Any ideas on repairing a (slightly!) blown motor board ?

[Trickysystems]

Malcolm,

I just had a look in the latest handset I have and it has the values stamped on the ferrites and  an extra diode. Picture attached.

102 ferrite is 0.15 ohm at DC and 1,000 ohms @ 100MHz and 1.5A max current.

For a test, you could probably use a low ohm resistor or even short it.  As this is the interface board to the rest of the hand set, it could be tested in isolation.

Hope this helps.

Richard

[malc-c]

Thanks for the suggestion.  Does FB4 connect to the 3rd pin from the top as viewed ?  If you look at the image I posted I have a a good trace between the 4th pin form the top and pins 7&8 (+12v in from the mount) so no need for the diode.  I have also noticed U6 next to the crystal is a little charred but there are no pads or legs, so guessing its a BGA type package - but its very small ??

I've spent best part of  5hrs play with it and can't see any possible fix - even checking continuity for both Euarts on the 80 pin 18F Pic to the 6 pin connection cable header at the top of the main board doesn't ring out...  I was hoping that with some delicate soldering of some fine wire I could just take the TX/RX  directly out to the daughterboard

[malc-c]

 

unlike the ferrite bead, 3/4 of U6 remains !!

I'm also interested in what the small three wire connector does as the USB to serial chip would be used if the handset was to be used as a pass through.  The fact you could have either an EQDIR cable connected to the PICs serial port, or use the USB port suggest that the TX and RX must be linked, and that three wire could be the TX/RX and GND from the onboard USB/TTL serial chip.   Anyway, its almost 1.30am so time I retired to bed !

 

My gut feeling is that whatever vaporised the Ferrite bead may have also blown the PICs Euart and I lack the skill, tools and eyesight to remove the 80 pin package, fit a replacement and then program it !

Edited January 25 by malc-c

[Trickysystems]

Malcolm,

Just a little more snooping between my two hand controllers. The added diode in one simply is a replacement for D4, so maybe just an upgraded component for extra current? The IC that is vaporised on your controller has AJ3 as the marking and appears to be a DFN8 package. Without checking the tracks, I am uncertain of the actual type of device or the function. Note that the other hand controller I have does not have that IC present. I'll try and figure out the differences between the two circuits. Perhaps it might be possible to bypass it completely? J8 on the controllers appears to be the ICSP for the PIC. I'll check further to confirm if that is the case.

If the controller does not prove repairable, then an alternative is to remove it and put in a wireless or BT adapter in place of it to control the motors.

Richard

[malc-c]

Thanks for the images, and ID of the part U6.  Like a lot of parts in Chinese electronics those that do have labels are often non traceable.  But even then if I could find out what it is I don't have the tools to replace it.  My logic was that regardless of which port is used the TTL signals end up at one of the two USARTS on the processor, so once a suitable place was found I was going to see if adding a couple of wires directly from the pins of the handset port to the pics serial port pins.  However looking at the datasheet for the PIC18F85J10 I can see that whilst the chip is a low voltage package, the digital pins are 5.0v TTL level tolerant

[MikeAa]

So excessive current flow caused the bead to fail, therefore  something else on the board is faulty?

[malc-c]

2 hours ago, MikeAa said:

So excessive current flow caused the bead to fail, therefore  something else on the board is faulty?

That's my guess too.  Something could fail, allowing a new path for the current to flow, which then has a domino effect allowing higher voltage to reach low voltage components.

[Trickysystems]

The two controllers are very similar, with some rearrangement of components. The blown IC  is replaced by a resistor on the other board. The only SMD device in the same package, with AJ3 marking, was 3AA02/34LC02. This is a Microchip 2K I²C Serial EEPROM. If this is what the device is, I'm not sure what it is doing. Pictures attached of each controller. R18 replaces the IC on the other controller, and some other resistors are added or omitted between the two controllers. Both controllers are version 5, yet slightly different. Both have the USB connector on the hand unit.

J8 pins connect to the correct pins on the PIC to enable ICSP.

It is looking like a better option is to replace the hand controller, or use a BT or WiFi adapter. Controllers seem to be available on the second hand market from time to time.  Although my engineering past makes me want to know what went up. All ICs and transistors are encapsulated with smoke, so that is the indicator they have failed, at least that is what I was taught...

