| Author |
 Topic  |
|
|
Kaninbjerget
Denmark
18 Posts |
 Posted - 08/07/2024 : 11:43:39
|
Hi
I bought the GMC-800 as it was on offer on Amazon. Firmware revision 1.09.
First impressions were great. Screen is nice and clear and nice to have all that info at a glance. I really like the countdown to power save. Now I'm familiar with the shortcuts I like them too.
1st bug
However there's especially one bug that really bugs me. "Fast Estimate Dynamic" is really buggy. I can have a static object (one of those infamous thorium filled Chinese wands) sitting on the table against it.
My 300S also set on dynamic mostly fluctuates +/- 0,1 uS/h when settled down - on rare occasions +/- 0,2. That is expected behavior.
The GMC-800 on the other hand. It should read about 4 uS/h depending on placement (beside or below, even give or take a mm in distance to tube result will vary noticeably). It it not really stable. It mostly fluctuates around the expected reading but within any given short timeframe (10 sek or so) there's also huge variations as high as almost 6 uS/h and as low as 2.30 uS/h. LED flashes and it clicks like expected - no odd thing noticed there. So there's certainly a bug in algorithm. A little better if choosing fixed intervals although at 15 sec fluctuations are still a little larger than I'd expect.
2nd bug
Another thing. When choosing to read out dose rate (press left most button until it speaks) results dose rate to momentarily increase. Like it's putting a lot of strain on the processor. It still click while it speaks and although clicks apparently are registered but it is as if that timespan just collapsed into a singularity - if you understand.
3rd bug
Last bug - readout of dose rate. 3.11 uS/h wil read out as "three point eleven...". 3.45 will read out as "three point fortyfive...". However any number .3 - for instance 3.32 will read out as "three dot three two...". Same goes for tens (34 CPM reads out as "three four C P M") and I'm pretty sure it goes for thousands as well.
Hope You'll investigate these bugs. Especially the first makes an otherwise really nice device on the border of useable. The second and third - probably mostly relevant for someone with special needs.
1st suggestion
Pressing and holding the down key (third button) goes to settings. That is really nice. How about pressing and holding the second button (up button) to go to dose rate screen? (Perhaps even configurable)
2nd suggestion
When the fast estimate bugs is fixed it could be nice with an estimated precision on the dose rate like on the 500+. Perhaps even the timeframe it is based on - especially informational when using Dynamic setting.
3rd suggestion
Last suggestion. Timed count. Something I really love on the 300 series. It is a really great way to rate objects against each other as it greatly enhances precision as no need to estimate a constantly varying number (I usually do 10 min timed counts)
Hope You'll investigate the bugs and consider suggestions 
Thanks in advance! |
|
| Reply #1
Kaninbjerget
Denmark
18 Posts |
Posted - 08/18/2024 : 01:39:04
|
I'm a bit surprized that nobody responded. No one else with GMC-800 ?!
I did a 3 minute video on the issue and comparing it to the 300s using one of my strong nano wands.
As can be seen the 300s has no issues except a tendency to overshoot at first but thereafter readings remains stable within expected fluctuations.
On the 800 however readings fluctuate between 3.something and 5.something uS/h. Even as low as 2,8 uS/h - even after a minute and more at which readings should have had time to stabilize. From its clicking I don't observe any oddities - but readings are if not all over the place then quite spread out.
Is a new firmware in the making? It's relatively unusable as it is right now.
https://youtu.be/tkkfH3qrxJk
 |
 |
|
| Reply #2
EmfDev
2250 Posts |
Posted - 08/19/2024 : 10:22:20
|
| Hi Kaninbjerget, sorry it looks like i missed commenting on this post. we are now working on the dynamic estimate to be the same as the 300S. For the timed count and accuracy display, we will check with the team. |
|
|
| Reply #3
Kaninbjerget
Denmark
18 Posts |
Posted - 08/20/2024 : 00:57:09
|
quote:
Originally posted by EmfDev
Hi Kaninbjerget, sorry it looks like i missed commenting on this pose. we are now working on the dynamic estimate to be the same as the 300S. For the timed count and accuracy display, we will check with the team.
