EEVblog 1699 - uTimer PROJECT

[EEVblog]

Part 1 of the µTimer project, a versatile benchtop lab timer.
Choosing the LCD, and getting two sample LCD's running. One using a demo board, and one that needed to be hacked onto a Raspberry Pi.

00:00 - Benchtop Lab Timer Project
01:19 - It's all about the choice of LCD
03:14 - ERC19624-1 192x64 Transflective 4" LCD
03:44 - Reflective vs Transflective LCD's
05:01 - ER3806 Font Chip
06:06 - Demo Board
09:10 - LCD Bias Angle and Viewing Angle Explained
11:45 - What is THAT? LCD Bleeding?
13:19 - Power consumption measurement
15:48 - Trying another LCD, the ERC19624-6. What difference does a dash number make?
17:15 - Bodging on the LCD connector to the breakout board
20:34 - The pin offsets and lack of routing is annoying...
22:10 - Reprogramming the 8051 demo board, it should be easy, right?
26:32 - Entering The Matrix
29:40 - More soldering
30:33 - The Programmer Software doesn't work without a HEX file...
32:17 - All of that for nothing...
33:40 - Let's try this Arduino / Raspberry Pi ST7525 LCD code
35:36 - Getting the Raspberry Pi running and installing the BCM2835 C library
38:56 - Getting the Make to work on the source code.
40:01 - Make Clean should work...
41:16 - Winner Winner Chicken Dinner
41:10 - Power consumption measurement. This one is supposed to be much lower...
42:58 - Changing the LCD contrast in software
45:10 - Is it better or worse than the first LCD?
46:37 - Conclusion

Links:
https://www.buydisplay.com/4-inch-low-cost-white-192x64-graphic-cog-lcd-display-st7525-spi
https://www.buydisplay.com/4-inch-cog-192x64-display-graphic-lcds-module-ist3020-black-on-white
http://www.airspayce.com/mikem/bcm2835/
Font Chip: https://www.buydisplay.com/download/manual/ER3806-1_Datasheet.pdf
Custom LCD videos: https://www.youtube.com/playlist?list=PLvOlSehNtuHvmK-VGcZ33ZuATmcNB8tvH

[Whales]

Awesome vid   Please look at mounting options next (rear pressure, adhesives, etc); I'm currently having a grumpy time with some small-bezel colour LCDs, but I'm also trying to have them waterproof :| so beyond your scope.

Not being able to run ./lcdtest after running make: this is the fault of unix voodoo and your USB stick having a windows (probably fat32 or exfat) filesystem.  The problem was solved when you moved the files to your raspberry pi's (probably ext4) filesystem.

On windows: a "program" is a file that ends in .exe, .com or similar.  Double-clicking or typing the name of said program will run it.

On unix-likes: a "program" is a file with an "execute permission" set for it.  Files with this permission are typically displayed as green when you type ls

My example:


From your vid:


Fat32 can't store this special bit for each file, so no files are ever executable on fat32 filesystems (technically this isn't always true, but close enough for most users).

Don't ask why it's done this way, it's lots of historical multi-user-unix-mainframe stuff combined with a fear that people will accidentally run programs by typing their names.  That's likely also the reason we are forced to type "./programname" instead of just "programname" to run a program in the current folder -- what if someone put a program called "ls" on a usb stick that was a virus?  Still, very annoying, especially if you're not familiar with the problem or are used to DOS.

[EEVblog]

Not being able to run ./lcdtest after running make: this is the fault of unix voodoo and your USB stick having a windows (probably fat32 or exfat) filesystem.  The problem was solved when you moved the files to your raspberry pi's (probably ext4) filesystem.

Wow, thanks, so it wasn't me.
I was puzzled by this, but copying the files worked so I just carried on.
Not sure what prompted me to copy the files and try again though, just a hunch I guess.

[chilternview]

Why not colour OLED than black&white LCD? I hate theose old fashioned LCD displays.

[Whales]

OLED and monochrome LCDs are magnitudes different in terms of power consumption.  Also a reflective LCD is set and forget, whilst an actively lit screen requires brightness controls (or annoyance).

@3 minute mark: Dave wants this battery powered and to show a clockface on it when not in use.  I suspect this is partly motivated by removing the need for power supplies and worrying about isolation (compare the annoyance of using an oscilloscope vs a handheld multimeter).

[mikeselectricstuff]

If a manufacturer is offering  a RasPi demo for their product, it would save so much time if they just provided a full SD card image, so no RasPi knowledge is needed.

[chilternview]

OLED and monochrome LCDs are magnitudes different in terms of power consumption.  Also a reflective LCD is set and forget, whilst an actively lit screen requires brightness controls (or annoyance).

Fair enough. I'd rather have the higher power consumption and shorter battery life with a nice display, but I guess others may not.

[typematrix]

Hi , 

I have a combined display Raspberry pi library which supports the  ERM19264 / UC1609 controller ,

 I think the UC1609 is compatible with ST7525 but have not looked into it in detail , Someone at buydisplay told me ages ago according to my notes.

