Z80 LQFP44 pinout

[luiHS]

 
Hello.

Can someone confirm if the pinout of the Z80 LQFP44 on the data sheet is correct?
I have replaced a Z80 PLCC in a design with a Z80 LQFP, and it does not work. The rest of the scheme is identical, only the Z80 was replaced.

The original that works is a Z84C0006VEG in PLCC socket, and the new one is a Z84C0010AEG in LQFP44.
I attach screenshots of the pinout of the data sheet, and both in my Eagle schemas.

[Andy Watson]

I cannot directly confirm the equivalence, however, I did use the very same datasheet to reverse engineer these diagrams:
https://www.eevblog.com/forum/repair/ap-series-power-supply-(ap60-50)-(wayne-kerr-farnell-hitek-aps)/msg1733669/#msg1733669
All I can say is that all the pins on the LQFP44 package were in accoradance with where I expected them to be.

[wilfred]

I couldn't see you made an error. Why don't you draw *WAIT next to *BUSACK and *BUSREQ to avoid having a wire run into the chip symbol?

[obiwanjacobi]

Are you sure the pin (number) mapping in the footprint is correct?
The NC's are on different pins/pads between these packages...

[2c]

[luiHS]

The footprint I made for Eagle is correct. Today Zilog answered me by email, they sent me the same document I already had, a PDF with the "Product update, errata for Z8400 / Z84C00 NMOS / CMOS Z80, CPU devices".

I do not know what the problem is, I checked the scheme, the footprint, I replaced the Z80 with a new chip in case it was damaged. Now I'm going to see if I replace the RAM chip, a 6264, in case I had burned it, but I would be very surprised because it is all solder in the oven and I have never burned anything.

The eprom 27c010 is fine, because I put it on the other board with the Z80 in PLCC and it works perfectly. I have also measured continuity in the socket PLCC32 of 27c010, and everything seems to be fine.

It seems like a design error or the RAM is damaged, I can not think of anything else. But checking the footprint and the schema it seems to me that everything is fine. If Zilog confirms that the pinout is correct, I run out of ideas.

 

[luiHS]

 
I replaced the RAM (6264), but everything remains the same. I also resold the welds of all the chips and measured continuity in all the signals, everything is fine, but it does not work. I even raised the supply voltage, since my board has a schottky diode in the positive that caused the voltage to drop somewhat. Adjust my source, to measure 5 volts in the supply voltage of the Z80.

I also checked that there are no short circuits between pins, everything is fine.
I can not think of anything else, the only thing left for me to think is that the pinout of the datasheet is wrong, it is not what they indicate.

Another thing I thought, but I would be surprised if it was that. In the circuitry of my board is used 74HC chips, this works perfectly with the Z80 in PLCC, I do not think it is different with the Z80 in LQFP and it needs 74LS or 74HCT, for TTL compatibility.

I'm about to leave the circuit as was before, with the Z80 in PLCC mounted on a socket. It's a shame because in LQFP it's much smaller, it does not need a socket, and my Pick and Place Neoden4 machine can mount it.

What bothers me is not knowing why it does not work, just for the fact of changing the format of the chip, when the rest of the circuit is the same. Intuition tells me that the pinout is wrong, but I do not know how I could prove it.

 

[wraper]

I would try bodgwire CPU in PLCC and see if it works.

[luiHS]

I would try bodgwire CPU in PLCC and see if it works.

You mean to wire by hand, pin to pin the LQFP Z80 on the PLCC board?

Something like that I also thought, although it is a very laborious job, they are 44 pins, less some that are not used. Soldering 40 wires by hand, from an LQFP44 chip to the pads of a PLCC in the PCB, I see it too complicated.

Maybe making a small PCB with the LQFP soldered and from there to the main board, but I do not know how to connect those two boards, so that it's simple.

If this worked, it would indicate an error in the LQFP main board, but I would be very surprised, because the scheme is the same, copied from one board to do the other, simply replacing the Z80 PLCC with an LQFP.

I could make an adapter with the LQFP44, and to weld it to the PLCC pads, make the adapter with castellated mounthing holes. I think that is what I'm going to do, but I suspect that it will not work and that will confirm that it is an erroneous pinout.

