Is CGM better with MM 670 than with the 630?

I’ve been on the Minimed 630 for roughly a year and a half or so. I’m on the wait list to get a 670, thanks to their free upgrade program, now that my insurance finally accepted the upgrade.

But the CGM is driving me nuts on the 630, so I’m kind of afraid to plunge into closed-loop territory when I get the new pump. When my 630’s CGM is working correctly, it’s an awesome thing that helps me.

But sometimes, it’s just dead wrong. For example, last night around 1am, it woke me up, alarming that I was at 70. This was the second night in a row of this, and I didn’t trust the sensor. Sure enough, I checked and my blood sugar was actually 170. I didn’t calibrate (any time I calibrate when the sensor is that far off of reality, the pump rejects the calibration).

So over the next four hours, my pump continued to give false alarms until it finally decided to shut off my pump at 5 am. I got up, re-tested my sugar, took a correction bolus, shut off the sensor and turned my basal delivery back on, and went back to sleep. I turned it back on after my shower, calibrated again (my blood sugar had actually gone up). The calibration was successful. But then about 11:30 am, it again alerted on low. I tested and was high, and it rejected the calibration.

I killed the last sensor 2 or 3 days early in the same cycle of false lows ending in 2 failed calibrations in a row and the pump ending the sensor early. I’ve learned that calling Medtronic after the half-way point usually means a waste of time. But about every 6 weeks, I get to call their support line (they have great support people, really), and explain that yet again I’ve fried a sensor. They mail me out a new one and I go on about my business.

But I’m scared to turn my bolus delivery over to a sensor system that’s so frequently out of calibration with reality. I’ve asked Medtronic folks a couple of times, and their response is “Oh, sure, the new sensor is better!” but I feel like the sales folks aren’t really listening. And the support folks are trained to deal with specific inquiries that aren’t that general. My doctors don’t really know enough to answer me either.

The two-hour window between inserting the new sensor and the first calibration means I have to be careful how I plan for changes, so these unplanned failures throw things off every time.

I guess I’m just frustrated. And nervous. I’m careful to calibrate when the pump isn’t showing arrows indicating BS changes, or when I have active insulin from bolus deliveries… But I keep having problems with failed calibrations and/or wildly inaccurate sensor readings. I don’t really know why.

Sometimes, I know after the fact, because there will be blood at the sensor site when I change it out. But most of the time, there’s no way to know.

Is the sensor you are using the “Enlite”?

As far as accuracy goes, the Guardian 3s really are as good as claimed, in my experience. I wore one in tandem with my Dexcom 5 in the ten days or so prior to going on auto mode and they were equally accurate, maybe a slight edge to the Guardian. I also experienced fewer bluetooth dropouts with the Guardian.

This is not to say there aren’t other issues. Re calibration, there is that two hour “warm up” window, but Dexcom has a warm up period as well. A bigger problem is that the Guardian stops giving readings when a calibration is required, which is annoying, especially if you’re not running in auto (where it makes more sense). Insertion, with all the taping and whatnot, is much more of a pain, though if you’re used to Enlites I think it’s the same. People don’t necessarily do all that by the book, but while I was in auto mode training they were very particular about it.

If you search my posts I’ve got a couple up about the 670G from last summer/fall, when I started–lots good Q&A there. Full disclosure: I went off auto in mid-December–just wasn’t a good fit for me. So I’m not a big booster, but happy to share my experience good and bad!

ETA: I never used Enlites, so don’t have personal experience of them, but everything I’ve read or heard put me off of them–used Dexcom with Minimed Paradigm for 4 years, and was skeptical when looking at the prospect of going to the 670G. But as far as I can tell they really are a different beast from the Enlites.

1 Like

@DrBB - I thought your write-ups were very good. Seemed quite devoid of bias and very matter-of-fact.

Excellent reading material for anybody seriously considering the 670G.

