Gemini 1 - a bit of a mystery

[pmrid]

I have lost RS-232 communications between my PC and Gemini 1. Worked fine a week ago. Not today.
We had a storm in between, local lightning strikes, driving rain etc and the Losmandy gear was set up in the Obs - plugged into the wall but not turned on. The storm was so sudden I couldn't get to the Obs in time to unplug everything. Damn.
The Gemini had lost date and time - I have had to reset that - but other defaults seem to have remained - location etc.
I've checked the integrity of the connecting cable and have installed the USB-RS232 cable on a different USB port and also installed a second USB/RS232 but neither seems to be able to connect.
Maxim reports that it can't find the Gemini.
I've opened the Gemini and checked that the RS232 connection is sound - no cracked solder points etc and no sign of any burnt chips etc.
Not sure where to go from here.
Any suggestions?
Peter

[JohnH]

Hmm, Gemini oddness. First port of call is change the battery. Not sure if it will help or not, it might.

[pmrid]

Thanks John. Yes, I'm going to do that first. It's the 'chicken soup' recipe - as the old joke goes - "It may not help, but I can't hurt."
Peter

[wasyoungonce]

Peter do the usual's (http://www.gemini-2.com/faq.html):
Check SRAM voltage is above 2.8VDC
Continuity check RS232 cable.
Check RS232 IC has not entered sleep mode. See the FAQ's #11 for this.

I replaced a RS232 IC (as well as many others) on David's Gemini so it can be done easily, cheaply, as a Last resort. It's more than likely the IC is in sleep mode.

Brendan

[pmrid]

Thanks Brendan.
I'm getting 2.93 VDC across the CMOS battery. I imagine that is a reasonable result and probably does not warrant any replacement at this stage. I just turned the Gemini on to check that the battery was retaining the date/time/location etc data and it was all there. I also double checked that the PC connection was still failing. Sadly it was.

So what is left is either a sleeping IC or a dud.

I have identified the relevant cap (C21) coming off Pin7 of the IC. The FAQ says the side closest to the IC is negative and that closest to the RJ22 is ground. Does that equate to the positive side for the purpose of my multimeter? I gather the task here is to power up the Gemini and see if I have voltage =~5VDC. Is that it?

Peter

[wasyoungonce]

Hi Peter,

The Battery voltage is not the SRAM voltage.

You really need to check the voltage at the centre terminal of D1 (or SRAM pin 18) and any earth. Even though your SRAM values were ok the voltage may be low and I have noticed sometimes the SRAM can keep values on low voltage sometimes not.

So you must check SRAM voltage is above 2.8VDC.

As for the C21 voltage...you are measuring across C21. It won't matter which side your leads are on. One way you'll get +5V (if ok) the other -5V (if OK)...aka putting the probes across the wrong way just reads the voltage in reverse potential.

But they measure it across C21 with the negative probe on the left side (power connectors on bottom, hand set cable on left) of C21. The left side ..closest to RJ22 (actually RJ11, 4P4C in Australia) is the earth side, see attached pic. The red arrow is C21 earth side.

Yep your looking for ....-5V or if leads wrong way +5V (see below).

Brendan

Just an edit:
I measured David's on bench and found 0V powered up across C21. You need to plug in a working RS232 lead (or your Gemini BT box:lol:) and connect to the device (edit edit: actually I didn't need to connect, I just plugged in a RS232 to USB and it worked) to see the -5V. Also the earth I marked above is correct but since it's a -ve charge pump you'll see -5V DC or if you have the leads the wrong way +5V DC.

[pmrid]

In my blissful ignorance of all things electronic, I have assumed that the voltage across pins 16 and 32 of the SRAM chip (as per FAQ 15) was the equivalent to the reading of the CMOS battery. My error. The 2.93 VDC I got was the reading across the SRAM chip. So I assume that is a satisfactory reading? Yes, no, maybe??

Peter

[wasyoungonce]

Hi Peter yep 2.93 across SRAM is fine.

Just check C21 with an RS232/USB (or Gemini BT) connected. You do not need to electronically connect to the device, just a physical connection should get the RS232 IC running. Looking for -5v or +5V across C21 (with Gemini powered up).

