Ubicom SX Embedded Controller Clocks

Summary

See also:

• http://romanblack.com/pic2clk.htm 2 systems for switching a PIC or SX between two different speed clocks. A dual speed RC version and a RC / XTAL version.

Stephen Holland says:

There are several ways to get a clock... A few of the more popular ones are listed below.

• Crystal Oscillator - the 4-legged can. There are thru-hole versions which are generally called 'full size' (rectangular) or 'half size' (square). Various companies offer surface mount as well. Generally, they are a good solid clock, but also can be very power hungry. A quick search of Digi-Key found the CTX MXO45-50.0000 (Digi-Key # CTX121-ND), which consumes ~40mA at 5V. This is on top of the power consumed by the crystal driver within the SX, which is still required. This type of oscillator is harder to find at lower voltages, and as it is a hard driver output, they generally cannot accept higher than 7V on the output pin. This is one good reason why the oscillator must be removed from the OSC circuit when the SX-Key is programming or debugging (it puts Vpp = ~12.5V on OSC1). Only requires to be connected to OSC1 only.
   
• Crystal - The most accurate type of clock source. To use on the SX, it will need to be a fundamental mode. Most uControllers require fundamental mode anyway (3rd overtone is possible, but with the required LC filter circuit to be stable, it becomes too expensive), the reason it is mentioned specifically is that relatively few companies offer fundamental mode crystals above ~40MHz. Abracon (www.abracon.com) is one such vendor, as has always been quick with samples for the SX. Check out their ABL (HC-49), ABL2 (HC-49 low-profile), ABLS (AT49 SMT), ABLS2 (AT49 SMT low-profile), ABM5 (SMT ultra miniature) and ABMM (SMT micro miniature). Crystals, because of their high accuracy, are really required for applications where timing is absolutely critical.
   
• Crystal Resonator, Ceramic Resonator or just Resonator - A little more drift (less accurate) than a crystal, but smaller and sometimes cheaper. Again, relatively few companies offer fundamental mode resonators above ~40MHz. Murata is a good source of resonators, and offer versions with and without built-in load caps in various packages. The names for the various families Murata offers are (the info I have is a couple years old, but should still be good): Thru-hole without caps - CSA, thru-hole with caps - CST, SMT without caps - CSAC, SMT with caps - CSTCC.

Questions and answers:

As the SX device can operate at frequencies up to 75 MHz, are there recommendations for proper oscillator circuit layout to minimize EMI?
 

It is typically recommended to reduce the power and ground loop area. The long supply lines with the relatively large areas that these lines surround may form an effective antenna. At the frequencies present, an unacceptable level of interference may be radiated and may result in unstable oscillation. A grounded area under the SX device must be connected to the Vss pin. In addition, the ground area should be tied to the ground plain with multiple vias. This ground area ensures that the major part of the field lines emanating from the SX are concentrated between the SX and the ground level.

Since the SX28AC power and ground pins are on the opposite side of the package from the one for the oscillator pins, one needs to evaluate the effect of the larger loop created as a result of connecting the oscillator capacitors to the ground. This large loop will result in larger amount of current flow in the oscillator resonant circuit. The creation of a ground plane underneath the chip (mentioned above to reduce the power line loop) will help shorten the loop effect. In addition, it will be useful to run the capacitor ground connection close and in parallel to the clock input and output signals. If ground area underneath the chip is not available, the capacitor should be directly connected to the SX ground pin through a short and independent trace. The shape of the oscillator signal should be observed with a low capacitance FET probe to identify any ringing. The oscillator signal should look smooth. Impedance matching techniques can be employed to smoothen the oscillator signal.

See: EMI
 

I am using an external clock source to drive the SX. How do I do ISP of the SX?
 

You need to have circuitry to disconnect the clock-source from OSC1 when the SX-Key or SX-ISD is plugged into your system. It should be noted that 12.5V is present on OSC1 during device programming. That circuitry could be as simple as a jumper terminal.
 

I have a application for the SX18/SX28 processors where I need to switch between two processor clock frequencies. I plan on doing this by having two external crystal oscillators and multiplexing logic to switch between them. Can you tell me if I need to take any special precautions to prevent glitching on the clock line during the switch.

For instance suppose I generate an extremely short clock cycle that is beyond the speed range of the processor during the switch, is this permissable, or will it cause the SX to malfunction?
 

This will be very tricky to do. You will have to ensure that there is no phase shift when switching the clocks. We have had several people design with USB chips that have low power modes, with the SX running from the clock source provided by the USB controller (switching between 4 or 12MHz) and it did work just fine.
 

Is the internal oscillator usable above 4Mhz and if so can you supply technical information about using it?
 

The SX internal oscillator only supports up to 4MHz operation.
 

