who have a lot of experience repairing Gemini computers. In the US, I know Michael Herman is excellent and extremely helpful. David Partridge does repairs in the UK and Brendan Smith does them in Australia. There’s some discussion about repair options in this thread. There is also a list with contact info here.

To start, I did some cleanup on my Gemini printed circuit board (PCB). It had some grass bits and dirt in some of the corners. That cleaned up well with isopropyl alcohol and Q-tips. After that, I used my hot air gun and a pair of anti-static tweezers to remove the old serial interface IC. I then used the hot air gun to remove the C19, C21 and C44 capacitors. I used flux, solder wick and a soldering iron to remove the excess solder from the circuit board. After another round of isopropyl alcohol & Q-tips, the Gemini board was fairly cleaned up.

The old RJ22 port was next. I used some cutting pliers to remove the tips of the plastic barbs sticking through the bottom of the circuit board. I flipped the board over and cut away most of the connector. This left the 4 pins still soldered in the circuit board.

With just the 4 pins sticking through the circuit board, I used a soldering iron, flux and tweezers to remove them. After the pins were out, I used a solder sucker and some more solder wick to clean up the holes.

I supported the Gemini board and the new RJ22 port so that the port sat flush with the board. I soldered one of the leads, checked it for alignment then soldered the others.

The MAX3226EEAE+ serial interface chip was next. My solder paste is near the end of its life, so I bought some really cheap SSOP-16 chips and some blank SSOP-16 breakout boards to practice & test first. That went well, so I used the hot air reflow gun and the solder paste to solder the MAX3226 onto the Gemini board.

The solder paste I used is amazing at flowing solder to just the pads & contacts, as long as you don’t use too much and get it hot enough (but not too hot, of course!). I know the solder paste looks messy below (and there are better ways to apply it), but it works great. The main thing you need to watch out for is tiny solder balls. I did some cleanup with a tiny pick, X-acto knife and some alcohol-dipped Q-tips when I was done.

The capacitors were next. These are all 0805-sized capacitors and with some practice, a small soldering iron tip and thin solder, that size is pretty straightforward to solder by hand. The process goes: clean the pads, tin one pad, hold the capacitor with tweezers and heat that solder pad to attach one side, then solder the other side. I can never make them look as good as the automated assembly machines, but they’re still good joints and they work well. (I love my lit magnifying glass for this work.)

After the last capacitor was on, I did some more cleanup (watch out for those solder balls!), installed the CR2354 battery, and secured the PCB to the case.

Smoke check time. I connected the Gemini to a power supply and hand controller then fired it up. Including the expected CMOS reset warning, everything worked as expected. Next, I connected a computer to the (new!) RS-232 port and connected with ASCOM & NINA. It’s alive!

After a few hours of testing in the shop, I did a couple hours of testing at the telescope under the night sky. It all went well and the mount seems to be back to 100%. Here’s an 8x speed video of a slew test I did using NINA and multiple slew-to-AZ/EL commands.

After that testing, I added the DB9 serial optical isolator and tidied up the cables. With the optoisolator, initially connecting from a computer to the Gemini is a little more fussy. I’m not sure if it’s the optoisolator, the USB-to-serial adapter, or the MAX3226’s AutoShutdown Plus feature causing this, but to get a 100% chance of a valid ASCOM connection I have to do this:

• Start with the mini PC on and the Gemini off
• Disconnect the USB-to-serial adapter from the mini PC
• Disconnect the other end of the serial cable from the RJ22 port on the Gemini
• Boot the Gemini & wait about 10 seconds (I’m not sure the wait time matters)
• Plug in the USB-to-serial adapter and the other end of that cable to the Gemini RS-232 port
• Click connect in the ASCOM Hub on the mini PC

I need to do more testing with this to figure out the connection pain root cause. I only connected a few times before I installed the optical isolator, so that’s worth digging into more. It should be easy to see if it’s the Gemini (or new MAX3226EEAE+) by doing a GPS position request in the Gemini controller to wake up the serial port then initiating an ASCOM connection. I could also try different USB-to-serial adapters.

Anyway, as long as I connect using the above method it always works. Once the Gemini computer is connected through the ASCOM driver to my mini PC, it works perfectly.

In the last handful of nights I have over 26 hours of imaging on my repaired Gemini. It’s great to have it back!

Disclaimer: I paid for all items above with my own money. Some links above are affiliate links. As an Amazon Associate I earn from qualifying purchases at no additional cost to you.