Brendan

[multiweb]

You also have to make sure the transfer rate is the same in between PC and Gemini or they won't talk. I know that mine resets to 9600 bauds and my laptop was set 19200. I got caught once.

[wasyoungonce]

You know...that's odd. I can connect & communicate from 4800bps all the way to 57600kps.

edit:

oh ok yep if I use my RS232/USB then coms rate is 9600bps but I always use my Bluetooth device, which can connect at differing COMs rates. But that is the BT COMs connection data transfer rate not the Gemini RS232 port rate so yep you are spot on Marc.

[pmrid]

I just fired it up and got a virtually zero voltage across C21 - but I hadn't put a load on the RS232 - so I'll go back and do that again.
Peter

[multiweb]

I can connect at different speeds too but I have to match the COM speed in the control panel with the GEMINI rate. My old laptop runs Win XP. 9600 is ok so is 19200. Anything over and I sometime get drop outs. Then again I have a very long USB cable. (10m)

[pmrid]

Thanks Marc. I'll check that. But in view of the tests I just did, I don't think that's the answer.
Peter

[pmrid]

OK. Done that. Zero volts.
So it's either dead or sleeping.
If it was sleeping, according to the FAQ (if I have understood them) then it ought to have woken up when I tried to connect to the Gemini from the PC - via a USB/RS232 dongle. Or am I missing something?
Peter

[wasyoungonce]

No Peter your do'in fine. I can wake my Gemini RS232 by just connecting it to the lappy, I don't have to connect via Gemini ASCOM.

I fired up, measured 0V DC across C21 then connected them to the lappy (didn't fire up Gemini ASCOM yet) and both woke up ok from this with -5V DC.

I don't have a U6 (MAX3221) IC but they are available from Element14 quite cheaply, refer to my document on this, Gemini 1 parts availability (http://tech.groups.yahoo.com/group/Gemini_Users/files/Gemini%201%20PCB%20parts/) (on the Gemini Yahoo site).
What you need to do is identify the IC (your PCB version may differ from mine) is the same part number (as per my document) and maybe someone near you is good with a soldering iron & ESD procedures.....maybe a good TV repair place?

Or if you want you can send it to me. Problem is that generally these IC's are killed off thru ground loops and you may have other IC issues? Hard to tell. Also it's a hassle sending it all the way to Mexico city (Melbourne), well for you it is.

I have most other Gemini IC's on hand except the encoder PIC pre-scalers and quadrature motor encoder IC's and EPROMs. They have firmware and thus need to come from Rene or Losmandy or such like. I'm waiting for an IC from Rene ATM, U11...all the way from Germany.

What is it with Queensland'ers and Gemini's anyway???? You're Killing them...or does someone have a bad cable that's being passed around?:lol:

(edit) You can send it to Peter Ward...maybe he is best to comment on this.

Brendan

[pmrid]

5.42 volts across C21. That's with a live load in the RS232 connection. I used one of those SkyWire devices plugged into my IPad.
Strangely though, the software in the iPad didn't recognise the Gemini
So, we have a live chip that wakes on load. For this exercise, I had disconnected the hand pad - because another device was present to control the Gemini.
Now, I'll reconnect it and try a normal load from the PC.
No chnge. Gemini doesn't respond. So it would seem the IC is OK and is showing correct voltage.
The hand controller functions all seem to work fine. It will slew to a target, track and do all the usual things.
Seems the problem lies elsewhere.

Peter

[wasyoungonce]

Peter...give me a few minutes as I'm tracing out the U6 circuit on the PCB. I've also read up on the MAX3221 specifications.

The fact the charge pump works does give some confidence the IC is ok however it still may not be. More than likely it's ok.

You have to check your cables for pin to pin continuity? Did you try the Gemini BT and it's cab;le?

Brendan

[wasyoungonce]

Hi Peter

Ok I've traced all the pins. Nothing special about this IC circuit, pretty much follows the manufacturers layout specs.

Rj11 pin 1 is not connected.
Rj11 pin 2 is earth.
Rj11 pin 3 (RS232 Rx in) goes to pin 8 of U6.
Rj11 pin 4 (rS232 Tx out) goes to pin 13 or U6.

