That program looks good, Sid. I'll try it out this afternoon. With your device, it might help to include user prompts, i.e.,
"Testing BS2--"
"Ready for you to connect the tester to pins p0--p7 now"
PAUSE ' then test 0-7
"Ready for you to move the tester to pins p8--p15 now"
PAUSE ' then test 8-15
Don't forget the BS2e and BS2sx in the second SELECT statement.
It would be simpler for the protocol just to make a 16 pin tester!
I don't know, I think for my own hackerish mind, I would keep the program stripped down KISS and show the raw numbers for each of the pins. Let me decide for myself whether it passes or fails. The raw numbers are self-documenting, in that bad pins stand out like a sore thumbs among the ones that are behaving properly. Maybe you need to distribute both the turnkey version and the hacker version!
With regard to possible interaction between pins, the result would depend on how a pin has failed. With the round robin tester for example, if pin p3 fails open, there is no effect on the test of p4. However, if p3 happens to fail as a short circuit to Vdd, then p2 and p4 will also appear to fail. The "groundless" tester is actually more robust in response to that situation. We just have to remember that there are lots of different ways that a pin can "fail".
Re your optoscope results. I am puzzled why the p0 voltage does not go all the way back up to Vdd at the end after the "RCTIME 0,0,wy". It looks like it maxes out at about 4 volts, and the same thing at the beginning of the trace, why is it at 4 volts and not 5 volts?
Tracy, you always come up with improvements to my perfect programs !
I added the suggested prompts, plus I included a "Press any key when ready" thing.· I have attached Rev C.
Regarding the Optascope display, I do not really understand it either.· It is almost as though the pins were floating until they are taken low or high by the program.· You set DIRS and OUTS but the state of the pins is never set until the program does it.· I was hoping you would come up with your usual brilliant analysis.
I'm not suggestinng the program will perform any better, it will simply follow Parallax formatting guidelines that many customers find easy to learn from.· I'll also take advantage of Conditional Compilation so that it will work the same on any BS2 module.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Jon Williams
Applications Engineer, Parallax
Dallas Office
So Tracy, have the three of you decided which test circuit works best?
The original R/C Network from pin to Vss, the R/C Network with a common bus but not tied to Vss, or the Round-Robin R/C Network...?· I'm seeing benefits to each, but no definitive answer on which would be best.· As I said, I am waiting for a final solution to build my own...I ordered a bunch of Caps & Resistors, since 1KΩ, 10KΩ & 100KΩ are my most commonly used here, and I am almost out...Same with .1 µF, .01µF & .001µF Caps...Well, for Cer Disc anyway...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage
Knight Designs 324 West Main Street P.O. Box 97 Montour Falls, NY 14865 (607) 535-6777
If I interpreted Tracy's previous comments correctly, I think he is inclined toward the RC network with the ground bus but not tied to Vss, mainly because of its robusticity.· I'm satisfied with it·and it is the design I'll use.
I don't think it· makes much difference whether you use 10K/.1uf or 100K/.01uf.
I IM'ed you but you weren't there [noparse]:([/noparse]
Overall, I think the groundless configuration that Sid's serendipity discovered will be the best bet. "R/C Network with a common bus but not tied to Vss". However, when you construct it, leave a tie point where you can attach a jumper to Vss or Vdd if you need to. There might be rare situations where making that connection would help to resolve a problem (But if a Stamp is that far gone, you'd better convert it into a refrigerator magnet anyway!)
As to component values, 10K/.1uf or 100K/.01uf or 1M/0.001, it should not matter. I prefer the last one, because it is easier to get _small_ components that are also good precision. The 220 ohm pin protection resistor is also a good idea, if it is something that people are going to touch with their fingers (having shuffled across a carpet). Also, the 220 ohm protection resistor is a good idea with the 0.1uf capacitors, because the protection diodes have to absorb some current from the cap when the pins are driven to opposite states.
Jon, did I understand correctly that you were going to re-do my program?
