Do I even WANT a new pump?

Coming up on officially 4 years since my last new pump, a 670G that sucked so bad it went back into its box in my D drawer and hasn’t been out since. I reverted to my good old reliable pager-style Paradigm 723 and G6, and my A1Cs have been in the 5.8-6.3 range. But my 670G warranty expires in March, so of course I’m getting pinged by Medtronic and Tandem both about signing up for a shiny new replacement. Part of me just instinctively says “Oh goody, new D tech, let’s go!” Kinda like buying a new car.

EXCEPT, when I think about it I’m just kind of un-thrilled about either choice. Partly I think it’s because my experience with the 670 soured me on the whole AID (Automated Insulin Delivery) experience. Here’s what’s been running through my mind at the prospect:


  • +Free upgrade to 780 (when it’s released) lets you set a lower algorithm target now (100, vs 120 for the 670)
  • +I’m used to how Mdt does things, including reservoir & Mio insets.
  • +Whole new algorithm, supposedly better than 670.
  • -Guardian Sensor (new CGM might be better, but it’s still in development)
  • -No longer talks to ContourNext!!! Have to switch to less accurate BG meter. Why did they do this??? This may be the deal-breaker, oddly enough. You have to calibrate Guardians 2x/day.


  • Control-IQ. People seem to like it, but more recently I’ve been seeing gripes from people who were already getting low-6 AICs and aren’t happy with results on this system. Fixed target isn’t flexible enough. Can anyone comment/confirm?
  • Rube-goldbergian reservoir/inset system. Changing sets seems like a PIA, unless it’s compensated by other advantages.

AID in general:

  • We locked up your pump! Just trust the algorithm! Doesn’t work so good? Oh well. That was my feeling about the 670, anyway. Does anyone feel like that’s no longer operative? Especially interested to hear from anyone who was getting ~6 A1Cs before going on Control IQ and is still satisfied!

Your experience with automated insulin delivery is so different than mine. As you know, I use the DIY Loop and the wide-open access to all the important control levers for someone who understands well what those settings do, makes, I think, all the difference.

For me, these last four years have been my best glucose management in 37 years, the lowest A1c’s, the highest time in range, and the lowest glucose variability. It’s done all these phenomenal things at the same time requiring less of my attention; it’s made a huge contribution to my quality of life.

I think the commercial designs target the large demographic who value simplicity and just want the pump to take care of business and let them get on with their life. Unfortunately, for people like us who have investigated how these things work, our knowledge when confronted by locked-down simplicity frustrates our ability to produce an optimal glucose outcome for us. We don’t like vanilla.

If I were forced to consider a choice between the Medtronic and Tandem offerings at this time, I would likely choose Tandem. I base this preference primarily on Tandem using the Dexcom CGM and Medtronic sticking with their Guardian sensor. I’ve never used a Medtronic sensor but I am influenced by all the negative comments I’ve read here over many years about their troubles with this CGM.

I read an account on another forum about someone who was in the Beta Bionics insulin-only iLet trial. It is the ultimate in simplicity. You set up this system just by telling it how much you weigh and when you eat, you announce the meals as small, medium or large. Other than that, the system does everything else.

When I read the numbers that this person was experiencing, I realized that I would have a hard time living with it. The person in this trial had an average BG in the high 140s and a standard deviation in the 40s. These metrics would not inspire me to trade my control for simplicity.

I do think, however, that both of the choices you’re considering are much better than the 670 you valiantly tried to make work. But they have not yet reached the performance people in the DIY community enjoy.

You may find this video featuring Lane Desborough, current CEO of Nudge BG, an interesting one as he discusses his long experience at Medtronic and Big Foot. He examines the choices made by glucose management systems designers. I found it thought provoking. Fast forward to the 8:00 mark for the formal start.

1 Like

There are ways to get a lower range on tandem. You can adjust basal rates and sensitivity. There is a discussion here about that very thing.
I’m very happy with my tandem. I could never rely on Medtronic sensors well enough to use their loop.

I still make corrections on my pump all the time, but I like knowing I really don’t have to.

If it was me at the end of my warranty, I would wait for the tandem sport. It was supposed to come out by now.
It’s half the size. And controlled by your phone.

I have still my trusty 523 and I really liked it, however I’m much happier with the control iq once I got my settings set up.
It took me a good 2 months tho to get it right.

You will likely hate the weirdness of filling cartridges,because it’s so easy on Medtronic.

