Vacuum Fluorescent Display Driver

[bostonman]

Oh okay, I should have realized it. I was focused on thinking both N and FN were DIP.

My attention was: new IC model 518 is the PLCC and old model is DIP; and both DIPs came in a N or FN option.

[bostonman]

I bought six of these for $36


https://www.ebay.com/itm/325686864607?epid=1718442701&itmmeta=01HZR9DBRFN0HEZH2TY8GH04JX&hash=item4bd4732adf:g:ISgAAOSwrwhkgmQ~&itmprp=enc%3AAQAJAAAA4FfSigCHd%2BTuVaEiHXE3KyQWntr7hqnMIKSg%2FwznHO12HJJR0Qn7tAxoMFVzqWnRDp%2FkTSjs2fWfhEITunRwa5YDvekl0h8tAXzDDJqU024tCBm7YPYL4R9al7UGSn00R4OeBPXPjtPor%2BmgsMXdAHkELnkU5S9CKWcMkM26PSh80hHtk690RWQXfBgqY6y4EmdeYyt4b%2BFTtL2guDFE7GNfduuJOj7KJhENRB3%2F5Mz4MOR6QFsB8YPlQU%2BbdK5%2FQNRNP8yogNicZDDQ7WkVS%2BB91v51X1pWePSqbKyd0iPI%7Ctkp%3ABk9SR6a8tYn-Yw

[bostonman]

I'm still trying to make time for a daughter board, so this is why the thread hasn't been updated.

[bostonman]

I've completed the schematic (it's nothing spectacular) and the PCB (see attached).

Rather than draw all the connections on the schematic, I used ports. Also, I may have wasted time creating a PCB with both ICs since ideally the surface mount replacement one can't be used since it keeps blowing, however, it opens the option for using the correct one should a need arise.

The schematic has some issues. The IC pins aren't labeled input or outputs, and all the ports are facing one way. The PCB shows pin 1 on the VFD being at the wrong end.

Unfortunately it took me a bit to re-familiarize myself with PCB layout along with creating components. Once I was knee deep, I just kept the errors as is.

Earlier I ordered the PCBs (I'm waiting to see if the company finds any issues).

[bostonman]

I wired my daughter board and installed the SN75518N IC.

From what I can tell, the display works fine (it was a bench test and a functional). Also, I re-powered the unit four or five times without the IC blowing unlike the HV518.

My board is designed to take either a PDIP or PLCC. In a previous message I stated if the PDIP SN75518 doesn't blow after re-powering, I'll try the HV518 (again) to confirm it isn't a compatible device to use.

I'm reluctant to push things by removing the 75518 and installing the HV518, but, if others are interested in me confirming the HV518 shouldn't be used as a replacement, I'll gladly perform the experiment.

I'll upload a few pictures in a bit, but wanted to provide an update on things.

Update: I injected 1.234567890MHz into the counter.

[pqass]

That looks very nice. 
I'm glad it worked out in the end.

[bostonman]

Thanks.

The mechanical error is that I wanted the spot for the PLCC a bit more to the right so it sits between the two holes on the front panel board. Also, it really didn't need to be that big but wanted the ability to have both IC packages to be available.

Ironically, the header pins fit in the VFD holes perfectly, so I got those dimensions correct, but they were getting soldered to the VFD pins once I soldered it back onto the board, so worst case I could bend them in place.

The signal wires are all 24AWG and I used 20AWG for the 5v, 38v, and GND.

This week I plan to check the power switch (it was sticking and maybe it needs to be replaced) and reflow the solder on the DC/DC since I had removed that at one point.

After all that, I'll reassemble it and hope my board doesn't hit any obstacles. If it all fits, I'll upload the gerbers so others can have them if need be.

[bostonman]

I just noticed the top part of the '3' is diagonal in both digits. All the other digits have the correct segments, so I assume this is part of a fancier method of showing '3', but wanted to confirm this isn't an error.

[pqass]

Apparently, you're the lucky owner of a ROM with the fancier font. 
But it looks like they reverted back.

[bostonman]

Phew, my heart skipped a beat when I saw the diagonal for the '3'.

