New ZX Devices Now Available

[dkinzer]

We have introduced six new ZX devices, all based on the ATmega1284P. The processor is like the ATmega644P except that it has an additional 16-bit timer (Timer3) and sports 16K of RAM. The new devices are as follows:

VM Mode:

• ZX-24r $59.95
• ZX-40r $39.95
• ZX-44r $39.95

Native Mode:

• ZX-24s $59.95
• ZX-40s $39.95
• ZX-44s $39.95

On the new devices, Timer3 is used as the I/O timer to regulate time-based I/O routines. This change allows simultaneous use of certain I/O functions (like PulseOut() and InputCaptureEx(), for example) that is not possible on ZX devices with a single 16-bit timer. The availability of two 16-bit timers also allows generation of four simultaneous 16-bit PWM signals.

With the introduction of the new devices, the existing mega644P-based devuces (ZX-24p, ZX-40p, ZX-44p, ZX-24n, ZX-40n and ZX-44n) have been reduced in price by $10.00. Moreover, the mega644-based devices (ZX-24a, ZX-40a and ZX-44a) are being discontinued. The latter set of devices have been discounted for clearance, limited to stock on hand.

[dkinzer]

You may notice that the ZX-24r and ZX-24s are built using a new board design that uses a surface-mount crystal. This was possible because neither of these devices need the external serial EEPROM and removing that part made space available for the crystal.

The ZX-24n will also use this new board design. The ZX-24p, however, will continue to use the previous design because it requires the serial EEPROM for program storage.

[kranenborg]

Hello Don,

Great to see the new devices introduced. In addition to the extended RAM I assume that all devices have 128Kb Flash as well, as opposed to the 64Kb of the earlier 664p devices. The first part of the ZX.40r and ZX-40s webpages are not completely consistent on that matter as they still mention 64Kb Flash (but the table mentions 128Kb)

/Jurjen

[dkinzer]

The first part of the ZX.40r and ZX-40s webpages are not completely consistent on that matter as they still mention 64Kb Flash (but the table mentions 128Kb)

Thanks. The web pages have been updated. For VM mode devices the user code space is 60K while it is 124K for native mode devices.

[spamiam]

What a great enhancement! I presume the P-code is written to the flash memory. How much faster is it to read data from flash than from the SPI EEPROM? Does this translate into meaningful increases in program execution speed compared to the EEPROM-based devices?

-Tony

[mikep]

How much faster is it to read data from flash than from the SPI EEPROM? Does this translate into meaningful increases in program execution speed compared to the EEPROM-based devices?

The mega128, mega1281, and mega1280 based ZX devices already use this facility. Reading ZVM bytes from flash is 2-3 times faster than using EEPROM but is not as fast as native mode programs which do not need to interpret an intermediate language.

[dkinzer]

How much faster is it to read data from flash than from the SPI EEPROM?

Using the rather simplistic b = b + 1 test the ZX-24p rates 175,000 instructions/second compared to 270,000 for the ZX-24r and 2.9 million for the ZX-24s. The overall difference will not be that stark in a "real" application that is waiting on I/O, etc. but the raw speed can make a difference in specific code sequences like toggling an I/O pin, etc.

[dkinzer]

Using the rather simplistic b = b + 1 test the ZX-24p rates 175,000 instructions/second [...]

By way of comparison, NetMedia specifies the BX-24p, using essentially the same test, at 83,000 instructions/second which I have confirmed. Other than the VM firmware, the only difference affecting performance is that the BX-24p runs at 1/2 the clock frequency so the normalized rate for the BX-24p would be 166,000. The extra performance of the ZX VM series is due to careful design, coding and analysis that we did very early in the development process.

[mikep]

We have introduced six new ZX devices, all based on the ATmega1284P.

Congratulations on bringing another ZX device to market!!

The corresponding Oak Micros devices are called ZX-24ru (VM mode) and ZX-24su (native mode) and are very similar to the existing ZX-24pu and ZX-24nu devices respectively. The only difference is the mega1284P and the removal of the 64K EEPROM. The ZX-24ru and ZX-24su devices are currently undergoing test and will be released in a few weeks. Look for the announcement.

Our mega128, mega1281, and mega644p based devices (ZX-128e, ZX-128ne, ZX-1281e, ZX-1281ne, ZX-24nu, and ZX-24pu) have been reduced in price by $10.00 and will be discontinued when the existing stock has been exhausted. All our devices continue to be supported by Oak Micros.

[liam.zbasic]

I will have a few ZX-40A devices leftover for a long time to use as "guinea pigs". Will the v3.0.2 compiler and future versions still support the ZX-40A?

[dkinzer]

Will the v3.0.2 compiler and future versions still support the ZX-40A?

Yes, for the foreseeable future. We are still supporting the ZX devices based on the mega32 that were discontinued in April of 2008.

If someone needed ZX-40a or ZX-44a devices and could not use the mega644P-based devices, we can make them available on a special-order basis. The same may be true for the ZX-24a although they're more difficult to manufacture on a special-order basis.

The existing stock of mega644P-based devices (ZX-24a, ZX-40a and ZX-44a) have been significantly discounted for clearance.

[liam.zbasic]

What does the "foreseeable future" mean, one month, 6 months, one year, 5 years? I'm also considering purchasing the ZX-1280n in the next few months. But if you think you might offer higher end versions such as ATXMEGA or 32-bit AVRs, then I'll wait. Exciting products. Thanks.

[kranenborg]

Don,

Concerning the new devices, could you think of providing a low-voltage (2.7V at 7.3MHz) and low-power variant of the new ZX-40s as well, i.e. a ZX-40sl ? This would enable easy interfacing with the increasing number of 3.3V peripheral devices available without the need for level shifting (since most of them do not take 5V). The native mode devices are still extremely fast, even at half the original clock speed.

Regards,
Jurjen

[dkinzer]

What does the "foreseeable future" mean

It means that we aren't committing to any specific time frame but that we see no reason to discontinue support. Our plan is to continue to have the compiler and IDE support the discontinued devices and to release updates to the VM and ZX Library for discontinued devices. That does not necessarily mean, however, that new functionality that is introduced after discontinuation will be implemented for the discontinued devices. For example, the flow control capability that was recently added is not implemented for the mega32-based devices (because there isn't enough unused Flash space).

f you think you might offer higher end versions such as ATXMEGA or 32-bit AVRs, then I'll wait.

We are always considering other base devices and we expect to introduce several new ZX devices before the end of the year. One or more xmega-based devices would probably be the first to be released as we have some currently undergoing testing.

This is the conundrum faced by all technology consumers: do I buy what is available today or wait an unknown period of time to buy a new product with unknown, but likely better, capabilities.

[FFMan]

> This is the conundrum faced by all technology consumers: do I buy what is available today or wait an unknown period of time to buy a new product with unknown, but likely better, capabilities.


summarised by the phrase - the only good time to buy technology is yesterday and tomorrow !