Looks to me like Brymen didn't so much as add a serial protocol as they just simply added an IR LED to the data line that goes to the LCD controller. That's why it is a "weird" protocol. It was never designed to be used that way. It was designed to only transfer a "bit mapped" image to the LCD controller. When you power the meter up for RS232 (see 3:54) it likely only changes the bit rate to 9600 bps from the optimal rate it normally starts up in. This way they can say it has serial PC connectivity but it only cost less than a buck to add the IR LED and virtually nothing else to allow that mode. This "RS232" mode could very well be a test mode designed into the main meter IC to allow easy factory testing and/or prototyping without the LCD. IMHO.
+ElmerFuddGun - That being said... looking at the HY2613 LCD controller data sheet, it would be possible to monitor the SCL and SDA (pin 15 and 16) signals with a micro such as a PIC and add an IR LED (cut the traces to D13 and add jumpers to the PIC). Should then be easy to convert the serial LCD data to 9600 bps data format that the BM257 already has or better yet, output it as a proper serial string using ASCII instead. Might be a fun project!
+ElmerFuddGun A good idea would be to add some code to this MCU to add smart hold capability (as in flukes). Just attach an extra button to the meter somewhere, and make controller read the display. After the press of that button the MCU will then wait until display readings stabilize, and short hold button pad via mosfet.
+SaNjA2659 - Sure but that's getting a fair bit harder to do than my relatively simple IIC to serial converter suggestion. The hold button MOSFET mod wouldn't be too hard (find and scrape the PCB traces and solder the wires) but drilling a hole in the case and hacking in a button gets messy and hard to do without looking like crap. And I'm not a fan of a mod like that to the case that may compromise safety. It also gets a bit more complicated because you might need to use a PCB (or protoboard) now that you have added a MOSFET where as the simpler converter suggestion could likely be done with just the PIC glued to the meter PCB (dead bug style) and using a few jumper wires and a single resistor to the IR LED pad. That being said I would be interested in seeing yours if you decide to do the mod.
Almost every multimeter IC with optional RS232 support uses a protocol like this. It's not *literally* tapped off the LCD controller. It's just simpler to implement it this way, and it's kind of universal.
Are you sure about you conclusion? Could it be that pin 4 is configured as voltage input and the same trace is connected to the led pad? Where does the other pad connect to? Look like it vanishes in a via some way down.
Really unconvincing reasoning there. That empty pad just happens to be located exactly in the same spot where the IR led should be? And it being empty just happens to be a happy coincidence? and it being right next to the LCD controller, emitting LCD protocol, another happy coincidence? Just probe the damn thing and be sure! maybe it's connected to a different pin in through a different layer of the PCB? maybe there's an undocumented mode to that pin? not probing it to check is just lazy.
What about the other pin of the diode ? I mean you seem to say it "connect to a rail" and show the fat trace with a via, but AFAICT, it doesn't connect to that fat trace at all, it connects to a small trace right next to it.
+Sylvain Munaut You are right. I must be going blind. It does to the other diode below it, and then into a via. Doesn't make sense to be an IR LED. The BM257 has a drive transistor for the LED.
+Sylvain Munaut yeah! I just commented about that too... I've tried to trace it on an image (using the other side available at farm2.staticflickr.com/1699/24180627643_bea96ffbe5_b.jpg and farm2.staticflickr.com/1448/24689654432_726f1ff591_b.jpg) but got lost beneath U4.... edit: nevermind... I zoomed in and saw that it went to the other diode as well...
+Diego Spinola I traced and beeped the bottom diode pad to the left pin of J1 and a bunch of pins on U4, U5, before I gave up. BTW, I couldn't find data sheet for U4/U5.
For multimeters like the UT61E, IR comm mode is configured by pulling a uC pin low or disabled by leaving it floating (internal pullup). Maybe this chipset also needs a pulldown/pullup on a pin to fully enable IR.
just a guess: but maybe on the DMM with the IR out they are just abusing the serial out of the display controller that is normally meant to be daisy chained with another one. that would explain the strange "protocol". If this is true then the IR-Signal should not be modulated like it normally shoudl be on IR-transmissions. Maybe something similar can be found on the other one but it seems that it is a I²C based controller. But maybe it is possible to just hook a IR-LED to the SDA-Line (via some transistor) and recover the clock-signal from the data to reconstruct the transmitted data.
