Transolve P6K and Altus Metrum Telemega: Mammas don't let these two play together!

[cberrya6e]

I am new to the Forum and have something to pass along. I had an interesting "event" several weeks ago. I was launching a brand new 4" AGM-33 Pike with redundant dual deploy electronics onboard. The altimeters used were a Transolve P6K and an Altus Metrum Telemega and ground testing had shown no issues. Pad side, with the rocket on the rail, I armed the Telemega and received the proper aural indication that continuity was good and that the altimeter recognized PAD state. Next, I armed the P6K and it beeped once or twice before stopping and I heard some sort of a hiss from within the rocket. I initiated my escape sequence and started running. Before I got more than 10' from the beast, ejection charges fired and the rocket deployed its recovery guts around the pad.

Analysis after this showed that the P6K had fired both the APOGEE and MAIN ejection charges (not the Telemega). After a series of emails with both Bdale Garbee (Altus Metrum) and John Fleischer (Transolve) and discussions with others, the initial potential culprits were identified as being the arming switch (Schurter rotary type) or interference between the 70cm band transmitter on the Telemega and the P6K. I performed some simple tests at home and could not induce any sort of a brownout/power cut using the switch; further review and discussions shows that these switches are sold and used by a number of high-profile vendors in our hobby and I felt that this was not the real problem. Furthermore, I realized that my ground testing prior to the launch attempt was always with the ebay in the horizontal position. In this state, the Telemega is in an IDLE state and is in a receive mode awaiting comms from a ground station. Once it is in the PAD mod (vertical), the altimeter goes into full transmit and does not receive inbound comms. Using LEDs in the ejection charge connectors, I was able to repeat 100% the results that the launch attempt revealed - once the Telemega and the P6K were both on and booted, the P6K would fire both LEDs. The start sequence did not affect the outcome - the results were identical regardless of which device was started first.

To fully close the loop on this, I took my ebay to a lab at Rice University and ran some tests there. By connecting oscilloscope leads to the P6K ejection charge outputs across a 1 ohm resistor and using a spectrum analyzer targeting the 435 Mhz band, we were able to conclusively show that the P6K does not filter out the RF from the Telemega and the current induced by this energy is enough to cause the P6K to "see" flight conditions that lead to triggering of the ejection charges. Here is a screenshot of the scope for one of the tests:

View attachment 173794

I forwarded these results to John Fleischer and after some discussion we both agreed that this was a problem and that the Telemega and the P6K should not be situated in the same ebay.

I like the simplicity and the robustness of the P6K. But I am new to the telemetered side of the hobby and this was a bit of a shock for me. Given that the Telemega is relatively new, I thought I would post this to alert others to this issue. I will continue to use my P6Ks, but not on any of my radio birds without having the two altimeters separated within the rocket body.

Regards,
Chuck Berry
Houston, TX
TRA L2 #14528

 

[dixontj93060]

Thanks for the report.

 

[Salvage-1]

This is the type of report we need. Sorry about the problems, but you have saved others from the same fate.

Needs stickied

 

[hball55]

Thanks for the heads-up, I have a P6K in one of my birds . . . no redundancy though. I wonder though, if a rocket with electronics transmitting on the next pad could trigger my ejection charges. If it happens, at least I will know why.

 

[GregGleason]

Thank you for posting this on TRF.

And welcome to the forum Chuck!

Greg

 

[RFMan]

Thanks for the report. This is an important illustration, not just for the particular models involved, but also for the entire notion of using multiple onboard electronics that include a radio transmitter. It has been a bit of a luxury to omit susceptibility testing from rocketry electronics product release testing, but as payloads increase in capability and complexity, that may no longer be true. At any rate, it's a good wakeup call for thorough testing.

 

[cberrya6e]

As a follow-up, the work at Rice was done with Chris Harris' leadership (Houston Tripoli rocketry and electronics guru par excellance). His take on this was that the P6K lacks a ground plane in the PCB design and uses through-hole components, making this altimeter susceptible to RF interference, particularly one as relatively powerful as the Telemega. I passed these observations along to John and he agreed with the conclusions. Obviously, the appeal of the P6K is that it is a simple kit that has a robust fundamental design, but changes to make it more RF-tolerant would move the price point of the altimeter and change the feasibility of it being a kit - Surface Mounted Devices (SMDs) are a trick to solder (see the Eggfinder...). The Telemega appears to get along with Stratologgers, which have those features incorporated - ground plane, SMDs.