 

[malc-c]

That's very interesting.. I wonder what that small 8 pin package does.  It does seem that the same via's are all connected to one side, with the same trace running to the second from left pin on U8.  I wonder which revision came first?

What strikes me is that major manufactures of graphics cards make active schematics and component ids available, but try and get the same from Synta... you have more chance of finding fairy dust !

Anyway, this handset will sit in the draw until someone else ends up cracking a screen or something that can be fixed by swapping parts over. 

 

[Carbon Brush]

Late into the discussion, and only a couple of photos to look at. But maybe relevant.

Circuit designers sometimes omit over voltage suppression/protection components.
Either they don't think about them, or a bean counter somewhere argues about bill of materials cost and the designer tries to save 10p.
Then in the real world it is realised there are more failures than expected. If not warranty claims, they are loss of company image/status when they get known for unreliable product.
Oh - maybe overvoltage or reverse voltage protection is a good idea after all.

The 'extra' diode mentioned by @Trickysystems may be a 1N5401 (3A rated) diodes in series with the supply. Providing reverse voltage protection.
The tiny 8 pin device U6  could well be a zener protection network. These are made by several companies and find their way onto USB ports.
They would help on logic level RS232 ports.

HTH, David.
 

[malc-c]

David, thanks for the info... I wonder how many designers have gone back to the accountant and said "told you so" after such practices are employed.  

[Trickysystems]

Of the two controllers I have, the earlier one has the AJ3 IC (U6) and the later one just has a resistor, at least based on when they were purchased, but the actual controller delivered may just be a question of stock on hand at the time.

The controller without the tiny IC 6 has R8 omitted, R10 installed (150K), and a new position, R18  (100K) installed.

Having a closer look at the circuit, PIC pin 6 (RG1/TX2/CK2) connects to  pin 5 (2A) of U8 (74HCT2G125), so is an input. The output of U8 then goes via R18, which is a pull up resistor, to a through hole where small IC U6 would be mounted on the other controller. The other output from U8 (1Y) appears on another through hole next to it. As this appears to be a multi layer board, it is very hard to identify where the tracks actually end up.

This could mean that the PIC is OK, and also U8. Repairing the issues on the daughter board may allow the main board to power up. The outputs from the burnt U6 would need to be identified and direct connected, so possibly fixing the controller?

As for the extra diode soldered to one of my controllers, it simply replaces D4 and so is not extra, just heavier duty.

Richard

[malc-c]

Richard,

I did consider a direct connection to the pins of the RJ45 socket, but with an 80 pin package, and not knowing if the firmware is written to use com port 1 or 2 (or should that be com port 0 and 1?) and the fact that I don't have anything thin enough to solder I shelved that.  But now you have done your best to trace out the path, at least to points where the solder mask can be scrapped back to expose the tracks and enable a ware to be soldered might give me something to work with. - Many thanks

 

[mimmo]

On 23/01/2024 at 00:15, malc-c said:

Just to clarify the port on an EQ6 / NEQ6 that has a D9 port does not run the serial lines at true serial voltages.  The TX and RX lines are TTL level, 0 to 5v.  The problem is that next to the data lines there is +12v and GND so what you have probably done is basically shoved 12v across the data lines and blown the EUART port of both PIC micro controllers.

The only way to fix this is to follow the steps in this thread to program two new 16F886 microcontrollers, desolder the existing ones, being carful not to lift any pads and damage the tracks, and then solder the newly programmed microcontrollers in their place.  Prior to Christmas I could have sold you a mainboard that already had the repair undertaken as I had several spares due to other members kindly donating blown boards to me after they had already gone out and purchased a new replacement control board.  However I no longer have any as they have all gone.

The other problem is that as you are in Italy the cost of shipping the faulty board for repair will be expensive as it needs to be insured and tracked.  Due to issues I've had with sending replacement micro controllers  and a control board outside of the UK I can only take in boards for members based in the UK mainland.

Replacing the microcontrollers isn't that difficult, but unless you have or can find someone with a PIC programmer who can program them for you it's pointless to try.   The HEX code for an EQ6 has been attached (assuming the forum software allows that).  

If a repair is not possible then the only alternative is to replace the old board with a new one.  If you do choose that option, then please consider donating the faulty board to the exchange program and send it to me by the cheapest postal service (no need to send it insured or tracked as its faulty anyway, should it get lost in transit).