Sid
When you finalized the technical elements (including the test circuit), I will reformat using "The Elements of PBASIC Style" that we [noparse][[/noparse]mostly] adhere to at Parallax.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Jon Williams
Applications Engineer, Parallax
Dallas Office
Thanks, the two testers arrived in the mail yesterday. The one with the 0.1uf monolithic capacitors and 10k resistors certainly looks a lot less "clunky" than the one with the _huge_ 0.01uf film capacitors and 100k resistors.
Are the small capacitors X7R ceramic? That is usually +/- 10% tolerance. I wouldn't recommend anyone use Z5U or Y5V ceramic capacitors, because their tolerance and temperature coefficient is awful. COG ceramic would be best, at 5% tolerance, but it is only available readily in sizes to 0.001uf, which would take a 1Mohm resistor to get the same time constant.
For the resistors, it would be possible to save yourself some soldering and board space by using one SIP resistor pak, with 9 resistors in one. E.g., Digikey part # 770-101-R10k. But the board has a nice feel as it is. (schranderheid correcte intelligenter?). Regular resistors, you can buy them anywhere.
Do you have a lot of these boards to sell, and I wonder, do you plan to sell them assembled, or as a kit?
Oh, with regard to the ~3.8 volts output from the optascope trace you posted a few messages back. We were wondering why it did not get all the way up to 5 volts.
That might be explained by the input impedance of the optascope. It forms a voltage divider with the test circuit, and having the optascope connected would change the time constant.
Glad you got the boards, Tracy.· Thought you might like the little one better.
Thought about the SIP resistors but they are too wide to fit the board.· I only have 8 columns of holes on .100 centers.
I used Y5V for prototyping but am switching to the X7R for the finished product.· You are correct - they are +/-10%.· Would like to have 5% but they are not available in a .1uf.
My grandaughter took some pictures of the tester for me.· I changed my avatar and also have a picture on my webpage, so we will see what happens.· I've been told the price is too low, but it's good enough for me.
Didn't do anything bout the ground wire.· The pad is there but I could not think of any possible reason someone might need a ground - as you said, if the Stamp is that bad, frame it!
·· Sid, I know you were anxious to see the results of my new BS2 Tester...For those that forgot, I had built a 16 pin tester on a Parallax Proto Board, but as I am now aware they are discontinued, I didn't want to waste my board on just a tester.· So I took it apart.· Today my spare parts (More Caps) came in, and I was able to re-build the unit.
·· I had been hoping for 10% tolerance caps, but alas, they sent me 20%...SO, I decided to go ahead and build it anyway.· Like Sid's & Tracy's prototypes, this one is build on a Solder-Ring Perf Board.· The one I used is Radio Shack Cat# 276-150.· A little larger than necessary, but the board is $1.75, so no biggie.
·· Tracy, I took one of your suggestions this time and used a SIP resistor pack (I have tons of them), instead of the individual resistors I used before.· I also switched from my original specs to Sid's specs for the resistors and capacitors, using .1 μF Caps and 10KΩ resistors, instead of the .01 μF Caps and 100KΩ resistors I was using before.· No particular reason, except continuity, I guess.
·· The only difference with mine is, like before, it tests all 16 pins without changing (Call me lazy).· And Tracy, there is a facility for connecting a wire lead to the target test board into Vdd or Vss, if I should choose to.· The connector has only 16 pins, and plugs into the X2 header on the Board Of Education, and into X4 and X5 (independently) on the BS2P40 Demo Board, allowing me to test every type of BS2, and ALL pins.· Even the AUXIO pins on the BS2P40 (Which I have the most of).
·· Attached are 3 pics...·· 1) Picture of the tester (Minus it's wire lead).·· 2) Picture of the tester mounted on a Rev. A BOE.·· 3)·Screenshot of test results from BS2 on said BOE.· And by the way, Sid's code works great!· Thanks to Sid, Tracy & Jon for the info.· Enjoy!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage
Knight Designs 324 West Main Street P.O. Box 97 Montour Falls, NY 14865 (607) 535-6777
My tester was built on a PC board about 1 inch square.
I looked at your readings - you have about a 14% spread, which if you are using 10% components is still in range.· However, I still maintain that even if you have .01% components you are going to get a spread which will vary from pin to pin.· You got a high reading on pin 15.· Can you reverse your tester 180 degrees and see if the high reading follows the pin or follows the tester.
