FEDEVEL
Platform forum

, 01-14-2025, 12:30 AM

Hello, I wanted to try and complete my first hardware project that’s been shelved for a very long time. It’s for a Nixie Clock, and I need help finding out a way to go about creating a level translation from 3.3v from my ESP32-S3-WROOM-1 microcontroller to a high-voltage shift register 12v.

I’m not sure if Im reading into this datasheet correctly, but it looks like the chip im using has some strict requirements and needs a maximum of 1 μA for highs and -1 μA for lows. It also allows clock lines to run at 8MHz so I would need something that keeps up with that too.

I see a lot of people who decide to drive them with a BJT inverter circuit (as shown below) but I’m really confused because whenever I try to simulate it in LTspice it seems like the clock speed is way too fast and the low logic current is nowhere near a microamp.

Would anyone here have any recommendations on how to construct such a driving circuit that meets this criteria? I feel like I may be overcomplicating things or forgetting something and I wanted to ask around.

HV Shift Register Chip Datasheet:
https://www.microchip.com/en-us/product/hv5530

Reference Schematic from Another Project:
https://github.com/GadgetReboot/HV5222_Breakout/blob/main/Sch_HV5222_Breakout.pdf

(This example breakout board uses an arduino hence the 5V vin in the spice simulation)

QDrives , 01-14-2025, 08:49 PM

That the shift register can handle 8MHz, does not mean you have to **do** 8MHz.
You can go a lot slower.

Robert Feranec , 01-15-2025, 06:15 AM

exactly as @QDrives says .... do you really need 8MHz? Is that what you are going to send out from ESP32? If this is used to control the clock segments, you should be able to run it at much lower speed

, 01-16-2025, 01:06 AM

Ah forgive me, you’re both right. I was originally using the HV5222 which is from the same family and they had 8MHz as a recommended condition so I may have gotten the two mixed up. I can definitely pull that down to something as low as the Hz range if its just to change the digits.

Do you also think that I should still be concerned about what I highlighted for the criteria for IH/IL? Specifically with the reversed current for IL.

I don’t believe the behavior of this switching circuit consistently holds a negative value for a low-logic level (There are only negative dips in the transition from high to low as shown above).

QDrives , 01-16-2025, 04:57 PM

That is in nA.
Simulations have these kind of 'problems'. Noting to worry about.

, 01-21-2025, 07:25 PM

ESP32 Nixie Tube Project Review

, 01-21-2025, 07:28 PM

Back again, and I’m at the part where I have a full project architecture laid out.

Currently trying to create a power budget for it and I wanted to know if I’m going about it correctly.

So far, I’ve been building a table of individual components with an estimate of their power consumption.

Since, my design has multiple power sources, according to the conservation of power law wouldn’t I just need to find the total consumption for each sub-rail and then from that, select each power source based on the requirements of their respective loads?

(e.g. Treat Σ 3.3V * Rail Current Draw, Σ 5V * Rail Current Draw, and Σ 12V * Rail Current Draw separately to draft up requirements)

Wanted to see if creating this sort of High-Voltage/Low-Voltage mix power network from a USB-C VBUS was reasonable. Excuse the newbish question, just want to be very certain is all.

QDrives , 01-21-2025, 08:53 PM

For a linear regulator (e.g. LDO) the input current = output current + Iq (quiescent current)
For switching regulators (e.g. fly-back) input power = output power / efficiency.

, 01-21-2025, 11:32 PM

Thanks for looking over again, I really appreciate it.

I’m aware that the regulators/LDO have special equations regarding their power I think I maybe was more trying to ask if the way I calculated the consumption for each power source is correct (P=VI for each component per rail)?

Apologies if I worded it weird.

Like should I just base the pout / iout values in the equations you just mentioned based on the all the current/power consumed for the specific rails each device powers?

QDrives , 01-21-2025, 11:35 PM

As I tried to explain, linear (LDO) is **current** whereas switching (fly-back) is **power**.

QDrives , 01-21-2025, 11:38 PM

So if you are using linear, it does not matter what the power is, the current consumption * input voltage is the power required.
For the 170V and 12V you can use power and efficiency.

, 01-21-2025, 11:49 PM

Okay I think I seem to understand a little bit clearer.

, 01-21-2025, 11:49 PM

If power isn’t relevant for selecting an LDO but current is, for 3.3V I should take the current demands into consideration and power for everything else when searching for components then?

QDrives , 01-21-2025, 11:53 PM

If you consume 340mA on 3.3V, you consume 340mA on 5V.
If you consume 0.1W at 12V, it is 0.1W on 5V.
Neither of these take Iq and efficiency in consideration, nor does it have any other consumers on the voltages.

QDrives , 01-21-2025, 11:55 PM

340mA @ 3.3V = 1122mW, 340mA @5V = 1700mW.
0.1W @ 12V = 8.3mA, 0.1W @ 5V = 20mA

, 01-22-2025, 09:47 PM

Right, but that just suggests that I would need to add the contributions that a power source causes on both sides of the rail to everything else consuming in my budget to keep track of total consumption for each rail/the circuit no?

As you mentioned before, the LDO creates a contribution to the total current drawn on both the 5V rail and 3.3V rail.

Which should be equal to:

Load current on 3.3V rail + (IQ from LDO)

Also shouldn’t the switching regulator(s) downstream from 170V->12V->5V since I have the two cascaded.

So if I were go back to your example, if I wanted 0.1W at the 170V rail, and the datasheet says I have an efficiency of 90% corresponding to whatever load I need on the 170V rail, then I’d need 0.111W coming in from the 12V line into the Flyback.

And then for that 0.111W output from the switching regulator I would have to repeat the calculation to find the power I need from the 5V rail.

QDrives , 01-22-2025, 09:57 PM

Correct.