1 Like

Yes. The Enlite sensor. I tried the previous generation, back around the 530 series pumps (I think? it may have been the pump before my 530?), and never refilled the sensor prescription even once. They were just too unreliable to ever use at all. So the 630’s Enlite is superior to that at least.

Sof sensors were so bad I got a refund on the entire CGM from MM, except for the cost of the sensors that I used. The Enlites were “better”, but not better enough for me. I’ve complained about them numerous times on this site so i won’t belabor the point.

1 Like

Gary Scheiner posted this comparison/review

I started on the 670, with the Guardian 3 sensors, about a month ago. I’m finding the sensors to be remarkably accurate during the day and consistently, inaccurate at night. I will get a nice, steady reading around 115-120 all night long however a test, either during the night or when I wake up, shows I’m actually 170-180. Of course I have no idea if I’m that high/steady ALL night but, it’s certainly discouraging. Medtronic suggests I try the sensor in my arm rather than stomach. Not sure how I’ll manage the insertion with one hand - - quite a process - - but, I may try.

1 Like

I’ve gone through and read several of your posts. Wow, very insightful. Yet a little unsettling. I can’t help but envision the pump speaking in HAL 9000’s voice. “I’m afraid I can’t let you do that, Dave.”

Not soothing!

I also have long suspected my various ratios and settings are not right, but have never been able to tweak them to the point where they felt right to me. Maybe the 670 will “fix” that for me.


If I had a complaint about their accuracy, this would be it. They do seem to have a bias toward reading low overnight, which is a real problem if you have severe DP, as I do. Because the pump can’t respond to data it doesn’t see. I tried getting up and doing a test/calibration at 3-4 a.m. to convince it to get ahead of rising a.m. BGs, but my DP really only starts to kick in around 4-5, and to head it off you need to be ramping up the basal well ahead of time. I managed to catch it a few times, but cripes, if I’m up jabbing my finger at 5a.m. I’m basically up for the day anyway. And I don’t need a pump so “advanced” that I have to keep getting up at stupid-o’clock in the morning to try and keep it on track. They say it “learns your pattern,” but this pattern is about as predictable as any I have, and I saw no evidence that the algorithm was anticipating it. Of course it starts trying to correct once it has the data, but that’s after the effect has already kicked in, and corrections are a sl-o-o-o-o-w way to fix that, especially with this pump.

After ~3 months I’d seen some improvement overall but I just couldn’t get past this issue, with only two parameters I could change and one of 'em (AIT) already maxed out. Whereas in standard pumping I’d long ago gotten my basal dialed in to where my fasting a.m. BGs were regularly in the ≤ 120 range. So I had to ask what I was getting in return for all the stress and hassle of trying to flog auto mode into working for me, and to be honest with myself I had to say the answer was… not much, really. I really had hoped for better, and I don’t like throwing in the towel–I’m jealous of people who don’t seem to be having these problems, but for me the results just weren’t there for the amount of effort it was demanding.

1 Like

Sadly I’m arriving at the same conclusion but, am determined to try longer. I’m hoping I can at least work out the sensor issue so I can use the setup in manual mode. I’m so disappointed and feeling more than a little deceived (just naive?) to have bought in to the hype/marketing.

Jut curious . . . . does the DP show on your graph or do you also get a beautiful, steady picture?

Well all of the preceding comments are very interesting as they seem to mirror the problems I had with Enlites yeah I believe you all are talking about Guardian. The Enliteswould always go low when I laid down. So unacceptable. it would alarm and alarm and alarm and it was all fake

I’ve been wearing the G5 since September 12th and if it ever goes low it is just a small amount like maybe 6 to 10 points. Nothing like the Medtronic sensors which could go more than 60 points Lo habitually.

My wife didn’t have as much trouble with Enlites as I did.