On my little PC board tester, I had a spread of about the same, and I was using +80/-20 caps.
Anyway the thingie works, and it is a good quick diagnostic tool that is handy to have around.
·· I reversed the tester, and attached a screenshot of the results...Also I wanted to say, for anyone who isn't sure they could build their own tester, I would recommend one of Sid's units...The code works, and the tester could save you pulling out your hair down the road.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage
Knight Designs 324 West Main Street P.O. Box 97 Montour Falls, NY 14865 (607) 535-6777
I ran the program, StampTesterRevC, using both the tester you sent me and also my round-robin tester . Below are the results from 5 different stamps (BS2 to BS2p, one of each tested):
(A)=(0.1uF Y5U caps & 10kohm 1% resistors) groundless
(B)=(0.001uF 2% film caps & 1M 1% resistors) round robin
One thing that jumps out is how different the numbers for the BS2p are. They are much higher than the other Stamps. (Remember, the program is using a multiplier appropriate for each Stamp to convert to microseconds, so they should all be the same.) Maybe the BS2p I happened to use was a fluke, but I see that Chris posted very similar numbers for his BS2p, also high. I wonder if we have the correct multiplier for the BS2p? Or maybe something else makes them higher. There are differences among the other Stamps, but as far out as the BS2p numbers.
Second observation: The tester with higher precision components does in fact result in a tighter spread of values. Yes, there are going to be differences from pin to pin on one Stamp, but the tester components are a big factor.
BS2 MAX MIN SPREAD MAX MIN SPREAD
(A) 1440 1314 126/10% 356 332 24/7%
(B) 1294 1280 14/1% 338 334 4/1%
BS2sx MAX MIN SPREAD MAX MIN SPREAD
(A) 1342 1215 125/10% 357 326 24/7%
(B) 1224 1204 14/1% 363 357 4/1%
BS2e MAX MIN SPREAD MAX MIN SPREAD
(A) 1324 1202 122/10% 348 324 24/7%
(B) 1204 1190 14/1% 356 352 4/1%
BS2pe MAX MIN SPREAD MAX MIN SPREAD
(A) 1339 1218 125/10% 347 326 24/7%
(B) 1198 1204 14/1% 352 349 4/1%
BS2p MAX MIN SPREAD MAX MIN SPREAD
(A) 1606 1455 151/10% 439 405 34/8%
(B) 1450 1431 14/1% 440 436 4/1%
Maybe change this line in the program....
IF wx>1200 and wy>300 THEN
to
IF wx>1100 AND wx<1700 AND wy>250 AND wy<500 THEN
That clamps down the acceptable range from both sides. I gave it a little more tolerance. I ran into a couple of "fail" messages that should not have been there because the number was just under 1200 or under 300.
Tracy, the numbers from my BS2P are about the same as your BS2P (b) readings, maybe a tad lower.· I did have a couple of high readings so I reversed the tester 180 degrees and the high readings followed the tester so I figured it was the tester components.· I've got some 10% caps on order and it will be interesting to compare the numbers from both testers on the same Stamp.
I thought about clamping the high and low readings but decided not to do it because a faulty component - or a bad solder connection - might indicate a failure when the pin was actually OK.
From what I get, it appears that the time increments for the different Stamps are pretty close to correct.· I get about the same readings from the BS2, BS2E, BS2pe and the BS2p.· Maybe I can verify this by observing the duration of pulsout 100 on all Stamps with my Optascope.· I should get 200us for the first three and 80us for the BS2P.
·· My screenshots of test results were from a standard BS2, not a BS2P...However, if you would like, I do have several BS2P40's and I could run the test on one of them...In fact, as soon as I get a free moment today I will do that.· I have had an experiment running on my BS2P40 Demo Board for the last week, but it's pretty much done, so I will run the test in a little while.· I will post my results on here.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage
Knight Designs 324 West Main Street P.O. Box 97 Montour Falls, NY 14865 (607) 535-6777
I just built another tester, this one with SMD resistors.· On my BS2P24 I got a range of 1475 to 1540, which is about a 4% spread.· From this, it would appear that the timing·period -.8us - of the BS2P varies from chip to chip.