[wraper]

You mean to wire by hand, pin to pin the LQFP Z80 on the PLCC board?

I mean PLCC on LQPF board. Though the way as you wrote might be useful too. I don't think it would take more than an hour to me.

[luiHS]

You mean to wire by hand, pin to pin the LQFP Z80 on the PLCC board?

I mean PLCC on LQPF board. Though the way as you wrote might be useful too. I don't think it would take more than an hour to me.

Ok, I understand what you want to say. Although solder 40 wires, one by one by hand, it seems quite hard.

I have design these two small adapter boards, to connect the LQFP Z80 to the PLCC board, using THT 1.27mm connectors to solder to the PLCC SMD pads. If my main board with the Z80 in LQFP is badly designed, and that is really the problem, with this small adapter the LQFP must work on the board that now has the Z80 PLCC installed (this one works).

One adapter board is according with datasheet, the number 1, and the number 2 is pin to pin.

[obiwanjacobi]

You could use the typical Z80 test setup (*) to reverse engineer the pinout...

*) Use a clock source (NEN555), connect all D-lines with resistors to ground and pullup all interrupts (BUSREQ/INT/WAIT etc)
https://z80project.wordpress.com/2014/02/09/z80-test-circuit/

[luiHS]

You could use the typical Z80 test setup (*) to reverse engineer the pinout...

*) Use a clock source (NEN555), connect all D-lines with resistors to ground and pullup all interrupts (BUSREQ/INT/WAIT etc)
https://z80project.wordpress.com/2014/02/09/z80-test-circuit/

Thanks, if my adapters fail, I'll try your idea. But if the pinout is all wrong, I think it will also be difficult to locate where each pin is actually, it may even be that neither the clock signal is where the datasheet indicates, or the interruption signals.

The truth is that I miss so much a major failure in the datasheet, and there are no complaints. Unless the Z80, nowadays they are practically not used, or use the DIP to replace them in original vintage boards, because both the DIP and the PLCC work without problems with the pinout of the datasheet.

[wraper]

The truth is that I miss so much a major failure in the datasheet.

99,9% chance there is no error in the datasheet. Frankly it's stupid to blame the datasheet without even doing any measurements. It could be simply erratic pinuot in your design. For example it may look fine as a symbol, but some pin in the footprint does not match. Did you even try replacing CPU, where did you buy it?

[luiHS]

The truth is that I miss so much a major failure in the datasheet.

99,9% chance there is no error in the datasheet. Frankly it's stupid to blame the datasheet without even doing any measurements. It could be simply erratic pinuot in your design. For example it may look fine as a symbol, but some pin in the footprint does not match. Did you even try replacing CPU, where did you buy it?

All tested, and the scheme is the same that I used with the PLCC, in fact I copied the same Eagle project that I made with Z80 PLCC and I simply replaced it with the LQFP. I have already checked the LQFP footprint pin assignment many times in my design, and everything is in accordance with the data sheet, I put the schematic with its pins above.

And as for measurements, the clock signal is correct, I also see signals on the data and address bus, replaced the RAM, the Z80 and the ROM (PLCC32). Now, with this adapter, I will know if my design is incorrect or if the pinout is incorrect.

As for the fact that the datasheet is wrong, it would not seem so strange to me, that a single pin is wrong and may be enough to make it not work. Also taking into account that the Z80 in its LQFP version is likely to be used little, I would not be surprised if nobody claimed it. In DIP and PLCC its easy to be used currently to replace in old designs.

[luiHS]

This is the footprint pin assignment of my LQFP44, according to the datasheet. I have also checked that the order of the pins and their location is correct, as it appears in the datasheet.

This Wednesday or Thursday, I will receive from JLCPCB the small adapter PCB board to check the LQFP in my PLCC board, to know if really the problem is a wrong design of my board or the problem is the pinout of LQFP.

I have thought, if using the Z80 LQFP, the additional circuit chips should be TTL compatible 74LS, instead of the 74HC that my board now installs (and that works with the PLCC). But I find it unlikely.

Another possibility is that some of the circuit chips are burned, but it seems strange because I soldered everything in the oven. I also went through all the welds of the 74xx, RAM and ROM chips by hand, in case there was any defective solder, I soldered them all again.

