REPLACEMENT FOR MM5369 TIMEBASE CHIP

[craig-1]

Dear NUTS & VOLTS readers. I need help to find an chip that seems to be discontuned. It's a timebase chip # MM5369 that uses a color TV burst crystal and devides the frequency of the crystal 3.579.454Hz. and gives out 60Hz. It works by deviding the color burst frequency by a 17 stage devider to get the 60Hz out. I checked the distrubers MOUSER, DIGI-KEY, JAMECO, AND COULD FIND IT. I do rember in ELECTRONICS NOW magazine Q & A colume thay had an 2 chip substitute for the MM5369 month and year I don't know. I hope someone can help me in this request. If someone can find that article in the magazine, please send me a photocopy or E mail back. THANKS

[jimandy]

Try eBay. I see some for sale there in the "buy it now" category.

[L. Daniel Rosa]

Do you need the colorburst frequency output as well? If not then just about any 8 pin microcontroller could do the job as a single chip replacement, and with just about any crystal frequency. If you choose one with ADC onboard, you could give it voltage and temperature compensation as well.

[Edd]

“It works by deviding the color burst frequency by a 17 stage devider to get the 60Hz out.”

As a direct power of 2 binary derivative , it won’t compute…seems like there is a little bit of divide by n and Reverse Polish Logic divisional gating also incorporated within.
Checking my voluminous inventory of goodies, it seems that I have one stored away…MM5369.. of ’72 vintage in an 8 pin plastic DIP, also there is the 100~ output version available.
If you need one…au gratis… let me know, and I will dig in and pull it for you, I haven’t used it in 33 years so I guess it would be safe to say that I will not need it in the future.

73's de Edd
[email protected] .........(Interstellar~~~~Warp~~~Speed)
[email protected].........(Firewalled*Spam*Cookies*Crumbs)


I got a jaywalking ticket. I asked, "Is this going to go on my record, or can I just go to a Walking School and have this taken off ?"

[craig-1]

Does anyone knows PIC microcontrollers that can make the microcontroller give out a 1Hz & a 0.1Hz squarwave to drive a frequency counter. In other words I need a good TIMEBASE to replace the MM5369 chip with a microcontroller chip. Does anyone have the softwear to do it? Does anyone have a working circuit to do this?

[philba]

any PIC can do this. The most minimal circuit (xtal, caps, mclr pull-up resistor and the pic) is all you need. the software is very very simple but I doubt there is anything available based on that exact crystal frequency.

If this is a one-off project, why not just buy one on ebay or take edd up on his offer.

[rshayes]

First, a colorburst crystal will probably not give you an accurate 60 Hz timebase. The colorburst frequency is not an even multiple of 60 Hz, 1 Hz, or .1 Hz. You might be able to pull the crystal to the proper frequency, but this is not very predictable and may not be stable.

The MM5369 was basically a mask programmed divider chain. The data sheet also implies that operation at 10 volts may be necessary with a colorburst crystal.

A field programmable logic device such as the Xilinx XC9536 or some similar device can be programmed to do this function. These are available from DigiKey as well as other sources and the software needed to program them can be downloaded from the Xilinx web site (www.xilinx.com). The current version is large and requires either Windows 2000 or XP. Earlier versions are smaller and can run under Windows 90. The 3.3 version tends to lock up if it is run while Norton Antivirus is active. If yoy disable Norton while running the program there are no problems. The 3.3 version is also modular, which can reduce the download size. The 3.3 programming module can also run on a separate computer if desired. The actual JTAG programming file is quite small and can often fit on a floppy disk for transfer. The intermediate files are numerous and much larger but the JTAG file contains the final programming information.

[jimandy]

In re-reading your initial post, you were looking to generate 60hz pulses using the 5369 chip.

You could generate 60 hz pulses by picking off a signal from your ac line (via a transfomer) and shape them with a Schmitt trigger. May not be as accurate as you want it but it's simple and real cheap.

[philba]

You will never get a 100% accurate time base from anything since there is no 100% precision. an 80 ppm crystal (pretty typical) will give you pretty good but if its not a multiple of your desired frequency, you will have to play some games to get it close. In theory, you can get within 1/2 a clock period of the desired frequency using a microcontroller. Use the 16 bit timer with external osc input. Do multiple timings to accumulate the right amount of time. You can get within a few cycles of a precise count. That error is going to be noise when compared with 80 ppm (which is +/- 80*3.579454). For a somewhat less accurate but easier approach, use the cb xtal as the uC clock.

Theory is one thing, though. Getting the code to produce the exact period with the colorburst frequency is tricky and requires instruction cycle counting. I've done it (for other frequencies) and it's a PITA. If you are determined to use a microcontroller, why not just use a crystal that is a multiple of 10 hz?

By the way, the color burst freq is, by definition, a multiple of 1 hz and .1 hz.

