RRC2-mini malfunction mystery

Typically I would go straight to the source for this question, but with Jim in limbo with the flooding in CO I thought I would post my problem here and hear your thoughts on this.

Last week my RRC2-mini in my Vindicator Jr simultaneously ignited the main and drogue charges while sitting on the pad. I switched on the altimeter (twist and tape), it went through the start-up sequence, and then started beeping normal continuity for the main and drogue charges (3 quick beeps). After about 15 seconds of beeping the beeps stopped and then 5-10 seconds later the charges blew. There were no beeps from the altimeter after it blew the charges.

Looking a little embarrassed, I collected my rocket parts, returned to my table and took the AV bay apart looking for a short or something. I didn't see anything so I powered up the RRC2 without charges connected and it appeared to power up normally and I let it beep for about a minute - no problems. So, I tried again - figured it must have been a transient short. I got the rocket back out on the pad, turned on the altimeter and it did it again. Exact same sequence of events.

This rocket has launched at least 12-15 times with this altimeter prior to this event. Wind was minimal that day 2-5mph max. The battery was a new Duracell 9V. It still reads 9.15V on my meter.

At home I've connected wires to the main and drogue terminals to simulate an e-match being connected and the altimeter just behaves normally. I can't get it to fail again - but I don't trust it now either. I haven't connected actual e-matches to it at home - maybe that's next.

Any thoughts???? :confused2:

Is the launcher wireless?

How many holes in the av bay?

Do you have a tracker in the rocket?

-Kevin

troj said:

Is the launcher wireless?

How many holes in the av bay?

Do you have a tracker in the rocket?

-Kevin

Click to expand...

It was not a wireless launcher
3 holes in the AV bay (Don't recall the size off the top of my head - followed the online Darkstar Jr instructions)
I had a 900MHz GPS tracker connected near the nose cone - not in the AV bay (Homemade - Derek's Rocketracker kit/plan)
- I had not previously tried to launch it with this tracker - first time in this rocket

farsidius said:

I had a 900MHz GPS tracker connected near the nose cone - not in the AV bay (Homemade - Derek's Rocketracker kit/plan)
- I had not previously tried to launch it with this tracker - first time in this rocket

Click to expand...

For what it's worth, I've flown an RRC2-mini in about the same configuration with the same radio that Derek is using -- or least the previous generation XBee 900 Pro XSC, which did have lower xmit power (100 mW instrad of 250 mW on the newer S3B modules.)

I'd try doing a ground test with the tracker in proximity to the RRC2 and light bulbs instead of charges. The RRC2 is known to be RF-sensitive to high enough power levels.

It's possible you need some shielding on your charge and power leads on your RRC2 Mini, to protect it from the 900Mhz transmitter. Twisting the leads may help, as well.

-Kevin

Thanks for the prompt feedback. I'll play around with RF interference tonight and see it that does it. I hope it's that simple.

farsidius said:

Thanks for the prompt feedback. I'll play around with RF interference tonight and see it that does it. I hope it's that simple.

Click to expand...

The RRC2-mini is indeed RF sensitive... What's known is that the 2W Garmin Astro dog collar in proximity has created issue. I concur with the other's assessment that it may be interference, however there's not been any reports of 900MHz ISM transceivers causing false positive baro interference... try without the tracker initialized and ensure that there's not another instability in the baro system chain that's causing you grief. It's pretty coincidental however that this operation manifested on your first launch with the tracker.

I can be reached via email now as I have established temporary internet connection using a neighbors sat dish... no phone or DSL service is probable in the near future.

Farsidius, I can't help with your problem, but I love your little cartoon!

It's worth pointing out that I've flown a 100 mW 900 MHz tracker with the RRC2-mini without problems, but I use Q2G2s with leads between 2 inches and 8 inches. Are your leads longer?

BTW, the cartoon is from https://xkcd.com

mikec said:

It's worth pointing out that I've flown a 100 mW 900 MHz tracker with the RRC2-mini without problems, but I use Q2G2s with leads between 2 inches and 8 inches. Are your leads longer?

Click to expand...

Lead length, and relative orientation of the various components can be a factor.

-Kevin

I'll let you know what I find tonight. Glad you have some communications available to you. My wife's family is all in Colorado and it was a little frightening to get their first hand accounts of that flooding.

MWC said:

The RRC2-mini is indeed RF sensitive... What's known is that the 2W Garmin Astro dog collar in proximity has created issue. I concur with the other's assessment that it may be interference, however there's not been any reports of 900MHz ISM transceivers causing false positive baro interference... try without the tracker initialized and ensure that there's not another instability in the baro system chain that's causing you grief. It's pretty coincidental however that this operation manifested on your first launch with the tracker.

I can be reached via email now as I have established temporary internet connection using a neighbors sat dish... no phone or DSL service is probable in the near future.