Chris, from putting them side by side, it looks like your result on the BS2p is very close to the BS2. I'll try it again here with a couple more Stamps when I get a chance. Which version of BS2p do you have? Mine was old, probably a version 1.2, if that makes any difference.
On to the RCTIME command, the units of period are
2.0 microseconds for the '2 and '2e
0.8 microseconds for the 2sx
0.9 microseconds for the '2p
1.89 microseconds for the '2pe
That is what it says in black and white in the help file, but I don't know if I trust it, in particular, the 1.89 figure for the BS2pe. This should be exactly 2.5 times the value for the BS2p, which would make it 2.25 microseconds, not 1.89. Or if the BS2pe figure is right, then the BS2p figure should be 1.89/2.5 = 0.756.
On the subject of uncritical acceptance of the numbers, the 0.9 figure for the '2p and 1.89 for the '2pe.... The PULSOUT command is the same on both the '2p and the '2sx. that is 0.8 microseconds per period. However, on RCTIME, the documentation makes the '2p slower than the '2sx, at 0.9 microseconds vs 0.8 microseconds, whereas for the PULSIN command the '2p is faster, 0.75 microsecond vs.0.8 microsecond. Go figure. I guess that is where they landed when the loops where programmed in turbo mode.
I heard that one of the tweaks in the version 1.4 BS2p is a small change in the timing of the PULSOUT command, to make "PULSOUT pin,1" have exactly 1/2 the time of "PULSOUT pin,2". The '2p unit of period should be 0.8 microseconds, the same as the BS2sx. The first loop had been a few mciroseconds short. {The manual (version 2.0b) says 1.18 microseconds, which was way wrong.} Anyway, Sid, if you measure these on your Optoscope, please also do RCTIME directly by framing it with a HIGH-LOW on another pin.
Tracy,·my BS2P40 version is 1.3, however, I do have about 7 of these, and I have only been using the one on the Demo Board for Prototyping...I could see what the others are and test one of them for any differences.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage
Knight Designs 324 West Main Street P.O. Box 97 Montour Falls, NY 14865 (607) 535-6777
I tried to capture the RCTime on my Optascope but it ws just too fast.· So I switched to my HP 275mHz scope.· From 5VDC down to 1.4 - charge time - I measured about 1150us.· From 0 to 1.4 - discharge time - I measured 320us.· These realtime figures translate to 1277 and 355, which is pretty close to what I see on my on my BS2P using the tester and the #Select/#Case parameters for the different Stamps.· Not as precise as I had hoped, but I don't think I can do any better.
Comments
"Testing BS2--"
"Ready for you to connect the tester to pins p0--p7 now"
PAUSE ' then test 0-7
"Ready for you to move the tester to pins p8--p15 now"
PAUSE ' then test 8-15
Don't forget the BS2e and BS2sx in the second SELECT statement.
It would be simpler for the protocol just to make a 16 pin tester!
I don't know, I think for my own hackerish mind, I would keep the program stripped down KISS and show the raw numbers for each of the pins. Let me decide for myself whether it passes or fails. The raw numbers are self-documenting, in that bad pins stand out like a sore thumbs among the ones that are behaving properly. Maybe you need to distribute both the turnkey version and the hacker version!
With regard to possible interaction between pins, the result would depend on how a pin has failed. With the round robin tester for example, if pin p3 fails open, there is no effect on the test of p4. However, if p3 happens to fail as a short circuit to Vdd, then p2 and p4 will also appear to fail. The "groundless" tester is actually more robust in response to that situation. We just have to remember that there are lots of different ways that a pin can "fail".
Re your optoscope results. I am puzzled why the p0 voltage does not go all the way back up to Vdd at the end after the "RCTIME 0,0,wy". It looks like it maxes out at about 4 volts, and the same thing at the beginning of the trace, why is it at 4 volts and not 5 volts?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jon Williams
Applications Engineer, Parallax
Dallas Office
I added the suggested prompts, plus I included a "Press any key when ready" thing.· I have attached Rev C.
