Author Topic: GWInstek SFG-1013 Displays not working correctly - FIXED  (Read 2684 times)

0 Members and 1 Guest are viewing this topic.

Offline tron9000Topic starter

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: gb
  • Still an Electronics Lab Tech
    • My Hack-a-day project page
GWInstek SFG-1013 Displays not working correctly - FIXED
« on: April 06, 2018, 08:30:13 am »
Hi guys,

I acquired two SFG-1013 sig gens. Both of them had "display Broken" written on them. Plugged them in a sure enough their displays were broken.

One has 2 digit's that won't come on - am thinking that the CC/CA transistors for those digits has gone (if multiplexed), or no supply to them is not present, the whole digit is dead, or worst case: the micro-controller for the display has blown outputs. Also some of the button don't work.

The other one is really weird. On all digits only the numbers 2, 5 & 6 are faulty, yet punch in 888888 from the number panel and all segments light, see attachments. Does anyone know where to even start or come across this sort of fault before? Am I even right thinking it'll be a multiplexed display, could it be charlieplex?

I've also included a picture of the insides, as it can be seen on the board at the back, there are many SOT transistor packages.

I've had a look on the internet for the fault, but the only hit I've found with regards to this problem is an amazon review of someone complaining of this fault.

Thanks
« Last Edit: April 13, 2018, 08:03:08 am by tron9000 »
Partsbox.io - orangise your parts!
"If you're green you can only ripen. If you're ripe you can only rot!"
 

Offline Chris56000

  • Frequent Contributor
  • **
  • Posts: 896
  • Country: gb
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #1 on: April 06, 2018, 08:27:24 pm »
Hi!

The displays look like standard 0.56" common-anode types with a standard 10 pin arrangement, the middle pins of each display will be the common-anode connection.

The first thing to do is to make sure all the displays are in good order - the Chinese 7-segment modules have a habit of becoming intermittent or partially dead on one or more segments or the common anode!

To do this you need a digital multimeter with a diode-test range (arrow-and-line symbol) - do you have one of these?

If you have a suitable meter, switch it to the diode-test range and disconnect your generator from all power. Place the red lead from the meter on the central pin (third pin) of each display and briefiy touch the black lead from your meter on each other display module pin in turn - if the display is common-anode you will see a faint glow from a working segment when the black lead connects to a segment element.

If, by doing so and you find some that seem to be dead/different from adjacent digits, mark the side/top of each module with a red dot from a marker-pen to identify it and let us know which they are with a suitable picture!

If the displays are common-cathode they test exactly the same way except the black meter lead would need to be connected to the middle common pins and the red lead to the other pins.

If you can't get a glow when using the middle pins of each segment-module, it's possible you have a board with non-standard segment module pin-outs - in which case you'll need to remove one of the front-panel PCBs so we can see what display modules you have and identify the pin-out for you!

Test all your segment-display modules first from each unit and report back - I do NOT advise you attempt to un-solder or change any of the display modules, transistors, etc., etc., at this stage - we need to see some really good close-up pictures of the board first!

It is possible you may have a yakky micro but yakky lead-free soldering is far more likely to be the problem - sometimes it gets so bad that completely depopulating the board, cleaning and re-tinning all the pads and re-soldering is the only way to get reliable operation!

Dead LED segments could possibly cause the micro to misbehave if the segments are multiplexed together with the control-buttons, which is sometimes done in these Chinese designed instruments!

Chris Williams
« Last Edit: April 06, 2018, 08:32:26 pm by Chris56000 »
It's an enigma that's what it is!! This thing's not fixed because it doesn't want to be fixed!!
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 5795
  • Country: ca
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #2 on: April 06, 2018, 10:18:31 pm »
i would slowly pry out the socketed ic, had troubles in the past due to bad contacts in theses socket types, they tend to deform with time,  pry out the chip and reseat it 2 times, it will move out any oxydization dust  etc ... i would start with that
 
The following users thanked this post: tron9000

Offline tron9000Topic starter

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: gb
  • Still an Electronics Lab Tech
    • My Hack-a-day project page
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #3 on: April 09, 2018, 08:27:39 am »
Thanks guys

I do have a digital DMM with diode check/continuity mode (not my first rodeo, but i know sometimes you have to ask), it is a cheapo B&Q one, but its only temporary.

