Semi-disposable micro-ballooning experiences wanted

[Nick McCloud]

As much as it pains me to suggest such an idea, but having a tracker that doesn't matter if it's not retrieved is sort of starting to make sense. I can sort that out, BakTrak (ProMini, RFM98, random uBlox clone from Ali) is less than optimal at present but can be refined.

This would allow well-briefed school teachers to do a launch at short notice without all the extra that a full scale launch brings. And if it all goes wrong, no major loss.

Does anyone have any experience with the smaller balloons that are well inside the 2m burst legal limit. Do they burst and when?

Note, I'm looking to burst not float - part of the experience for students is the frisson of the retrieve - getting the launch right to get the payload back - most likely with some memory of some sort to capture data. Plus the McD's visit on the road trip.

[Mika Köching]

Alright, I'll try to keep my message short for now but would be happy to write up some more info if you find it interesting.

Exactly this is my point of interest in HABing.

I am developing trackers myself. I am currently working on a sub 15g tracker including everything except the balloon. It uses an STM32wl with integrated LoRa Radio and an atgm336h(probably the best for balloons).

Before that I was using an ESP32-S3 and an rfm98w but that is quite a bit more expensive. My current design with the stm32 costs around 10€/tracker when hand assembled (about 1hr of work but everything is 0201 smd so it takes some training).

I bought a box of 250 30g Balloons from Hwoyee. They are awesome and have a burst diameter of around 1.5m. With Helium and an ultra-lightweight tracker, they should be able to reach 25km. I have already tried them with a modified RS41 (around 30g) and it reached 21km before it burst.

Alternatively, there are 36" latex balloons (26g) on amazon that work OK. With 40g payload and about 1.5m/s ascent rate they can reach anywhere from 9 to 15km. I did a mass launch of 9 of those within an hour. The quality varies from balloon to balloon but they are cheap and easily available.

You might want to take a look at my blog at dl8mik.darc.de

73s 

Mika, DL8MIK

[Steve]

Pawan 100g were quite popular a way back for this as their stated burst diameter was 1.7m IIRC - I think 25Km was typically achieved (with light payload)  - although if you do the math the diameter would have been over 2m at that point (I think balloon companies generally under spec balloons).   The Hwoyee 100 has a rated burst diameter of 2.0m - so with neck and payload antenna would be well over the 2m linear dimension of the small balloon regulation.  Burst would probably be above 30Km with light payload.  Totex do a 100g with burst diameter 1.98m - so again too big with neck and payload antenna (I have a few of those too).  I had wondered if there was a way to invert the neck (like a belly button) and place the neck, payload and antenna inside the balloon - as a way of limboing under the regulation and achieving maximum altitude .  I flew a few Pawan 100g a ways back also see JoAl flights (raspberry pi PITS based with camera).    https://www.flickr.com/photos/16828840@N07/albums/72157632399831961/    https://www.flickr.com/photos/16828840@N07/albums/72157631375086030/

I'm currently playing with Hwoyee 30g balloons - which Mika & Co seem to manage 20Km (Happy to ship a few to anyone in the UK that can report back altitude achieved).  Hwoyee also do a 50g balloon - which I'm guessing would do 25Km.  I asked Hwoyee about this quite recently and they said the only other balloon they do is an 65g - but this is very much a "special".

30Km is a real aim for me on HAB flights as anything much less isn't HAB IMO.

    Steve

--
You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ukhas/73be843e-15a6-40f4-8008-1add626b62dfn%40googlegroups.com.

Virus-free.www.avg.com

[Mika Köching]

Oh, one addition:

I think it's quite fun to launch really small balloons as well when there isn't that much wind as those can easily be collected afterwards.

I have a lot of these "punch balloons" (those annoying things with a rubber band). They weigh 8g and cost under 0,30€. With a 14g tracker I was able to reach 7km altitude. That's an altitude that's already partially interesting for some research.

I think if you safely want to stay under 2m including payload, you should go with the HY30. The neck isn't very wide though. I have a permit here in Germany so length doesn't really matter but my goal is to get usable data so I have quite a bit of string attached.

[Nick McCloud]

Thanks both, I'll be looking at the detail after the launch tomorrow because no doubt some of the project team will want to get their kids/grandkids involved and the current full height PitS project doesn't fit with the world we live in so well.

