Opinions: Generating negative bias voltage

[jpanhalt]

My GLCD requires a separate negative bias supply. It needs to be sequenced on after Vdd is on and off before Vdd is off. Initially, I am using a Max851 that has a shutdown pin, mainly because I had one in my parts boxes. It requires 4 capacitors plus the chip. I am unclear of its decay time when it is shut down, but will measure it.

Eventually, I will consider using a PWM pin from my 16F1519 chip to power a charge pump of two capacitors and two diodes. I may also include a negative voltage switch or drain resistor to switch/drain the capacitors producing Vee. New Haven Displays says having Vee applied without Vdd will destroy its displays.

What is your opinion on using a simple charge pump versus a dedicated chip for that function? Has anyone here used a negative-voltage switch to control the sequence?

John

 

[Ian Rogers]

You surprise me.... Newhaven display's normally have their own Vee generator on board ( The ones I use have )... I have also used a display with no Vee generation, I used the max851(-17v) I didn't give the timing a second thought.... I think the Vee has to be there for a bit to damage anything... The on time isn't a problem as the Vee caps have to charge (as you said ), but placing a 100k from Vee to earth may be a good idea.

 

[jpanhalt]

I am using the NHD-12864MZ (https://www.electro-tech-online.com/custompdfs/2012/02/NHD-12864MZ-FSW-GBW-L.pdf ).

All of the datasheets and safe-handling instructions related to NHD's devices without on-board Vee show a timing diagram. This application note is quite explicit on the question:

Source: https://www.electro-tech-online.com/custompdfs/2012/02/LCD_Contrast_Voltage_Circuits.pdf

The order in which the power supplies are applied to an LCD, power sequencing, must be considered when designing an LCD bias power supply. The power sequencing requirement can be summarized by stating that theV (V ) must never be present withoutV also being present. If this condition exists, even for a short period of time, the display may be permanently damaged. The desired power on sequencing for graphics type LCDs with an external controller is shown in Figure 6. For graphics type LCDs with a built in controller you can ignore the “signal” line as this is taken care of in the controller at power on time. For character displays only the V andV (V ) lines need be considered.

I was concerned that the timing (particularly the off sequence) would be difficult in practice to duplicate. I wrote a note to technical service asking for clarification, particularly for the term "for a short period of time," and received this reply from a Mr. Khan:

Excerpt from e-mail to tech service:

1) What sequence is required for start up? What is best?

You should first apply VDD, then V0 (supply for contrast).

2) What sequence, if any, is required for shut down? In particular, are there any requirements with respect to display signals?

When turning the display OFF, it is recommended to first turn off V0 and then VDD. However, you can turn both of them off at the same time. The display signals are handled by the on-board controller during power on/off.

As a result of those technical recommendations/instructions, I decided to provide and/or control both Vdd and Vee/Vo for the GLCD from a pins on the PIC (current is less than a mA, except for the LED backlights). That way, they can be turned off in sequence before pulling the plug, so to speak. I was/am concerned that if one simply pulls the plug, Vee may stay low longer than Vdd is high due to slow discharge from its capacitors.

John

 

[Ian Rogers]

As I said... I never had an issue.. My systems turn off and on all day... as long as the max851 is powered from the same rail, they should work perfectly.

As a side note, if you put a scope on all the rails you'll see that V0 is referenced against the Vdd rail not the ground, they dissipate together quite uniformed.

 

[jpanhalt]

Yes, I will be looking at the decay curves.

Do you control the Max851 with a pin or just tie the enable high?

John

 

[Ian Rogers]

High... Same rail as the LCD power.

 

[jpanhalt]

Oops. I raced back here to correct my error. I believe tieing the shutdown high will shut down the 851 chip. At least that is how I interpreted the datasheet. I have never used one. Of course, I will test that too. CORRECTED BELOW

On re-reading for the n^th time, it seems tieing high activates the chip. I was reading "SHDN" as meaning disable, not enable.

John

 

[Mike - K8LH]

A 128x64 display that requires an additional power supply? Yuch!

John, may I ask how much your display cost? How does it compare to $36 for this **broken link removed**?

Regards...

 

[jpanhalt]

$25 at DigiKey (http://search.digikey.com/us/en/products/NHD-12864MZ-FSW-GBW-L/NHD-12864MZ-FSW-GBW-L-ND/1701268). It completed a small order to get me over the minimum. Actually, I didn't know anything about displays and posted on ETO before ordering. I suspect there must be a reason to leave the Vee out, but I haven't found it yet. Maybe it is in my wallet?

I tend to go with DigiKey when I can. It gives excellent, fast service and got shafted by a consent decree to charge sales tax on everything. It is like rooting for the underdog, if you will.

But back to my original question, is there any reason not to use a PWM pin and a couple of diodes and capacitors instead of a stand alone chip for generating the bias? When I simulated it, I used a 330 Ω resistor at the input. There were spikes to 13 mA at the resistor for each pulse, but they were very short lived. Average current was only a couple of mA; voltage was -3.96V @1200 Hz. Output looked decent.

Noise from those spikes is one thing that concerns me.

John

 

[jpanhalt]

I guess it is too late to edit/correct post#7. By experiment, my Max851 is OFF when SHDN (pin4) is tied high and ON when it is tied low.

Double-negatives are quite confusing.

John