I’m pretty sure you can get a trial of the tandem if you wanted.
It doesn’t seem to me the Medtronic has changed too much since your currently not being used model except the algorithm

Well,I guess my thought is, do you need a new pump? I mean in the 20+ years I was with Minimed/Medtronic, I had 3 pumps. They worked, there wasn’t a lot that changed so I just used them for a very long time.

My other thought is you need to think about what drives you nuts about your current system and what do you love. If you love the ease of infusion site changes and that is a deal breaker, than Tandem might not be a good choice. It takes about 7 minutes to do it for me after using Medtronic that took a few minutes. It wasn’t a concern for me and now after years, it is my new normal. Do you like easy or are you ok with some work? I want and need easy so I went Tandem. No blood testing! And alarms are so much less! I really pull it out when I eat and every 3 days for site change.

It was a learning curve for me at the beginning, learning a whole new system. But it is now my new normal. I had my basal, carb and correction factors nailed before I switched. My A1C stayed pretty much the same 6.2-6.5 range and it hasn’t changed since the switch. My time in range has increased a lot! And my standard deviation is lower. And all this came with a lot less work for me. After 50 years, I want easy!

I start the iLet study tomorrow. Insulin only but I will be doing the study this time instead of just the week to help train people doing the study. There are new pumps on the horizon so if things are going great, maybe you can wait.

No perfect answers here. You just need to think about what you love and hate about your current system. Maybe your current pump is the way to go.


Sally’s suggestion may be your best bet.

I enjoy my Tandem pump most of the time. I wish the algorithm would allow you to lower the glucose goal. Like all medical device manufacturers, Tandem is over-cautious. The big scary hypo keeps the setting locked in I guess. I have found that I can set my basal a little higher than needed and keep it on sleep mode all day to get my glucose below the preset 112 mg/dL. Once I level out, I’m usually in the upper 90’s doing it that way.

I’ve seen Tandem users refer to this before, so let me check if I’m understanding correctly. With the 670G you have “manual mode,” which I’m guessing is the same as what Tandem calls “sleep mode.” When you set up the pump, you program it with a standard basal profile that it uses when it’s not running in automode. So if auto fails for some reason (actually a lot of reasons), or you switch it off deliberately (because it’s driving you nuts), the pump doesn’t just shut down, which could ruin your day, but defaults to running as a standard pump. With the 670 the two modes are completely walled off from each other—your profile settings have zero impact on how it behaves in loop land. I assume that’s the same for the Tandem? Something I read wasn’t actually clear about that so it’s one of the questions I have.

No. When C-IQ is turned off, it is manual mode.
To use Sleep or Exercise modes, and basal suspend/auto bolus, you have to turn on C-IQ.

Sleep mode can be used or not, set on manually, or have day/time periods to have them on automatically.


This was posted by Chris Wilson, the administrator of the Tandem group on Facebook. I was particularly pleased to read that Tandem has already filed for an update to Control IQ that will allow more user customization and I assume a lower target range.

“ Notes from Tandem’s Feb 24 investor earnings call. Quotes are from the transcript, and are all by John Sheridan, Tandem’s CEO

I’m going to assume that most of you don’t care about the business numbers, but those were good. Here’s what y’all probably do care about:

CIQ v2 (for lack of a better name): “Last year we submitted a regulatory filing to the FDA for the first enhancements to Control-IQ, to allow for greater personalization, and certain rates and ranges, and to add an indication for Admelog.”

Mobile Bolus: “we are preparing to launch our app based mobile bolus feature for the t:slim X2 in the first half of 2021.” and “We will be offering this feature on iOS and Android”

It sounds like Mobile Bolus will require the update to CIQv2: “The pump software update, with our new mobile bolus feature will be the fifth update that we’ve offered to our customers, since first launching the Tandem Device Updater in 2016.”

t:sport delayed slightly, still expected before the end of 2021. This is a result of some things that Tandem found during human factors testing:

“In the late December, we completed a collective analysis of these studies and it was our opinion that some of the opportunities identified during the testing did not meet the high user experience standards that we’ve committed to providing to the diabetes community and have built our reputation on”

“Internally, one of our core values is no short cuts, and this is a fitting opportunity to work by those standards and refine the design until it’s right to ensure its safe and intuitive to use. None of the changes we plan to make are particularly challenging. It will just take extra time to update and test the software, and confirm our new designs to additional human factors testing. As a result our final goal has shifted from the third quarter to the fourth quarter, and we’ll provide further updates as we have more clarity on the timeline.”

Dexcom G7 integration: “Based on the FDA’s interoperability initiative and the timing of Dexcom’s clearance, our goal is to commercially launch our Control-IQ technology with the G7 sensor within one quarter following their receipt of FDA clearance.”

