NOT a compression low!?!

Got my CGM leash yanked twice last night for false alarms:

I was 99% sure they were erroneous—I mean, just look at those curve lines. I’ve been at this a LONG time and I know my patterns. No-carb dinner, so no lingering bolus insulin, not to mention the first alarm was like 6 hrs after dinner. So “compression low” methinks. EXCEPT when it yanked me awake I was on my left side, opposite the shoulder I have my sensor on. “Musta rolled over in my sleep,” I guessed, though that’s pretty unusual for me bc I have arthritis in my shoulders that wakes me if I’ve been on one side too long and makes the process of rolling over pretty awkward and painful in its own right. It’s not something I can easily manage without coming fully awake, more’s the pity. Still, I couldn’t figure out what else it might have been, given that abrupt a drop and the fact that I didn’t feel at all low, so I cancelled the alert and went back to sleep. Very emphatically ON MY LEFT SIDE.

So all’s well until just before my usual get-up time when it went off again. SEVERE LOW this time, breaking-glass alert sound, iPhone notification showing 50 with a straight down arrow. Now I was absolutely on my LEFT side, position unchanged from 5 hrs earlier, and wasn’t even fully asleep this time but had been dozingly hoping for a brief return to dreamland (hey, I’m retired, I don’t need to be up yet). I have my basal rate set to compensate for DP starting around 3am, but that’s not going to produce a drop like that. Sure enough, when I did get up and do a finger stick the meter said 88, which is a bit below avg for me on rising but not remotely a hypo reading.

So chalk this up to “Weird sh** that happens with T1” I guess but I’m now pretty uncertain what to do about it if anything. It’s a 2-day old sensor, so should be out of the erratic phase, but I’m really not looking forward to another night like this. Guess I’ll just have to keep an eye on it all day and if it seems like a bad-sensor issue I’ll just have to swap it out and call Dexcom for a replacement. Fortunately they’re pretty generous about that.


Those are consistent with “recalculated data” as AndroidAPS terms it. The sensor/transmitter reports a significant change in BG (normally a drop for me); so the downwards part of the two drops you show. Readings continue for some time (maybe 4-6 readings, 20-30 minutes) then the transmitter notices that the system is borked and “recalculates” the recent values; the receiver/app will report the new values (as above) and not 'fess up to the change. xDrip+ reports the old values (or it used to, I haven’t checked recently) and AndroidAPS takes the new values but displays a “warning triangle”.

It’s a feature; G6 systems have done this since inception. I’ve always referred to it as “rewriting history”.

Here’s a relatively minor example of “recalculated data” from last night. This is what xDrip+ got from the Dexcom G6 app:

The red “low” reading was at 23:30 and reads 68mg/dL. The G6 app also transmits to AndroidAPS and it downloads to NightScout. It shows the “recalculated data”:

The big red dot in the middle is my entry, because I felt my BG going low so I did a fingerstick at 23:15 (52mg/dL) then did a “hypo” temporary target and entered the sensor reading at (from NightScout) 23:31. You can see that the preceding reading is higher. The actual readings are:

Time 23:15 23:20 23:24 23:30 23:34 23:39 23:44
xDrip+ 84 84 74 68 72 75 80
NightScout 84 81 75 71 72 76 80
FingerStick 52

You can see this is a very minor re-write (and it was actually in the wrong direction; my FingerStick BG has always been lower than my G6 BG recently). Nevertheless this is what I’ve seen previously where the G6 pretty much drops out then recovers after maybe 30 minutes.

1 Like

I, too, see these “lows” that are not compression lows. They do mess with your sleep as the only dependable way of dealing with them is to interrupt your sleep and do a fingerstck.

I write them off as sensor anomalies. Some sensors (or sensor body locations) are more susceptible to this phenomena and I have no explanation or advice on how to handle. I have a suspicion regarding manufacturing inconsistencies but have nothing to back up that.

I’ve also observed this in new, middle aged, and old sensors. Sometimes this is the early symptom of an ultimate early failure. I can’t live without the Dexcom CGM, so I just put up with it.


I get those fake lows sometimes around day 8 and not from compression. I would like to be part of a study for G7 on them working past 10 days.

I wonder if it has to do with my unique biochemistry. My last sensor lasted10 days out of pure stubbornness and lots of calibrations.

