EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: motocoder on September 01, 2014, 02:05:53 am
-
I purchased a couple of 40x2 character-mode Noritake VFDs off of eBay. My intent is to use them to replace an LCD in a small project my son and I are working on.
99.9% of the time, the display should be off. It is only when someone is interacting with the device, triggered by press of a large pushbutton, that I want the display to come on. The device is powered with a wall wart, so while battery life is not a concern, I don't want to waste power if not necessary. In addition, I know these VFD have a limited lifespan, and I'd like to maximize that.
Having not worked with a VFD before, I am not sure what I should do when the device is idle, and the VFD should be blank. Should I just use a MOSFET to switch power to the VFD? Or is just sending a blank display string sufficient?
Thanks
-
VFDs can last a long time -- most VCR screens display the time (albiet dimmed) for years happily. Make sure you are driving the filament with an AC waveform (you can use an inexpensive motor controller and a 555 timer to do this) to ensure molecules don't migrate only in one direction.
The string-interpreter probably only interfaces with the grid, not the filament, so sending a blank string will probably not turn the filament current off. IIRC this is the main life-limited part of a VFD [citation needed].
If the screen only needs to stay on for about 5 mins at a time, you could make a simple capacitor and comparater timer circuit. When the display is supposed to turn on, charge up a cap. Slowly drain this capacitor with a resistor. Compare the voltage against a potentiometer-set voltage, and while the cap is still charged enough, keep the filament driver powered.
-
The Noritake VFD is completely self-contained; all I have to do is supply +5V, GND, and serial data in.
Regarding a timeout, the micro controls all that - no need for any sort of RC based timer. I'm already doing that for the backlight on the LCD which the VFD is replacing.
So really my question is, should I put the extra few components in the circuit to allow the microcontroller to switch on/off power to the VFD module. It sounds like the answer to that is "yes", but I am just wanting someone more familiar with VFD to confirm.
-
So really my question is, should I put the extra few components in the circuit to allow the microcontroller to switch on/off power to the VFD module. It sounds like the answer to that is "yes", but I am just wanting someone more familiar with VFD to confirm.
Apologies - I don't have any long-term VFD experience to tell you whether or not it is worth it :D. Just in-case the VFD does already power itself off when displaying nothing, have you tried measuring its power consumption?
-
So really my question is, should I put the extra few components in the circuit to allow the microcontroller to switch on/off power to the VFD module. It sounds like the answer to that is "yes", but I am just wanting someone more familiar with VFD to confirm.
Apologies - I don't have any long-term VFD experience to tell you whether or not it is worth it :D. Just in-case the VFD does already power itself off when displaying nothing, have you tried measuring its power consumption?
No worries, I appreciate the opportunity to think it through with someone in any case. This modification to my project is more about me learning how a VFD works then anything else (although the high contrast display will be quite convenient), so some experimentation is fine.
I have not done any power measurements yet, but I will do that next. I did discover there is a 4-level brightness control, but i suspect that the filament current stays flowing continuously, and it's just PWM-ing the grid or segment (anode) voltage. I think given that the cathode current is essentially heating an element, it's not realistic for the device to shut that down when the display is blank. There would be some significant latency to heat the filament back up, and the VFD controller has no way of knowing when the display is going to become non-blank since that's an externally initiated event.
-
Most VFDs have a stepup converter that generates the high voltages for the anodes (depending on display size somewhere in the range of 30-70V). It also has a winding for powering the filament. Therefore it is unlikely that it can be switched off completely by software.
These stepup converters have a high inrush current, so use either a slewrate limited switch, or add some capacitors to the power rail when switching the module off completely.
The lifetime of VFDs is limited by 2 factors:
- The segments itself
- The filament
If a segement runs at high brightness level, the phosphor gets burned up like in any crt monitor This can be avoided by reducing the brightness or moving the display content around.
The filament (or more exactly, its electron emitting coating) loses its ability to emit electrons over time. A too low filement temperature leads to rapid aging. This is a main problem with vcr displays when the capacitors dry up und the filament voltage gets lower. If there is no current flowing, it will also degrade the electron emitting capability. Blanking the displays for longer times is therefore no usable solution. Switching the display (and its filament) on and off frequently also results in accelerated aging due to thermal cycling.
To get the best lifetime, display varying content and switch off when not used for a longer time (some hours).
-
A too low filement temperature leads to rapid aging. This is a main problem with vcr displays when the capacitors dry up und the filament voltage gets lower. If there is no current flowing, it will also degrade the electron emitting capability. Blanking the displays for longer times is therefore no usable solution.
This sounds like the cathode interface resistance issue which is well known in vacuum tubes when the heater is powered but the tube is held in cutoff for long periods of time like would happen in digital logic. The first vacuum tube digital circuits suffered from this and they developed special cathodes which did not have this problem.
The same problem crops up in CRTs when the user blanks the display for long periods using the intensity control in a misguided attempt to extend the CRT life and prevent phosphor burn. The solution is to cut or lower heater power as well. Typically a tube filament would be operated at half voltage to extend the life and also keep the warmup time short.