After more thought, and in the name of science, I'll install my last HV518 and see if it blows. This way I can confirm the socket is connected correctly and verify the chip blows up.

One odd thing I don't understand about this counter: I injected 1.2345...MHz and it decided to display 1,234,567.xxx Hz. Does a reason exist it displays it as Hz rather than place the decimal place in the beginning and display MHz?

[bostonman]

I'm very confused as the last brand new HV518 didn't blow; and will provide a summary of the steps that have occurred so far.

* Original issue: segments in the 4th and 5th (?) digits along with one segment in the 12th display were incorrect

* Un-soldered the original PLCC using hot air and cleaned the excess solder off the pads

* Installed a brand new HV518 using a hand soldering iron

* Powered unit, display worked (I didn't do a thorough test such as testing all the digits, just a basic test to see if it turned on), turned off power, turned on power, the display was blank, the DC/DC was getting extremely hot, and the +38V was (if I remember correctly) +13V

* Un-soldered the newly installed HV518 using hot air and cleaned the excess solder off the pads

* Installed a brand new HV518 using a hand soldering iron

* Powered unit, display worked (I didn't do a thorough test such as testing all the digits, just a basic test to see if it turned on), turned off power, turned on power, the display was blank, the DC/DC was getting extremely hot, and the +38V was (if I remember correctly) +13V

* Per the suggestion of users in this thread, performed several bench tests on the VFD (I will not list the tests as it will just be too long). All tests showed the VFD was good.

* Designed a daughter board to accommodate SN75518 PDIP , soldered it to the back of the front panel board, installed a PDIP SN75518 into the socket, powered unit, the unit worked great, re-powered several times, and it worked every time

* Removed the SN75518 from the PDIP socket, installed my last (brand new) HV518 into the PLCC socket, powered unit, and the display worked. As before, I expected the display to not work once I recycled power. I turned off power for about 30s, turned on, display worked (normally it wouldn't work). Recycled power several times either very quickly, sometimes left power off for a bit, etc.. and the display worked every time. I threw basically everything I could with recycling power and the display worked fine every time.

The only thing this daughter board has that isn't on the front panel board is an additional 0.1uF capacitor on the 5V line (the schematic I uploaded previously shows a 0.01uf). Besides this capacitor, the board is bare bones with a PDIP and PLCC socket.

I don't know why the PDIP isn't blowing now. My thoughts are that heat from the soldering iron when I was replacing it on the front panel board caused damage (but odd two chips would fail the same way) to the HV518 and why it kept blowing, two out of three brand new HV518s were faulty and I so happened to install the two bad ones, or the VFD had an issue that got resolved during bench testing. At the moment I don't have an answer as to what happened except it actually looks like an HV518 is a good replacement.

On a side note: the only odd issue I experienced after installing the front panel PCB back into the plastic face was the buttons and LEDs weren't correct. i.e. I'd push 50ohms on channel 1 and the display would show menu options for channel 1 instead. I put a little pressure on the ribbon cable and that didn't seem to solve anything, I recycled power, re-seated the connector on the main PCB, etc... and then things seemed to start working normal. A little concerning as I hope it's not a break in the ribbon cable due to all the moving around it's been subjected to these last several weeks. If the ribbon cable is bad, that's a whole ordeal as I'll need to un-solder the daughter board I just installed.

In any case, at the moment, the daughter board appears to fit fine, the display seems correct, holding my breath on the front panel buttons continuing to work correctly, I just ordered a replacement fan (the current one doesn't always start spinning at turn on), and will assemble the unit completely upon installing the new fan. After I'll calibrate the unit and, if the HV518 ever blows, replacing it with the PDIP will be fairly easy.

Update: I uploaded a photo of the daughter board (and PDIP) installed in the front panel. Note: I accidentally added this picture to the previous message by accident, but deleted it.

[BrianHG]

I'm very confused as the last brand new HV518 didn't blow; and will provide a summary of the steps that have occurred so far.

