CMOS can not drive LS input properly

[atferrari]

A PIC sends data to three CMOS shift register (CD4094) dasy-chain connected to output in parallel a 19-bit word which in turn is fed to five 4-bit adders (74LS283) for further processing.

With the CD4094 outputs unloaded, the desired 1s always show up where expected. No doubt about that.

But once the inputs of the 74LS283 are connected, some bits that should be "1" become "0". Others, when tested with the logic probe show a "pulse" indication instead of a steady high.

The most important: with everything running as described, if I disconnect the adders' inputs again, proper output of the CD4094s are restored on the spot.

I've checked as OK:

1- Proper grounding of Vdd pin and unused inputs.
2 - Proper decoupling of the CD4094. (.1 disk capacitors)
3- No shortcircuits.
4-No floating inputs.
5-Supply = 4.81V in all pins where required.
6-High levels= 4.8V
7-Low levels=0.1V
8-Noise in Vcc rail, 25 mV peak to peak. Same frequency as the clock used for the adders (4.1 MHz).

In the knowledge that one CMOS output is good enough to drive an LS input, what else should I check.? Comments, suggestions appreciated.

Agustín Tomás

 

[Roff]

If you look at the **broken link removed**, you will see that the A and B inputs are double loads. This is probably the cause of your problems. There are TTL and advanced CMOS parts with approximately the same functionality as CD4094.

 

[Nigel Goodwin]

My immediate thought is "why all the chips?".

What are the 4-bit adders doing that the PIC can't do for itself?.

 

[atferrari]

Problem solved - Moving ahead

Hi Ron,

Accidentally discovered that at two points there were almost microscopic shorts evident only when the board was slightly bent.

Thanks for your help anyway. I will revisit data about input/output (V and I) for CMOS and LS to retain some experience from this. Thanks Ron!

Hola Nigel,

It's a NCO taking data from a EPROM 2716. Give me time to see what I get at the end and then I will consider how to implement it with a PIC. Not sure if a could get 1 Hz resolution from 1 to 500,000 Hz. Perhaps running the PIC at 20 MHz?

This is a very long time / many times deferred project that I ow to myself.

Interested on your suggestion.

Agustín Tomás