U6 pin 9 goes to U1 (CPU) pin 71.
U6 Pin 11 goes to U1 (CPU) pin 72.

It's unlikely that the IC, U6, is dead if the charge pumps work. These charge pumps are used to transition voltage levels in the IC to usable levels for the circuit.

More than likely you have a bad RS232 cable! Did you try the Bluetooth module as a RS232 I/O (with it's own cable) and Gemini ASCOM?

Anyway lest see how you go with this. The pins are very hard to measure, you don't need to do this they are just put in for info.

Oh I have attached a diagram of the Gemini RJ11 socket designation and U6 connections.

Brendan

[pmrid]

Thanks Brendan. You're a great help. I haven't tried the BT dongle yet - mainly because the OS is Win7 64Bit and BT was always 'dodgy' with it. But I'll give it a whirl.
cheers
Peter

[pmrid]

OK. No joy with the BT. I have 2 of them here - both from you as you remember. Neither would connect to the Gemini. So itisn't a able issue either.
I thinkGemini 2 is looking better all the time.
Peter

[wasyoungonce]

Hi Peter:

do you use a Gemini GPS? If you do and that communicates ok then U6 should be ok.

Another thing to check, as Marc said, is the Handset RS232 data rate is set to 9600bps (in the Gemini set-up menu) and ensure the USB/RS232 port is set to the same.

Further to this ensure the USB/RS232 converter (or BT) port settings are:
bps: 9600
data bits: 8
Parity: None
Stop bits: 1
Flow Control: None

If these are ok then it's starting to look like an bad U6 then (even though the charge pump is ok). Tom Hilton has stated many times that the charge pump is a good indication of a working or not IC....most of these IC's fail due to ground loops....specifically this fail is attributed to lightening.

Brendan

edit:
One final thing to try...do you have a computer with a proper RS232 serial DB9 port (or someone else with a proper RS232 port)? There are many variations of RS232 and maybe there is a degrade on the IC or someth'in?

[pmrid]

Cheers mate. Yes, tried all these things. Port setting are right. And the device is not recognised on a true COM port either - and yes, I did try the Gemini BT and no, it didn't say hello.

Where to from here mate?
Peter

[wasyoungonce]

Ok, looking like all options tried, need to replace the IC or maybe the CPU. I'm pretty sure it can be resurrected.

You can send it to Losmandy, or try Tom Hilton (Gemini Yahoo group) or Peter Ward...though I'm not sure he my just on fwd it to Losmandy (I don't know could be speaking thru my hat!)....or I can do it....or maybe someone local?

Also maybe ask Tom on Gemini yahoo forum if there is anything else to try.

I do not have the IC on hand, I can get one in a matter of a few days. I do have the CPU (new spare) but I do not think it's that.

The IC change is really just a few minutes (although I need to fully test it) ...David is coming down here for a boat show (in March?) maybe send it with him.

Back to you Peter.:lol:

Brendan

[pmrid]

That sounds like a plan. I'll have a yarn with David. I'll let you know how I get on.
cheers
Peter

[wasyoungonce]

Peter do you want me to order a few of the ICs in advance?

If you do, I need to to ensure they are the same IC's as mine, that is they should have MB3221I as their markings.

Also a a pic of your PCB so I can see what version...the IC's are different sizes on older PCB vers.

Brendan

[Brundah1]

Glad mine was a fairly straightforward fix!

No lightning involved there.:)
D

[pmrid]

G'day Brendan.
It's nearly 3 years since the last time we did this but it's on again. Lost RS232. We had a few close strikes in a storm and despite my being nearly fanatical about disconnecting things, I slipped up and it seems you have lost any claim to being the certified village idiot.
Can I ask you to have a look and see what can be done?

Peter

[wasyoungonce]

Hi Peter...the dreaded ground loop causing RS232 nasties. No problems...I have more than enough of those ICs. Just need the main PCB ...as long as it's well packed (in an antistatic bag & box etc) ...or send in it's case.

Brendan

[h0ughy]

Gemini Dr Brendan to the rescue:thumbsup::D