EQ6-NCP.hex 46.5 kB · 4 downloads

Hi Malc I found some pic programmers for a few euros 15/20 euros, are they safe or better left alone.
Thank you

[Trickysystems]

On the subject of programmers, the PIC programmers I use are pictured. The K150 needs the config file updated to program 16F886 PICs. The PICKit 3 works better for me although the required sockets need to be purchased. There are PICKit 3.5 devices advertised on the Internet, but these weren't designed by Microchip. I'd stick with PICKit 3 and use the standalone software available free from Microchip.

The latest programmer is the PICKit 4, but for our purposes, the PICKit 3 will work and is cheaper. Make sure it has the logo.

[malc-c]

On 27/01/2024 at 18:21, mimmo said:

Hi Malc I found some pic programmers for a few euros 15/20 euros, are they safe or better left alone.
Thank you

Without any details it's impossible to say.  The 16F886 has been around for some time so should be supported by most programmers.  I use the PicKIT2 (now superseded by the PicKIT3 and PicKIT4 ).  You will also need an adapter to hold the 28pin SOIC package.  This is my setup I use to program them, the clam socket and PCB was kindly sent to me a couple of years ago from a fellow SGL member.  

 

Prior to that I was using a 28 pin SOIC breakout board and a sprung clothes peg 

 

 

Edited January 28 by malc-c

[mimmo]

On 29/1/2024 at 00:12, malc-c said:

Senza dettagli è impossibile dirlo. Il 16F886 è in circolazione da un po' di tempo, quindi dovrebbe essere supportato dalla maggior parte dei programmatori. Utilizzo PicKIT2 (ora sostituito da PicKIT3 e PicKIT4). Avrai anche bisogno di un adattatore per contenere il pacchetto SOIC a 28 pin. Questa è la configurazione che utilizzo per programmarli, il socket a conchiglia e il PCB mi sono stati gentilmente inviati un paio di anni fa da un altro membro della SGL.  

 

Prima utilizzavo una scheda breakout SOIC a 28 pin e una molletta da bucato 

 

 

Ok thanks

[mimmo]

On 23/1/2024 at 00:15, malc-c said:

Giusto per chiarire che la porta su un EQ6/NEQ6 che ha una porta D9 non fa funzionare le linee seriali a tensioni seriali reali. Le linee TX e RX sono di livello TTL, da 0 a 5v. Il problema è che accanto alle linee dati ci sono +12v e GND, quindi quello che probabilmente hai fatto è stato sostanzialmente inserire 12v attraverso le linee dati e far saltare la porta EUART di entrambi i microcontrollori PIC.

L'unico modo per risolvere questo problema è seguire i passaggi di questo thread per programmare due nuovi microcontrollori 16F886, dissaldare quelli esistenti, facendo attenzione a non sollevare alcun pad e danneggiare le piste, e quindi saldare i microcontrollori appena programmati al loro posto. Prima di Natale avrei potuto venderti una scheda madre che era già stata riparata poiché avevo diversi pezzi di ricambio perché altri membri mi hanno gentilmente donato schede bruciate dopo che erano già usciti e avevano acquistato una nuova scheda di controllo sostitutiva. Comunque non ne ho più perché se ne sono andati tutti.

L'altro problema è che, essendo in Italia, il costo di spedizione della scheda difettosa per la riparazione sarà elevato in quanto deve essere assicurato e tracciato. A causa dei problemi che ho avuto con l'invio di microcontrollori sostitutivi e di una scheda di controllo al di fuori del Regno Unito, posso accogliere solo schede per membri con sede nel Regno Unito continentale.

Sostituire i microcontrollori non è così difficile, ma a meno che tu non abbia o trovi qualcuno con un programmatore PIC che possa programmarli per te è inutile provarci. È stato allegato il codice HEX per un EQ6 (supponendo che il software del forum lo consenta).  

Se la riparazione non è possibile, l'unica alternativa è sostituire la vecchia scheda con una nuova. Se scegli questa opzione, considera la possibilità di donare la scheda difettosa al programma di scambio e inviarmela tramite il servizio postale più economico (non è necessario inviarla assicurata o tracciata poiché è comunque difettosa, nel caso in cui si perdesse durante il trasporto).

EQ6-NCP.hex 46,5KB · 6 download

Hi Malc, I've started checking the components of my eq6 board again. I found that the Skothy D1 diode gives 390 mV in one direction and 0 in the other.
instead the skothy diode d2 gives 0 both in one direction and the other, so it should be skipped, right?.
the diodes are SK34A smd. Can I possibly replace it with a traditional one?