This whole thing has opened my eyes. I now am asking questions about clashing between devices, something that I was largely blind to up until this point. It almost seems to beg for some kind of an online matrix of compatibility...

Regards,
Chuck Berry
Houston, TX
TRA L2 #14528

 

[JohnRLewis]

...The Telemega appears to get along with Stratologgers...

Has someone tested that combination? I am in the early stages of planning my next project, and I was going to use a Telemetrum as my primary, and a stratologger as my backup altimeter.

 

[cberrya6e]

At least one of our fliers here in Houston uses a Telemega with the Stratologger. I cannot speak to the Telemetrum.

Regards,
Chuck Berry
Houston, TX
TRA L2 #14528

 

[UhClem]

This certainly isn't the first case of EMI as I wrote up my experience in 2004.

I also turned up an old Info-Central article. Note that the tests performed here were only of electric field coupling. If an altimeter is located close to the transmit antenna then you cannot neglect the magnetic field.

Fully hardening an altimeter to EMI requires that it have a case. Ground planes will help to some extent but you really need to have a case and filtering on all external connections. If magnetic fields are a problem then you have to use special magnetic shielding materials.

But that level of protection might not be required. The problem is most likely at its worst in the analog signals between the sensors and the ADC. If those signal traces are kept short and provided with some shielding (ground traces parallel and close by) then the susceptibility would be reduced.

Or use digital sensors. Hoping of course that the sensors have decent EMI shielding inside them.

 

[RFMan]

Agree with the conductive case requirement for good EMI immunity (I'm an RF/microwave engineer). Getting a mostly solid conductor around a circuit, with feedthru filtering on the I/O, is typically good enough. The antenna is still an ingress point, but provides some frequency selectivity itself. A ground plane can provide one side of that case, and there are lightweight perforated shield cans which are surface mountable. Magnetic shielding is usually implemented with mu-metal, but there are two main issues with that: it's heavy (effectiveness is a function of thickness), and mechanical shock disrupts grain alignment accomplished during anneal after final forming. (Some folks consider that required annealing is a third issue )

 

[tbonerocketeer]

Has someone tested that combination? I am in the early stages of planning my next project, and I was going to use a Telemetrum as my primary, and a stratologger as my backup altimeter.

Flown them a bunch together with no problems. Couple higher altitude flights as well.

 

[1tree]

Out of curiosity, did you try placing filter capacitor across the leads at the P6Ks terminals? I don't know that I would trust this in flight, but I would sure be curious to see if it could be removed having seen many a filter capacitor in circuits. If it worked, I would definitely want to ensure that such would filter out much more power then the Telemega produces (such as the 5 watts from a handy talkie).

 

[ksaves2]

Same trouble with tracking transmitters. On pad ejection with both low power, 12mW and 100mW transmitters riding in the ebay with the antenna poking out the aft bulkhead into the sustainer.
I won't fly'em with transmitters anymore. Kurt

 

[bdale]

Furthermore, I realized that my ground testing prior to the launch attempt was always with the ebay in the horizontal position.

As you correctly surmised in our email discussion, this meant you weren't actually ground testing in flight mode RF conditions. Sadly, as many of us have discovered, lots of otherwise reasonable existing board designs just won't work reliably in proximity to an RF transmitter.

What I always suggest is to ground test by attaching bare e-matches (no BP!) to each pyro channel, turn everything on in pad/flight mode, and let things sit running for a while (by which I mean maybe an hour?) to ensure there are no unexpected events. Not 100% fool-proof, but this will catch the vast majority of RF immunity failings.

 

[troj]

As you correctly surmised in our email discussion, this meant you weren't actually ground testing in flight mode RF conditions. Sadly, as many of us have discovered, lots of otherwise reasonable existing board designs just won't work reliably in proximity to an RF transmitter.

What I always suggest is to ground test by attaching bare e-matches (no BP!) to each pyro channel, turn everything on in pad/flight mode, and let things sit running for a while (by which I mean maybe an hour?) to ensure there are no unexpected events. Not 100% fool-proof, but this will catch the vast majority of RF immunity failings.

