Supertex HV5522 supply current

[Pengouin]

Hi,

I intend to use Supertex HV5522 to build a Nixie clock with four IN-18 tubes. I found some schematics on the Internet. Some use 5V directly connected to HV5522 and some use higher voltages (12V) with a field effect transistor connected to a 5V MCU. I plan on using a PIC MCU running at 3.3V to drive my HV5522. The datasheet seems to indicate that I can use any voltage between -0.5V to +15 V, but the recommended voltage is 10V to 13V.

Does anyone uses HV5522 connected directly in 3.3? Is it working correctly?

Thanks for your help

Antoine

[Terry S]

The datasheet does not indicate an operating supply voltage of -0.5 to
+15v, those are the absolute maximum ratings.

Use the part within it's recommended ratings of 10.8 to 13.2 volts and
you won't have any issues. Otherwise, all bets are off.

It never ceases to amaze me when people use a part outside it's spec
and then wonder why they have problems. Would you use a lawnmower to
trim your hedge? Your weed whip to cut down a tree?

Terry

[Antoine Vanoutryve]

Hi Terry.

Indeed I found only one schematic using 12V​​. Most of schematics available on Google use 5V. That's why I preferred to ask. I will use 12V with transistors then.

Thank you for your reply.

Antoine

> --
> You received this message because you are subscribed to the Google Groups "neonixie-l" group.
> To post to this group, send an email to neoni...@googlegroups.com.
> To unsubscribe from this group, send email to neonixie-l+...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/neonixie-l?hl=en-GB.
>

[Adam Jacobs]

That is interesting. My guess is that there are lots of lazy engineers
out there that have figured out that the part works OK on 5v. That's why
experimentation is so important. If I was going to make a WAG, I'd bet
that it works at 5v although probably at slower speeds. Try it with 5v
if it is a huge savings in parts for your. Maybe it will work. If it
doesn't, then change it to 12v. (shrug) I'm assuming this isn't a
mission critical device or something that you are planning to sell, of
course.

I think that most people here don't have much experience with the
Supertex parts, although they do look neat. I have a tube of them but
haven't had a chance to try them out.

-Adam

[Terry S]

The part is fabbed in HVCMOS, so yes, it will continue to function at
some level down to where the CMOS transistor structures no longer
switch -- probably well below 3 volts. But...

The problem with using a part like that outside its spec is that you
simply can't predict what parameters won't be met. What if it affects
set-up or hold times? What if it affects output drive or input
impedance? There are simply so many things that can go wrong --
internal timing conditions that might not be met. As a hobbyist, you
might be forgiving of a misbehavior if it doesn't manifest itself
visibly on your clock display, or you find a work around for a timing
problem -- but what if this were a life critical medical or industrial
application, or a mission critical military app? As an engineer or
designer, your butt would be on the line should you use a part outside
its specs and someone gets hurt or dies, or your drone crashes...

Speaking of drones, Obama is in town today, about 6 blocks from here,
speaking at Honeywell. I bet I can't even go out for lunch today
without running into his entourage. Gonna be another vending machine
lunch I guess.

Terry

[Adam Jacobs]

I understand your point, Terry.. Which I think is the classic "Slippery
Slope" argument. You get used to cutting corners, before you know it
you're trying the same tricks at work where you design nuclear reactor
cooling systems, the systems fail due to your design choices that were
outside of specification, etc.

Luckily for me, I take mission criticality into account when I make a
design. A nixie clock project (for me) is foremost a fun chance to try
new things and sandbox some experimentation. Disclaimer: I don't sell my
clocks and I also don't use them in mission critical applications such
as drones or heart pumps. If something fails because of a design choice
that I made, then I make a note of it and change the experimental
design. The fact is that mission criticality is just another input in
the requirements phase of the design. Things like component count, size,
feature set, hazard analysis. Take the time to come up with some real
product requirements before you start drawing a schematic. Makes things
a lot easier.

-Adam W7QI

[Tobias]

Antoine

I am using the HV5522 on my clock and multimeter projects. Using both
SPI (for the HV5522) and I2C bus level shift.
The voltage for the HV5522 is the same that powers the boost
converter. I tested both 9 and 12V and it works just fine.
I based my design on this app note:

http://ics.nxp.com/support/documents/interface/pdf/an97055.pdf

The document describes a bi-directional level shifter but it works
just fine when communication is uni-directional like with the HV5522.
For the FETs I chose BSS138 as they are inexpensive.

Regards
Tobias

>  signature.asc
> < 1KVisualizarFazer download

[GastonP]

On Jun 1, 12:54 pm, Adam Jacobs <a...@jacobs.us> wrote:
> I understand your point, Terry.. Which I think is the classic "Slippery
> Slope" argument. You get used to cutting corners, before you know it
> you're trying the same tricks at work where you design nuclear reactor
> cooling systems, the systems fail due to your design choices that were
> outside of specification, etc.
>
> Luckily for me, I take mission criticality into account when I make a
> design. A nixie clock project (for me) is foremost a fun chance to try
> new things and sandbox some experimentation. Disclaimer: I don't sell my
> clocks and I also don't use them in mission critical applications such
> as drones or heart pumps.>

I'm more concerned when the cut corners are in "Friend or Foe"
detection systems :D

[Adam Jacobs]

The ED-209 scene in the Robocop movie comes to mind...

[Antoine Vanoutryve]

Hi Tobias,

This document is very interesting. Thanks. The design I saw on the internet (http://ad7zj.net/kd7lmo/hardware/nixie_page2.gif) is slightly different (he also use BSS138). I think this one is not bidirectional.

Even if it is not necessary, I prefer to use your design.

Antoine

[Tobias]

Antoine

Should have said you need to work with the resistors values to make it
work nice.
If you have an oscilloscope it is easy to look at the waveforms.
I did not test it for very high speeds yet.

I saw this design before (http://ad7zj.net/kd7lmo/hardware/
nixie_page2.gif) and I tried it already. It works fine. Only issue is
that the signals they get inverted. I.E. a high logic level on the
3,3V line will appear like a low logic level on the 12V line. If you
are programming the protocol you can work around this, but it is not
as simple as just using a library.

May I ask you a question? Are you going to use it to direct drive the
nixies or multiplex?

Tobias

[Antoine Vanoutryve]

Hi Tobias,

As suggested by Dieter,  I will not use multiplexing (You will find my current design below). I'll probably use the SPI library (if possible) to send data so your design is definitely better.

Antoine