Libre Integration (probably future product, not a currently-available Libre): “Our development efforts with Abbott Libre technology are also off to a strong start. It’s been a collaborative effort and we share a common goal to bring the integrated product with Abbott Libre’s technology to diabetes community as soon as possible in 2022. We intend to begin our commercial efforts with a U.S. launch and which Libre technology we initially launch with will ultimately depend on the timing of Abbott’s FDA clearance, supporting AID use.”


The best thing about Control IQ is overnight blood sugars. I can’t tell you the last time I woke up with a low or a raging high bs. Most days I wake up between 90 and 110.

If that was the only benefit, I would 100% make the decision to upgrade again.

As far as ControlIQ during the day, I’m kind of indifferent about it since I watch things pretty close. But it is nice if I’m in a meeting and having CIQ in the background doing its thing to keep me in target.

I had an A1C in the mid 5’s prior to ControlIQ and I’m still in that range a year later. I suspect a lot to do with better overnight numbers.

The cartridge change is different, I wouldn’t necessarily call it a PIA. Is it perfect, no. Most people who constantly complain about it want it to be like the Medtronic cartridge. It’s not and highly unlikely to ever change. If you would struggle with that I would say stay with Medtronic. For me not a big deal after a couple weeks.

I think a lot of people who used the beta version of Medtronic automated pumps are scarred by their experience. So I understand your hesitatance to change. ControlIQ just simply works.

1 Like

@DrBB you could always elect to forgo the algorithm on the tandem pump and just use it as a traditional pump. You have the option of toggling the algorithm on or off as it suits your needs


A few years ago I was probably one of the most pump averse people you could meet. I hadn’t ever been on a pump, didn’t want a pump, didn’t trust the technology. But, when Control IQ came out and I started hearing so much good about it from people on this forum and I was really struggling to maintain stable BG’s (I was still getting A1C’s in the 6’s but losing a lot of sleep and was just feeling drained) I decided to give it a go with the thought that if I didn’t like it I could go back to MDI.

Turns out I really do love Control IQ. I don’t even fiddle with my basal rates to get lower BG’s and I am still consistently getting mid 6 A1C’s. My Dexcom says my current GMI is 6.5 with a TIR of 86% with only 10% high and 1% very high, 2% low and 1% very low. My range is 70-180 mg/dl. My SD used to be mid 40’s generally and now it is usually high 30’s which isn’t as tight as some here can achieve but that’s not Control IQ’s fault, it’s simply that I don’t absorb insulin or digest food in the most reliable manner.

The only thing I find an annoyance with the filling process is that the fill tubing step takes quite a while, however I can still complete the whole site change in about 5 minutes. Oh, and my one major gripe with the Tandem X2 is the stupid alarm that starts going off if you stop insulin delivery for more than 15 minutes and doesn’t shut up until you turn insulin delivery back on (I want to take a long peaceful bath gosh darn it!).


I think this choice is going to keep getting harder in the coming years, with technology evolving so fast and new things coming to market so quickly. I think we’re going to see a little bit more buyer’s remorse when something brand new and exciting comes out just months after someone committed to 4-5 years with something else. Remote updates definitely help, but certain things require hardware and form factor updates, like for fitting dual hormone systems, Neither the t:slim, 770/780, nor Omnipod will ever be able to update such a pump if that proves to be the most effective system. I’m facing the same decision, being 3 months out from the end of my warranty.

Since Tandem and Medtronic reached an agreement to share patents last year, Medtronic is actually acquiring some of Tandem’s big selling points, such as mobile interface and remote updates… Making Medtronic more of a contender again after the 670G fiasco. (I STILL can’t figure out what Tandem got out of that deal, but I suspect we’ll see it in their new T:sport pump.) I’ve been trying to see how the the 780 is being received in Europe and looking at the criminal trials, and I’ve not seen anything near impressive enough to pry my Control-IQ out of my hands. I feel like the Medtronic system is trying to figure you out and force you into someone else’s pre-conceived notion of management. I think that’s comforting to doctors who want to ease their burden and treat everyone with the same formula… but it doesn’t jive at all with my Type A personality. Control-IQ isn’t trying to figure you out, but rather it does exactly as you tell it to. It makes it a little more difficult to set up, especially for those who haven’t learned to adjust their own settings. But you are 99.9% responsible for your happiness with the system. If someone is displeased, it is nearly always due to a bad settings problem. And there’s a glorious Hallelujah moment when it all falls into line. I’m open to improved options from other manufacturers, but I don’t see the 770/780 as a step forward.