[obiwanjacobi]

Where did you source the chips?
(looking for some myself)

[luiHS]

Where did you source the chips?
(looking for some myself)

If you mean where I bought the Z80, in Mouser

https://www.mouser.es/ProductDetail/692-Z84C0008AEG
https://www.mouser.es/ProductDetail/692-Z84C0010AEG
https://www.mouser.es/ProductDetail/692-Z84C0010VEG

[TomS_]

I have thought, if using the Z80 LQFP, the additional circuit chips should be TTL compatible 74LS, instead of the 74HC that my board now installs (and that works with the PLCC). But I find it unlikely.

Unlikely, the package has nothing to do with logic family compatability.

The Z84C00 is a CMOS device, not TTL, so be careful. Using TTL parts (e.g. LS) with a CMOS device that doesnt have TTL compatible inputs may lead to additional problems and more time wasted tracking them down.

Having said that, page 34 of the datasheet as per Mouser suggests that the inputs of a Z84C00 may already be TTL compatible (logic low is 0.8V max and logic high is 2.2V min).

Anyway, I have built a couple of projects using the DIP version of these parts and I have used HC logic without any issues.

[nctnico]

I'm seeing some open pins on the adapter board. That is asking for trouble. Another trick is to lower the operating frequency to rule out timing problems.

[TomS_]

I'm seeing some open pins on the adapter board. That is asking for trouble.

They are all no connects, or outputs (HALT, RFSH and BUSACK). Nothing to worry about. 

[nctnico]

I don't see whether the rest of the circuit the Z80 sits in should function with these lines not being driven. I have used the Z80 a lot decades ago and it is relatively simple to use. So there are 3 options here: faulty adapter, faulty chip or the circuit needs those extra lines connected.

A corner case may be that the original circuit's timing is marginal and the 10MHz replacement Z80 is responding faster than expected (compared to the 6MHz version).

[bson]

Something about your design might also be iffy, so the old (NMOS?) chip works but the CMOS one doesn't.

Do you have a scope?  If so, start by checking the reset pin (to make sure it's held in reset long enough for proper POR), then the clock input.  Then check the bus outputs to see that it tries to execute code.  Check the bus control signal levels.

[nctnico]

The original is CMOS too. Actually it seems you can still buy these new from Mouser in a PLCC package. That would have been cheaper than creating the footprint converter.
https://www.mouser.de/ProductDetail/ZiLOG/Z84C0006VEG?qs=sGAEpiMZZMvu0Nwh4cA1wSr8HV81hI8hMFMv9GP8MJI%3D

[luiHS]

 
Today I assembled and installed the adapter for the Z80 LQFP to PLCC. And it has worked perfectly.

I find it strange, because the project of the board with the Z80 in LQFP was a copy of the same project in PLCC. I can think of several explanations.

1.- In this adapter I used a Z80 of 6Mhz, and in the board that does not work I installed a Z80 of 10Mhz. I do not think that's the problem, I suppose a Z80 10 Mhz can work perfectly on a board with a lower clock signal, in my case 4Mhz.
ncnico commented that this could be the problem, I will try to install this Z80 6Mhz on the other board to see if it works, and that would explain everything, although it would seem very strange to me that this is the problem..

2.- For some reason that I do not know, for now, despite having copied the Z80 PLCC project to do the LQFP project, I must have made some mistake. But I can not think of which, because the rest of the scheme is identical, only the Z80 PLCC was replaced to an LQFP, and the pinout has already been revised several times, as well as the footprint.

Taking into account that on the board that the LQFP does not work, I have soldered all the chips again, I replaced the socket PLCC32 of the ROM and also the RAM, I do not see where the problem may be.

I'm going to install the Z80 6Mhz, on the board that does not work, and if it still does not work, what I'll do is copy the Z80 PLCC project back into a new one and make the changes from Z80 PLCC to LQFP. Now that the pinout of the Z80 LQFP is fine, with copying the adapter scheme to the new project and replacing the PLCC, it should work.

Common sense tells me that I most likely made some mistake in modifying the project to replace the PLCC with an LQFP, but I do not see where. The scheme is the same, the pinout is correct and the footprint is also checked several times.

[wraper]

There is also probability that PCB is simply defective. Once I've got a few PCBs from PCBway with defective vias.