@EEVblog2 : Dave could you check the other pin on the Diode? It might be used to drive the IR LED with inverted logic? I tried to trace it on the images of the back of the PCB(front of the multimeter) but got lost beneath U4...
I have a fluke 73 series 1. Been trying to figure out the protocol. This video is great for testing a different angle of attack. My UNI-T UT81B has a data USB cable. They made it to show up as a HID device and not a serial port. This makes it harder to sniff the data traffic.
looks like you have no luck with the lcd ctrl chip but there is still a chance with the multimeter chipset, if that was known and a datasheet could be found you might get lucky with that one outputting some serial data.
What a strange circuit. Could they be modulating the VLCD charge pump voltage to drive the IR LED? Glancing at the data sheet yields some interesting but vague tidbits. Some black magic in here. How does the BM257 drive it's LED?
+Justin “J” Lynn It just so happens that the 80Hz refresh rate on the display (page 19 of the English Datasheet) is 9600 baud / 120 and the dispay data is referenced to VLCD and VDD.
+Justin “J” Lynn Looking at the VDD trace from the display driver, it looks like it comes around back through vias and connects in just to the right of the C14 silkscreen mark with a thin little trace. ... hmm, probe time.
I have the Greenlee DM-510A, which is the same as the Brymen 257. I also have their PC cable. I’d be interested in checking out that protocol but the link is broken :/ any way someone could upload that to an alternate host if they have it?
Another way to try getting serial type data out of the BM235 would be to attach some small circuit to the I2C pins to drive an LED. Possible perhaps but not that clean a solution.
Please, just to amuse me... hook up the damn scope, it only takes a second... Also look to see if there is a 24c0x EEprom on the board, if so, there is a good chance that the calibration info is in the 24c02, along with a register, one bit of which determines if the chip is programmed for serial output or not. Toggle that bit and the serial will be available.
... and one final rant... the protocol is very similar to a whole bunch of other multi-meters. Almost industry standard I would suggest. Multimeter chips have pretty basic microcontrollers, so simply firing out the contents of the LCD registers is by far and away the easiest method of sending out the serial data if RAM/ROM is at a premium on your microcontroller. Let the PC host do all the thinking for you.
I feel like you're not exploring this fully. Clearly, that pin is not it. But have you looked elsewhere? Trace where the IR LED is connected on the BM257 and see if the chipset on the BM235 output anything there. Assuming the cipsets have the same chip form factor.
The charge pump side might be the power supply side. Indeed what about the other side of the led? That side might me pulled to gnd and contain data. just solder on a 1k resistor (to allow a voltage change) and measure the other side.
Why does it output the LCD pattern? It's not a true serial output. It's purpose is for manufacturing diagnostics. Since having a camera in the right place to observe the LCD segments they output a digital stream that indicates which segment is lit during testing and calibration. Thus verifying proper operation of the firmware.
+Kevin Bhasi - LOL... "BM235 vs BM235"... well you never know with some electronics and brands they can actually differ with the exact same model number! 8-P
I wonder if that weird arse protocoll was originally meant to be for transferring data to a wireless display without using a microcontroller and having to invent a protocol etc. That would make it easy. Just whack the data into the remote display without changing a bit and that's it. Thumbs up to a 6 minute video, shortest one for a long time ;)
Sorry but I may have misset something here. But what are the point with this video? You start with what if and conclude no not possible anyway and then nearly no comparation except a talk about a unknown chip and orientation of two soldering pads. "Dave please explain" :-)
Well looking at the datasheet for the LCD driver, you could read the 4 COM lines - the data from these could be used to build the required values for the protocol. Not nearly as simple as the original idea.
Looks to me like Brymen didn't so much as add a serial protocol as they just simply added an IR LED to the data line that goes to the LCD controller. That's why it is a "weird" protocol. It was never designed to be used that way. It was designed to only transfer a "bit mapped" image to the LCD controller. When you power the meter up for RS232 (see 3:54) it likely only changes the bit rate to 9600 bps from the optimal rate it normally starts up in. This way they can say it has serial PC connectivity but it only cost less than a buck to add the IR LED and virtually nothing else to allow that mode. This "RS232" mode could very well be a test mode designed into the main meter IC to allow easy factory testing and/or prototyping without the LCD. IMHO.
+ElmerFuddGun - That being said... looking at the HY2613 LCD controller data sheet, it would be possible to monitor the SCL and SDA (pin 15 and 16) signals with a micro such as a PIC and add an IR LED (cut the traces to D13 and add jumpers to the PIC). Should then be easy to convert the serial LCD data to 9600 bps data format that the BM257 already has or better yet, output it as a proper serial string using ASCII instead. Might be a fun project!
