Is there a resource that lists products which are similar to a particular item?

Or, can someone assist me directly.

I am looking to do a project which will require several 12V lamps to be lit (possibly 80). Unfortunately the PIC I will be using will be too slow to use PWM and a transistor setup, so I was thinking that something like the M74HC374 but 12 volts would be suitable.

I could use the M74HC374 and a transistor on each output, but it would be nicer to have it in a single chip. I will need to do around 80 outputs.

To sumarize, I need a device which will latch on or off, taking logic/5V input an enable signal, and be able to either output or take in 12V. Current would never exceed 0.5A per pin (I reckon).

 

why will it be too slow?
pics can be pretty nippy...
have you considered multiplexing?

 

A PIC is too slow.?

Why? are u going above 20Mhz ?

 

A PIC is too slow.?

Why? are u going above 20Mhz ?

4MHz for some of the.. retro pics

 

A 74HCxxxx IC is designed to use a 2V to 6V supply. Maybe its maximum allowed supply is 6V or 7V but certainly not 12V.
Any transistor can be fed from a 74HCxxxx IC through a current-limiting resistor and its collector load can have any supply voltage it is rated for.

 

The 74C374 will work up to 15V and has an absolute maximum rating of 18V. Unfortunately at a 12V supply it requires a minimum of about 10V for a logic high. This low logic level to high logic level can be done by a converter such as this.

Of course this setup is slower than the 74HC374 but it should be fast enough for your requirements.

 

The 74C374 will work up to 15V and .... low logic level to high logic level can be done by a converter such as this. (ADG3123)
...

Thanks, that is almost exactly what I wanted.
(don't get me wrong, it isn't your fault there isn't a single chip for the job!)
Although, I think from the datasheet (and the fact farnell believes it is still in production, which makes purchasing them easier) that the 74c373 is more suited (374 appears to have a clock based enable, where the 373 is a single enable signal).

4MHz for some of the.. retro pics

Bang on, phase 1 was going to be a 16f84, as apart from the lack of speed, it meets my requirements (and I have several needing a job).

why will it be too slow?
....
have you considered multiplexing?

I am not sure, but isn't the largest pic something like 48 pins? I am having to multiplex anyway.
I will be using a single 8 bit port for 'output' doing a copy from an internal 'array', then using other pins for a 'select'/enable sequence.

However, at some point, I am going to have to provide communication. Exact method I have not yet decided upon. Whether it is with another PIC, or with an external memory device. It is likely that this communication could cause unwanted delays in the pwm output. It is going to be difficult to split up meaningfull comms to be once per 'byte' of lamps, and not cause delays if it were once per 10 bytes (for 80 lamps).

By using latching outputs, I can mitigate the effects of these delays.

 

Why do you think a 12V device would be better? You can drive something like ULN2803 with a 74HC374. The 74HC part can be driven by your PIC, if you are running it off 5V.

 

Why do you think a 12V device would be better? You can drive something like ULN2803 with a 74HC374. The 74HC part can be driven by your PIC, if you are running it off 5V.

The lamps (output) are 12v
Your method will mean almost twice the components I reckon (my initial thought was 80 transistors on the outputs, your part would be a similar concept, it is effectively 8 transistors in 1)

you are intending pic->*multiplex*->74hc->uln2803
for 80 outputs, =10x uln2803, 10x 74hc, 1x pic (assumes multiplex on 5v logic)
whereas the 74c would be
pic ->adg3 (or uln2803 I guess)->*multiplex*->74c
so 80 outputs = 10x 74c, 1x adg3, 1x pic plus (1x adg? + transistor ) for multiplex

But, then again, your part would make the multiplex cleaner.

 

You might want to consider using 74HC595 or 74HCT595 instead of the 374's:
http://www.nxp.com/documents/data_sheet/74HC_HCT595.pdf
They are a bit more handy to use.

Investigate TPIC6B595's:
http://www.ti.com/lit/gpn/tpic6b595
They are capable of 500mA peak sink current, 150mA continuous per output. Don't use these if you are going to exceed that specification, however looks like just what you need at the moment.

Beware that incandescent lamps have a very high start-up current. You should apply a constant PWM to all of the lamps to keep the filaments on the edge of glowing; that will extend the life of the bulbs. If you cycle them hot/cold, they will fail very rapidly.

 

You might want to consider using 74HC595 or 74HCT595 instead of the 374's:
http://www.nxp.com/documents/data_sheet/74HC_HCT595.pdf
They are a bit more handy to use.

Investigate TPIC6B595's:
http://www.ti.com/lit/gpn/tpic6b595
They are capable of 500mA peak sink current, 150mA continuous per output. Don't use these if you are going to exceed that specification, however looks like just what you need at the moment.

Beware that incandescent lamps have a very high start-up current. You should apply a constant PWM to all of the lamps to keep the filaments on the edge of glowing; that will extend the life of the bulbs. If you cycle them hot/cold, they will fail very rapidly.

Thta's a cool part. I didn't even realize they existed. However, it's a shift register. Isn't our OP looking for a D register, like TPIC6273? This doesn't have output enable, but my understanding is that the MUXing is going to be in front of the D registers.

 

The lamps (output) are 12v
Your method will mean almost twice the components I reckon (my initial thought was 80 transistors on the outputs, your part would be a similar concept, it is effectively 8 transistors in 1)

you are intending pic->*multiplex*->74hc->uln2803
for 80 outputs, =10x uln2803, 10x 74hc, 1x pic (assumes multiplex on 5v logic)
whereas the 74c would be
pic ->adg3 (or uln2803 I guess)->*multiplex*->74c
so 80 outputs = 10x 74c, 1x adg3, 1x pic plus (1x adg? + transistor ) for multiplex

But, then again, your part would make the multiplex cleaner.

74C parts cannot source or sink more than a few milliamps.

 

Beware that incandescent lamps have a very high start-up current. You should apply a constant PWM to all of the lamps to keep the filaments on the edge of glowing; that will extend the life of the bulbs. If you cycle them hot/cold, they will fail very rapidly.

74C parts cannot source or sink more than a few milliamps.

Thank you both for your answers.
You are right Ron that the 6273 is more accurately allied to what I was looking for. (OE is of little use to me on this task).
However, as I am still at a pre-protoype stage, the 6b595 means that although I would transfer data slower (minimum of 16 clock cycles I guess, for the serial output), it could save me 8 pins of parallel output on my PIC. As this is a latching situation, this would afect my ability to control the PWM to some extent, I would have to calculate it between the transmissions/other comms.

I have several projects on the go, so will think on this further, I have downloaded the datasheets. I will update the thread when I have a solution (I have a physical socketting issue I need to overcome)

Thanks for the info on needing to PWM the lamps. I had heard of this before, but forgotten. I know it is best practice, but given that the current solution is plugging a 12v transformer into a wall, this may be a low down priority.