* Original issue: segments in the 4th and 5th (?) digits along with one segment in the 12th display were incorrect

* Un-soldered the original PLCC using hot air and cleaned the excess solder off the pads

* Installed a brand new HV518 using a hand soldering iron

* Powered unit, display worked (I didn't do a thorough test such as testing all the digits, just a basic test to see if it turned on), turned off power, turned on power, the display was blank, the DC/DC was getting extremely hot, and the +38V was (if I remember correctly) +13V

* Un-soldered the newly installed HV518 using hot air and cleaned the excess solder off the pads

* Installed a brand new HV518 using a hand soldering iron

* Powered unit, display worked (I didn't do a thorough test such as testing all the digits, just a basic test to see if it turned on), turned off power, turned on power, the display was blank, the DC/DC was getting extremely hot, and the +38V was (if I remember correctly) +13V

* Per the suggestion of users in this thread, performed several bench tests on the VFD (I will not list the tests as it will just be too long). All tests showed the VFD was good.

* Designed a daughter board to accommodate SN75518 PDIP , soldered it to the back of the front panel board, installed a PDIP SN75518 into the socket, powered unit, the unit worked great, re-powered several times, and it worked every time

* Removed the SN75518 from the PDIP socket, installed my last (brand new) HV518 into the PLCC socket, powered unit, and the display worked. As before, I expected the display to not work once I recycled power. I turned off power for about 30s, turned on, display worked (normally it wouldn't work). Recycled power several times either very quickly, sometimes left power off for a bit, etc.. and the display worked every time. I threw basically everything I could with recycling power and the display worked fine every time.

The only thing this daughter board has that isn't on the front panel board is an additional 0.1uF capacitor on the 5V line (the schematic I uploaded previously shows a 0.01uf). Besides this capacitor, the board is bare bones with a PDIP and PLCC socket.

I don't know why the PDIP isn't blowing now. My thoughts are that heat from the soldering iron when I was replacing it on the front panel board caused damage (but odd two chips would fail the same way) to the HV518 and why it kept blowing, two out of three brand new HV518s were faulty and I so happened to install the two bad ones, or the VFD had an issue that got resolved during bench testing. At the moment I don't have an answer as to what happened except it actually looks like an HV518 is a good replacement.

On a side note: the only odd issue I experienced after installing the front panel PCB back into the plastic face was the buttons and LEDs weren't correct. i.e. I'd push 50ohms on channel 1 and the display would show menu options for channel 1 instead. I put a little pressure on the ribbon cable and that didn't seem to solve anything, I recycled power, re-seated the connector on the main PCB, etc... and then things seemed to start working normal. A little concerning as I hope it's not a break in the ribbon cable due to all the moving around it's been subjected to these last several weeks. If the ribbon cable is bad, that's a whole ordeal as I'll need to un-solder the daughter board I just installed.

In any case, at the moment, the daughter board appears to fit fine, the display seems correct, holding my breath on the front panel buttons continuing to work correctly, I just ordered a replacement fan (the current one doesn't always start spinning at turn on), and will assemble the unit completely upon installing the new fan. After I'll calibrate the unit and, if the HV518 ever blows, replacing it with the PDIP will be fairly easy.

Update: I uploaded a photo of the daughter board (and PDIP) installed in the front panel. Note: I accidentally added this picture to the previous message by accident, but deleted it.

Warning, the HV518 has a specific required power-up and power-down sequence order.  If you do not follow it, you may blow up the IC or draw absurd current on the Vpp with excessive heat building up.  You will need to read the data sheet, and, if you haven't already made accommodations to follow the instructions laid out, you might not be able to reliably use the HV518.

If you do not follow the documented sequence and timing, the functionality and/or death will be random.

[bostonman]

Thanks for the reply.

This isn't a new design, this chip blowing has been an ongoing issue if you read through the thread. The discussion of power on/off sequence has been mentioned also.

Now suddenly it's no longer blowing even though it's in the same circuit.

[bostonman]

I've done several re-power tests and the unit works fine every time. It remains a mystery why this IC failed when I soldered it onto the front panel PCB, but works fine now in a socket (keep in mind the display worked originally, just had some bad segments).