+ElmerFuddGun A good idea would be to add some code to this MCU to add smart hold capability (as in flukes). Just attach an extra button to the meter somewhere, and make controller read the display. After the press of that button the MCU will then wait until display readings stabilize, and short hold button pad via mosfet.
+SaNjA2659 - Sure but that's getting a fair bit harder to do than my relatively simple IIC to serial converter suggestion. The hold button MOSFET mod wouldn't be too hard (find and scrape the PCB traces and solder the wires) but drilling a hole in the case and hacking in a button gets messy and hard to do without looking like crap. And I'm not a fan of a mod like that to the case that may compromise safety. It also gets a bit more complicated because you might need to use a PCB (or protoboard) now that you have added a MOSFET where as the simpler converter suggestion could likely be done with just the PIC glued to the meter PCB (dead bug style) and using a few jumper wires and a single resistor to the IR LED pad. That being said I would be interested in seeing yours if you decide to do the mod.
You should be able to use the High Low function to get the same result.
Almost every multimeter IC with optional RS232 support uses a protocol like this. It's not *literally* tapped off the LCD controller. It's just simpler to implement it this way, and it's kind of universal.
put the probe in to check at least
Can't you trace where the IR LED connects and find the same component pin on the BM235?
Are you sure about you conclusion? Could it be that pin 4 is configured as voltage input and the same trace is connected to the led pad? Where does the other pad connect to? Look like it vanishes in a via some way down.
Really unconvincing reasoning there. That empty pad just happens to be located exactly in the same spot where the IR led should be? And it being empty just happens to be a happy coincidence? and it being right next to the LCD controller, emitting LCD protocol, another happy coincidence? Just probe the damn thing and be sure! maybe it's connected to a different pin in through a different layer of the PCB? maybe there's an undocumented mode to that pin? not probing it to check is just lazy.
Welcome to EEVblog!
What about the other pin of the diode ? I mean you seem to say it "connect to a rail" and show the fat trace with a via, but AFAICT, it doesn't connect to that fat trace at all, it connects to a small trace right next to it.
+Sylvain Munaut You are right. I must be going blind. It does to the other diode below it, and then into a via. Doesn't make sense to be an IR LED. The BM257 has a drive transistor for the LED.
+Sylvain Munaut yeah! I just commented about that too... I've tried to trace it on an image (using the other side available at farm2.staticflickr.com/1699/24180627643_bea96ffbe5_b.jpg and farm2.staticflickr.com/1448/24689654432_726f1ff591_b.jpg) but got lost beneath U4....
edit: nevermind... I zoomed in and saw that it went to the other diode as well...
+Diego Spinola I traced and beeped the bottom diode pad to the left pin of J1 and a bunch of pins on U4, U5, before I gave up. BTW, I couldn't find data sheet for U4/U5.
For multimeters like the UT61E, IR comm mode is configured by pulling a uC pin low or disabled by leaving it floating (internal pullup). Maybe this chipset also needs a pulldown/pullup on a pin to fully enable IR.
just a guess: but maybe on the DMM with the IR out they are just abusing the serial out of the display controller that is normally meant to be daisy chained with another one. that would explain the strange "protocol". If this is true then the IR-Signal should not be modulated like it normally shoudl be on IR-transmissions. Maybe something similar can be found on the other one but it seems that it is a I²C based controller. But maybe it is possible to just hook a IR-LED to the SDA-Line (via some transistor) and recover the clock-signal from the data to reconstruct the transmitted data.
@EEVblog2 : Dave could you check the other pin on the Diode? It might be used to drive the IR LED with inverted logic? I tried to trace it on the images of the back of the PCB(front of the multimeter) but got lost beneath U4...
It doesn't seem to be connected to the positive RAIL!
Dave did you mean to title it BM235 vs BM257? It says 235 vs 235.
+Gud1 M8 I do believe so...
I have a fluke 73 series 1.
Been trying to figure out the protocol. This video is great for testing a different angle of attack.
My UNI-T UT81B has a data USB cable. They made it to show up as a HID device and not a serial port.
This makes it harder to sniff the data traffic.
looks like you have no luck with the lcd ctrl chip but there is still a chance with the multimeter chipset, if that was known and a datasheet could be found you might get lucky with that one outputting some serial data.