I just started looping with AAPS and these are driving me crazy!

Day 4 of my sensor, so it should be settled in, but seems erratic. Thankfully AAPS shows me a 15 minute ∆, not just 5. But since I’m watching the system like a hawk (I only have low glucose suspend available at the moment, a whole new system to learn which I don’t trust yet, and lots of old bad habits to un-learn), I keep eating panic carbs when I’m under 100 and see -15 or -20 since the last reading. And then like 15 minutes later, Dexcom is like, “oh, haha, nevermind… Psyche! Made you eat carbs for nothing, now you get to dose.”

I’ve never noticed erratic jumps like this before unless it was the end of a sensor’s life.

Slightly off topic but nevertheless… I moved to AndroidAPS from Omnipod Dash with the G6 because I couldn’t get the O5 (we are not waiting etc). The Dash is pretty much an open loop system (i.e. no closed loop option) so it is the same as AAPS for the first few days (depending on how you game the entrance exam :wink:

What I found with the open loop is that it was actually a whole lot easier and more flexible. I would use it on its own, without the closed loop option. Perhaps the difference is that I have had years of experience of Dash+G6 and I had to manually check the Dexcom app then transfer the results into the Dash PDM; an open loop for sure but a lot of steps. With AAPS the loop was still open but all I had to do was click “OK” when it wanted to do something. I also really liked the “superbolus”; it did something in one click (one checkbox) that would have taken a convoluted use of temp basals on Dash.

I didn’t see a lot of potential lows, but then I have a habit of picking up a slice of fruit if I feel I might be going low and I still do that with the closed loop (oref1, using SMB which pretty much replaces the superbolus).

I do see and, recently, have seen massive errors in the G6; it reads at least 20mg/dL high for me. I suspect this might be because of diet changes I ended up doing at the same time; I’m only eating enough carbs to keep me out of ketosis and I have to wonder if the G6 doesn’t start reading high if it sees ketones.


If @DrBB doesn’t mind is hijacking his thread…

My issue is that I’m a TERRIBLE diabetic. I don’t count carbs, I don’t pre-bolus, and I don’t test settings. I’m just knowledgeable enough to figure things out with the information the system gives me, and somehow I found magic with Control-IQ and Trulicity. With Control-IQ, I’m able to largely guesstimate my boluses, and the pump uses what I need of it and then suspends the rest. The Trulicity slows my digestion to minimize post prandial peaks, but it also works really well with Control-IQ s “sleep mode” to cover those lingering slow digesting carbs and protein rises. Anytime I’m predicted to go over 120 it adds more insulin, and the timing works out perfectly. Unfortunately, my pump is out of warranty and misbehaving, and I don’t want to commit to 4 more years of old tech.

I’ve been eyeing DIY looping for a long time, and i think AAPS will be to my benefit once I get to that shining UAM pot of gold at the end of the rainbow… (And once I get used to these giant, honking Omnipods. How do kids wear these???) But for now, I’m stuck trying to figure out how much carb is actually in my portion of scratch-made food with no labels because COB seems really critical, how long should I extend those carbs for since I do everything slower than normal, and when is the right time to microdose all the insulin Control-IQ would have automatically spread out for me? I can’t just calculate a bolus and deliver it when I eat, or I’ll bottom out in 30 minutes. And without basal modulation available to me yet, I can’t program the equivalent of an extended bolus. So now, every time I see one of those glitchy data plunges, I panic that I mis-timed the insulin, the carbs weren’t absorbed yet, and the ship is going down fast! If it were a -5, I’d chill and let the LGS do it’s thing. But I know those -20s need treating right away, because there’s no room to safely maintain that sort of momentum.

I’m just jumpy and uncertain about a lot right now. Thankfully I’m 3 days down out of 5 in this objective, so I see the light. And I have actually gotten better over these last few days, so I must be adjusting. Sitting at 97% TIR today, which is way better than the 67% my first AAPS day. But that’s also using the full 70-180. I used to be able to do 98% 70-160 reliably with Control-IQ. (Almost always edge a toe across the line, one way or the other, after dinner.)


I had one just like this with the G6 on Sunday. It just bottomed out, sent errors everywhere in the stratosphere. Took 2 hours to recover.

@John_Bowler This hasn’t been my experience, and I’ve pretty much been “in ketosis” since 2018.