From what I can tell, the unfortunate part is that the ribbon cable is flaky because the front panel buttons don't work correctly, and, if I wiggle the cable, the display kicks out. When the front panel doesn't work correctly, pushing buttons reveals a completely different function. An example would be 50ohm/high impedance goes through the menu for (I believe) selecting which channels to measure frequency (which is the button all the way to the left). Maybe the IC for this is failing too, but I haven't been able to manually duplicate the problem. It seems the problem occurs sporadically and then vanishes. Maybe it's when things are cold, I'll have to try it tomorrow after it's had all night to cool.

Unfortunately I'd have to remove my daughter board along with trying to remove the front panel PCB from the face plate which is a pain if I want to replace the ribbon cable. At this point it's too risky to dismantle again, it's hard for me to leave it as a potential failure, but may need to bite the bullet and leave it.

Maybe if time permits in the future I'll tackle replacing the ribbon cable.

I've pieced together a package with the gerbers to make the daughter board along with dimensions of my board. I'm uncertain if this is the best thread to post it as the title doesn't direct anyone to find a replacement daughter board, so should I start a new thread; maybe in the 'repair' section? If anyone has a github account, maybe they can upload it for me.

[bostonman]

Guess I couldn't leave well enough alone.

Today I tried removing the ribbon cable and thought I had enough room to lift the ribbon cable end under the VFD. After realizing the connector was hitting the VFD, I stopped.

Long story short, I reassembled everything and it seems the work I did on the ribbon cable (I re-soldered all the pins and hoped the solder ran up into the internal area where the wires are pressed in and/or it burned off any film that was causing intermittent open connections).

Unfortunately I saw segment C wasn't lighting and thought it was just a bad solder connection since I had to remove my "daughter" board. I measured the segment C pin on the VFD and it was clocking. After I realized I also lost segment M and N (and possibly P and Q which are the M and Hz - but I haven't confirmed this).

I looked under the microscope and it looks like the glass is cracked on the VFD backside (right around where I was pushing up on the ribbon cable connector end), however, the display lights and all the digits work (except missing segments).

Is it possible I cracked the glass but didn't break the vacuum since the display still works? At this point I think the VFD is garbage since the issue appears internal.

[bostonman]

I removed everything again and looked at the VFD under the microscope. Unfortunately the crack was real as a slight flick with the tweezers caused the section to break away.

It looks like the crack happened at a solid point on the glass and didn't make it to the vacuum section.

At this point I have a decision to make: give up and not have a frequency counter or buy a used VFD off eBay for $100

By any chance, has anyone ever made a plug in replacement display such as a LCD? I'm wondering if anything exists that would accept all the inputs and can be programmed to display specific segments based on the inputs.

[PCB.Wiz]

By any chance, has anyone ever made a plug in replacement display such as a LCD? I'm wondering if anything exists that would accept all the inputs and can be programmed to display specific segments based on the inputs.

Can you get access to all the signals ?  Maybe a Pi PICO could sniff the lines, if they are (or can become) 3v3 levels ?

Years ago, we did a CPLD sniffer design that looked at the display interface link on a custom printer, and gave the customer remote information.

[pqass]

No need to sniff, the protocol has been documented and an alternate display solution has been implemented.

It's too nice of a counter to just abandon at this stage. 

Your choice is an ebay/aliexpress VFD replacement (likely slow, sketchy yet easy route to take)
or an OLED/other display to fit into the same spot with the protocol converter described in the link above.

[bostonman]

That alternate display looks cool, but, I have to admit, the VFD does fit the design well.

Most likely I'm not willing to pull the trigger on altering the design to swap display types, so most likely I'll go with one from eBay or Aliexpress.

I'm mad at myself for prying up on that ribbon cable connector. I was looking at the metal frame around the VFD and thought the distance was enough to lift the pins off the PCB and maybe solder the ribbon cable directly. I didn't give though to how far the VFD sits.

Someone on eBay has the VFDs listed as new, but the picture shows a display with different stuff on the front. I've sent a picture of mine and waiting for a reply. Someone else on eBay has "used" ones; both are similar in price.

I haven't looked on Aliexpress yet, but will.

[bostonman]

Due to being stubborn and wanting to see this project through, I purchased a replacement VFD.

