EPE datalogger, is it worth rewriting s/w, or should I scrap and start afresh?

[ermine]

Hi all

I have a 1999 vintage EPE 8-channel datalogger from the Magenta kit **broken link removed**.

Back in the day when I bought it, I used win98, and ISTR it served me well in characterizing some NiCAD batteries. The PC transfer software was horrible, some evil garbage using DOS Qbasic and inline 8086 assembler (on the PC) to directly access the Com port registers. You can't do that these days as Bill Gates has decided writing directly to memory is a bad thing. The RS232 data transfer is in some unfathomable binary format, so catching it with hyperterminal and postprocessing doesn't work for me.

I have three questions:

• does anyone know of somewhere that has brought this device/program into the 21st century - Magenta haven't, or at least they're not letting on.
  
• Is a PIC 16F877 a reasonable thing to use these days? It looks OK to me, or is there a reason to not use it? I want serial output not USB, I am looking at using this project as a ready made PCB kit and rewriting the program from the ground up to get rid of the TASM dependencies. I don't know what the EPE designer was on but writing Micropchip assembler in anything other than MPASm seems downright weird to me.
  
• Any alternative single channel PIC dataloggers out there capable of sampling every second for up to a day? 8-bit resolution would do me, though 10 is nice. I want rs232 out, and I don't want to program in C or basic or anything proprietary. A datalogging multimeter for < £60 would do the job but I've never seen one. I don't want to run a PC all the time, the aim is to reduce power usage here

My application is to characterise the current drawn/put into a battery fed from a solar panel to get a feeling for what sort of things I can run.

I was going to hack the software which is available from **broken link removed** under 8-channel analogue data logger, basically swapping out the binary out for an ascii rendition I could catch as text with hyperterminal and munge in Excel.

Unfortunately, the program is written in some bizarre variant of non Microchip MPASM, and after correcting the obvious differences and reprogramming the chip has bricked the Magenta datalogger, though the reset function and text worked so it is soft-bricked not hard-bricked If I can get that working with replacement code then the Magenta kit is a nice source of instant PCB kit and parts for a reasonable price, even though I think it is iffy to flog a product that requires a PC that is over 10 years old without warming people up to that dependency...

 

[Nigel Goodwin]

You can convert the wierd assembler (TASM) to MPASM using the PIC Toolkit on the EPE website.

TASM was used because the author (incredibly bizarrely) had never heard of MPASM until I brought it to his attention a long time back now.

It should be pretty easy to rewrite to use a more sensible serial format - but if I recall correctly, the stored data is actually packed in the EEPROM to make more room - as ten bits doesn't fit very well in 8 bits. So instead of wasting 6 bits of every 16, the data is packed to utilise the space. I presume it's transformed in that form, in order to make the transfer faster?

 

[Wp100]

Hi,

Well you must be one of the few like me and Nigel who still prefer Assembler and also do not want to do everything via USB.

I think you would find doing the project afresh with the 16F877, or the easier to use 18F4520, would be very rewarding and give you a system you can easily add onto later.
Nigels Assembler tutorial has examples of many of the functions you will be needing eg rs232, ADC , Displays etc

Data storage is something you will have to consider - so larger external eeprom chips may be the best option, connected via spi or I2c; many would say use an sd card but have seen virtually no complete assembler examples.

 

[Nigel Goodwin]

Data storage is something you will have to consider - so larger external eeprom chips may be the best option, connected via spi or I2c; many would say use an sd card but have seen virtually no complete assembler examples.

The EPE example uses banks of I2C EEPROM's.

 

[ermine]

Thanks guys, ideas much appreciated.

You can convert the wierd assembler (TASM) to MPASM using the PIC Toolkit on the EPE website.

I did try that, but it barfs bigtime on my PC, even after loading the VB6 runtimes into system32. Though you have given me an idea there, which is to run it on an old W98 laptop to see if I can win the HEX file. The code is pretty rough, but I have identified the section I want to break out. From what I gathered, the data is sent as a high byte low byte pair from two file registers hand over hand implying it's been demultiplexed within the PIC, but:

So instead of wasting 6 bits of every 16, the data is packed to utilise the space. I presume it's transformed in that form, in order to make the transfer faster?

That could explain why I couldn't make any sense of the data dump, and favours the toss the old code and treat the kit as a hardware kit. I can live with wasting the six bits per 16 for an easy life

and also do not want to do everything via USB

USB drivers, yack, proprietary stuff that becomes unserviceable after 10 years . Plus another application is at a remote site, I might want to TX the data out and I don't want to lug computers around. OK, I admit it, I'm a bit old skool there

Sounds like so far the 16F877 hasn't been deprecated/EOL'ed like the 16F84, so persevering with the existing hardware means I can get another Magenta kit if I want another one. The price is right for the parts and saving a hardware design PCB layout job. I have more confidence in the hardware design than the curious software...