670 Automode & maxing Out?

I’ve been staying in the high range on auto mode, for the last 6 weeks. At first I assumed it was just bad programming. Then we’ve had a minor job crisis, which caused extra stress.

I’m trying to get my blood glucose in range, but it seems like it isn’t working well. I’m running in the low 200s most of the time, then I eat and I’m crabby. My hubby pretty much hates my moodiness at this point. I adjusted my active insulin time to 2 hrs and my insulin ratios to 1:8 and 1:9 in the early evening, I don’t know how much lower I can go?

And should I be thinking about type 2 meds to supplement insulin at this point??? I’m a bit frustrated…and still stressed. :flushed:

I’m at a nice weight for my size and frame. And I take taekwondo 3-4 times a week, but I probably eat more carbs than necessary (esp at lunch, like hamburger and fries or sandwich with small baggy of chips) but try to limit most of the time.

I’m sure you know that fries are evil.


If you don’t know how to respond, that’s fine. But I don’t really feel good on a permanent low carb diet. I don’t need to lose weight and I don’t eat more than 45 carbs per meal ever. Usually not more than 20 except when I happen to eat out at lunch, which isn’t all the time…but was yesterday as I left work after a noon meeting to take my kids to the dentist, so I ate a drive through burger. The point of the question was if my active insulin time is maxed to 2 hours and my ratios are less than 1:10 is that a lot of insulin and should I consider mentioning Symlin to my dr? Not should I change my diet because that’s fine. Let’s not be rigid and only eat the same thing every day, that’s not a good quality of life in my opinion.

1 Like

One thing about the 670 is that it’s very hard to get it to be aggressive about correcting highs, and a bit feeble about pre-bolusing as well. This was one of my difficulties with it–in my case starting high first thing in the day due to dawn phenomenon, not getting low enough by lunch time, which is when I tend to eat more carbs, then spiking, staying too high all afternoon, etc.

One way to deal with it is to just drop out of auto for a while so you can administer your own correction doses in order to break the cycle. I think this is preferable to what a number of people do, which is to use “phantom carbs”–telling the pump you’re having carbs you aren’t actually eating, or overstating the ones you are–as that distorts the information the algorithm will use going forward in calculating its behavior.

1 Like

I recall you had mentioned the Duration of Insulin Action (I think that was the setting) appeared to have much more of an impact than might have been obvious?

What was the most aggressive setting you had tried on your pump?

(Or am I just recalling this totally incorrectly?)

Also I was wondering. On the 670G, when you drop out of Auto, are you able to turn the Auto right back on in say 3 hours or does the pump not allow that rapid of a “back and forth” in terms of auto mode?

You can turn it back on immediately, but it will ask for a finger stick.

The 670g, as DrBB mentioned, is not aggressive about correcting highs. Once you go high, it seems to take forever to get back “in range”. It struggles with high fat meals (hamburger/fries). Adjusting the active insulin time does help.

They call it AIT (Active Insulin Time), but it’s the same thing. I had it dialed all the way down to 2.0, same as the OP. It’s the parameter with the most effect on how aggressive the pump can be, and is pretty critical to understanding how the whole system operates. Because it isn’t really doing a “basal” delivery the way most pumpers are used to thinking of it. You’re getting a continual stream of continuously-adjusting “microboluses,” and each one takes into consideration the effect curve of each previous blip in the stream to avoid stacking–or not so much avoid it as calculate how much stacking it’s going to allow. You’re setting how much overlap you’re going to get. That’s why it’s better to think of it as just a dial you can turn, rather than an actual, empirically tested effect curve as the name implies. It ends up being a very different number from the one you use when NOT in auto mode. And 2.0 is the minimum.