Click to expand...

cvanc said:

Farsidius, I can't help with your problem, but I love your little cartoon!

Click to expand...

One of my favorites. Perfectly describes my attitude towards the unexpected - probably explains why I do what I do.

mikec said:

It's worth pointing out that I've flown a 100 mW 900 MHz tracker with the RRC2-mini without problems, but I use Q2G2s with leads between 2 inches and 8 inches. Are your leads longer?

BTW, the cartoon is from https://xkcd.com

Click to expand...

One lead is only about 3", the other is probably 1.5 ft (I push it aft towards the motor)

I'll test lead length as well. Thanks again for all this input. I love problem solving.

-brant

farsidius said:

One lead is only about 3", the other is probably 1.5 ft (I push it aft towards the motor)
-brant

Click to expand...

3" is about 1/4 wavelength at 900Mhz. Try lead lengths that not close to multiples of 8.33cm.

I haven't done the math, but my gut feeling is that the length of the igniter leads is probably not an issue. The impedence of the igniters is so low that it would take a lot more than the induced power from a 100 mW transmitter to fire the igniter, even with a Q2G2 (all-fire of around 200 mA at 3.7V, I fire them with 1S LiPos all the time). What is more likely is that somehow the RFI is inducing interference with the much more sensitive circuitry on the controller. A few mW of induced current could play havoc with high-impedence CMOS circuitry.

cerving said:

I haven't done the math, but my gut feeling is that the length of the igniter leads is probably not an issue. The impedence of the igniters is so low that it would take a lot more than the induced power from a 100 mW transmitter to fire the igniter,

Click to expand...

That's not the theory of the malfunction. The lead being whose length is close to a 1/4 wavelength will act as a 1/4 wavelength transmission transformer, transforming the low impedance of the match head into a open circuit at 900mhz. This allows relatively high RF voltage to be generated at the FET input in the presence of a strong 900Mhz electric field. This can cause noise on the analog traces or in the baro sensor which the altimeter can interpret as a launch detect, with random events to follow.

I wonder if anyone has done any analysis on this with a scope. I can see that if you have an analog baro sensor and a high-impedence ADC then it may be possible to induce enough voltage to fake out the ADC. A 20-bit ADC at 5 volts gives only about 5 uV per LSB, that's pretty low. I would think that a ground plane underneath with grounded guard bands around the trace, and keeping the trace as short as possible would pretty much eliminate the RFI risk, though.

jderimig said:

That's not the theory of the malfunction. The lead being whose length is close to a 1/4 wavelength will act as a 1/4 wavelength transmission transformer, transforming the low impedance of the match head into a open circuit at 900mhz. This allows relatively high RF voltage to be generated at the FET input in the presence of a strong 900Mhz electric field. This can cause noise on the analog traces or in the baro sensor which the altimeter can interpret as a launch detect, with random events to follow.

Click to expand...

cerving said:

I wonder if anyone has done any analysis on this with a scope. I can see that if you have an analog baro sensor and a high-impedence ADC then it may be possible to induce enough voltage to fake out the ADC. A 20-bit ADC at 5 volts gives only about 5 uV per LSB, that's pretty low. I would think that a ground plane underneath with grounded guard bands around the trace, and keeping the trace as short as possible would pretty much eliminate the RFI risk, though.

Click to expand...

Two words: Antenna Theory.

The length of the trace between the analog baro sensor and the ADC would probably be way too small to be an effective antenna at 900MHz, probably 2mm or less, so I doubt that RFI-induced interference with the baro reading could be a cause of this problem. That's why I'm saying that it would be interesting to scope it out. I've only seen anecdotal comments about telemetry transmitters causing interference with altimeters; obviously they must be able to coexist since there are several vendors that make combo units, such as Altus Metrum. I haven't heard of anybody recommending any remedial RFI shielding or other countermeasures, nor do I see any being used by any of the vendors.

cerving said:

The length of the trace between the analog baro sensor and the ADC would probably be way too small to be an effective antenna at 900MHz,

Click to expand...

The induced voltage is not at the ADC trace. Some MMC pressure transducers and their integrated circuitry is susceptible to RF creating noise on the outputs. That is a known and well know and repeated observation. Any external leads (power, ematch etc) to the board if acting as an antenna will act as a port of RF into the board. Its will be on the power and ground planes and sometimes not in phase.

Often the issue is not the field effect on the board itself but the interaction of the board and its interfaces. That is why EMI certification is never done at the board level but must be done at the system level.

I haven't heard of anybody recommending any remedial RFI shielding or other countermeasures, nor do I see any being used by any of the vendors.

Click to expand...

That is because RFI shielding is futile if the board isn't designed with good EMI best practices. Not all altimeters are sensitive to RF, some may be.

obviously they must be able to coexist since there are several vendors that make combo units, such as Altus Metrum

Click to expand...