Once I get time will remove the board from the unit, give those segments a check and if they are socketed, give them a couple of reseats. Will post pictures too, y'know in case your interested? ;D

Partsbox.io - orangise your parts!
"If you're green you can only ripen. If you're ripe you can only rot!"
 

Offline tron9000Topic starter

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: gb
  • Still an Electronics Lab Tech
    • My Hack-a-day project page
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #4 on: April 10, 2018, 12:17:18 pm »
I had a another poke around last night.

Managed to get the panel off and tested the segments of the unit that had the digits that wouldn't work at all. Tested the segments as Chris56000 described and all segments work on the non-functioning digits and is a CC 7-segment display with the CC pins being the middle on each side of the display.

You'll see the digits aren't socketed.

To get the panel off: remove the knob of the encoder and undo the screws holding the PCB to the panel and unhook from main board, that's it

Good way to determine which pin is CC/CA and if your display is mulitplexed: all segment pins are common on each digit, so each of those pins should be connected to the same pin on the other digits (e.g. ‘a’ on digit 1 is connected to ‘a’ on digit 2, etc, etc). The ones that aren't connected to other digits are either CC or CA and that can be determined if it can be traced to any nearby transistor or resistor to a supply rail.

Once I found that I managed to trace out the following circuit

Now I made an educated guess about some things here: That the transistors were NPN, it’s the only way it could work from the connections. They both had the same markings on them.

The uC is a Winbond W78E052C40P

Pin 6 of this uC is connected to part of PORT1 which is a bi-directional I/O port with internal pull-ups

Interesting way they have one pin to do 2 things. Notice that the switches S1214 & S1209 are connect to the same pin. They must switch the pins function from output with driving the digit to input when that pins doesn’t need to drive the digit.

I only traced one switch (it was getting late & cold), but if i had to guess that the when the pin changes to input, and if S1214 was pressed it would go low and probably trigger and interrupt or something, not sure how it would know which one was pressed, would have to have look beyond S1209.

Anyway going back to the problem,

If any of the transistors had gone S/C then the digit would still work (of sorts). So looks like the output of the uC might be to blame. I checked the pins on the uC’s socket and no signs of poor connection or corrosion.

Also S1214 is the pad for the ‘1’ button, which is also not working and tested D1220 (D1), which was ok, yet more evidence against the uC. So either that pin on that port has truly gone or lost pull up? I could tag a pull up onto that trace and see if it fires up, but I’ve never heard of a uC loosing its pull-up?

Anyway have a look through all the pictures below if you’re interested.
« Last Edit: April 10, 2018, 12:27:03 pm by tron9000 »
Partsbox.io - orangise your parts!
"If you're green you can only ripen. If you're ripe you can only rot!"
 

Offline Chris56000

  • Frequent Contributor
  • **
  • Posts: 896
  • Country: gb
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #5 on: April 10, 2018, 09:09:22 pm »
Hi!

The circuit sketch of the digit–drive for the c.c. connection gives a clue for testing – reconnect the front panel to the remainder of the generator's circuit and switch on.

Use your meter on 10V or similar, connecting the black lead to a convenient chassis point,  and test each pin on the red lead in turn.

The exact d.c. voltage reading you'll get depends upon the duty–cycle of the digit–drive enable pulses from. the micro, but they should all be approximately the same, about +1.8 to +2.8V.

Trace back along the two transistors from each digit cathode that's missing to the micro using the schematic you've drawn already as a guide ;– this will be the digit–drive enable pin – each should be about the same. Any pin with a much lower voltage by comparing with a lot digit is suspect, whiilzt a high voltage may indicate an open–circuit digit drive transistor.

Unfortunately if the micro outputs turn out suspect I'm afraid there's little more you can.try, and if swapping the micros swaps the faulty digit patterns with it then it's it game over for both units, unless you can obtain another micro from.GW Instek!

If the faulty digit patterns do NOT change when you swap the micro, the fault is elsewhere on the PCB and you'll need to go through all the remaining transistors, diodes, etc., etc., on the board very carefully!