Two observations:

1. I've an STM32WL5J board I use for LoRaWAN sensors and a board for the Heltec HT CT-62 which is an ESP32 equivalent to the STM32WL - it has an ESP32-C3 + SX1262 and it is fully Arduino'ed so somewhat simpler for junior coders to add some sensor data & recording. I'll be evaluating both and thank you Mika for pointing me towards that GNSS module.
2. It's good to know others are trying similar things so Steve, I'll be in touch about balloons to buy.

PS, whilst "real" HAB is way up there, even 20km is double Everest / airliners. Where do we draw the line? Even the Karman line has the US vs RoW definitions (80 vs 100km). The school projects get bogged down because I've been "selling" the full height magic 100,000' deal which requires the full setup. Getting students / a school started with mini-launches which are much more manageable & DIY seems the way to go. Once they have an active club with a couple of years experience they can graduate on to a full height full science launch.

[Steve]

The ATGM336H is a great, cheap module - but the manufacturers state 18,000m (59,000ft) as the maximum height in the datasheet - has someone tested it above that and got it to work?   Great for pico ballooning though (which as yet haven't got that high).

Personally I draw the proper HAB line at about 85,000ft  (highest air breathing aircraft sustained flight) 

Steve

--
You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/ukhas/df6ecda2-9622-4f1e-a338-6315a7072cb4n%40googlegroups.com.

Virus-free.www.avg.com

[Mika Köching]

The atgm336h work fine over 18km and even support ublox binary. That means they also work with wenet.

Beware of the clones though, there seem to be some on AliExpress. I just buy them from lcsc whenever I order there.

To view this discussion visit https://groups.google.com/d/msgid/ukhas/1e717c80-6d89-4dac-85f6-b072db835937%40randomaerospace.com.

[Mika Köching]

And the Stm32wl also works with the Arduino IDE and Radiolib. It's quite easy to port tb tracker esp32 to the stm32wl. It's actually a bit simpler because you don't have to configure any pins for the radio.

[Steve]

Interesting - I guess you have to ignore the manufacturer spec and either fly them or test them with HackRF.

The GP02 is based on the same chip as the ATGM336H (the AT6558) so if the ATGM336H works >18Km its a good guess that the  GP02 works too.  Certainly worth a try.

The newer SIM65H also lists the maximum altitude as 18,000m.

    Steve 

To view this discussion visit https://groups.google.com/d/msgid/ukhas/CALEL9Z28M%3De9oEPe6EFY_NVGXCVNcawp0ED%2B3PsNH1yHN3rhSw%40mail.gmail.com.

[Nick McCloud]

The ATGM336H is a great, cheap module - but the manufacturers state 18,000m (59,000ft) as the maximum height in the datasheet - 

Beware of the clones though, there seem to be some on AliExpress 

Hmmm, are these clones of a clone of the uBlox 6? Anyhoo, that would possibly explain why BakTrak stops at 18km - but works fine when it gets back down so I'll stop fiddling with it and fly it tomorrow in it's Woodi config. And then find a better solution.

And the Stm32wl also works with the Arduino IDE and Radiolib.

I don't have problems with STM32WL/RadioLib and Arduino - although I normally use PlatformIO given I am the primary LoRaWAN tester for RL ...

But the average random youth learning to code in C/C++ tends to get on better without all the extra STM32duino libraries - the ESP32 is just a bit more 'normal'.

Certainly worth a try.

I'll still have some CUSF balloons left & some random hydrogen - can organise a test payload subject to some low stress CAA permission action.

 

The newer SIM65H also lists the maximum altitude as 18,000m.

I wonder if many of these are copy n pasta of other data sheets because they don't know what they have.

At £3 a pop from Digikey, the GP02 bears investigation and it would be feasible to fly a bunch of them with different antenna configs. 

 

[Mika Köching]

Well, Platform IO is supposed to work with arduino framework but I couldn't get it to work yet. I'll have to do some further investigation.

What gps module did you use on your bak track? I flew the atgm336h-5nr32 from lcsc successfully over 18km and it worked absolutely fine. When you send the ubx binary configuration command for the mode (you'd have to change that for ublox modules to have them work over 18km) it doesn't do anything also so that isn't even an option.

[Steve]

Does the questionable uBlox 6 respond with the correct UBX ACK packet when you apply the set flight mode command?

I have cheap uBlox 6 module from aliexpress - but it does respond the correct UBX ACK - so I have assumed it isn't a fake. 