Why can't I find a 75 or 100MHz resonator?
 

External oscillator support for crystals or ceramic resonators is limited to 50MHz as this the highest frequency offered in a fundamental mode oscillation (third overtone is the other choice, but not recommended as it requires the use of a LC filter in the oscillator circuit, which can be quite large, expensive and difficult to set up).
 

We recommend 50MHz fundamental mode crystals from Abracon (abinfo@abracon.com, (949)448-7070). For low-profile thru-hole, use ABL-50.000-F (3.5mm height) or ABL2-50.000-F (2.5mm height). For low-profile SMT, use ABLS-50.000-F (4.2mm height) or ABLS2-50.000-F (3.3mm height). Please refer to Abracon's Web site for specifications (http://www.abracon.com/).

Richard Ottosen [rottosen at IDCOMM.COM] says :

A third overtone crystal on the [SX] can interfere with the oscillator pins when they are used to program the SX. The inductor of the filter for third overtone operation can short the programming voltage and damage the SX or the programmer! [ed: Use a large value cap (0.1uF or so) in series with the inductor to remove the un-necessary DC short.see http://www.saronix.com/pdfs/128/128.p1.pdf]
 

The Murata 50 MHz resonator (Murata CSTCV50.00MXJ0H3) is recommended to be used with SX devices. This is a newer and better part.
 

Frequencies higher than 50MHz require the use of a TTL Clock Oscillator as the clock source. For example ASL or ASF family from Abracon or PLL IC (ICD2053 from Cypress). At these frequencies the SX should be configured at the highest gain settings, OSCHS2 or OSCHS3.
 

Can I run two or more SX devices from one clock source?
 

This clocking scheme should work: 1ST SX running from CRYSTAL oscillator ( XT , HS1,HS22,HS3- depends on Crystal frequency), OSC2 output of 1st SX used as clock source for 2nd SX. 50 Mhz crystal from ABRACON can be successfully used with all SX devices.
 

Also:

• General Clock information

Ted Inoue says:

I had been having problems with intermittent problems with some of the boards I'd developed using the SX chips. They'd work fine using the SX-Key, but when using a resonator they'd either fail to start up at all or would only do so randomly.

After troubleshooting and help from those on this list (thanks guys!) I tracked it down to improper reset on power-up. I'm pretty certain that the chip was slowly coming on as the power cycled up and was unstable.

There were excellent suggestions about time delay startups, RC filters etc., but since my boards were already built, I needed a software solution. After reading the spec sheets carefully, I tried using the brownout reset fuse, which I'd disabled previously. I enabled the 4.2v setting and instantly all my boards began working 100%.

There's very little information that discusses the brownout option, and it's so useful that I thought I'd pass this along. It's made the difference between having a (commercially) useless product and having one that I can comfortably put into users' hands.

Questions:

• ryan at flyingrobots.com asks:

  http://ucsub.colorado.edu/~patterrr/IndStudy/IndStudyHome.htm Hello,
  I'm working on a project that has 16 SX28 ucontrollers in it. Each chip is clocked by a 20 mHZ surface mount resonator, all of which were purchased from the Parallax website. My problem, is that when the system is powered up, some of the ucontrollers don't respond correctly to asynchronous communications. Therefore, I checked the clock frequency, and found that the clock frequency (provided from the 20 mHz resonator) of about 1/4 of the chips is about 34 mHZ! It seems totally random; each time I power-up the system, different resonators are running at this higher frequency, and sometimes they all start up correctly. The rise time of my power supply is about 16ms, and I've placed a 1 mohm resistor across the oscillator inputs of each of the SXChip/Resonator pairs. Has anybody seen a problem like this? Any ideas? Thank you very much!
  Regards, Ryan

  James Newton replies: Try changing the osc drive settings in the programming of the sx28's e.g. from HS to XT2 ryan at flyingrobots.com replies: "Thanks James! That solved the problem that has been keeping me from finishing this project for days! Thanks a lot for supporting the SX List - it's very helpful!" +

  +

  +

• gregb at laserlab.com asks:

  I was under the impression that Ubicom had abandoned support
  for the SX processors. Furthermore, they seem to be hard to
  get, and quite expensive. I do have an SX-KEY and SX-Tech
  development system, but have been shying away from the SX chips
  because they seem to be fading away.
  I welcome comments.

  James Newton replies: Nope, Ubicom supports the SX chips nicely, thank you. There have been some shortages, but they are available and the cost is not bad at all considering what you get. I don't think they will fade away until there is another microcontroller than can run at 75 MIPS!+

  James Newton replies: Update: Ubicom has passed the SX chips to Parallax, Inc. (makers of the popular BASIC STAMP) and so the future of the SX is assured.+

  +

Interested:

  .