Chris Williams
It's an enigma that's what it is!! This thing's not fixed because it doesn't want to be fixed!!
 
The following users thanked this post: tron9000

Offline tron9000Topic starter

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: gb
  • Still an Electronics Lab Tech
    • My Hack-a-day project page
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #6 on: April 11, 2018, 08:24:29 am »
Hi!

The circuit sketch of the digit–drive for the c.c. connection gives a clue for testing – reconnect the front panel to the remainder of the generator's circuit and switch on.

Use your meter on 10V or similar, connecting the black lead to a convenient chassis point,  and test each pin on the red lead in turn.

The exact d.c. voltage reading you'll get depends upon the duty–cycle of the digit–drive enable pulses from. the micro, but they should all be approximately the same, about +1.8 to +2.8V.

Trace back along the two transistors from each digit cathode that's missing to the micro using the schematic you've drawn already as a guide ;– this will be the digit–drive enable pin – each should be about the same. Any pin with a much lower voltage by comparing with a lot digit is suspect, whiilzt a high voltage may indicate an open–circuit digit drive transistor.

Unfortunately if the micro outputs turn out suspect I'm afraid there's little more you can.try, and if swapping the micros swaps the faulty digit patterns with it then it's it game over for both units, unless you can obtain another micro from.GW Instek!

If the faulty digit patterns do NOT change when you swap the micro, the fault is elsewhere on the PCB and you'll need to go through all the remaining transistors, diodes, etc., etc., on the board very carefully!

Chris Williams
Thanks, will give that a try, obviously if the fault follows the uC in a drifferent PCB - its the uC, so will try that first as its quicker and more definitive. Good job I found 2 then!
« Last Edit: April 12, 2018, 08:21:41 am by tron9000 »
Partsbox.io - orangise your parts!
"If you're green you can only ripen. If you're ripe you can only rot!"
 

Offline tron9000Topic starter

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: gb
  • Still an Electronics Lab Tech
    • My Hack-a-day project page
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #7 on: April 12, 2018, 08:45:55 am »
So I took the display board out of the second one last night. Let's call them Units A and Unit B.

So I took the uC out of unit A and stuck it in Unit B’s display board and vice versa.

My intention here was to check if it was actually the uC’s that were toast. If it was, then probably give up and move on.

Connected unit A’s display board back into unit A with unit B’s uC in it….power on…


WORKS!
All buttons, all digits on all displays, encoder, checked signals at output through a range of frequencies and waveforms (50OHM output at 3MHz squarewave is a little distorted, but the TTL signal worked well), loaded the output up with 50ohms and measured the signal again: no distortion or loading issues.

Tried Unit B with Unit A’s uC in its display board: ALMOST SAME RESULT! Only thing that let it down was the default DC offset was out by about -300mV. Tried looking for some adjustment, but could see any!

This has Left me a bit stumped!

The only thing I can think that happen is the contacts on the uC socket were bad as Coromonadalix suggested, and removing/reseating cleared it up. Then how come I was getting continuity when tracing the circuit out?

Thanks Chris56000 & Coromonadalix (you were right and sorry I miss read your post originally) for your help with this.
« Last Edit: April 12, 2018, 08:52:20 am by tron9000 »
Partsbox.io - orangise your parts!
"If you're green you can only ripen. If you're ripe you can only rot!"
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 5795
  • Country: ca
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #8 on: April 12, 2018, 10:55:03 am »
We have 10 years old machines who comes back because of the cpu socket, i remove the socket and solder the cpu directly on the pcb.

The socket somewhat deform with time, and bad contact happens, even reseating the ic doesn't last long.

The socket become "roundish" look a like, not totally square anymore  ...

good to know it worked for you :)

 

Offline tron9000Topic starter

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: gb
  • Still an Electronics Lab Tech
    • My Hack-a-day project page
Re: GWInstek SFG-1013 Displays not working correctly
« Reply #9 on: April 13, 2018, 07:57:30 am »
thanks and apologies again for not reading your post more closely. :-+
Partsbox.io - orangise your parts!
"If you're green you can only ripen. If you're ripe you can only rot!"
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf