Fluke 8060a refurb... possible mainboard-to-micro hack and oled conversion

[ogdento]

Hey all,
I've really enjoyed reading all the 8060a repair threads on this forum and over at modemhead's site...

I've got an 8060a that I think has a bad mainboard-to-micro elastomeric strip - everything's been cleaned many times with 99% ipa.  The lcd is probably okay and  is the older style without the integrated polarizer.  It uses the old, roundish lcd-to-micro strip and has some faded segments but the real problem is the micro exhibits signs that it can't talk to the mainboard (hung boots, locked continuity, general wonkiness)

I was considering replacing the elastomeric strip with a soldered-in, pluggable connector and then doing an oled conversion (dmitri on modemhead's site did an awesome led conversion in "a-very-unique-fluke-8060a").  But before I do any permanent damage, does anybody have an extra micro board and a mainboard-to-micro elastomeric strip?  If another elastomeric strip works, I can skip soldering on a connector... and I'd like to have an extra micro board to prototype my oled conversion (I could solder wires onto the lcd pads and still have my original to go back to).

I don't have many spare parts but I do have new bale/stand and an inner top shield to trade if that's interesting to anyone.

[JdaleH]

I know this is an old post, but did you do the display conversion?  I have three parts units, so I should have a processor and strip if you need it.  Regards, Dale KS4NS

[ogdento]

Hey Dale,
I didn't get very far with it... Vindoline generously sent me a micro board which I soldered up and made into a breakout board of sorts, but something was up with the processor and that board wouldn't run.  I'm not sure if I fried it doing the soldering (I like to think I do decent solder work!), or if it wasn't working to begin with... but I couldn't get a pulse out of it.  I've since been distracted with other projects but would like to get back to it.

Two other "display rework" options did occur to me:

1.  Instead of using an oled - that requires something to read/interpret/translate the current lcd drive signals - it might just be easier to do an smd led display made out of pcb layers.  Modemhead wrote up a post on his site about a meter Dmitri reworked with LEDs, but my idea is to use the pcb manufacturer to make a 3 board display sandwich... bottom board with pads for the leds/segments, middle board with routed slots for the segments, and the top board for a diffuser mask.  Then I'd have to solder in a bunch of leds and maybe add some glue for extra diffusing, but I bet it would work.  I saw somebody build a 7-seg clock display this way on hackaday.

2. Discard the current micro board and build/program a new one that drives an oled instead, but that would require reverse-engineering the signals to/fro the mac/micro and then re-writing the software to do all the calcs and tests and gahh I've already got a headache.  I never did ask Dave Taylor if the source code for the micro chip was ever found (if it would even be able to be shared).  I think he mentioned it the amp-hour pod-cast, and that his friend that wrote it had unfortunately passed.

I do love these meters though... and I still think it'd be cool to have a cheap easy display replacement.

[retiredcaps]

but the real problem is the micro exhibits signs that it can't talk to the mainboard (hung boots, locked continuity, general wonkiness)

The electrolytic caps in the 8060A leak from the bottom bung and are usually NOT visible by looking at them.  If they have leaked, they can cause problems like the 8060A not booting, etc.

I suggest taking a very close look at all the electrolytic caps and possibly desoldering some to check the bottom bung.

[JdaleH]

I'm looking at another option; super thin green 7-digit LEDs that are now available.  Those plus a few single LEDs might work fine.  I think I should be able to squeeze them in the space alloted.  A black overlay would be required like Dimitri's.  Dale

[ogdento]

Hey retiredcaps!  Thanks for the reply, but the meter itself is fine - I was just swapping out the micro board. 

and Dale, this thread has a picture of what I was doing initially...

https://www.eevblog.com/forum/testgear/fluke-8060a-refurbish-questions/msg1971275/#msg1971275

and the following hackaday thread is what I was talking about with the multiple layer pcb sandwich display... I was going to try it with only two layers, skipping the top "diffuser" board that he's got, and trying hot-glue or epoxy right in the routed-out channel as a diffuser.  my top board was going to mate with the elastomer (to make it removable) and would have the channels for the leds, while the bottom board would have the leds on it.  I think this would be the simplest and most readily reproducible way to go, I drew a sketch of what I'm talking about but can't find it! 

https://hackaday.com/2016/08/23/charliplexed-7-segment-display-takes-advantage-of-pcb-manufacturers/

[ogdento]

Hey JDale,
I realized I posted at the same time you did and missed yours... I swa some thin-ish 7-segs but didn't find anything thin enough, which ones did you find?  That would greatly simplify things!

[ogdento]

Here's a couple of sketches... there's nothing to scale and I didn't work up a pcb or schematic, these are just general ideas...

This is what the display would look like (using a thin vinyl/plastic top mask that would be cut-out for the words):


This is the top board I envisioned that would mate (via pads on the underside) with the elastomer:


The bottom board would have the leds soldered onto it and probably use castellated edges to solder it to the underside of the top board:


After the leds were fitted, and the two boards connected, I'd fill the led "channels" with glue/epoxy

But the thin 7-segment display idea would make this a lot less tedious... I could just route out giant square holes in the top board so they'd fit.

[ogdento]

at risk of , I found the mock-up of my OLED idea.  Beware it's a total smoke-show... just an OLED connected to an arduino so I could find a workable font and fiddle with layout.  It doesn't sample any lcd lines or do anything useful  



Note that it shows "Ohms"... I was going to sample the function selection lines as well so I could show Volts(ac/dc)/Amps(ac/dc)/Ohms/etc (a stock 8060 does not show this).  In the end, as I alluded to above, I don't think trying to sample and translate lcd drive lines is going to work all that well.  But it would look friggin cool.

[JdaleH]

The thin LCDs I found were by Kingbright.  However, they had NO decimal point.  Bummer.  Now, Kingbright has some lcds that are reverse mounted within holes in the PCB, just what the doctor ordered for your concept.  I was just looking at homebrew coke can aluminum stencils that could be used for the top layer.  I am concerned about lcd heat on anything else.  https://www.instructables.com/id/Easy-Aluminum-Solder-Mask/  Make aluminum stencils and paint/anodize the aluminum black.  I have cleaned up one of my existing displays and will copy the clear circuit layout to see where each pin goes.  Have to hold the light just right and the clear circuits will reflect.  My thoughts are to reuse the zebra pad and elastomeric piece, so there will be no additional wiring.  The pcb will just replace the lcd.  Pretty straightforward except I will probably have to convert from ExpressPCB to KICAD to enable the hole cuts.  Plan B is that I am going to repair the original display by replacing the back polarized sheet and mylar backing.  I have ordered the materials to give that a try.  I think that, as long as the liquid crystals haven't leaked out, I should be able to restore it.  Won't know 'till I try.  My first career in my youth was commercial art, so I am very comfortable with an xacto knife.  Dale

[ogdento]

Hmm, I'll have to take another look at the 7-segment ones... I just checked the Kingbright website and I'm pretty sure that was where I was looking, but it was a while ago so maybe they've gotten thinner.

Good point about redoing the polarizer and mylar backing... I did carve open one of my bad displays after seeing a post by modemhead and it worked pretty well.  I didn't have any mylar on hand so I just put some white paper behind the glass... it's not great but it works.  Post some pics when you get yours fixed up!

And I'm pretty sure I took a photo of one of these displays that showed all the circuit lines on the glass... I'll post it when I find it.

[ogdento]

super ghetto white paper backing on the lcd...

[ogdento]

Dale,
I found a sketch that I worked up after looking at some scans Dmitri sent me, so I entered it into Kicad and attached it here.  While I was entering the schematic I remembered that we can't use 7-segment displays that have a common anode or cathode, since the 8060's multiplexing scheme won't work with those.

The multiplexing scheme is pretty straightforward - this is taken off the schematic in the manual:

pin  H1    H2
---------------
1    H1
2    cont  bar
3    rel   audible
4    A0    B0
5    G0    C0
6    E0    D0
7    F0    DP0
8    A1    B1
9    G1    C1
10   E1    D1
11   F1    DP1
12   A2    B2
13   G2    C2
14   E2    D2
15   F2    DP2
16   A3    B3
17   G3    C3
18   E3    D3
19   F3    DP3
20   1/2   1/2
21   BT    NEG
22   K     HZ
23         H2


Dmitri used more resistors than I did (he had one per anode whereas I have one per shared anode - this might not work!)... so his "pin" 3 connection (for the rel and audible indicators) had the LED cathodes going to H1 and H2, a resistor on each LED anode, and then connected to pin 3.

Here's my schematic (I have not tested any of it yet):

[IconicPCB]

There is a problem with Your schematic.
The LEDs are current driven devices, LCDs are voltage driven devices.

Consider the junction of four LEds in the right most digit where top and middle LEDS join to two other LEDS along the right side of the digit. There is no return path for the current.

Your adaption has to accept the individual segment drive voltages  and EXOR them with the back plane drive signal to decode the LED current drive.
You can then use either common cathode or common anode depending on whether the current is sourced or sunk.

[ogdento]

Hey Iconic, thanks for noticing that... the junction for A/G/E/F should have been connected to H1 (at the line that comes off of the REL, CONT, K), but I goofed - and then repeated the mistake on all digits!



I haven't tested this or even checked any of the signals yet - I need to put a meter on a scope and look at H1/H2 and the pins.  In the docs from Dmitri the H1/H2 signal was a 3v peak-to-peak square wave (from 2-5 volts) at 156Hz, assuming I translated it correctly.  I assumed H1/H2 were 180 out of phase so that when H1 was high, H2 would be low... so if pin4 was high for 1 H1/H2 cycle and H1 started high, H2 would be low - sinking current and illuminating B0.  Then when H2 went high, H1 would go low - sinking current and illuminating A0.  Does that sound even remotely plausible?

[ogdento]

Before I go any further I really need to check the signals at the meter's display connector.

I added a shot to this thread of what I soldered up last year (thanks to a donor card from vindoline), but it didn't work... I'm pretty sure it worked with another micro board without a display so I'm not sure if I broke the board vindoline gave me.  The SM-4 is static sensitive and I do use a wrist strap... could I have killed it with my soldering iron?  It's an un-grounded weller 25-watt job from the early 90s, and I've read about zapping stuff with un-grounded irons but I'm not sure how often it happens in practice (and i've soldered hundreds of static sensitive devices with it over the years).

[IconicPCB]

The way the LCD drive works is typically as folows.

Firstof all DC may be used but it is not ikely to be benevlent to the LCD dispaly.

Usuall approach is a bi phase drive ( as you have noted ).
The backplane is driven with one phase while the active segment is driven 180 degrees out of phase such that the liquid crystals are driven by an alterating potential between the backplane and the segment electrode.

Inactive segment electrode is driven by the backplane signal. This voltage between the segment and the backplane is decoded with an EXCLUSIVE OR  gatewhose output is LOW for inactive condition and HIGH for segment on condition. From here on the segment activity is processed to drive the LED display.
So in a nutshell six quad EXOR chips to decode LCD signals followed by an LED display drive(r).

[JdaleH]

I'm confused.  How did Dmetri do this without the extra chips? http://mrmodemhead.com/blog/a-very-unique-fluke-8060a/  I think I will concentrate on completing the template and leave it to others to do the pcb.  I have attached a draft of the display overlay done in Coreldraw.  This is the original as best I could reproduce it.  Everything will need to be altered to allow for laser or knife cutting.  I'll finish it up and post it as a cut file for future use.  I hope someone will pick up the baton and do a pcb.  Regards, Dale

[ogdento]

Hey Dale/Iconic,

Now I get what you're saying about the XORs   but I don't think we need them yet... I'm more or less copying what has already been done by Dmitri, and he used only LEDs and resistors.  But he selected extremely low current LEDs so they could be driven by the SM-4 without separate drivers.

With that said, if this works out and there's ample room on the board it might make sense to do a rev 2 that has more robust drive circuitry - but adding anything like that is going to increase current draw and might affect the meter operation.  I'd also be curious if this driver-less design hastens the expiration of the SM-4 chip in any way - if it draws the same current then I think it might be safe (famous last words!)

My original goal in all of this was to come up with something that is simple to produce... Dmitri's display was quite labor intensive as he glued/wired each individual led and resistor and custom milled his own face plate.

[ogdento]

Hey Dale that looks great! 

One thing I'll point out is if we're using LEDs I don't think we can have the same long continuity bar... we'd probably need 3 or 4 leds to make it that big, which relates to Iconic's point that we really don't have proper drive circuitry here.

It looks like Dmitri only used 1 led to show the continuity bar (i think it's the blue one between the continuity and audible indicators)

[JdaleH]

Yes, that was one thing I planned to change.  This first step was to get everything in position on the replacement pcb.  Also, the LED segments need to be made smaller.  When finished, it will be very similar to Dmetri's. 

Given the previous conversations, we need to find some very low current LED like Dmetry used.  Dale

[IconicPCB]

I think any green LEd will do. Note the quoted current is average value.
The eye will repond to low average value but reasonably high peak value.

In fact as long as the back plane pin can sink/ source milliamps of current then direct drive of LED is possiblewhen segment drive and backplane drive signals are not in phase and the current is limited to a low but usable value.The LED reverse voltage needs to be observed so it does not failunder reverse bias conditions.

[ogdento]

Thanks Iconic - this is good information!

I was finally able to find a datasheet for the SM-4a... it was in a 1990s sharp microcomputer book.  I'll give it a look tonight.

[ogdento]

Ooops, I forgot to attach the datasheet.  This is for the SM-4a... I'm not sure if they're the same, but it does show the voltage/current ratings for the segment and backplane pins

[JdaleH]

Here is my draft cut file for the display overlay.  I don't have a cutter, so I can't see if it is correct.  Maybe someone who has a metal or vinyl cutter can give it a go to see if I have everything right.  You can look at the pdf file.  The eps file is zipped.  Could I get a .sch file for your schematic?  Dale