You beat me to it.

My ground testing of RF interference (Garmin Astro, specifically) indicates that you need to let the setup "bake" for over an hour -- I have a specific combination I've had fail within 10 minutes one time, and take over an hour another. Exact same setup both times, down to the leads, batteries and wiring. I use christmas tree bulbs as my test "match" when I do this.

My recommendation is two hours, and do it multiple times.

Back to the original post, I'd be interested to find out how far away a transmitter has to be before it doesn't confuse the P6K -- with it being that susceptible, I have to wonder if someone flying something putting out watts of power (Garmin Astro, again) on the next pad over might confuse it.

-Kevin

 

[G_T]

https://www.lessemf.com/mag-shld.html - perhaps this could be of use. I was starting a search for small mu-metal boxes but came across this first.

I do have a severe dislike for the lack of shielding in our typical rocket installations. I typically have run a telemetry system along with another altimeter, though I try to keep the antenna as far away as practical from other electronics. I've flown two telemetry systems (different bands, one a prototype) plus other electronics before. In that case there is the additional concern of the signal from one telemetry system modulating the pre-amplified signal to be sent by the other system, resulting in out-of-spec signal(s) being transmitted.

Chuck, very nice catch on this issue! You've probably saved a few mishaps down the road.

You know, it's sort of too bad we have to put our ear to the rockets in many cases to hear the beep patterns to discern if an armed altimeter is working correctly...

Gerald

 

[ksaves2]

Two years ago, in the NAR magazine, there was an article about the converted dog trackers causing maleficence with specific altimeters/deployment electronics. It named names. It was recommended to place the tracker in the nosecone if at all possible away from the ebay and laminate a metal
plate either on the nosecone bulkhead or the forward ebay bulkhead. Also twisting all wire leads if possible I've been told can help.

Now I had a P6K deploy the apogee charge on ascent with a 100mw tracker. Mach delay was set but really not needed because it was a longneck 4 inch rocket with a 38mm motor. Wasn't going to go that fast anyways. The early deploy zippered the sustainer royally. The main blew somewhere about apogee and the upper bay survived without a scratch.

Built a new sustainer and used a 12Mw tracker with the antenna going through a tight hole in the bulkhead aft . Stupidhead here assumed with such low power, no problem. Stupidhead was wrongo!! Put rocket on pad to fly, turned on altimeter, bent down to insert igniter and beeping stopped.
I jumped up, was going to try to get the key in to shut off the P6K but the apogee blew on the pad as I stood back. Ran over to the upper bay now on the ground and it blew before I could attempt a shutoff. Well, at least I knew this second charge test was "spot on". Aside from dinged paint, no damage. I swapped out the P6K for an AIM and ditched the tracker. Heck it won't fly above 3k' anyways but I got a good lesson on Rf/altimeter interference.

I now follow Kevin's and Bdale's advice to a "T" . But even with that much care there are no guarantees.

I will say I've stuffed a 12mW Beeline GPS in a Wildman Jr riding with a Raven 2 without any trouble whatsoever. The Beeline rides on top with the
antenna projecting tightly through a form fit hole in the forward bulkhead. I use a cardboard tube that the AT first fire igniters come in and wrap it with duct tape for flame protection. Use clay to hold the tube in place and stent the antenna so it doesn't get smushed by the parachute.

The parachute bay is painted with yellow non-metallic paint. The sustainer with red metallic. I flew the rocket with a J-330 and it simply disappeared
up high. No one saw any of the events but it was easily seen from the APRS radios that the apogee charge did it's job. The descent slowed further at
the main deployment altitude but still no rocket/main chute was seen. It was only seen on the Garmin mapping GPS and the laptop tracking program.
I just set the Garmin to "navigate to" and by golly followed the path to where "automagically" the rocket was lying. I wouldn't have had any idea where it was otherwise.

The low powered Beeline trackers seem to play well with deployment electronics (except the P6) but I wouldn't have the antenna of any tracker inside
the ebay with the deployment electronics. Kurt

 

[CarVac]

Two years ago, in the NAR magazine, there was an article about the converted dog trackers causing maleficence with specific altimeters/deployment electronics. It named names. It was recommended to place the tracker in the nosecone if at all possible away from the ebay and laminate a metal
plate either on the nosecone bulkhead or the forward ebay bulkhead. Also twisting all wire leads if possible I've been told can help.