The thing that seems to set Tandem’s AID apart from the other systems (including Dana’s CAMAPS FX that isn’t even in the States, but not including pipeline products that haven’t published data yet, like iLet and Tidepool’s commercial Loop)) is that Control-IQ is the only one offering a “sleep mode”, or any other option with tighter parameters. Thus, Tandem is the only one who divided TIR specs between daytime and nighttime. All three averaged mid-60 percent TIR either all day or daytime-specific hours (technically Tandem had the highest value, but it’s not an entirely fair comparison because they also started with the lowest average initial Ha1c), but Control-IQ annihilates the competition upping the nighttime average TIR to 80%, and thus more overall TIR. And of course so many of us just keep running those tighter parameters through the day, too. Note, those averages include the best and worst values, and it’s not indicative at all of the best you can personally achieve with them.

You’ve expressed a few confused comments about the T:slim.

To start with, you have to set one or more personal profiles. (You can have a profile specific for weekday, weekend, exercise, sick day, steroid shot, or whatever else you want that you can name with a T9 keypad. I just use 4: 50%, 100% and 150% of my “normal” rate, and one set to zero so I can suspend the pump without it yelling at me.). Each profile is divided into timed segments and for each and every timed segment you set a specific basal rate, target blood sugar (can’t be set lower than 110 if Control-IQ is turned on, can go down to 100 if not), correction factor, and insulin: ratio. You choose one of your profiles to be active.

There is a toggle switch for Control-IQ. You literally turn it on or off with the flick of a virtual switch. If it is off, you have zero pump automation, including no hypo suspend. It’s a "dumb"pump that is only as smart as your settings. You will still see your cgm on the screen, but aside from that, it delivers insulin exactly as you programmed in you active personal profile.

If you turn Control-IQ on, basal rates will sometimes be adjusted. It will use your correction factor primarily to determine how to do so. You can even chose to turn Control-IQ off during the day when you want to micro-manage yourself and turn it on at night to take advantage of the peaceful sleep.

While Control-IQ is turned on, you’re given a few extended parameter options called “activities”, one for sleep and one for exercise. If no activity is selected, the pump does nothing so long as you’re between 112-160. The exercise activity raises the actionable zone to try and keep you between 140-160, while the sleep activity lowers it to 112-120. Control-IQ will always reduce your basal if you’re below 112, but you can still maintain lower with adjustments to basal and correction factor settings. Control-IQ also gives 60% of a calculated correction bolus is you BG exceeds 180 in all BUT sleep mode.

If you watch the Facebook groups, people are always celebrating their Ha1c results because they’re getting low ones for the first time ever. You’ll see lots of 5’s and low 6s. I’ve gotten 5.8 on my last two tests, but Dex estimates me at 5.4ish. Before I started on Control-IQ, Dex was estimating mid-6, but lab values came back 7 and higher.

Yes, there are a few limits to the parameters, but they can be compensated for with other settings. Also, they’ve promised to open up those parameters and give more customization with the pending software update. I guarantee you they come down to a target BG of at least 100, if not better. There’s just no way they’d cede this point to Medtronic when it’s such a bone of contention.

I would say there’s awesome compensation. That rube-goldberg contraption is part of their micro-bolusing feature, which makes the pump more accurate and safer than others, because the reserve of insulin is completely isolated from you by an intermediary chamber. It’s also critical to the T:slims small size, because it eliminates the need for that giant syringe-style plunger.

And the reality is that it’s the lamest Rube-goldberg device ever. There’s no moving parts whatsoever in the cartridge. It’s a clear plastic bag in a black plastic box. You jab insulin in through a hole at the top. There is absolutely nothing DIFFICULT about it. It’s just DIFFERENT. Do it a few times and it’s second nature. If you’ve ever done MDI with vial and syringes instead of pens, it should feel perfectly familiar. You’re just stabbing a piece of plastic instead of yourself.

My problem is that I’m really interested in both Tandem’s upcoming T:sport pump (half the size, no screen, completely controlled by my phone, and with the latest update of Control-IQ? Sign me up!) and the iLet. I was really hoping the T:sport would be out this summer when my warranty expires, but looks like it’s been pushed back again. Though sounds like it was for good reason. The trial participants aren’t allowed to disclose much, but the few I’ve spoken to said they were happy to go back to their T:slim, mostly just because it wasn’t as comfortable/refined of an experience. I’m happy to hear they’re working on that. And until this week, I had written off iLet completely for this round of decisions. I really wanted to see how more people fared with the dual hormones and the new algorithms that used them. But now I’m hearing so many trial participants speak so highly and call it their favorite, that it’s become much more desirable.