They can coexist but you have to design them to coexist (or depend on luck). I believe the Telemetrum did indeed have a board revision because of an RF interference issue which caused resets.

cerving said:

I haven't heard of anybody recommending any remedial RFI shielding or other countermeasures...

Click to expand...

There was a pretty extensive article about the RRC2 being confused by a Garmin Astro (4 watts EIRP IIRC at about 150 MHz) in an article in SPORT ROCKETRY in 2009. The author fixed it by adding shielding to the nose cone mount of the Garmin.

Regardless of how it's happening (the RRC2 doesn't have any ground planes) and how to diagnose it (a scope is more likely to confuse things further IMHO because of coupling into the leads, and the signals we're talking about are very low) a controlled experiment to see if the tracker is the issue at all would be the first thing to do.

First: I realized last night when I recreated the flight set up that I lied about the placement of the GPS tracker in this rocket. The tracker was hung from the eye bolt beneath the AV bay, not from the nose cone. On my other HPR setups I mount it in or below the nose cone. But this rocket has a very small payload section that is barely big enough for the main chute and the nose is sealed (was my first dual deploy rocket - wasn't thinking of trackers and such when i built it).

Method:
I took two pieces of left over ematch leads, cut one to 3" for the main charge and the other to 22" - the length needed to reach from the connector on the bottom of the AV bay to the top of the motor mount in the booster. I connected the leads to the appropriate terminal blocks, twisting the free ends of each to close the circuit. I placed the GPS tracker at the base of the AV bay. This pic shows the setup.



Using this layout last night I powered up the altimeter bay 6 times each with and without the tracker turned on. I also tested the altimeter with the tracker placed about 12 " above the AV bay simulating where it would be if I could hang it from the nose cone.

Result:
With the GPS tracker powered on below the AV bay, five of six tests ended with the Altimeter malfunctioning. For the failures, each time the unit would power up normally, then after 5-10 seconds of charge continuity beeping, the beeping the would stop and the altimeter light color would change from green to red. At this point I would turn off power to the altimeter. This pattern is identical to what I heard when it was on the launch pad (I couldn't see the altimeter light when it was on the pad, and of course, there was no exploding charges throwing nose cones and body tubes into the air).

With the GPS tracker off and placed below the AV bay, there was no malfunction of the altimeter during my six tests.

With the GPS tracker on and placed 12" above the AV bay, there was no malfunction during six consecutive tests.

Conclusions:
My simple experiments would support the idea that there is interference to the altimeter from the GPS tracker when the tracker is powered on and in close proximity to the lower ematch lead.
I could probably do some additional testing regarding proximity (off to the side of the AV bay, next to the 3" match lead, various distances from the bay and leads) but I think I've answered my initial question. I guess I'll have to find a different location to mount the tracker in this rocket. Maybe I'll have to open up the nose cone.

Thanks for all the discussion on this. It's always fun to find the cause of an unexpected event, especially when the event was a little embarrassing.

-brant

Brant, thanks for this follow on testing and report. Helpful for all to learn.

+1 on that.

dixontj93060 said:

Brant, thanks for this follow on testing and report. Helpful for all to learn.

Click to expand...

farsidius said:

My simple experiments would support the idea that there is interference to the altimeter...

Click to expand...

Thanks for doing this. If you're in the mood to do more tests, it'd be interesting to know if using a ahorter lead on the lower charge would show the same results. I don't think you need the apogee charge very far from the bottom bulkhead and you could easily use shorter leads (though maybe you're putting it where it is to avoid blowing up the tracker.)

mikec said:

Thanks for doing this. If you're in the mood to do more tests, it'd be interesting to know if using a ahorter lead on the lower charge would show the same results. I don't think you need the apogee charge very far from the bottom bulkhead and you could easily use shorter leads (though maybe you're putting it where it is to avoid blowing up the tracker.)

Click to expand...

I use the longer length to put the charge near where the motor ejection charge would be. On this small rocket, I usually keep the motor ejection as back up to the altimeter controlled charge. Thus all my "packaging" for the apogee event is the same.

I will do some more testing in a few days. I can test different lengths of wire. I want to be really comfortable with the distance of the tracker from the altimeter.

Make sure your ematch test leads have continuity, a pair to an open will be different than a pair to a short as far a RF susceptance goes. You may not want to test into an actual short as it may damage your altimeters outputs. Ideally you should use real ematches to simulate the actual state while protecting your altimeter.

Try twisting the longer match leads and see if it makes a difference.

It would be interesting to try your test with other altimeters. If it was me I would not fly this altimeter with any trackers.

terryg said:

It would be interesting to try your test with other altimeters. If it was me I would not fly this altimeter with any trackers.

Click to expand...

Or only in applications where the e-match leads are not long and laying across the tracker. I have used the RRC2-mini hundreds of times with trackers with no mishaps.