    Steve

--

You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/ukhas/81998a67-36a1-4f8f-9bcf-8c9247dfd0a4n%40googlegroups.com.

Virus-free.www.avg.com

[Sasha Tim]

Yes, I've personally flown the ATGM336H up to 28km on a real balloon so far. Other than a minor hiccup midway through (it reported 32km in one packet, then jumped back down to the proper altitude), it worked great. 

To view this discussion visit https://groups.google.com/d/msgid/ukhas/1e717c80-6d89-4dac-85f6-b072db835937%40randomaerospace.com.

[Nick McCloud]

On Monday, 1 September 2025 at 20:48:27 UTC+1 Steve wrote:

Does the questionable uBlox 6 respond with the correct UBX ACK packet when you apply the set flight mode command?

Don't know, BakTrak is a copy & pasta of HABduino that was built in a hurry and has since been rather neglected. Only two people receiving it yesterday and whilst it stopped at 12,000m, it did it's job at the landing in providing another target circle - all three (LoRa, RTTY and BakTrak on LoRa) in the same part of the large field. We were ten minutes out from it so could be sure of where to go and reacquired signal as soon as we got within a km of it.

But as I have some interest being shown in HAB on a small scale - albeit not to your definition of >85,000' - this is something worth revisiting. And gets me away from only being able to do a one or two monolithic launches that are a scheduling nightmare. As people advance, they can then try a full height launch.

On Tuesday, 2 September 2025 at 14:34:19 UTC+1 Sasha Tim wrote:

Yes, I've personally flown the ATGM336H up to 28km on a real balloon so far. Other than a minor hiccup midway through (it reported 32km in one packet, then jumped back down to the proper altitude), it worked great. 

 Always worth putting in some code to filter the unexpected sensor/GNSS glitches - if it's out by a certain percentage, take the reading again &/or take several readings & use the median or more advanced stats to filter out the wrong 'uns.

 

[Nick McCloud]

On Monday, 1 September 2025 at 18:03:49 UTC+1 Mika Köching wrote:

Well, Platform IO is supposed to work with arduino framework but I couldn't get it to work yet. I'll have to do some further investigation.

For most platforms, PIO and Arduino is perfect - so much better than the Arduino IDE - particularly when debugging RadioLib.

For STM32WL we have a port of RadioLib for STM32CubeIDE which is a pain due to the way the compiler blocks the global use of various C++ structures. We have tried STM32Wl/RL with Arduino IDE. I can have a go at PIO, Arduino framework & STM32WL at some point soon.

 

What gps module did you use on your bak track?

It's label says it is a u-Blox Neo-M8N and the label looks real but ...

I'll do a bit more research to see what it actually is, but it's not super important, I'd rather prioritise the ATGM336H as its price is pretty attractive.

For me, 0806 is what we use for boards we aren't trying to make super small and the Seeed Wio/LoRa-E5 module is big enough that a couple of 0806 caps isn't going to add much to the space. We are looking at building our own SX1262 so we can power optimise it for the EU market when we want to use a different MCU. That may well end up with some smaller components!

For the purposes of HAB or should that be MAB (medium altitude), I'm thinking of two versions: A heavier DIY, buy the modules and use some Vero or a thin mounting PCB to join it all up. And a custom PCB with a GNSS module, MCU & SX1262 as a pre-built version. 

PS, love the parachute instructions - fits nicely with my objectives! 

[Mark Jessop]

If the radio chip supports it, I'd highly recommend implementing Horus Binary on your tracker.

This will let you send more custom telemetry and have it be properly displayed on the tracker (and grafana plots).

In terms of accessibility, we now have browser-based decoders for Horus Binary (and Wenet) at https://horus.sondehub.org/ , which can use either SSB audio input, or directly connect to a RTLSDR from the web browser (Chrome/Chromium-based only). We also have a little helper setup links, e.g.:

https://horus.sondehub.org/#freq=434.2&mode=horus&spacing=270&name=AREG%20Horus%20Binary

73

Mark VK5QI

--

You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/ukhas/0b55477b-9833-4562-821a-2756a8bd95e5n%40googlegroups.com.

[Mika Köching]

Horus seems to be quite problematic on LoRa chips. The sx1278 works alright but anything newer make a really dirty signal that can't get decoded most of the time. I'll do some more experiments with the settings but it doesn't seem like, it will ever work. Maybe it's time for something new based on LoRa. I personally don't like LoRa but it works so well it kind of outwheighs all the negative points. 