But how long can I wait without a warranty for these things to work through the pipeline?? I’ve never had any catastrophic pump failures, but I have had a few warranty replacements for minor things. Being so rural and isolated, 24/7 customer service and over-night shipping is big peace of mind. I also don’t know if an out-of-warranty pump affects my Dexcom warranty. Sensor issues are the most common reason I call Tandem. Twice now they’ve sent me a new pump when troubleshooting transmitter issues… because somehow, replacing the $7,000 pump comes before replacing the $125 transmitter in their flow chart. And does the 10 day sensor warranty become null and void if my primary data collector is invalidated?

I might vary well chose the T:slim again for my next pump, but I want to make that choice because I still feel it’s the best… Not because I ran out of time!


Super helpful, Robyn, that filled in a lot of my blanks, especially about the Sleep Mode thing. I had a vague impression it wasn’t the same but most remarks about it made no reference to the MdT way of doing things so it was really hard to tell.

That’s the biggest concern. I had so looked forward to trying out AID with the 670, went into it expecting great things, and then it was like having a robot endo continually saying “NO! You can’t do that! SHUT UP! Hands OFF!” The idea that Tandem is looking at making it more user-controllable down the road is also encouraging.

1 Like

When you set a different profile called 150, do you only change basal rates? Or do you also adjust the correction factor?

1 Like

No, but that’s because I’m already running aggressive (lower number value) in my 100%. I don’t actually use that profile often enough to really fine tune it. I could probably get away with a teensy bit lower, though.

Basically my tested and “true” correction factor is 1:46. My 100% profile behaves happiest with my correctionn factor set to 1:42, though. If I go much lower, my standard deviation jumps up because I wind up in an over-correction roller coaster. Generally the lower correction factor makes Control-IQ BETTER at preventing lows, but at 1:40, I actually see more lows before it rebounds back up. So even on my aggressive 150% profile, I’m hesitant to drop it any more.

I’ve only really used it once when I got a cortisone shot. (turned out WAY better than the pre-IQ shots, but not perfect but any means), maybe a few times for sick day type stubbornness. Usually I use it for when I expect indulgent eating, or some kind of grazing situation where I don’t know what/when I’ll be eating or what went into it because I didn’t prepare it myself and I know my bolus guesses will be garbage. It’s my holiday/party profile.

I’ve got a lot more experience with the 50% profile. That’s my “exercise mode ain’t gonna cut it while I’m doing manual labor on the farm” profile. That one I have turned the correction factor up just a titch to 1:44. It’s still SLIGHTLY more aggressive, to help ward off the lows, but keeps me more stable. I don’t need quite as aggressive of setting since I use it combined with exercise mode usually, which is usually enough to keep me in range.

I actually call my profiles LowIQ, IQ, and HighIQ.
I get to be brilliant when I’m snacking my way through an evening with company and saving room for dessert on the holidays, and a “dumb” farmer when I’m working away in the field. It makes me laugh. :slight_smile:


Tandem really doesn’t do that. There are only 2 settings it’s stubborn about, and only when Control-IQ is active. The 110+ target BG, which is changing SOON, and the DIA is currently locked at 5 hours, which I think was probably the right choice for Tandem. They changed the insulin decay model so much that we’d all be in a world of hurt if we tried to carry over our old settings of 2.5-4ish hours… And people would. I think that’s too much of a change for the typical med team/doctor/patient to comprehend. Maybe once the 5 hours gets accepted as the new normal, they’ll open that one up for adjustments, too.

And no matter what the settings are, you can always over-ride the system and deliver whatever insulin you want (assuming it’s less than the max bolus setting you chose yourself), without it throwing a temper tantrum and kicking you out.

It might tell you something like, “due to IOB, no correction has been added” and you accept and clear the message. All it means is that it didn’t autofill a dosage into the bolus wizard because you have more IOB already than needed for a correction. But absolutely nothing stops you from manually selecting and delivering a dose anyway.


Kinda funny that the AIT (=DIA) setting on the 670 is one of the few things you can adjust, and actually the most powerful one in terms of getting it to be more or less aggressive over all (not sure if that’s true with the new algorithm). I regarded it as basically allowing the microbuses to “stack,” so if you set it lower than the actual time insulin takes to dissipate in your metabolism you’d get faster corrections, or at least not quite so glacial ones. I eventually took mine all the way to the 2hr lower limit, but it wasn’t enough.

Had to look that up… very interesting use of reference.


I learned about them early. One my favorite games as a kid was MouseTrap!, where you built the absurd mouse-catching machine.