Regarding the Optascope display, I do not really understand it either.· It is almost as though the pins were floating until they are taken low or high by the program.· You set DIRS and OUTS but the state of the pins is never set until the program does it.· I was hoping you would come up with your usual brilliant analysis.
Sid
Sid
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jon Williams
Applications Engineer, Parallax
Dallas Office
Sid
The original R/C Network from pin to Vss, the R/C Network with a common bus but not tied to Vss, or the Round-Robin R/C Network...?· I'm seeing benefits to each, but no definitive answer on which would be best.· As I said, I am waiting for a final solution to build my own...I ordered a bunch of Caps & Resistors, since 1KΩ, 10KΩ & 100KΩ are my most commonly used here, and I am almost out...Same with .1 µF, .01µF & .001µF Caps...Well, for Cer Disc anyway...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Knight Designs
324 West Main Street
P.O. Box 97
Montour Falls, NY 14865
(607) 535-6777
Business Page:·· http://www.knightdesigns.com
Personal Page:··· http://www.lightlink.com/dream/chris
·
I don't think it· makes much difference whether you use 10K/.1uf or 100K/.01uf.
I IM'ed you but you weren't there [noparse]:([/noparse]
Sid
As to component values, 10K/.1uf or 100K/.01uf or 1M/0.001, it should not matter. I prefer the last one, because it is easier to get _small_ components that are also good precision. The 220 ohm pin protection resistor is also a good idea, if it is something that people are going to touch with their fingers (having shuffled across a carpet). Also, the 220 ohm protection resistor is a good idea with the 0.1uf capacitors, because the protection diodes have to absorb some current from the cap when the pins are driven to opposite states.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Sid
Sid
I only know the Dutch word for serendipity, it is "schranderheid".
Now you will be able to find the meaning in english
Klaus
Sid
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jon Williams
Applications Engineer, Parallax
Dallas Office
Sid
Thanks, the two testers arrived in the mail yesterday. The one with the 0.1uf monolithic capacitors and 10k resistors certainly looks a lot less "clunky" than the one with the _huge_ 0.01uf film capacitors and 100k resistors.
Are the small capacitors X7R ceramic? That is usually +/- 10% tolerance. I wouldn't recommend anyone use Z5U or Y5V ceramic capacitors, because their tolerance and temperature coefficient is awful. COG ceramic would be best, at 5% tolerance, but it is only available readily in sizes to 0.001uf, which would take a 1Mohm resistor to get the same time constant.
For the resistors, it would be possible to save yourself some soldering and board space by using one SIP resistor pak, with 9 resistors in one. E.g., Digikey part # 770-101-R10k. But the board has a nice feel as it is. (schranderheid correcte intelligenter?). Regular resistors, you can buy them anywhere.
Do you have a lot of these boards to sell, and I wonder, do you plan to sell them assembled, or as a kit?
Oh, with regard to the ~3.8 volts output from the optascope trace you posted a few messages back. We were wondering why it did not get all the way up to 5 volts.
That might be explained by the input impedance of the optascope. It forms a voltage divider with the test circuit, and having the optascope connected would change the time constant.
-- Tracy
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Post Edited (Tracy Allen) : 8/23/2004 1:14:25 AM GMT
Thought about the SIP resistors but they are too wide to fit the board.· I only have 8 columns of holes on .100 centers.
I used Y5V for prototyping but am switching to the X7R for the finished product.· You are correct - they are +/-10%.· Would like to have 5% but they are not available in a .1uf.
My grandaughter took some pictures of the tester for me.· I changed my avatar and also have a picture on my webpage, so we will see what happens.· I've been told the price is too low, but it's good enough for me.
Didn't do anything bout the ground wire.· The pad is there but I could not think of any possible reason someone might need a ground - as you said, if the Stamp is that bad, frame it!
Sid
·· Sid, I know you were anxious to see the results of my new BS2 Tester...For those that forgot, I had built a 16 pin tester on a Parallax Proto Board, but as I am now aware they are discontinued, I didn't want to waste my board on just a tester.· So I took it apart.· Today my spare parts (More Caps) came in, and I was able to re-build the unit.