The ones on Aliexpress had the getter (?) on the bottom and the pins were bent in an L shape, so I was reluctant to purchase them. A seller on eBay had one listed as "new", but the picture showed different stuff on the screen. After sending a picture of mine per their request, I got a reply "Hey Dear" can you send the part number.

Due to such an odd reply, I purchased one listed as used. It's an entire front panel PCB, so it's plug n' play, however, the driver IC may fail in the future, so I'm thinking of removing the VFD.

It's expected to arrive the end of the month.

Meanwhile, I'll investigate the alternate display which looks pretty cool. If I crack this replacement VFD, it doesn't work, or fails in the future, I'll resort to the alternate display.

Will this alternate display work while the VFD is also connected and working? I'm wondering if I can have dual display. Seems like it would work since it just needs the signal lines.

[PCB.Wiz]

.... After sending a picture of mine per their request, I got a reply "Hey Dear" can you send the part number.
Due to such an odd reply, I purchased one listed as used.

That's a standard generic Chinese Vendor greeting 

[pqass]

Will this alternate display work while the VFD is also connected and working? I'm wondering if I can have dual display. Seems like it would work since it just needs the signal lines.

Since it just tees off the SPI lines, it should work concurrent with the VFD (just don't connect VFDSIN as that is already driven by the exiting SN75518).

The project linked above aims to interpret the segment content (ie. turn segment combinations into ASCII chars) but you can probably just make it dumb (no MCU); just use 5 x 74HC595 + 18 x 330R + 2 x ULN2003 ('595s are good for 4mA per segment; or just use another 75518 which is good for 25mA but can't drive the common cathodes directly) and drive a 12 digit x 14-seg LED display.  Maybe start with a 3 x quad display like a common cathode version of this one with a (flex?) PCB overlay of SMD LEDS between digits (for the dots and comma), and a back-lit mask for the annunciator.

See this idea and this one too.  You might even cram it into the same space as the VFD.

[bostonman]

(just don't connect VFDSIN as that is already driven by the exiting SN75518)

I dislike the node naming on the schematics. VFDSIN implies it's an input to the VFD driver, but it's actually an output.

The links you provided are nice, but remember, this display also has other names displayed too such as the channel number, rise, MHz, etc...

The alternate display solution with the Adriuno seems the most practical. Let me figure out whether I'll get this unit working with the replacement VFD. If so, I'll keep this alternate design on the side. If not, then it's off to use the alternate design (or maybe implement it and keep both should the VFD fail in the future).

[bostonman]

Looks like I finally got this fixed; and it fought me the whole way.

The replacement VFD arrived, it was sold as an entire front panel PCB and could have been a "drop in" replacement quick fix. The display worked, but I feared the IC would fail over time. The replacement front panel PCB was in good shape and the ribbon cable (I assume) hadn't been flexed too much.

The route I decided to take was: remove the VFD, remove the IC, solder jumper wires from the connector so I can connect my daughter board, reinstall the VFD, solder my daughter board onto the front panel PCB, solder the jumper wires to my daughter board.

At this point the display worked except one G line wasn't soldered, so this had to be fixed. After I inserted the PCB into the front and noticed two LEDs weren't working.

Being stubborn, I had to fix this, so I removed the PCB again front the front panel, removed two LEDs from my old PCB, soldered them onto the replacement front panel board, and then it worked.

At least now in the future if the driver IC fails, I can just swap without any soldering (assuming the other IC that controls the LEDs doesn't fail too).

Now it's fully assembled and I just need to adjust the calibration screw on the rear.

The only odd thing the counter seems to do is not utilize the first digit or few few digits all the time. I'm sure a reason exists, but seems all twelve digits should be displayed.

[bostonman]

I somewhat began creating an external display based off the Github link provided - just to have an alternate display if I ever need it.

Although I haven't purchased the display, I know the parts to order, however, I'll begin installing a connector (already arrived) on the rear panel so I can connect/disconnect the external display.

It's suggested to put resistors in series with the data lines (and maybe 5V too) to prevent shorting. Besides resistors, should I put TVs on all the data lines to ground to prevent ESD issues?