As far as extending the lifetime of the vacuum fluorescent display, since you only have access to the module and not the display itself, I think it will come down to either removing the +5 volt supply or blanking or operating at the minimum brightness level when not being used.
If the warmup time is insignificant and the display will not be needed for long periods of time (days or more), then I would switch the +5 volt supply. Otherwise I would blank the display or operate it at the lowest intensity level.
-
The device is a food (treat) dispenser for a pet. So it is typically triggered at most a few times a day, and in the interim the display could be off.
Regarding the slew rate, the Noritake data sheets all mention potential startup problems if the voltage ramp-up is too slow. So my thought was to use a big bypass capacitor on the source rail, followed by a p-channel mosfet gating this supply to the VFD +5v input. This would buffer the micro somewhat from the impact of the inrush current to the VFD, and provide the VFD with the fast ramp-up it wants.
When the user presses the local "dispense" button on the machine, the display will be switched on. Once on, it will remain that way for a minute or two, and perhaps I'll come up with some movement of characters to prevent burn-in. This is a significant concern because most of the displayed text is constant. After a few minutes of inactivity, the mosfet will be switched off and the power to the VFD cut.
-
Regarding the slew rate, the Noritake data sheets all mention potential startup problems if the voltage ramp-up is too slow. So my thought was to use a big bypass capacitor on the source rail, followed by a p-channel mosfet gating this supply to the VFD +5v input. This would buffer the micro somewhat from the impact of the inrush current to the VFD, and provide the VFD with the fast ramp-up it wants.
I would be be really careful doing this. Using a fast switch between the regulated output and the load can cause significant overshoot do to inductive ringing so if the VFD module is sensitive to voltage spikes on its supply line, it may be damaged. The old Flash memories which required switched external programming voltages suffered from major problems with this:
http://www.linear.com/docs/4203 (http://www.linear.com/docs/4203)
http://www.linear.com/docs/4124 (http://www.linear.com/docs/4124)
Turning a 5 volt regulator on and off via shutdown should provide safer results and be fast enough. Alternatively, include a bulk decoupling capacitor like a 10 microfarad aluminum electrolytic or slightly smaller solid tantalum at the VFD which is probably a good idea anyway.
When the user presses the local "dispense" button on the machine, the display will be switched on. Once on, it will remain that way for a minute or two, and perhaps I'll come up with some movement of characters to prevent burn-in. This is a significant concern because most of the displayed text is constant. After a few minutes of inactivity, the mosfet will be switched off and the power to the VFD cut.
I bet there is an application note which discusses how to use VFDs for long operating life somewhere.
-
Regarding the slew rate, the Noritake data sheets all mention potential startup problems if the voltage ramp-up is too slow. So my thought was to use a big bypass capacitor on the source rail, followed by a p-channel mosfet gating this supply to the VFD +5v input. This would buffer the micro somewhat from the impact of the inrush current to the VFD, and provide the VFD with the fast ramp-up it wants.
I would be be really careful doing this. Using a fast switch between the regulated output and the load can cause significant overshoot do to inductive ringing so if the VFD module is sensitive to voltage spikes on its supply line, it may be damaged. The old Flash memories which required switched external programming voltages suffered from major problems with this:
http://www.linear.com/docs/4203 (http://www.linear.com/docs/4203)
http://www.linear.com/docs/4124 (http://www.linear.com/docs/4124)
Turning a 5 volt regulator on and off via shutdown should provide safer results and be fast enough. Alternatively, include a bulk decoupling capacitor like a 10 microfarad aluminum electrolytic or slightly smaller solid tantalum at the VFD which is probably a good idea anyway.
When the user presses the local "dispense" button on the machine, the display will be switched on. Once on, it will remain that way for a minute or two, and perhaps I'll come up with some movement of characters to prevent burn-in. This is a significant concern because most of the displayed text is constant. After a few minutes of inactivity, the mosfet will be switched off and the power to the VFD cut.
I bet there is an application note which discusses how to use VFDs for long operating life somewhere.
I can control the turn-on rate of the MOSFET if need be. Regarding the app note, I did look through the Noritake web site for something like that, but did not find it. I posted a message on their FaceBook page - will see if I get a reply there, and if not I'll send an email to the sales engineer who helped me previously,
-
Turning a 5 volt regulator on and off via shutdown should provide safer results and be fast enough.
It depends on the regulator. I had problems using a 5V 2A stepdown regulator supplying enough current for a VFD to start. If the voltage regulator has some foldback characteristic, it does not even start up.
Maybe an OLED display is better suited for the application: It can easily switched on and off without any problems and consumes much less power.
-
Maybe an OLED display is better suited for the application: It can easily switched on and off without any problems and consumes much less power.
Yes, other than the expense/size, it would no doubt be better. Still, it's just a one-off project, and I have two of these VFD that I got relatively cheap off of eBay. I may just leave the VFD on and spend the extra time on some sort of a balancing / screen saver routine for the software. The Noritake site lists fairly long life for the display, with the only caution being about uneven display wear across segments.