, 01-22-2025, 10:04 PM

Alright thanks once more! Now back to planning again :)

, 02-16-2025, 01:28 AM

Ok, so I just finished up a preliminary of the full clock schematic for this alongside some of my calculations for the power budget.

The schematic itself is the file named "output" if anyone could take a look to see of some things stick out I would greatly appreciate it!

, 02-16-2025, 01:29 AM

ESP32 Nixie Tube Schematic Review

QDrives , 02-16-2025, 08:04 PM

Where does the +12V come from?

, 02-16-2025, 09:24 PM

From the schematic? I'm trying to create that rail from a PD sink controller IC (bottom of this page)

, 02-16-2025, 09:26 PM

also for some reason, the way I originally printed my schematic misrendered everything and took away all of the negated symbols from some of my signals feeding into the ICs, this should be less confusing to look at (the reference links for some of the designs I incorporated including the pd sink should be there too)

QDrives , 02-16-2025, 09:45 PM

The PD sink controller does not create 12V from 5V now does it?

, 02-16-2025, 09:59 PM

If you’re referring to the 5V line at the start of the connector, I made a mistake but can try to remedy it by adding a buck converter from the 12V rail to 5V so there isn’t a conflict.

, 02-16-2025, 10:00 PM

I probably overlooked it despite being in my original plans :(

, 02-16-2025, 10:04 PM

Or actually alternatively I can use the 5V out on the IC itself

QDrives , 02-16-2025, 10:08 PM

Then the previous question remains: where does the +12V come from?

, 02-16-2025, 10:21 PM

Well, I thought you could use a chip like this by connecting an appropriate PD charger to the receptacle and then having the chip negotiate then ouput power based on a resistor configuration (in my case i set it to provide 12V and 1A).

Is there something else that I’m missing out? I thought it was a simpler chip to use for such an application and I was referencing it’s EVB as a demo.

https://youtu.be/xSR1v241LG0

QDrives , 02-16-2025, 10:29 PM

Just because you have USB C, does not mean that the other end can/will deliver.
If you get 12V, you need to make 5V from that.

In your original block diagram, you had another boost converter from 5V to 12V, right?

QDrives , 02-16-2025, 10:30 PM

Getting 2...3A @~5V is more common/easy than 12V. Even most USB A connector systems will be able to do that.

, 02-16-2025, 10:52 PM

I did have it in my diagram initially and I can try to boost up to 12V from the default 5 @ 3A that I can nab from the 5k1 resistors USB-C CC lines, doing away with the PD sink controller entirely.

But are you suggesting that it’s just unwise to use a power delivery solution altogether in the case that a negotiation fails and shutdowns all of my rails as a consequence?

Wasn’t exactly sure if it was a matter of me not understanding that I couldn’t make 12V the way I setup in my schematic *at all*, and if in that case, I’m curious to know why the USB-C receptacle matters or how you would get it to work, or that you were suggesting was just a poor/unneeded design choice because of the elements of uncertainty for a successful negotiation.

, 02-16-2025, 10:52 PM

The latter is well in the realm of possibility as I tend to over complicate things.

QDrives , 02-16-2025, 11:05 PM

With USB A, as far as I know, there is no CC. So no possibility to high voltage either.
If you get 12V from USB (C), you need a regulator to get 5V. The voltage step together with the current you mention, that would need to be a SMPS buck.
As you calculated about 10W total power, that would make 2A @5V. Most USB adapters will be able to that. Not exactly 5.0V, there will be quite some tolerance.

As for the negotiation... In my opinion you should not draw more power than the source allows you to take.
Old USB standards had 100mA and 500mA. Be nice and if the source is only 500mA, do not turn on the most power hungry parts. Notify the user that the power is not sufficient (somehow).
Do things get damaged? Probably not.

If this is for hobby and generally a one of, then just go for the PD solution and get a USB adapter with it that can deliver it.
When commercial, as I stated, be nice.

, 02-17-2025, 12:55 AM

> Old USB standards had 100mA and 500mA. Be nice and if the source is only 500mA, do not turn on the power is not sufficient (somehow).

Could you elaborate on this a little more, if I’m attaching a PD sink to my USB-C wouldn’t I not have to worry about the standalone range and just be elevated to the 20V/5A max range in the table and just need to worry about finding a wall plug to properly supply that?

, 02-17-2025, 12:56 AM

> If this is for hobby and generally a one of, then just go for the PD solution and get a USB adapter with it that can deliver it. When commercial, as I stated, be nice.

Also yeah, this is a little hobby project of mine so I can afford to play fast and loose a little.

From how I am interpreting your comments, to keep the PD sink with the USB-C receptacle that I have would I need to do the following:

1. Add a SMPS buck to create 5V from the 12V (I just checked and the 5V output LDO on the PD chip IC itself thats too low so Ill do that instead).
2. Be sure that I get a 3.1 PD power adapter/cable to get a source that can supply everything in within reasonable margins? Was thinking of just using a 45W adapter.

QDrives , 02-17-2025, 03:40 PM

There is no 12V mentioned in this document. There is 15V.
https://www.ti.com/lit/slyy109

, 02-18-2025, 08:28 PM

That is just a primer document for explaining USB PD as a concept and thus, it only mentions standard usb profiles as reference.

When looking doing some research, I found that 12V support does exist as an "optional" fixed voltage and it isn’t as commonly supported but still possible to get out from some chargers.

https://www.reddit.com/r/UsbCHardware/s/TQDAIuEJ7U

I’ve crossposted this project specifically to the USB Hardware community on Reddit and identified a few wall blocks that could give me 12V/3A which would be the best profile for my specific situation.

https://www.reddit.com/r/UsbCHardware/s/MqwsW2FGz8