[Nick McCloud]

On Friday, 5 September 2025 at 05:24:50 UTC+1 Mika Köching wrote:

Horus seems to be quite problematic on LoRa chips. The sx1278 works alright but anything newer make a really dirty signal that can't get decoded most of the time. I'll do some more experiments with the settings but it doesn't seem like, it will ever work. Maybe it's time for something new based on LoRa. I personally don't like LoRa but it works so well it kind of outwheighs all the negative points. 

In theory you push your bytes in to a LoRa chip and the settings are the same at both ends, the same should come out the other end. There are some differences in sensitivity (better for the newer SX126x) and a few corner cases as exceptions to catch us out. But most of my stuff is using LoRaWAN libraries so I've not done any LoRa to LoRa with mixed modules for ages. 

Which vendor modules were you using? 

It may be that the output stage on your Tx was borked. Perhaps start another thread on this topic and share some code on GitHub - I've got a LOT of different cheap LoRa modules plus all the Semtech dev boards that I can test with.

[Steve]

I might be wrong (and stand to be corrected) but in my brief look at the Horus / radiolib code it looked like it was jumping the SX12xx frequency registers - which does not strike me a great.  In the back of my mind I recall someone who had been very clever and generated 2FSK RTTY on the SX12xx by calculating the appropriate  data stream to feed into LoRa - perhaps something similar can be done for 4-FSK. Or maybe I just dreamt that.   Even if it could be done I think its likely that the symbol rates would be different from standard Horus.

I'll be flying a LoRa flight shortly - using my own tracker based on a Pi Pico, Arduino TinyGPSplus and the  sandeepmistry / arduino-LoRa libraries.  I'll be using Daves PITS LoRa gateway to receive and upload to Sondehub.

    Steve

To view this discussion visit https://groups.google.com/d/msgid/ukhas/1a72671a-e3c4-4582-988e-fa58dee24503n%40googlegroups.com.

Virus-free.www.avg.com

[Mark Jessop]

Yeah, my personal preference for a HAB modulation is a modulation where you don't need any *modulation specific* receive hardware - e.g. a LoRa chipset.

That's why I built Horus Binary instead of using LoRa chips on my payloads. I wanted anyone with a SSB receiver, or any kind of SDR to be able to receive it. Nowadays, you can hams with SSB receivers to horus.sondehub.org and they can pipe in their audio and voila. Or anyone with an RTLSDR can plug one into their phone and receive it. 

I also didn't want to use a modulation that was *proprietary* - owned by a single company, and that had the potential to go after anyone doing a reverse engineered implementation of it (e.g. gr-lora). I have a whole bunch of opinions on LoRa and its used in the amateur radio bands, but I'll leave that for some other time...

The FSK transmit functionality of the LoRa chipsets is pretty hobbled, usually a smaller FIFO, etc. One of the issues is that we have to modulate 4FSK by re-programming either the frequency, or preferably a fine PLL-offset register at the symbol rate. That's how we can get the small frequency shifts that keep Horus Binary fairly narrow. I suspect that it should be possible to get it working, but it might require more experimentation with other registers.

Horus Binary does modulate fine with FSK/MFSK specific ships like the SiLabs 446x series, and you can even do the modulation using the internal 4FSK modulator rather than having to do the reprogram-the-fine-offset-register-at-100Hz thing that we do on other chipsets. I suspect those chips could probably even *receive* Horus binary, but I haven't tried this. 

73

Mark VK5QI

--

You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/ukhas/a1ba7102-966c-468e-a314-998c26364c6an%40googlegroups.com.

[Mika Köching]

I made a custom tracker with the stm32wle5 so without any module. I don't really have the time to look into this any further.

I'm currently working on a sensor boom which seems to be more complicated than the entire rest of the tracker.

[Mika Köching]

Regarding the si4463: Definitely interesting. I have one I wanted to use for Wenet fast video transmission but I never got that working. I might try receiving Horus sometime in the future.

[Sasha Tim]

I’ve had some luck with the Si4432 and Horus Binary - you can get the chip for about $1.6 USD from Chinese suppliers, a bit more expensively from Digikey and Mouser. Some shifts don’t work correctly (not sure what the root cause of this is, I suspect RadioLib) but making it slightly more wideband helps. 

--

You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/ukhas/85afb5cf-4b88-4882-8de7-ef40530968c9n%40googlegroups.com.