A second foot note - isn't the AC period adjusted throughout the day? I recall reading that it does wander around a bit though the cumulative error is small.

[Robert Reed]

Philba
You are thinking right about power line frequency. The actual "60 Hz " frequency can vary as much as 0.2 Hz due to load demands and power grid variations - for the short term. But for the long term average it is surprisingly accurrate.

[rshayes]

The most basic rational frequency in the television standard is 4.5 Mhz. Interesting enough, this frequency is also a connection between NTSC and PAL.

The NTSC scan rate is 4.5 MHz divided by 286. This resurts in a repeating decimal which is rounded off to 15,734.265 Hz. This is further divided by 525 to get the frame rate of 29.94 Hz (also rounded off). The field rate is twice this or 59.94 Hz.

The color subcarrier is 455/2 times the horizontal scan rate, or 3.579545287 Mhz. Again, this is a rounded off repeating decimal.

If you divide 4.5 MHz by 288, you get the PAL horizontal frequency of 15,625 Hz. Dividing this by 625 will give you the PAL frame rate of 25 Hx.

Both NTSC and PAL can be digitized with a 13.5 MHz sample rate, which is 3 times 4.5 MHz.

The division ratio in the MM5369 is 59,659. To get an even 60 Hz would take a crystal frequency of 3.579540 Hz. This will require pulling the crystal high by about 5 Hz. This might be workable since the crystal is probably ground for a slightly high frequency so that it can be pulled down to the correct frequency by the circuits in the television receiver. The long term stability might be less than that of a crystal that is operating at its design frequency with only slight adjustment with circuit elements.

<small>[ November 20, 2005, 04:41 PM: Message edited by: stephen ]</small>

[jimandy]

(robert reed)

But for the long term average it is surprisingly accurrate.

I would think the long term would be exceedingly accurate as there must be some stable common 60 hz time-base reference to synchronize the myriad number of generators feeding the grid. Wouldn't they (the generators) all have to be within a half cycle of each other so that phase adjustments could be done somewhere between the nodes in the grid?

[craig-1]

I want anybody who has the magazine ELECTRONICS NOW in the year late 1998 or 1999 in the Q & A colume thay had a circuit that was a drect substitute for the MM5369 chip. The replacment circuit used two IC chips, had a switch that switched between 50Hz or 60Hz. All I ask if someone has ELECTRONICS NOW please look it up in Q & A. I do remember seening it in there, but at that time I didn't need it!!!!! I don't think it was in POPTRONICS.

[k7elp60]

I have been using a substitute for the 5369 since 1998 that works real well. I use two versions with two different microprocessor crystals. They both use the 4060 CMOS oscillator divider as the first circuit. In both circuits the output of the 4060 is 300Hz. The second circuit uses either a 4017 or a 4022.
They both have the crystal connected to pins 10 and 11 of the 4060. The crystal has a 1M resistor in parallel with it. There is one capacitor from pin 10 to ground, and one capacitor from pin 11 to ground. The capacitance of the two capacitors in series equals the load capacity of the crystal. The values can be adjusted to trim the crystal on frequency. I have used a fixed value for one and a trimmer for the other. I have also use npo ceramic capacitors and sellected them to trim the crystal.
If using a 4.9152Mhz crystal and pin 3 of the 4060 out the frequency is 300Hz. Pins 8 and 12 are tied to ground.
If using a 2.4576Mhz crystal and pin 2 out of the 4060 the frequency is also 300Hz.
I use the 2.4576Mhz crystal mostly because the data book lists the maximum frequeny for the 4060 is 4Mhz @ 10 volts. By using the lower frequency crystal I can use lower Vdd for the 4060. With the crystals I have been using at 2.4576Mhz I am using 68pf and 56pf for the capacitors.
The second circuit uses a 4022 with pin 14 input of 300Hz, pin 2 out 60Hz, pins 8and 13 to ground and pins 4 and 15 tied.
Pin 9 of the 4060 is the crystal frequency out.
I have been using the circuits described as timebases for digital clocks. I adjust the crystal freqency while monitoring the frequency on my frequency counter. At room temperature the clocks vary only a few seconds with WWV in a month.
When using the 4017 pin 14 is the 300Hz input, pins 8 and 13 to ground, tie pin 15 to pin 1 for 60Hz out or pin 15 to pin 5 for 50 Hz out. Pin 3 is 50Hz or 60Hz out. If you need 100Hz then the 4017 or the 4022 can be configured to divide by 3.

[Robert Reed]

Jimandy
Exeedingly accurate IS surprisingly accurate. Any power plant generator that is coming on line or back on line is phased to the power grid (isolated at the time) within 30 degrees, if memory serves me right. At that time the generator is closed into the grid and is automatically locked into the grid frequency and follows that by a phenomina I beleive is called 'motor effect'. As far as long term accurracy goes--what is the definition of accurracy? But it seems to keep the electric wall clock in tune, so I would call that surprisingly accurrate