In fact, while doing 3 day fasts and running ketone levels as high as 6 mmol/L, the G6 remained remarkably accurate.

1 Like

Thanks, that’s really helpful to me.

1 Like

Had another plunge yesterday mid-afternoon after posting this. Totally out of the blue–I was lying down but on my back, no compression. Finally broke down and called Dexcom Support. They said “take it off, start a new one and we’ll send you a replacement.” Hate doing that with no time to pre-soak, but if it’s a faulty sensor it’s no use to me. So I went ahead. And here’s what happened last night:

Multiple alerts, leash getting yanked repeatedly, and this time I had had to bolus for dinner so couldn’t be sure it wasn’t for real and just acknowledge and ignore. You can see I broke down and calibrated at 4am, even tho the techie had said “Don’t calibrate for the first 24 hrs.” But criminy, what else can I do? Yank it off and start ANOTHER???

And ok, some of those excursions were prolly compression. I opted for belly placement this time, which is usually less of a problem that way, and dang it I just do need to alternate sides to get through the night, and I haven’t had this degree of false alarms before, ever.

Here’s the thing. For years I’d heard of “compression low”, but rarely if ever experienced it. Arm, belly, left-side right-side, usual sleep habits, and this just wasn’t something I worried about. But it has started to be more frequent for me in the last year or six months, still intermittent enough that it wasn’t screwing with my sleep habits. And this recent episode is giving me a twinge of that old “T1 Claustrophobia”—that Oh crap I’m STUCK with this feeling when you realize you’ve got some issue that isn’t just a one-off but something more systemic you’re going to be dealing with. Distant echo of when I first got dx’d (1983). A very mild but not happy little taste of PTSD.

I thought maybe just a bad sensor, but the current one is from a new box with a different LOT number. Wondering if it’s some alteration in production where they’re just generating sensors that are way more compression-sensitive than they used to be. Which would suck.


Could it be the transmitter? I’d check some of the lows with finger poke (not necessarily calibrate, just cross reference).
Or you could have suddenly become mor insulin sensitive — that HAS happened to me a few times.

1 Like

The Tslim 2 is not really old tech quite yet. When you look at the two new pumps coming out if thy ever get here.
The tslim 3 has magnetic charging. Wireless updates But nothing else really changes.

The mobi is that very small patch pump that’s worn more like a pod and controlled 100% by your phone.
I’m all about it when it comes. I bought my tslim 2 years ago and they said I could get a free upgrade if it came out in a year or less. So yea still waiting and my warranty is half done.

Every time I purchase a pump a better one is just on the horizon. It’s just always going to be like that.


I hands down believe the X2 is the best pump on the market still. You won’t get any disputes from me there. Their constantly growing numbers are proof that others feel the same. It’s a fabulous product and still changing/improving lives.

But it was developed circa 2013. That is old for electronics technology. There have been massive evolutions since then in chipsets, battery composition and performance, Bluetooth communications, processing power and the critical ability to cool those elements, and so much more. Electronics today are much more capable than they were in 2013.

I’m even okay with “old tech” in the case of medical devices. Dependability is way more important than new and shiney. No disputes from me there, either!

What you’re overlooking, though, is that the biggest improvements in the next gen pumps have nothing to do with wearability or how they charge. (I actually think wireless charging is a bonehead move when Qi caps out at 15W and USB-C can easily do 150W. Anything is better than the current micro-usb, though!) It’s all about the processing power and how capable they are for handling the AID algorithms. There’s a lot of discussion right now surrounding the computing power necessary to run these forthcoming systems and how inadequate the current tech is for it.

Tandem did an amazing job of packing extra hardware into the X2 so it could grow into it’s future pipeline projects. But we’re reaching the end of that line. COVID put us at least 2 years behind on their timeline. There just isn’t much more room under the hood for it to grow into emergent AID technologies. They have promised one more big update for the X2, with “more customization options”, but they haven’t elaborated on what that means. I’m hoping that includes the ability to adjust the glucose target, but even that may be too complicated.

There are guaranteed to be updates and functionality given to the X3 and not the X2. This timeline might not concern most people, when we’re taking about arguing a few more years onto the wait. I get that. We’re used to slow progress in these things. But I think it’s irresponsible to not at least have it be part of the discussion when people are considering their options. Especially when Tandem’s reputation has largely been built on their successful history of exciting updates, and new buyers won’t likely see the same type of growth. For someone buying an X2 now, there will be major automation advancements in the near future they can’t have. We’re close enough that it might be the difference between access to a fully closed loop or not.

