NAND only freq. counter

[GK]

I'm vegetating a bit here at my electronics bench ATM as I wait for parts on back-order and project funds are on hold for home maint. stuff. This is what I knocked up yesterday afternoon as a diversion from my major projects.

I'm going to need a frequency counter to monitor the synth for my 630m TX in my exciter rack. I have a bunch of these old ZM1050 neon display/indicator tubes, which I figured would make for a nice visual readout. These rare indicator tubes were originally designed as a nixie substitute as when nixie tubes first appeared there was a dearth of economical H.V. switching transistors available to drive them. The ZM1050 has a trigger electrode for each neon light and can be driven from logic-level signals as low as 5V - no expensive HV transistors required. Unfortunately for Phillips/Mullard, shortly after they put the ZM1050 to market affordable H.V. transistors arrived and their unique little neon indicator device went instantly obsolete.

So I wired up a B.C.D. counter and a decimal decoder to test the tubes out, but with a twist. A uC would be the most practical way to implement a freq. counter, but that is pretty BORING. So I did it entirely in  NAND gates - 74HC00 to be precise. It's pretty basic. There are just four pure-edge-triggered slave/master DFF's configured as a ripple counter, that resets to 0 on count 10(bin). The B.C.D. to decimal decoder is of the weighted "pyramid" configuration, which is the most economical on NAND gates that I am aware of.

Note that a counter and display decoder for a freq. counter would require a latch between the B.C.D. counter and the decoder. I didn't prototype the latches as that would have required an additional four 74HC00's and I only had one.

Unfortunately most of my ZM1050 are stuffed. They were originally salvaged from a seismic chronograph and ran for years 24/7. Most of them no longer trigger properly. So to implement my freq. counter I have decided to use jumbo 7-segment LED displays instead, which means that now I am going to have to embark on the comatose task of wiring a B.C.D. to 7-segment decoder, which unfortunately will take a fair few additional NAND gates.

I intend to build the counter 100% entirely with NAND gates. Even the high-gain signal input amplifier and be implemented with NAND gates forced to operate in analogue mode. This will be a 6-digit counter with a fixed gate period of 1s (the 630m band is 500 kHz). After this is spun off I'll do an other counter for my H.F. band(s) exciter entirely in NOR gates.

[GK]

...which means that now I am going to have to embark on the comatose task of wiring a B.C.D. to 7-segment decoder, which unfortunately will take a fair few additional NAND gates.

Done. Now I just need to implement leading digit supression I/O. LED-segment drive capacity will be implemented by putting gates in parallel if necessary.

[GK]

OK, I've completed the logic design for a complete single digit counter card and have it 95% drafted. These cards are just cascaded for how many number of digits the counter needs to have. There will be a second circuit board for the counters clock/timebase, signal amplifier input and control.

I managed to fit it into 33 74HC00 packages. Two gates are unused for 130 NAND gates in total. The 7-segement outputs are active low, for common anode displays. This is just a rough view of the complete circuit logic diagram, I still have to finish drafting it and then thoroughly check it over before starting the PCB layout.
 


EDIT: there is a wiring error in the diagram above for the display decoding of the "e" segment.

[Thor-Arne]

Nice 

Hope you post a image the pcb 

[Vgkid]

That will be impressive when finished.

[GK]

Hope you post a image the pcb 

I stuffed it onto a super home-brew friendly single-sided layout with a 100 thou grid and 50 thou tracks. The etch tank is heating up. It's a pretty big board though. This counter is going to require a 4U rack case.

[Thor-Arne]

Impressive that you got to fit all that onto a single sided board. 

[GK]

There may be a wire link or two 

[rob77]

There may be a wire link or two 

1 or 2 means 12 or 21 ? btw... really nice project 

[Thor-Arne]

I expected some links, still impressive.
Sometimes I struggle to get counter circuits to fit on a two layer board, even with wire links.

[ConKbot]

Nice work, lots of Karnaugh maps and boolean math? Jeeze its been a while since I could say Ive done any work like that.

[GK]

Nice work, lots of Karnaugh maps and boolean math?

Just a little of the latter.

[GK]

This evening I got the 14 IC's of the B.C.D. counter loaded. The scope shot shows the counter B.C.D. outputs cycling 0 through 9.

[Alex Eisenhut]

Nice scopes. What's with that gap in the 545's side panel?

[GK]

Nice scopes. What's with that gap in the 545's side panel?

The side panel is just be***snip***...follows: 




EDIT:
Oops, just killed this post and erased almost all of the content. Wanted to quote a section of it but clicked "modify" instead.


   

[phenol]

How about using linearized 74HCU04 in the frontend?

a year ago i too made an all discrete CMOS logic freq 8-digit counter, but it wasn't just NANDs that I used. the frontend consists of a broadband discrete transistors amp feeding a 74AC132 gate. it's all ac coupled and the lower freq limit is in the low 100s of kHz...but it's sensitivity is <8mV up to 210MHz. This is the point where the 74AHC74 divider after the '132 squarer craps out.

[Alex Eisenhut]

Nice scopes. What's with that gap in the 545's side panel?

The side panel is just bent slightly. I'll flatten it out one day.

I was hoping you'd know what this is for. I don't see it in any pictures of other scopes, and it's not on the mechanical diagram.
It looks like something hooks in there.

[GK]

How about using linearized 74HCU04 in the frontend?

Well, I just have this weird desire to do it all in NAND gates.

[GK]

Nice scopes. What's with that gap in the 545's side panel?

The side panel is just bent slightly. I'll flatten it out one day.

I was hoping you'd know what this is for. I don't see it in any pictures of other scopes, and it's not on the mechanical diagram.
It looks like something hooks in there.

Dunno. My 545 panel doesn't have any holes in it where the bend is and none of my other 5XX series scopes have one of those.

[GK]

I love it when I solder all of the bits to the board, wire it up and it just works 

[zapta]

Well, I just have this weird desire to do it all in NAND gates.

Real men do it with AND gates.

BTW, I like your PCB design, it has an artistic beauty.

[harrimansat]

You can use pnp germanium transistors...

[GK]

Well, I just have this weird desire to do it all in NAND gates.

Real men do it with AND gates.

That would prove challenging.

BTW, I like your PCB design, it has an artistic beauty.

Thanks. In the rush to get the layout to PCB I missed putting a radius on a number of the 90 degree bends. That's since been fixed.

[GK]

You can use pnp germanium transistors...

Nice old design!

[GK]

Here is the completed schematic for the counter board., now that it has been built and tested. I am currently laying out the PCB for the timebase.