·· I had been hoping for 10% tolerance caps, but alas, they sent me 20%...SO, I decided to go ahead and build it anyway.· Like Sid's & Tracy's prototypes, this one is build on a Solder-Ring Perf Board.· The one I used is Radio Shack Cat# 276-150.· A little larger than necessary, but the board is $1.75, so no biggie.
·· Tracy, I took one of your suggestions this time and used a SIP resistor pack (I have tons of them), instead of the individual resistors I used before.· I also switched from my original specs to Sid's specs for the resistors and capacitors, using .1 μF Caps and 10KΩ resistors, instead of the .01 μF Caps and 100KΩ resistors I was using before.· No particular reason, except continuity, I guess.
·· The only difference with mine is, like before, it tests all 16 pins without changing (Call me lazy).· And Tracy, there is a facility for connecting a wire lead to the target test board into Vdd or Vss, if I should choose to.· The connector has only 16 pins, and plugs into the X2 header on the Board Of Education, and into X4 and X5 (independently) on the BS2P40 Demo Board, allowing me to test every type of BS2, and ALL pins.· Even the AUXIO pins on the BS2P40 (Which I have the most of).
·· Attached are 3 pics...·· 1) Picture of the tester (Minus it's wire lead).·· 2) Picture of the tester mounted on a Rev. A BOE.·· 3)·Screenshot of test results from BS2 on said BOE.· And by the way, Sid's code works great!· Thanks to Sid, Tracy & Jon for the info.· Enjoy!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Knight Designs
324 West Main Street
P.O. Box 97
Montour Falls, NY 14865
(607) 535-6777
Business Page:·· http://www.knightdesigns.com
Personal Page:··· http://www.lightlink.com/dream/chris
My tester was built on a PC board about 1 inch square.
I looked at your readings - you have about a 14% spread, which if you are using 10% components is still in range.· However, I still maintain that even if you have .01% components you are going to get a spread which will vary from pin to pin.· You got a high reading on pin 15.· Can you reverse your tester 180 degrees and see if the high reading follows the pin or follows the tester.
On my little PC board tester, I had a spread of about the same, and I was using +80/-20 caps.
Anyway the thingie works, and it is a good quick diagnostic tool that is handy to have around.
·· I reversed the tester, and attached a screenshot of the results...Also I wanted to say, for anyone who isn't sure they could build their own tester, I would recommend one of Sid's units...The code works, and the tester could save you pulling out your hair down the road.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Knight Designs
324 West Main Street
P.O. Box 97
Montour Falls, NY 14865
(607) 535-6777
Business Page:·· http://www.knightdesigns.com
Personal Page:··· http://www.lightlink.com/dream/chris
I ran the program, StampTesterRevC, using both the tester you sent me and also my round-robin tester . Below are the results from 5 different stamps (BS2 to BS2p, one of each tested):
(A)=(0.1uF Y5U caps & 10kohm 1% resistors) groundless
(B)=(0.001uF 2% film caps & 1M 1% resistors) round robin
One thing that jumps out is how different the numbers for the BS2p are. They are much higher than the other Stamps. (Remember, the program is using a multiplier appropriate for each Stamp to convert to microseconds, so they should all be the same.) Maybe the BS2p I happened to use was a fluke, but I see that Chris posted very similar numbers for his BS2p, also high. I wonder if we have the correct multiplier for the BS2p? Or maybe something else makes them higher. There are differences among the other Stamps, but as far out as the BS2p numbers.
Second observation: The tester with higher precision components does in fact result in a tighter spread of values. Yes, there are going to be differences from pin to pin on one Stamp, but the tester components are a big factor.
BS2 MAX MIN SPREAD MAX MIN SPREAD
(A) 1440 1314 126/10% 356 332 24/7%
(B) 1294 1280 14/1% 338 334 4/1%
BS2sx MAX MIN SPREAD MAX MIN SPREAD
(A) 1342 1215 125/10% 357 326 24/7%
(B) 1224 1204 14/1% 363 357 4/1%
BS2e MAX MIN SPREAD MAX MIN SPREAD
(A) 1324 1202 122/10% 348 324 24/7%
(B) 1204 1190 14/1% 356 352 4/1%
BS2pe MAX MIN SPREAD MAX MIN SPREAD
(A) 1339 1218 125/10% 347 326 24/7%
(B) 1198 1204 14/1% 352 349 4/1%
BS2p MAX MIN SPREAD MAX MIN SPREAD
(A) 1606 1455 151/10% 439 405 34/8%
(B) 1450 1431 14/1% 440 436 4/1%
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Maybe change this line in the program....