Now I had a P6K deploy the apogee charge on ascent with a 100mw tracker. Mach delay was set but really not needed because it was a longneck 4 inch rocket with a 38mm motor. Wasn't going to go that fast anyways. The early deploy zippered the sustainer royally. The main blew somewhere about apogee and the upper bay survived without a scratch.

Built a new sustainer and used a 12Mw tracker with the antenna going through a tight hole in the bulkhead aft . Stupidhead here assumed with such low power, no problem. Stupidhead was wrongo!! Put rocket on pad to fly, turned on altimeter, bent down to insert igniter and beeping stopped.
I jumped up, was going to try to get the key in to shut off the P6K but the apogee blew on the pad as I stood back. Ran over to the upper bay now on the ground and it blew before I could attempt a shutoff. Well, at least I knew this second charge test was "spot on". Aside from dinged paint, no damage. I swapped out the P6K for an AIM and ditched the tracker. Heck it won't fly above 3k' anyways but I got a good lesson on Rf/altimeter interference.

I now follow Kevin's and Bdale's advice to a "T" . But even with that much care there are no guarantees.

I will say I've stuffed a 12mW Beeline GPS in a Wildman Jr riding with a Raven 2 without any trouble whatsoever. The Beeline rides on top with the
antenna projecting tightly through a form fit hole in the forward bulkhead. I use a cardboard tube that the AT first fire igniters come in and wrap it with duct tape for flame protection. Use clay to hold the tube in place and stent the antenna so it doesn't get smushed by the parachute.

The parachute bay is painted with yellow non-metallic paint. The sustainer with red metallic. I flew the rocket with a J-330 and it simply disappeared
up high. No one saw any of the events but it was easily seen from the APRS radios that the apogee charge did it's job. The descent slowed further at
the main deployment altitude but still no rocket/main chute was seen. It was only seen on the Garmin mapping GPS and the laptop tracking program.
I just set the Garmin to "navigate to" and by golly followed the path to where "automagically" the rocket was lying. I wouldn't have had any idea where it was otherwise.

The low powered Beeline trackers seem to play well with deployment electronics (except the P6) but I wouldn't have the antenna of any tracker inside
the ebay with the deployment electronics. Kurt

I can attest that a Raven is mostly insensitive to RF. I flew one with a Telemini antenna wrapped around it. But I had tested that, so I wasn't going to be blindsided.

 

[troj]

I can attest that a Raven is mostly insensitive to RF. I flew one with a Telemini antenna wrapped around it. But I had tested that, so I wasn't going to be blindsided.

Adrian also regulary flies his with a BeeLine transmitter in the avbay, right along side it.

-Kevin

 

[UhClem]

You beat me to it.

My ground testing of RF interference (Garmin Astro, specifically) indicates that you need to let the setup "bake" for over an hour -- I have a specific combination I've had fail within 10 minutes one time, and take over an hour another. Exact same setup both times, down to the leads, batteries and wiring. I use christmas tree bulbs as my test "match" when I do this.

My recommendation is two hours, and do it multiple times.

This didn't help in my case. I was concerned about the potential for EMI so I packed everything up, attached e-matches, and let it sit. Nothing happened. Nothing happened on the first two flights either.

Back to the original post, I'd be interested to find out how far away a transmitter has to be before it doesn't confuse the P6K -- with it being that susceptible, I have to wonder if someone flying something putting out watts of power (Garmin Astro, again) on the next pad over might confuse it.

-Kevin

Unlikely because the field strength drops rapidly with distance. Roughly proportional to the distance squared so a transmitter 10 feet away produces 1/100 the field strength of something one foot away.

To make it even more interesting, it isn't only the antenna you have to worry about. That antenna is working against the ground inside the transmitter so that the entire transmitter must be considered as part of the antenna. And if you are close enough for it to be considered in the near field of the antenna, the field can do odd things so the distance squared relation may not apply.

 

[ksaves2]

This didn't help in my case. I was concerned about the potential for EMI so I packed everything up, attached e-matches, and let it sit. Nothing happened. Nothing happened on the first two flights either.



Unlikely because the field strength drops rapidly with distance. Roughly proportional to the distance squared so a transmitter 10 feet away produces 1/100 the field strength of something one foot away.