Was trying to figure out why it displayed a `C' on the LCD after powering on with the `HOLD' button depressed.
What a strange circuit. Could they be modulating the VLCD charge pump voltage to drive the IR LED? Glancing at the data sheet yields some interesting but vague tidbits. Some black magic in here. How does the BM257 drive it's LED?
+Justin “J” Lynn It just so happens that the 80Hz refresh rate on the display (page 19 of the English Datasheet) is 9600 baud / 120 and the dispay data is referenced to VLCD and VDD.
+Justin “J” Lynn Looking at the VDD trace from the display driver, it looks like it comes around back through vias and connects in just to the right of the C14 silkscreen mark with a thin little trace. ... hmm, probe time.
Open Drain controlled diode from the other pin?
Do you know what the purpose of the J1 header is? Might be for testing or external power, but looks like it could be used for comms.
I have the Greenlee DM-510A, which is the same as the Brymen 257. I also have their PC cable. I’d be interested in checking out that protocol but the link is broken :/ any way someone could upload that to an alternate host if they have it?
Another way to try getting serial type data out of the BM235 would be to attach some small circuit to the I2C pins to drive an LED. Possible perhaps but not that clean a solution.
Is there some place I can leave a good review for your multimeter delivery?
Hook a damn oscilloscope on those pads... You have plenty of them so SEE if it drives a led or not.
Please, just to amuse me... hook up the damn scope, it only takes a second... Also look to see if there is a 24c0x EEprom on the board, if so, there is a good chance that the calibration info is in the 24c02, along with a register, one bit of which determines if the chip is programmed for serial output or not. Toggle that bit and the serial will be available.
... and one final rant... the protocol is very similar to a whole bunch of other multi-meters. Almost industry standard I would suggest. Multimeter chips have pretty basic microcontrollers, so simply firing out the contents of the LCD registers is by far and away the easiest method of sending out the serial data if RAM/ROM is at a premium on your microcontroller. Let the PC host do all the thinking for you.
Take a look here if you want to learn a bit more... sigrok.org/wiki/Multimeter_ICs#Cyrustek_ES51922
I feel like you're not exploring this fully. Clearly, that pin is not it. But have you looked elsewhere? Trace where the IR LED is connected on the BM257 and see if the chipset on the BM235 output anything there. Assuming the cipsets have the same chip form factor.
Where is the sensor for picking up the electric field ? Tracking data says my meter left Sidney ! Thank you for the video ! tjl
Let's see... Nothing a camera pointed at the LCD and raspberryPI running some OCR code that converts it to serial data, could not fix. Lol
The charge pump side might be the power supply side. Indeed what about the other side of the led? That side might me pulled to gnd and contain data. just solder on a 1k resistor (to allow a voltage change) and measure the other side.
Why does it output the LCD pattern? It's not a true serial output. It's purpose is for manufacturing diagnostics. Since having a camera in the right place to observe the LCD segments they output a digital stream that indicates which segment is lit during testing and calibration. Thus verifying proper operation of the firmware.
In fact there my not even be a LCD hooked up to the PCB while the ATE is calibrating the meter.
please fix the title
+Kevin Bhasi - LOL... "BM235 vs BM235"... well you never know with some electronics and brands they can actually differ with the exact same model number! 8-P
I orderd my BM235 yesterday and had the same idea. Because I saw the unequipped footprint on the teardown photos. =P
maybe it does have a secret serial output hack it just doesn't have anything to do with the pad.
I wonder if that weird arse protocoll was originally meant to be for transferring data to a wireless display without using a microcontroller and having to invent a protocol etc. That would make it easy. Just whack the data into the remote display without changing a bit and that's it. Thumbs up to a 6 minute video, shortest one for a long time ;)
Sorry but I may have misset something here. But what are the point with this video?
You start with what if and conclude no not possible anyway and then nearly no comparation except a talk about a unknown chip and orientation of two soldering pads.
"Dave please explain" :-)
Well looking at the datasheet for the LCD driver, you could read the 4 COM lines - the data from these could be used to build the required values for the protocol. Not nearly as simple as the original idea.
This video needs renaming Dave. Thanks for this tho!
High margin junk, promoted by Dave.
Bless your heart.
👍
BM235 vs BM235
Hahahha...epic at the end you said...it doesn't work...LOL..but Nice try..
Messy construction. Sorry. I refer my Fluke.
Hack something else ;)