Well, this feeds my misgivings about talking about my own experiences because I’m afraid I sound like I’m bad-mouthing the thing and I don’t really mean to do that. I certainly don’t want to put people off of it, because a lot of people are much happier with it than I was and YDMV as always. So “hype”–I dunno, maybe that’s fair, maybe not. But I do think it suffers from being such a one-size-fits-all design. There’s so little you can do to customize it to your own metabolic peculiarities because the algorithm is supposed to do all of that. I feel like if I could just have had some control over the target BG, for example, it might have helped. Or some kind of user-input capability that would allow you to force a more aggressive profile in the early a.m. to account for DP, which is a common and widely recognized issue, and in fact a lot of the discussion on the dedicated support groups on FB were about that. Enough so to justify having a “DP factor” built in somehow.

1 Like

I’m glad the closed loop tech has gone to market.

But like any new tech, future versions of it will find ways to make it better.

And that’s good for all of us.

I’m hopeful that the 670 works for me on Auto. We’ll see if/when it arrives.

A few take-aways I’d bullet from my experience:

  • “Basal” – To paraphrase Inigo Montoya, that word doesn’t mean what you think it means. They call it “microbolusing” for a reason–there isn’t the strong distinction experienced pumpers are used to in conceptualizing this stuff. In a way, it’s all-bolusing-all-the-time.
  • Active Insulin Time. This is the single most powerful parameter you can adjust (for why, see the previous bullet). Don’t think of it as an actual empirical measure of how long insulin is staying active in your system; just think of it as a dial you can turn to make the microbolusing more or less aggressive. I’d lay long odds that almost no one ends up with the same AIT setting under auto as the one they were using under manual.
  • IC ratio. This is the only other parameter you have any control over, and the only one that allows specifying different values throughout the 24-hr period. Theoretically it only applies to meal bolusing, but there’s speculation and some evidence that it may do more than that (for why, see first bullet). I had been experimenting with setting different values for this overnight to see if it would have an effect on DP, and I saw some indication it did, though things started to slide back again, and I finally ran out of patience.

As always DBEs are going to be very conservative about changing any of these parameters, because insulin==danger!, and there’s also this six-day thing, the interval during which the algorithm “decides” to integrate any new input and make it part of its configuration going forward–you don’t see the results the next day. So experienced pumpers may be shocked to feel like it’s a lot more like learning how to use a pump for the first time (I almost think it’s an advantage not to have had any pumping experience, since you won’t have assumptions about how things work that turn out to be erroneous). But at any rate it really is just as gradual to acclimate to and dial in, if not more so, as that was. Everyone says “You gotta be patient” and they’re right.

I mostly eat LCHF, especially at night, so unless I’ve broken pattern for some reason it’s usually a pretty steady curve up until ~4 a.m., when it starts ticking up at an accelerating rate. Like everyone else, I worry about overnight hypos, but I usually only get those when I’ve gone out to dinner or a party and had to guesstimate a big bolus for the occasional high carb indulgence or Wait, what did you say was in that dip again??? situation. One of the attractions for me was the “Never worry about hypos in your sleep!” promise, but it’s not as critical a problem for me as for some.

1 Like

Got it. Mostly my concern is that I show a steady line at +/- 115 all the way 'til wake up. The reality of 170 is confirmed by a BG test. It’s GREAT not to worry about nighttime lows but, confidence in accuracy will bring me most comfort. WIP.


…and there goes my Enlite sensor with two failed calibrations in a row. It’s only 2 days old. Yay, I get to call Medtronic support.

I’m so lucky!


Here in Luxembourg the CGM just got full coverage from the national health insurance. Until now we were blessed only with Libre. Anyways, the options given are Dex and Enlites, but Medtronic lobby seems to have worked well as the conditions are formulated something like “if you have a CGM compatible pump you get enlites”… as the only compatible pump on the market is MM.

Now, if I had not tried Enlites previously, I would have been super thrilled. But, having self financed Enlites for about a year in order to try, I KNOW that I’m never going down that road again.

1 Like