Let's say I have a a crystal resonator with a "strange" rating like 1.234MHz. Let's say I then I have an IC which needs a clock input at that frequency. And let's also say that you can't get can oscillators at that frequency.

I'm guessing I have a couple of options. I could make an oscillator circuit with some NAND gates, some resistors and some capacitors:



(Care of: http://www.z80.info/gfx/series2.gif)

But I was wondering if you can get that same circuit in an IC? It sounds like a generic pattern, crying out for a generic part.

I've also looked at programable oscillators. These look neat, with NVRAM to hold the last programmed frequency etc. The problem is they are all SMT parts, and I'm not so keen on them unless I have no other option. Plus SMT makes it hard to program them out of circuit.

Any thoughts are welcome as always.

And golly, it's quiet around here lately...

_________________
8 bit fun and games: https://www.aslak.net/

I don't know, the circuit you posted is just few inverters, two resistors and two caps. It doesn't really get much simpler - I imagine that if you really wanted something more integrated you'd just use an oscillator module at some common frequency. And if you want an uncommon frequency you'd have to adjust the capacitor/resistor values anyway (I think).

There are exceptions, but they tend to be adapted to a particular niche - like realtime clocks with integrated 32kHz crystals or those PLL-based frequency synthesizers.

But I'm sure if you want a couple of thousand units for a product, there are manufacturers that can make an oscillator module to specification.

3 bucks, gets you a 5V custom programmed DIP oscillator in any frequency...

http://www.digikey.com/product-detail/e ... -ND/387829

The full selection for 5V...

http://www.digikey.com/product-search/e ... ageSize=25

I use these all the time for odd VGA clock divisions.

Brad

Wow, that's awesome Brad, thanks! I wonder how to get one sent to the UK, without paying a fortune for delivery etc.

FWIW, I'm after a 3.58Mhz can. Not able to find one anywhere, though I have found the crystal.

_________________
8 bit fun and games: https://www.aslak.net/

NTSC color burst clocks and crystals (3.587MHz) are in abundant supply...

http://www.digikey.com/product-search/e ... ageSize=25

If the mail system wasn't so brutal here, I would drop one in an envelope for you, I have 2 trays full!

Are you doing a video project?
If so, you would need the PAL frequency of 4.4336MHz for your standard.

Brad

It's for a YM3812 AKA the OPL2. While I could, presumably, use any handy clock then I would not be able to utilise ready-made sound data. I'm guessing that even in the UK a PC fitted with an OPL2 sound card had it clocked at 3.587Mhz instead of the equivilent PAL frequency. Might be wrong though...

Thanks for your kind offer! Shipping is £12 at digikey. I'll have to see what else I need...

_________________
8 bit fun and games: https://www.aslak.net/

I wonder if the IC would survive a trip from my backwoods postbox to your place??

Are you using a WD65C02 to clock the sound chip?
If so, perhaps you could drive the 6502 with a 14.318 clock module, then just divide by 2 using a 590 or 7474 to gain the 3.58 clock.
14.318MHz cans are also commonly available.

http://www.digikey.com/product-search/e ... ageSize=25


Brad

I'm sure it would. I can paypal you the costs.

Erm. It's actually a 6809 based micro. Unfortunately I don't have any other parts on the board that are running a frequency in "binary range" of the 3.58 I need.

_________________
8 bit fun and games: https://www.aslak.net/

Ah, then may parts might not be suitable.
I only use CMOS parts, so my oscillators are not TTL.
I take it you would need the TTL variant?

Brad

The YM3812 datasheet says:

Input low < 0.8V
Input high > 2.0V

For the CMOS oscillator:

Output low < 0.5V
Output high > 4.5V

For the TTL oscillator:

Output low < 0.5V
Output high > 2.4V

So from this I'd conclude that the YM3812 is TTL (it doesn't actually say in the datasheet), and that it would accept either a TTL or CMOS output. I might well be ignoring current issues though. I'll get a CMOS part.

My understanding is that mixing CMOS and TTL is a bigger problem when TTL drives CMOS, since a high TTL output might not fall within the high threshold of the CMOS input. But again I suspect I'm simplifying things.

_________________
8 bit fun and games: https://www.aslak.net/

If you want to give it a whirl, I can wrap one up on a block of foam and see if it makes it to you.
Feel free to send your shipping address to me at...

atomic at atomiczombie.com

Brad

Mail sent.

_________________
8 bit fun and games: https://www.aslak.net/