Weather balloon inflation method

It’s nearly complete! I did my first weather balloon launch with hydrogen generated from rainwater this week. See video below

The PEM electrolyzer needs pure water (<1.0 uS/cm conductivity), so the rainwater passes through some deionizer filters on its way down.

Also nearing completion … powering the PEM with solar panels! The PEM takes 30Amps at 5V, and have been using two 100-watt panels from Lowe’s. Only problem is that taking the system down to 5V has been tricky, because even though the panels are labelled as 12-volt, their operating voltage is actually 20 volts. The only DC-DC regulators I can find are these here , and they are meant for 12V or 24V system. Since the panels output at 20V, I think the internal workings of the regulator gets confused and cannot decide between 12V or 24V … which leads to some really volatile Watt readings at the PEM.

Will bring the panels into the Makerspace either on Sunday or at the monthly meeting to see if anyone can diagnose what I’m doing wrong.


Sounds like what you need is a “buck” converter. A Buck converter will convert from a a range of higher DC voltages to a lower DC voltage, and can usually be tuned to support a variety of input and output voltages by adjusting trim pots on the board.

This one supports the input and output ranges you mention and can support 5 amps. Not a specific recommendation - just the results of a quick Amazon search.

It does have a pretty digital readout, though. Some minor assembly required (standoffs and cover bezel). Make sure you get the buck converter (high DC → lower DC), not the boost converter (lower DC -->higher DC) or the bulk-boost converter (higher/lower DC → lower/higher DC, but more expensive).


I’ve used buck converters in my security sytem to take a Dell power supply down to 5v. It works great.

1 Like

That’s the goal to get it down to 5V from the higher volts coming from the panels. That buck converter looks good with respect to range. Except to produce the needed volume, it needs 30 Amps as opposed to 5A.

My best luck so far has been to use products meant for Golf carting and RVs (36V), since putting two panels in series gives actual voltage 42V, which is a typical 36V system output.

Do pay attention to the fact that voltage falls off with current. The usual use case is connected to a battery with a low internal resistance compared to the panel. So a 20V, open circuit solar panel will pull down to battery voltage when connected to one, and as solar input changes, the output current changes. The voltage at open circuit falls off sharply as you increase the current draw. So a lot of devices that expect to see a battery see the solar cells as a worn out battery with a really high internal resistance.

DC to DC step down (buck) switching pwr sply. Spec says input range is 10 - 32 Vdc. Ckt doesn’t decide if 12 or 24 Vdc. Put a voltmeter across the pwr sply input while operating the PEM and check if input stays within range.

1 Like

I wonder if it might be better overall to grab an MPPT charge controller and store the power in some batteries, then using that to power the electrolyzer. It would at least smooth out power to the electrolyzer, and give you some modularity.

1 Like

The PEM electrolyzer is very flexible to current fluctuations and reacts quickly. Solar Batteries are very expensive, so the point is to cut them out of the system. I only need 6 hours of sunlight onto the 200W panels to get the needed amount of hydrogen.

To @artg_dms suggestion, the measured voltage from 1 panel was about 20V with the PEM load (24 open-circuit) before the 5V regulator. I haven’t tried with 2 panels.

It does indeed say 10 to 30V input but also says 12V/24V. Technically the 5V regulator does indeed drop the voltage, but it confusingly oscillates between 1 and 6 amps every 1.5 seconds or so.

This is what made me think something internally didn’t know how to handle 20V. If I was designing one of those things 20V could mean either a dead 24V battery or a super charged 12V one. But maybe I’m thinking about it wrong

1 Like

Ah, it makes sense to just go direct like you’ve done.

It’s likely the 12V/24V is just a marketing label, to make it easier to search for. I noticed that a lot when I was looking for similar products.

1 Like

@kobin Is there room for me to bring the finished product into the Science area next week to demo some launches outside? It measures 45 inches at its widest part.


Sure we can keep it in Science for a bit! Looks awesome!

Kobin Caddick
Science Chair
Dallas Makerspace

From: [email protected]
Sent: April 14, 2022 11:35 AM
To: [email protected]
Reply-to: [email protected]
Subject: [Science] Weather balloon inflation method

sunburntcat DMS Member
April 14

@kobin Is there room for me to bring the finished product into the Science area next week to demo some launches outside? It measures 45 inches at its widest part.


I’m out of town this weekend; if it’s going to hand out in science for a bit I’d love to come by and check it out next week.

1 Like

Hey Nick, I love your project!

For the rest of you to reference: Nick builds these systems to deploy in the field in Africa. They fill Weather Balloons which Nick’s helpers fly to gather data which is put on a Block Chain Network. Portability and Reliability are huge considerations!

We built a PEM Electrolyzer to separate Heavy Water for our Fusion Reactor. We were getting the same problem: it turns out the PEM Cell looks like a dead short to a Buck Regulator, so it cycles On/Off/On/Off… We fixed it by adding a Series Resistor of 20 Ohms (because we don’t care about efficiency–we are in the fusion business after all…)

This system DOES worry about effiency, so I’m guessing a Current Limiting Regulator is what it needs. Also, to improve output from the Solar Arrays, we might consider an aiming strategy, to point them directly at the Sun.

If you like, we can adapt our ESP32 and Driver Circuit to regulate the current from your Solar Cells to your PEM for maximum efficiency.


Yeah, we’re signing a contract with the University of Zambia to make it a commercial product. This is version 2 after university students in West Africa have been doing their launches with a similar PEM and soda bottle.

The PEM works fine on a 5V DC power supply. It was made by a guy in Utah and hits the advertised 1 Liter/minute mark when it’s cranked up to 40amps. And I’m pretty sure these solar panels work fine. It’s just these regulators are being finicky. I’ve got two more we can play around with once it gets to the Makerspace.

Anyone got a pickup truck? I could also rent one and drive it up to MS this Saturday. An extra set of hands would be great too

1 Like

My Forester’s back hatch measures 42" x 32", will it fit through the door? I wouldn’t recommend strapping it to the roof rack…

It’s exactly 70 inches at its tallest point. I got a Sub Crosstrek as well (similar style, a little smaller), and it’s way too big for it. The wood is frame is a little wobbly too so wouldn’t trust it on a car rack haha. Not with the potholes and bumps on I35

1 Like

Zerowater filters (reasonably cheap at Walmart) result in water that reads 0ppm dissolved solids. Not sure if that’s helpful information.

1 Like

It is helpful. The rainwater has lots of deposits in it, so a 0 dissolved solid filter beats my current method of running it through fabric.

1 Like

I have a friend who has done electronic design work and is very familiar with using solar cells. If you need some additional eyes on the issue, ping me and I’ll give you his FB profile.

1 Like

What we really need is a safety valve that will cut the solar circuit when the pressure reaches 90 psi. The PEM is rated for about 120psi and I don’t want to know what happens when hot, compressed hydrogen has nowhere to go.

By the way, the panels and wooden frame are all in the Science area as of last night. The step-down regulators are attached to the frame. My dad happened to be in town and lent a hand.

1 Like