Excellent. Thanks!
If the overshooting can be eliminated/minimized it'll be great. Not quite sure the fluctuations on the 800 is limited to Dynamic setting. Think I've seen some larger then expected fluctiations when using fixed intervals as well. I can try to experiment a little when I get the time.
Looking forward to hear when a new firmare is ready  |
|
|
| Reply #4
ullix
Germany
1171 Posts |
Posted - 08/20/2024 : 01:51:01
|
The "overshooting" is likely a consequence of this (in)famous FET-"Feature". It is a Bug in the real world!
Kill it by setting FET=60 (sec)! |
|
|
| Reply #5
Kaninbjerget
Denmark
18 Posts |
Posted - 08/20/2024 : 07:06:01
|
quote:
Originally posted by ullix
The "overshooting" is likely a consequence of this (in)famous FET-"Feature". It is a Bug in the real world!
Kill it by setting FET=60 (sec)!
But then you'd have to wait a full minute. Not exactly useful. If you want to know something "exact" - then that's what timed counts are for 
It's something in the algorithm that causes that tendency as it's rather consistent. If it was pure chance it'd not be that consistent. But I agree it's quite tricky to make such algorithms as what seems straight forward on paper isn't given so when you need to program it. You can do it very thorough and in theory you'd get perfect results - but puts a lot of strain on processor and partly memory. |
|
|
| Reply #6
ullix
Germany
1171 Posts |
Posted - 08/21/2024 : 00:46:24
|
| Yeah, 1 min is a long time. Using FET you get the first viable number already after 60sec! |
|
|
| Reply #7
EmfDev
2250 Posts |
Posted - 08/21/2024 : 10:56:38
|
| @Kaninbjerget please email me or support with your serial number then I can generate an update for your device. It should now have the same calculation as the 300S and others. |
|
|
| Reply #8
Kaninbjerget
Denmark
18 Posts |
Posted - 08/22/2024 : 00:05:21
|
quote:
Originally posted by EmfDev
@Kaninbjerget please email me or support with your serial number then I can generate an update for your device. It should now have the same calculation as the 300S and others.
Thanks a lot! I don't know your email so sent one to support 
I'll let everyone in here know when it's updated  |
|
|
| Reply #9
Kaninbjerget
Denmark
18 Posts |
Posted - 08/22/2024 : 10:50:17
|
Thanks a lot! Just updated to rev. 1.11
Can confirm that fast estimation is now working a lot better - much more stable.
Other bugs seem to remain though.
I've deleted the video as it's no longer valid. |
|
|
| Reply #10
Yanni2007
United Kingdom
1 Posts |
Posted - 09/03/2024 : 08:54:23
|
Where can this Rev 1.11 be found?. I tried to shoot a support email with my serial number, but I get no response.
If I cannot update then I guess I will return it. |
|
|
| Reply #11
Stargazer 40
USA
431 Posts |
Posted - 09/03/2024 : 10:11:16
|
| You might give them a bit to reply to your email. They can be very busy. |
Stargazer 40 |
|
|
| Reply #12
Stargazer 40
USA
431 Posts |
Posted - 09/03/2024 : 10:40:28
|
| Also, did you provide the number on back of meter or from the 'About' area under the menu. If the former you might also send them the lengthy one under the 'About' area of the menu. I sent both and haven't heard back in three or four days so far. They will get back to you (and me). |
Stargazer 40 |
|
|
| Reply #13
Michel2017
Germany
3 Posts |
Posted - 11/05/2024 : 22:18:44
|
quote:
2nd bug
Another thing. When choosing to read out dose rate (press left most button until it speaks) results dose rate to momentarily increase. Like it's putting a lot of strain on the processor.
The same thing happens with the History Menu. When entering the menu, it starts with the oldest record. To go to the most recent record one has to "double-click" the arrow-down button and hold the button. Then the history moves very 4 minutes until reaching the latest record. This seems putting quite some load on the processor that the counts shown afterwards increased significantly. Initially I thought the device is broken...but after reading this post, it seems a processor load issue. |
|
|
| |
Topic |
|
GMC-800 rev. 1.09 bugs - and some suggestions