[malc-c]

I would still replace it with a Schottky barrier rectifier diode that has the same 40v, 3A forward current (70A peak) rating.  It doesn't need to be a surface mount component, a normal radial through hold package will do so long as you can solder to the pads where the diode is mounted

[Jaimie]

Adding my own experience here at the request of @malc-c, who helped me get my Orion 14xxg back up and running! About a year ago, I got the scope out for the first time in several years... in my haste, I plugged the cables in wrong and on power-up got the "unable to communicate with mount" message on my SynScan handpiece. After some DM's to Malcolm, the diagnosis was that one or both chips on the Synta MC004 motor controller boards had failed. He sent me two programmed replacement chips, which I took to a local shop in the San Francisco Bay area (https://commbridge.com/) since I wasn't confident with surface-mount soldering. They replaced a blown diode (see attached photo) and both chips, and the scope is now miraculously working, thanks to @malc-c and the great folks at Commbridge.

Jaimie

[turboSlow]

This thread is chock full of amazing information. 

I've just purchased a used 400p synscan, and to my horror the motherboard appears to be fried like the others here. It's getting nominal 12v, however on D1 and U1 it is only seeing 0.5v. Doesn't seem to indicate the trend of bad chips. Any help getting this up and running would be amazing. I really wish we could just buy these boards outright...

[malc-c]

D1 is basically there to prevent damage if the polarity is reversed, however you normally get around 0.5 to 0.7v voltage drop across diodes so this could be perfectly OK.  U1 is the main voltage regular, so if you are only getting 0.5v on its output then this would suggest that it the regulator that has failed.  It's a 7805 and should have 5V on its output.

One thing to try is with the power disconnected check continuity through L3.  If the link (0 ohm resistor) has gone open circuit then you could replace it with a wire link as a temporary test to see if 5v appears on the regulators output.

If you get 5v on the regulators output but the handset still can's detect the mount then check the voltage on the supply pins of the PIC  (pins 19, 20 and 8).  If there's 5v there then it would suggest that the PIC has been damaged.

All this is assuming you've confirmed the data cable  has continuity between the pins at either end.

[malc-c]

Just to update this thread, and report a partial success 

"TurboSlow" contacted me via PMs and he agreed to let me take a look.  It was really puzzling as the 12v supply wasn't even getting form the barrel jack to the switch terminals, and that is a direct link.  

Postie turned up with a parcel yesterday.  I connected the two boards and turned the power on, there was no magic smoke or anything going pop.  Power was removed and the Fluke DVM brought out to do some basic testing.  Power comes in from the jack socket and goes direct to the switch.  But for some reason I couldn't get any continuity between the pins. now I have no idea if these boards are four layer like the HEQ5 / EQ6 boards, but I would assume so as its good practice to stick with a standard, and there were no obvious traces on the top or bottom of the board.

The large diode D1 was shot, so I replaced that with a standard 1N400x variant.  I had to use a through hole as I seem to have ran out of surface mount ones. Next I  removed the 7805 regulator from the board and tested it, it too was also fried, producing around 0.2v.  I rigged up an old TO220 package 7805 and fed a 12v supply into it and once confirmed 5v was on the output I soldered that to the board just to see if there was something drawing a few amps as a result of a direct short.  Nothing excessive was registered on the power supply.

Upon powering up and verifying 5v was being fed to the board I checked the supply pins on the PIC and it was getting 5v.  However I noticed that whilst not excessive, the current draw was around 300mA which is a lot for a bare board with no motors connected.  I then checked the board for any warm components and the PIC was getting hot.  Powered off and using a hot air gun removed the PIC. There doesn't appear to be an direct short or anything obvious on the board, and no other components got hot at the same time.  I placed the PIC into the programmer and confirmed it was dead as the programmer couldn't ID the device.

The rigged up 7805 was removed and a new 7805 surface mounted device soldered in place of the original.  I bridged the supply form the jack socket to the switch, and bridged the traces to the replacement diode D1.  I programmed a new 16F886 with the MC004 code and soldered it in place.  Hooked up the Az/RA  board, and my handset and connected power.  The handset went through the initialisation stages, but reported the "no response Alt/DEC" and then halted as expected.  The new PIC was running cold and wasn't overheating, and the fact I managed to confirm the AZ/RA board was in working order was a plus point as gave me hope.