IF wx>1200 and wy>300 THEN
to
IF wx>1100 AND wx<1700 AND wy>250 AND wy<500 THEN
That clamps down the acceptable range from both sides. I gave it a little more tolerance. I ran into a couple of "fail" messages that should not have been there because the number was just under 1200 or under 300.
I'm puzzled by the higher numbers from the BS2p.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
I thought about clamping the high and low readings but decided not to do it because a faulty component - or a bad solder connection - might indicate a failure when the pin was actually OK.
From what I get, it appears that the time increments for the different Stamps are pretty close to correct.· I get about the same readings from the BS2, BS2E, BS2pe and the BS2p.· Maybe I can verify this by observing the duration of pulsout 100 on all Stamps with my Optascope.· I should get 200us for the first three and 80us for the BS2P.
Sid
·· My screenshots of test results were from a standard BS2, not a BS2P...However, if you would like, I do have several BS2P40's and I could run the test on one of them...In fact, as soon as I get a free moment today I will do that.· I have had an experiment running on my BS2P40 Demo Board for the last week, but it's pretty much done, so I will run the test in a little while.· I will post my results on here.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Knight Designs
324 West Main Street
P.O. Box 97
Montour Falls, NY 14865
(607) 535-6777
Business Page:·· http://www.knightdesigns.com
Personal Page:··· http://www.lightlink.com/dream/chris
·
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Knight Designs
324 West Main Street
P.O. Box 97
Montour Falls, NY 14865
(607) 535-6777
Business Page:·· http://www.knightdesigns.com
Personal Page:··· http://www.lightlink.com/dream/chris
Comments. Tracy?
Sid
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Knight Designs
324 West Main Street
P.O. Box 97
Montour Falls, NY 14865
(607) 535-6777
Business Page:·· http://www.knightdesigns.com
Personal Page:··· http://www.lightlink.com/dream/chris
·
On to the RCTIME command, the units of period are
2.0 microseconds for the '2 and '2e
0.8 microseconds for the 2sx
0.9 microseconds for the '2p
1.89 microseconds for the '2pe
That is what it says in black and white in the help file, but I don't know if I trust it, in particular, the 1.89 figure for the BS2pe. This should be exactly 2.5 times the value for the BS2p, which would make it 2.25 microseconds, not 1.89. Or if the BS2pe figure is right, then the BS2p figure should be 1.89/2.5 = 0.756.
On the subject of uncritical acceptance of the numbers, the 0.9 figure for the '2p and 1.89 for the '2pe.... The PULSOUT command is the same on both the '2p and the '2sx. that is 0.8 microseconds per period. However, on RCTIME, the documentation makes the '2p slower than the '2sx, at 0.9 microseconds vs 0.8 microseconds, whereas for the PULSIN command the '2p is faster, 0.75 microsecond vs.0.8 microsecond. Go figure. I guess that is where they landed when the loops where programmed in turbo mode.
I heard that one of the tweaks in the version 1.4 BS2p is a small change in the timing of the PULSOUT command, to make "PULSOUT pin,1" have exactly 1/2 the time of "PULSOUT pin,2". The '2p unit of period should be 0.8 microseconds, the same as the BS2sx. The first loop had been a few mciroseconds short. {The manual (version 2.0b) says 1.18 microseconds, which was way wrong.} Anyway, Sid, if you measure these on your Optoscope, please also do RCTIME directly by framing it with a HIGH-LOW on another pin.
-- Tracy
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Knight Designs
324 West Main Street
P.O. Box 97
Montour Falls, NY 14865
(607) 535-6777
Business Page:·· http://www.knightdesigns.com
Personal Page:··· http://www.lightlink.com/dream/chris
·
Sid
My BS2P is V1.3