To make it even more interesting, it isn't only the antenna you have to worry about. That antenna is working against the ground inside the transmitter so that the entire transmitter must be considered as part of the antenna. And if you are close enough for it to be considered in the near field of the antenna, the field can do odd things so the distance squared relation may not apply.

Reminds me. A couple of years ago at MWP a 12 inch diameter 16.5' tall O powered project came in ballistic and buried itself past the fins in a cornfield.
Had two Adept 22's and carried a Garmin Dog Tracker in the same ebay albeit on the other side from the Adepts. The altimeters reset on ascent and the rocket wasn't seen again until a landowner found it. The dog trackers put out 2 watts on the 150Mhz MURS frequency if I'm not mistaken. That's a lot of Rf energy for the electronics to contend with.


If one wants to run dog trackers, best to put 'em in the nosecone or a shockproof carrier to ride on the sustainer shockcord as far from the ebay as possible. Testing and shielding the the electronics in the ebay is good insurance. Yeah, if one can hear by word of mouth of what electronics work with what tracker can help. The low powered Beeline GPS and I would suspect the Beeline NDB work fine with the Raven in close proximity. Kurt

 

[bdale]

https://www.lessemf.com/mag-shld.html - perhaps this could be of use. I was starting a search for small mu-metal boxes but came across this first.

I do have a severe dislike for the lack of shielding in our typical rocket installations.

Shielding can be a great thing, but it's actually really hard to get effective shielding "right". Shielding done wrong can turn a radiated interference problem into a conducted interference problem, and any conductor that needs to penetrate the shield is a potential source of interfering energy crossing the shield boundary. RF designers often understand this very well, folks designing non-RF electronics often just don't think about it at all, and/or don't have the training and experience to solve problems when they appear. That's not being pejorative, it's just what I've observed.

Because most Altus Metrum products include one or more radio subsystems, we design all of our products to work well in the same ebay in various combinations. We actually test this, too, because the most common way we test new products and firmware updates is to go fly a rocket with a known-good TeleMetrum board taking care of deployment in the same ebay with one or more "passengers".

You know, it's sort of too bad we have to put our ear to the rockets in many cases to hear the beep patterns to discern if an armed altimeter is working correctly...

Yep! Note that in recent firmware versions, Altus Metrum products beep out the igniter continuity tones over the radio when in pad state... :wink:

 

[Adrian A]

I can attest that a Raven is mostly insensitive to RF. I flew one with a Telemini antenna wrapped around it. But I had tested that, so I wasn't going to be blindsided.

There are a few factors that help to make the Raven relatively immune to RF interference 1. It uses a digital-output barometric sensor that has all of its analog signal measurement and amplification on one chip. I think the Raven was the first production altimeter to do this, but now it's typical for recent-vintage altimeters. 2. The back side is a mostly-continuous ground plane. It's not perfectly continuous, but it helps with shielding and 3. The small components and tight routing avoid copper traces that are long enough to be effective antennas for typical RF frequencies. These are just good design practices that are also followed by several other manufacturers.

The now-obsolete Featherweight Parrot provided an interesting test case, since it had significant RF susceptibility in its analog-output baro sensor. That sensor was a type that needed a separate amplifier chip to make it useful, and I think the RF was getting into the signal chain at a low voltage and then was amplified along with the real signal. Interestingly, the Parrot's baro sensor didn't have any trouble with RF unless deployment charges were connected. The ground plane effectively shielded the circuits on the board when it was by itself, but deployment charge wires acted like an antennas that snuck the RF energy into the board through the back door, so to speak, defeating mitigations #2 and #3. But with or without charges hooked up, the Parrot's (and Raven's) analog-output accelerometers aren't measurably affected by RF, because the output from the chip is at a high enough voltage that it doesn't need any external amplification, the output traces are short, and they're shielded by the ground plane and other components.

If you have an altimeter like the Raven that records all of the measured data it uses to make decisions, you can test it for RF susceptibility without needing to wait for hours. Just put your whole av-bay into a flight-like configuration with the transmitter on, get the altimeter recording, and then graph the altimeter data to look for variations. When the Parrot had trouble, the "beeps" of a tracking transmitter showed up as clear step changes in the baro data when the radio was transmitting.