Powered down and checked out D3, it showed normal readings for a diode, and  continuity between it and the pins on the handset connector check out.  Just in case the issue was with the PIC I removed it and checked out the communications using software on the PC, and it seemed that the  UART port appeared to have been damaged as I had no echo back.  So I got the last 16F886 I had in my box and programmed it with the code - then tested the PIC to confirm its UART was working and it was, so it was soldered to the PCB.  Everything was connected back up, and I got the same response... it just didn't want to play ball. 

I then got the logic analyser out to see if  the serial commands are getting through, and if the PIC is responding.  So connected two traces wires directly to the TX/RX pins of the PIC and did a test run.  It showed the PIC received the data from the handset, but no acknowledgement was transmitted back, and the TX lines remained high.   I can trace out the TX / RX lines and it all seems OK, but something is still taking the TX line (TX out from the PIC) down.  The problem is Skywatcher run the two serial ports on the PICs in series, using 1N4148 diodes to prevent the two TX lines transmitting at the same time.  If they followed normal convention and kept them independent it would make things simpler to debug !

I also noted the board got warm (not hot) in the middle where no component exist, so can only assume that there is some form of resistance internally, or one of the many small resistors or capacitors has broken down and preventing the TX line data getting through.  I was at a loss,  most other repairs have worked, but then they were just communication issues where they had 12v shoved up the serial port when 5v is the max and thus the Euart on the PIC gets blown, so simply swapping the PICs with newly programmed ones resolves the issue.  On boards that have had power issues such as reverse polarity supply its harder as there are so many components that can get damaged, and it just becomes unworkable to try and test every one.  I had tried three different PICs in the board and all do the same thing, so it was time to cut my losses and call it a day.

 

The kit has been repacked and is awaiting collection tomorrow morning.  Through out the PMs it seems "TurboSlow" will order the replacements form one of the retailers and if he faces the same issue as the OP in this long thread, is prepared to design and print his own housings so that he can use this huge 400P dobsonian.  I've asked him to run a thread on designing and printing the housings if that is the road he goes down as it could be useful to other people who face a similar dilemma  

[ben_mtl]

Greetings people !

I'm back with my motor issue (page 14 of this topic). Just to summarize quickly : I bought a second-hand dob 300 which ALT motor was blown and replaced by the previous owner. The issue is that the ALT motor is rotating in the wrong direction which makes the goto unusable. I'm sure that the problem comes from the motor board because for the AZ motor I have both the original and replacement one, and these 2 supposedly identical boards make the tube rotate in opposite directions.

So I finally took the risk of reversing the polarity of the motor. Hopefully it was reversible and did not damage anything but it doesn't solve the problem. It does reverse the rotation when I use the arrows of the hand controller but it's very sloppy and changing the motor speed has no effect. And as soon as I start any alignment, the tube goes up at full speed !! So I guess there's a mismatch between the motor and the encoder so there's something else to reverse, probably in the "MOTOR-IR" and/or "AXIS-IR" connectors but I'm not an expert with these things...

I was wondering if someone would have an idea looking at the schematic below ?

Thanks !

[malc-c]

The schematic isn't complete nor is it official... it was based on reverse engineering a physical board, but it covers the basics.  However if you note the connection detail around the motor connection is quite sparse with the two filter caps not connected.

From my recollection these mounts use  DC servo motors so replacing this with a true stepper wouldn't work.  Given the motor connection has just three pins, two connected to the outputs of the Dual H- Bridge L293 backs up this assumption.  Rotation of the motor is controlled by the L293 reversing which output goes to ground and which supplies power.  The schematic is incomplete in this area as it doesn't show any control connection on ENABLE 1, which needs to be logic high to enable the outputs OUTPUT1 and OUTPUT2.

Secondly, as this board has the encoder inputs this backs up the use of servo motors as otherwise there is no means for the board knowing how to perform a goto instruction as it has no other way of knowing where its pointing.  Now it could be that the signals form the encoders are either not getting through, or are getting through and the OP-AMP is failing ?

You mention old and new boards, it could be that they differ in the order they route the two outputs from the L293 to the pins on the motor connector which would result in reverse directions.

Not sure what else you could try other than fitting a genuine SW motor.