Disappointed in G4--Does anyone else find it less accurate than 7+?

A couple of months ago I finally was able to make the switch to the G4. I loved the 7+, always had amazing accuracy for up to 2 weeks. But with the G4 it seems like I need switch out the sensor every 5 days or so, because the values are way off (woken up twice last night with G4 readings of under 55, when my real BG was 110, then this morning had a reading of 184, tested before bolusing and found I was actually 87. I also get a ton of ???s when I know my bg is steady. This sensor is only 4 days old.

Am I doing something wrong? Is it a bad batch of sensors or is there something wrong with the transmitter? Although I love the new design, at this point I really want to go back to the 7+.

Which meter do you use to calibrate ?
It's important it is precise, if not accurate.
Its error has to be "stable" not once plus and the other minus otherwise the sensor gets broken references.
Read diabetesmine about a CGM - meter calibration nightmare:

Thanks for the recommendations! Garidan, I’m using the Freestyle on my Omnipod, which I also used on the 7+. I do calibrate when I’m stable, I read that the G4 is much more sensitive to this. Maybe it is a transmitter issue, because I’m having problems not just with inaccurate numbers but also with the ???s appearing after only a couple of days (usually when I try to calibrate to a number it doesn’t like, but also out of the blue when I haven’t calibrated for awhile…) I guess I’ll have to call Dexcom. (I was waiting till I started on a second box of sensors to see if the problems would be resolved, but have had the same issue with this lot.) Frustrating!

I used the 7+ system for three years and am now in my second year with the G4. I find the G4 more accurate. I recently had a sensor that was wildly inaccurate; it gave readings of 39 while my fingersticks were normal, like 100. It then swung and gave 300+ readings when my fingersticks indicated 150 or so. I stuck with it for about four days. When I pulled the sensor I noticed that a small amount of blood had leaked from the sensor insertion site and stained the sensor bandage.

I agree with garidan that calibration plays a key role. For the first 24-48 hours I think it's important to feed the system timely calibrations. During that time I like to calibrate only when the sensor reading is stable with a sideways arrow. Ideally I like to calibrate at low, medium, and high ranges, like 70, 100, 140 respectively. During days 1 and 2, I don't wait for the system to request a calibration, if conditions are stable and are at a low, medium, or high value, I will calibrate. After the first few days, I calibrate less often, but will do so if the Dex reading is off.

Due to the inherent inaccuracies of our fingerstick meters, whenever I calibrate, I use two fingerstick readings and then give the system the average. Of course, when the initial calibration is requested, I feed it the two numbers it asks for.

Thanks Terry, that’s helpful. I wonder if the G4 is more finicky when it comes to calibrations…I do try to calibrate while stable, but the readings have been so inaccurate it’s hard to tell whether I’m actually as stable as it’s telling me I am.

I’ve also noticed that in days 3 & 4 (the only days it’s been halfway accurate for me) it seems to really overemphasize drops and rises. I may be dropping from 160 to 110, but the Dexcom will tell me I’ve actually dropped into the 60’s…Same with rises. I never had that issue with the 7+, so I ran into trouble treating severe lows that weren’t actually lows at all. (I know I need to test first, but with the 7+ when the Dex was reading LOW my bg actually was extremely low and I’ve always felt safer eating before testing.)

G4 is more accurate and less variability in the readings.

I think ??? comes out not only when the receiver looses signal, but anytime the system doesn't feel "sure" of what the right number to show you is.
If according to its last calibration it senses you are at 200 and you give a new calibration of 100 I think it stays for a while in ??? to get enough samples to "estimate" a right number and "clear" its history.
Do you correct often with boluses ?