Its called Display_Lib_RPI on github.
https://github.com/gavinlyonsrepo/Display_Lib_RPI

Would be interesting to know if it works.

[iMo]

What would be a practical usage of such an uTimer in the lab? Perhaps - in a chemistry lab or a similar one lab where you have to cook something for, say, 360 minutes long..
In an EE lab you have to measure or watch time with picoseconds resolution these days, imho.. unless you want to measure the time you need to find and fix a shorted capacitor on a motherboard.. ..

[EEVblog]

What would be a practical usage of such an uTimer in the lab? Perhaps - in a chemistry lab or a similar one lab where you have to cook something for, say, 360 minutes long..
In an EE lab you have to measure or watch time with picoseconds resolution these days, imho.. unless you want to measure the time you need to find and fix a shorted capacitor on a motherboard.. ..

You're not thinking 4th dimentionally.

[EEVblog]

If a manufacturer is offering  a RasPi demo for their product, it would save so much time if they just provided a full SD card image, so no RasPi knowledge is needed.

I don't know why they offer that demo board, seems overly complex having to supply a different programmed MCU and I/O matrix for each display.
Just have an Rpi SD card image as you said, and supply the cable with the adapter board.

[EEVblog]

Why not colour OLED than black&white LCD? I hate theose old fashioned LCD displays.

OLED's are power hogs.

[mikeselectricstuff]

If a manufacturer is offering  a RasPi demo for their product, it would save so much time if they just provided a full SD card image, so no RasPi knowledge is needed.

I don't know why they offer that demo board, seems overly complex having to supply a different programmed MCU and I/O matrix for each display.
Just have an Rpi SD card image as you said, and supply the cable with the adapter board.

It's probably just a very old design they keep using because it's familiar - design resources in China are scarce

[chilternview]

OLED's are power hogs.

Can you quantify that? As it would be good to know. My iphone has a timer/alarm/clock, in colour, so why would anyone want a standalone timer? Jast asking.

[themadhippy]

My iphone has a timer/alarm/clock, in colour, so why would anyone want a standalone timer?

can you plug remote triggers directly into your phone? are you willing to leave yer phone attached too the test bench for sevral hours maybe even a day or 6? how longs the battery in your phone going to last without a charge?what happens when someone calls for a private  chat that means you need to leave the room? shall i go on?

[Doctorandus_P]

There are a few OLED DMM's out there and they tend to have a very short battery life. Like 10 hours or so, and Dave is attempting to get several hundred hours. There are also quite a lot of hacks of replacing OLED displays in these DMM's because they stop working after just a few years.

For the rest...
When I saw the Brannan timer on Dave's desk a few months ago I thought it was a quite cute thing, and assumed it was EOL somewhere in the '80-ies or so, but apparently it's still in production, and with the external inputs it costs around EUR 50:

https://www.brannan.co.uk/product/bench-top-timers-with-external-triggering-contacts/

My guess is that you're not going to get a long (years) battery life out of a high resolution dot matrix LCD, and replacing batteries a few times each year is a nuisance. Running it from a rechargeable bag seems a more sensible option to me. Plugging in an USB cord every now and then on your desk is not such a problem.  Popping in an 18650 reduces "charge time" to less then a minute.

Ultra low power is very nice in some situations, but if it's intended to use in your lab, then size and weight are much less of a concern. Standardizing on 18650 makes the most sense to me. Take the micro Current for example. On itself an interesting project but I would not buy it. Making an 10x 100x 1000x (differential) amplifier is quite simple and makes more sense to me. Pop in two 18650 cells and you get some decent run time even out of a power hungry opamp or inamp.

Also, small 1W to 5W SMPS circuits are cheap. From the DC-DC converters to "Mains to DC" blocks. Or even bog standard USB power supplies. Everybody already got plenty of those within arm's reach.

[Doctorandus_P]

For functionality and feature creep:

A lab bench timer really begs for feature creep. There are already a gazillion stop watches and simple timers, and once you have a uC with a decent display on a PCB with some power, you've already got 80% of the hardware for a very capable timing device.

I'm looking at the front panel of the Tektronix dc509 for inspiration right now.
GHz capability is a bit excessive for a "simple desktop timer", but any decent uC can measure up to a bunch of MHz accurately and timing intervals with better then us accuracy.

Another good source for inspiration are the user manuals of DIN-Rail Mount Multifunction Timer Relays. These have start , switch off and pulse functions. And of course, they have Relay outputs!

I very much like some simple start / stop push buttons such as on the Brannan timer, but a more capable timer surely begs for at least one rotary encoder. Both for setting intervals, and for scrolling through menu's. (A touch screen could work to).

DC509 has adjustable trigger levels. That certainly makes it more universal.

I'm also looking at the W1209. It's a china express cheap thermostat with 3 char of 7-Segment LED display. 3 characters is not much for a timer, but adding thermostat functions to a timer would be very nice.

Blinking a LED an/or a buzzer for when timer intervals or an alarm (temperature) has exceeded a threshold also seems a very sensible extension. When you're designing a product, keeping track of heatsink temperatures is a common task, and so is adding an ad-hoc fan for some additional cooling.