And honestly, I don’t know where the Mobi fits into that. Size, discreetness, and wearability were the prime design motivations. I haven’t seen anything about computing power enhancements. It’s a possible compromise I’m willing to make because I love Control-IQ, but desperately want a smaller form factor.

As to the free upgrades… Every time I’ve spoken with the “Tandem Cares” renewal department, they’ve been unable to comment about pipeline projects or free upgrades to them. They’ve flat out said that they’re not given any guidance on this, and that until the higher ups put it in writing, everything is just speculation. Leigh Vosseller (Exec VP and CFO) has said there will be upgrade opportunities, but nothing more. Nothing about who will be eligible. Nothing about if it’s free for anyone.


In my experience dropouts used to happen more frequently towards the end of the transmitter life. Alas the G6 app doesn’t report the battery voltage; xDrip+ reports the no-load voltage and the voltage under load (basically when the transmitter is actually transmitting).

Since the 8T transmitter (the first that stopped giving me accurate readings) I have not seen dropouts. I’m now on an 8X and haven’t seen any yet. The 8T hit the manufacturing expiry date while I was wearing it (but was still in warranty because I had used it within 150 days of receipt), I lost two sensors before the app said the transmitter had failed.

The 8X I’m using now also shows low readings (it’s low by about 30mg/dL; I’ve adjusted my profile settings to compensate), but I will replace it before the hourglass date on the packaging.

The next transmitter I will use will be an 8W, manufactured on exactly the same date as the 8X (but received two months later). If I run the 8W to 90 days I will be using that on 11/27/2022 which, fortunately, is still within the hourglass date. However I am not inclined to do that; the batteries really don’t last that long, maybe because they don’t or maybe because they sit in supply chain warehouses for six months or more. The battery is always connected; the transmitter has to monitor the sensor inputs to deduce when it has been inserted.

Compression lows can be caused by a body part on the sensor, yours or someone else, a pillow, a heavy or tightly rapped sheet, clothing or blanket, a compression of an ectremity where the sensor is, just to name a few. The reading jump is quite representative of a compression or lack of intrasitial fluid surrounding the sensor.

@DrBB I just removed a G6 sensor with this happening 4 times in the last 5 days. It is a first for me in the last 5 months of use.

So what happened? Once I removed the sensor I found the wire was bent slightly. and there was a slight blood spot on it. Now the sensor operated well most of the time, but it plummeted without warning and no it was not compressed.

I think it was the bent wire. but who knows I pulled it on the 10th day and it operated seamlessly about 98% of the time. I have to say all in all I felt pretty good about mine, but yes, I noticed the issue this time that i have not had since the switch over from Medt.

1 Like

I commonly see severe “fake hypos” that start as a deep drop on brand-new sensors in the first 6-12 hours if I don’t do a presoak. This is so common for me that I always do a presoak. Some of them can be as brief as 15-20 minutes bump-downs like you saw, but most are much more severe and show as a deep decline followed by “sensor failure” for several hours.

These fake hypos can happen after the first 12 hours too, but less commonly. Maybe I’ve seen them on the second day.

I’ve also seen “fake hypos” at 8/9/10 days. In one case the fabric of the Dexcom patch was still attached to me but had ripped, and the sensor wire was still in me but had pulled out a little bit.

1 Like

If you speak to sales, they have different talking points.
They told me initially that everyone with an in warranty pump could trade it in for a mobi for$ 1500 US. And that if your x2 was less than a year old you could do that for free.
There was no x3 in the news at that time. Those were selling points when I inquired in 2020.

The mobi is supposed to have its technology live on your phone and not on the pump. I have no way of knowing how that can be,I’m mean there are some hardware components that will age.

The point I was making earlier was that if a new pump came out today and you buy it, in 2 years there will be something better. We will always be in that reality unless you go with pods that are disposable anyway.
Tamdem probably has A T4 in early stages. Dex prob has a G8 on paper at least.

This happens every once in awhile. That’s why we keep those OG meters around with the test strips. Happened to me +/- 2 months ago. Dexcom screaming low BG, finger stick 102. Every once in awhile, we all need a little recalibration…