The one thing I never quite figured out is whether the I:C ratio worked the same way. Officially it only pertains to bolus calculation, but again, the distinction between basal and bolus ends up being more notional than hard-edged with this system–it’s all bolusing, all the time. And the one thing about this parameter (which is the only other parameter you can adjust) is that you can set different values by time of day. So out of desperation I started fooling around with setting some crazy-low ratios in the wee hours to see if I could chip away at that DP curve. Seemed like it was maybe helping a bit after a week, but then started trending back the other way, and I couldn’t tell why exactly–too many variables compounded by the fact that any change to a parameter takes days to integrate into the algorithm so you can see the effect, and I finally lost patience at that point. Came a morning with fasting BG back in the 180s and I found myself asking Remind me, for what benefit am I working so darn hard at making this work when I could get better results in Manual with a lot less bother?


Thanks DrBB. Yes, I feel like I’m taking crazy more insulin and wondering if I always underestimate my carbs or did I just not pay attention before I was using the 670. And it is super difficult to figure out because the algorithm does take several days of averaging to factor in the change, and then it’s gradual.

I found someone who made youtube videos about the 670G keeping you averaged high when you drop too quickly because it goes into minimum delivery (or just completely stops) when you are trending down, even if you are still over 200. That person suggested correcting in manual mode and then once stabilized going back into automode because otherwise it just yoyos. I just asked my pump trainer about that. She’s been checking in with me every couple of weeks to get me stabilized in the correct settings.

I rapidly learned to input absolutely every scrap of carb that I ate, even if it was only a few, and when I had to guess I would guess very much on the high side.

One thing to do is revert out of auotmode and use the pump in manual mode for a week. i am guessing what you will find is that your settings are aggressive enough, but that perhaps your carb counting is a little off. One thing I have been suggesting is that people use a telephone app for carb counting. If carb counting is the issue, prolonged highs can happen, if the total carbs are not calculated or if the portions are higher or lower than entered amount.

Note: I am a Medtronic ambassador. My opinions are my own. They did not pay me to say nice things about Medtronic devices or the company. OK, they sent me a shirt and a cup but even I am more expensive than that.


You know what? We found out where we are moving next with my husband’s company transfer (he was sort of suddenly and dramatically transferred this month, sent home to await the next post dramatic) and my blood sugars immediately dropped by 100 points all day the next day and have continued to be fairly stable in the low 100s now. Sad we are moving, but thank goodness we have a location and I’m not carrying extra anxiety! The high trend went on for so long that I thought it couldn’t be stress anymore and was just poor settings on my part.

I’m not sure why they excluded correction blouses. There’s no excuse. I have adjusted everything I can adjust. I just fasted for 7 minutes days in automode to see if the pump would adjust. It did not. My carb ratio is now 1:6 and my active insulin is 2.0. My sugars are 130 consistently. (Last 3 months) Not good enough for tight control. I wish my previous pump had not died. I had a lot better control. A1c 5.3 to 5.8. It’s not that now.

I think the general opinion on the 670 is that the people who struggled with tight control, A1c’s in the > 7.0 range, tend to like it the most, and those who had very good control, in the <6.2 range, are the least satisfied. The fixed BG target, 120, is equivalent to an A1c of 5.8, but that doesn’t mean you’re averaging 120; it’s just the point the algorithm aims for. In practice the average is almost necessarily going to be higher because of the strong built-in bias against anything resembling a hypo. The designers, FDA, and the medical profession generally are much happier seeing us run a bit high because the whole thing of letting us dose ourselves all the time with this dangerous hormone goes strongly against the grain of “First do no harm,” even though there’s no avoiding it. So if you want to run lower than they’re comfortable with, you’re going to come up against the roadblocks they’ve hardwired in.

In my case I found that even after 3+ months I was still working much harder at trying to coax and cajole the pump into doing what I wanted than I ever had to do in manual, for not as good results, and I finally had to ask myself what the actual point was. Some of the difficulty was running up against the limitations on adjustments you can make, but I think the larger problem was with the CGM. I think there’s a lot of variability, person-to-person, in how accurate a CGM can be, and for me it just wasn’t adequate to the task the system was demanding of it.