Heck, I'd certainly want a decent built in ADC. 16-bit with differential inputs and a PGA (measuring up to 50V or so) would be nice, but just a 10-bit ADC present in most uC's combined with an opamp and some CMOS switches to extend the range would already be very useful.

Adding a connector with a bunch of I/O pins seems very sensible. It can keep feature creep to a minimum while still allowing for extensions such as thermostats, relays, sensitive ADC front ends (for chemical experiments?) Maybe two connectors: One internal for "fixed modules", and one external for temporary or swap-able modules. (keep pinout compatible).

[Doctorandus_P]

Hacking, Open Source (+ Hardware) + Community?

If this is going to be a real life product. It's very likely go be hacked into and modified by quite a lot of people. To me it makes sense to turn this into a community project right from the start. Community projects need a project leader to give it guidance and a clearly defined scope. Dave can of course design whatever he wants. Define functionality and features etc...

But if the bare hardware is capable enough, then I'm sure plenty of people would want to contribute, and more interest if the hardware can do more. If there is a software framework with menu functions, then individuals can add specialized functions to the menu. If there is some (for example I2C) ADC defined, then people can add software functions to it.

I have thought long and hard of buying a 121GW, and the reason I did not was because the software had to be reverse engineered. In the end I bought an BM869s instead, as for my use, I prioritized accuracy and reliability over functions.

I'm not much use for writing software these days due to concentration problems, but I can put in a bunch of hours in PCB design and layout. Give me some type numbers of an uC and some peripheral IC's and I can work out schematic details. Give me a PCB outline and connector placement and I can fill in the rest. I'll put in some hours for free, but getting a prototype for (proven) hours put into it is an extra motivator. It's not really about the money, it's more about a return of favors and community building.

[SiliconWizard]

I think the Memory LCDs were mentioned, but possibly you rejected that because there wasn't one large enough for your project? There is. Several of us just here have used the 4" version which is 400x240 (IIRC) and has very good contrast and viewing angle. It doesn't seem to be available anymore from Sharp for some reason, but they have a bunch of new references to choose from:

https://www.sharpsecd.com/#/memory-in-pixel-lcds-product

I'd be surprised if you didn't find one that fits.

The downside is the cost, but I don't know if it matters in your project or not.

With that being said, a graphic LCD is great for flexibility, but why not consider designing your own custom LCD for this uTimer project? You've done this before, and I think it gives great results if fixed segments and symbols are fine.

[EEVblog]

I think the Memory LCDs were mentioned, but possibly you rejected that because there wasn't one large enough for your project? There is. Several of us just here have used the 4" version which is 400x240 (IIRC) and has very good contrast and viewing angle. It doesn't seem to be available anymore from Sharp for some reason, but they have a bunch of new references to choose from:

https://www.sharpsecd.com/#/memory-in-pixel-lcds-product

I'd be surprised if you didn't find one that fits.

The downside is the cost, but I don't know if it matters in your project or not.

Yes cost matters, and so does form factor. There just isn't a suitable form factor available either.
The only ones in a suitable size are square form factor and 320x240.

[EEVblog]

If this is going to be a real life product. It's very likely go be hacked into and modified by quite a lot of people. To me it makes sense to turn this into a community project right from the start. Community projects need a project leader to give it guidance and a clearly defined scope. Dave can of course design whatever he wants. Define functionality and features etc...

No thanks. It's a technical dictatorship or bust.

[BrianHG]

Why not colour OLED than black&white LCD? I hate theose old fashioned LCD displays.

OLED's are power hogs.

Maybe e-paper / e-ink displays.
Low-res monochrome wont be as cheap as LCD, around 6$-9$ USD, but very low power and excellent viewing contrast and angles.

Though, for minimum power, you would need to refresh the unit seconds digit every second, then, refresh the rest of the display only as needed.

[EEVblog]

Why not colour OLED than black&white LCD? I hate theose old fashioned LCD displays.

OLED's are power hogs.

Maybe e-paper / e-ink displays.

They are not designed for continuous and fast display updating.

[BrianHG]

Why not colour OLED than black&white LCD? I hate theose old fashioned LCD displays.

OLED's are power hogs.

Maybe e-paper / e-ink displays.

They are not designed for continuous and fast display updating.

No, it will not match the speed of a LCD.

You can get away with the seconds digit being a little blurry until you pause the timer.
You will not get 1/10th of a second to show until you pause the timer.
Since I do not know what else you want to display, I cannot say anything else other than painting black pixels is OK in speed, it is the erasing back to white which is slow.

The black & white ones arent too terrible in speed, it is the displays with shades of grey and multi-color which really need a lot of time to 'erase' the previous frame with multiple refresh passes to draw a new image.

[BrianHG]

Here is what the seconds would look like on a monochrome e-paper display:



It is better than I thought.  1/10 of a second may also be visible.

Note that his code is just erasing just the digit which is changing instead of a full paper clear command.