Sv: [CAD_CAM_EDM_DRO] DRO interface circuit
Posted by
J. J. Larsen
on 2002-10-14 14:44:56 UTC
Hi Larry
The 2 "outer" pins are plus and minus of battery and the 2 "inner" pins are
clock and data.
The output is a serial bitstream of 48 bit. The first 24 bit is an absolute
value with an offset value, the next 24 bits is a relative value. I have
found this information in the files section (long time ago), search for
"Hacking digital
caliper" or something like that. See below.
Jens
Denmark
<Date: Wed Dec 13, 2000 6:52 am
<Subject: Re: Hacking Cheap Digital calipers...
<oops had the 'section' descriptions reversed. below is corrected:
<Well, I figured out the serial format of cheap digital caliper I got
<on Ebay. Hopefully this will be a start to people wanting to make
<DRO's from this type of caliper and maybe the format is a common one
<(ie like Digimatic)
<It has and SPI interface with clock and data. The data "burst" has 2
<sections of 24bits data equaling 48 bits per packet. The first 24bits
<are shifted out on the rising edge and qualified on the falling edge
<by the reading device. Bit timing was approx. 13uS clock cycle
<yielding about 6.5us to detect the falling
<edge and sample the data.
<Section1:
<The first section of 24 bits is also shifted out lsb first. This
<24bit record is an 'absolute' binary count with a constant error
<offset, for example, when the caliper is fully closed (0.000) the
<count was 007FD3h(32723 decimal). By subtracting this value
<from the record then dividing by 5000h(20480decimal) you get
<the 'inch' readout.
<Section2:
<Bits are lsb first of a 24bit record. The 2nd section of 24bits is a
<'relative' 24bit count subtracted from 0xFFFFFF with scaling relation
<of 5000h(20480decimal) counts per inch.
<The mode buttons had no apparent affect on the data records sent out,
<other than freezing the 'relative' section, so the micro in the
<caliper must do a conversion by scaling for displaying the metric
<value.
<I wrote some quick and dirty Pic code to log data in hex format using
<Hyperterm. I cheated and used an emulator.
<I targeted a 16C558 running at 10Mhz, but may convert it to a 8pin
<part. I used 1/2 of an LM339 quad comaparator(cheap) to level shift
<the 1.5v logic to 5v for processing.
<Let me know what you think...
<-Lee
The 2 "outer" pins are plus and minus of battery and the 2 "inner" pins are
clock and data.
The output is a serial bitstream of 48 bit. The first 24 bit is an absolute
value with an offset value, the next 24 bits is a relative value. I have
found this information in the files section (long time ago), search for
"Hacking digital
caliper" or something like that. See below.
Jens
Denmark
<Date: Wed Dec 13, 2000 6:52 am
<Subject: Re: Hacking Cheap Digital calipers...
<oops had the 'section' descriptions reversed. below is corrected:
<Well, I figured out the serial format of cheap digital caliper I got
<on Ebay. Hopefully this will be a start to people wanting to make
<DRO's from this type of caliper and maybe the format is a common one
<(ie like Digimatic)
<It has and SPI interface with clock and data. The data "burst" has 2
<sections of 24bits data equaling 48 bits per packet. The first 24bits
<are shifted out on the rising edge and qualified on the falling edge
<by the reading device. Bit timing was approx. 13uS clock cycle
<yielding about 6.5us to detect the falling
<edge and sample the data.
<Section1:
<The first section of 24 bits is also shifted out lsb first. This
<24bit record is an 'absolute' binary count with a constant error
<offset, for example, when the caliper is fully closed (0.000) the
<count was 007FD3h(32723 decimal). By subtracting this value
<from the record then dividing by 5000h(20480decimal) you get
<the 'inch' readout.
<Section2:
<Bits are lsb first of a 24bit record. The 2nd section of 24bits is a
<'relative' 24bit count subtracted from 0xFFFFFF with scaling relation
<of 5000h(20480decimal) counts per inch.
<The mode buttons had no apparent affect on the data records sent out,
<other than freezing the 'relative' section, so the micro in the
<caliper must do a conversion by scaling for displaying the metric
<value.
<I wrote some quick and dirty Pic code to log data in hex format using
<Hyperterm. I cheated and used an emulator.
<I targeted a 16C558 running at 10Mhz, but may convert it to a 8pin
<part. I used 1/2 of an LM339 quad comaparator(cheap) to level shift
<the 1.5v logic to 5v for processing.
<Let me know what you think...
<-Lee