EEVBlog 121GW Discussion thread

[Brumby]

That's the Dave I know and love watching. I for one enjoy watching the meter development process and it getting tested by you along the way. No need for a media blackout and a code of silence being enforced. We members on these forums are mostly technical people and supporters of yours and appreciate being kept in the loop.

Bring us along for the ride as it develops, so we can all learn from the experience.

I'm sure Dave will be on top of that ... just remember, he does have other things to do.

[EEVblog]

That's the Dave I know and love watching. I for one enjoy watching the meter development process and it getting tested by you along the way. No need for a media blackout and a code of silence being enforced. We members on these forums are mostly technical people and supporters of yours and appreciate being kept in the loop.
Bring us along for the ride as it develops, so we can all learn from the experience.

It's not that easy, I have little visibility in what's going on at UEi, and I am not in a position to demand hourly updates from them.
And whilst it may sound great to have every little thing blogged, that does invariably lead to people trying to micromanage what's happening. That's bad enough when it's my own project and I'm in control of everything and can work on things instantly myself, but when you are dealing with a design team in another country on the other side of the planet, invariably with of priorities, it's even worse.

IMO your selling this meter's attributes short by not showing the features off as they are fixed. I say more videos highlighting what goes on technically to rectify the bugs. Don't discount the tremendous interest in this product. Make these early shortcomings its strengths by highlighting the fixes. I would have thought this is a video bloggers dream opportunity!

Those who seem to want to know nothing about the meters progress and are more than happy to wait weeks without an update can just view the thread once a month or so. That way all forum members/backers and supporters of the kickstarter project are all catered for. 

Food for thought, Dave?

This is not my first trip down the "publish as it happens" route, why do you think I essentially lost interest in the uSupply? Hint,  hundreds of people telling me to do this and do that doesn't help...
So no, it's not necessarily the "dream opportunity" you might think it is.

[exe]

Kids these days tend to love frequent updates, no matter how shitty the update, as long the manufacturer keep pumping out updates every week, they will be happy, just don't know why. 

Talking bullshit with confidence and calling other names doesn't add you credibility. Modern software development is not based on waterfall model anymore.

A good process development process starts with automated tests. But even if you do (semi-)manual testing I don't think it takes much time to test it. Besides that, let the user decide if s/she wants a hotfix that wasn't fully tested.

BTW, I guess you completely missed Dave saying that FW is a mess. I believe this is the main reason it's taking so much time to fix trivial problem, nothing else.

[FrankBuss]

They have tweaked the dimensions of the white plastic spring assembly and switch assembly I believe.

Does it get a new version number, or some minor revision I can see when I want to buy one for the mechanic changes? The bluetooth feature and integrated uCurrent etc. are really something I could use, but non-wobbly dial would be nice.

The autorange of the BM235 looks incredible. My BM257s is only as fast as the 121GW now. But no problem, if I need fast autorange, I can use my Fluke 8842A, which is at least as fast as the BM235. Sometimes fast autorange is important for me, e.g. when I sort a bunch of resistors after some breadboarding.

[BravoV]

Kids these days tend to love frequent updates, no matter how shitty the update, as long the manufacturer keep pumping out updates every week, they will be happy, just don't know why. 

Talking bullshit with confidence and calling other names doesn't add you credibility. Modern software development is not based on waterfall model anymore.

A good process development process starts with automated tests. But even if you do (semi-)manual testing I don't think it takes much time to test it. Besides that, let the user decide if s/she wants a hotfix that wasn't fully tested.

BTW, I guess you completely missed Dave saying that FW is a mess. I believe this is the main reason it's taking so much time to fix trivial problem, nothing else.

Thanks, point taken ... kiddo.

[Brumby]

But even if you do (semi-)manual testing I don't think it takes much time to test it.

Typical armchair expert.  "I don't think" indicates you have no practical knowledge.

Besides that, let the user decide if s/she wants a hotfix that wasn't fully tested.

This really shows how out of touch you are with practicality, let alone commercial reality.

[IanJ]

All I can say is kudos to Dave for the meter in the first place. It takes balls to go commercial with a product worldwide and there is a lot, lot more work across multiple areas that most folks don't see. Net result it consumes your life for a while.

I agree whilst it's tempting to rush out firmware updates fast it invariably ends up needing a follow-up, then another..........IMHO Dave/UEi's approach is deffo the way to go.......Fairly fast but more importantly gotta be thorough and well tested.

I think Dave and UEi have done a great job so far, the issues being fairly light and easily fixed. No show stoppers.
Can't wait to receive mine.

Ian.

[dcac]

Thanks for the video Dave, yes the 1.05 FW is noticeably faster than 1.04 and at least as fast as the U1282A. And as expected the Brymen 869 is quite faster but then it also only goes down to 10 milliohms resolution where the 121GW goes down to 1 milliohms.

It would be great if you could do two additional auto-ranging comparisons between 121GW, U1282A and U1272A (and perhaps the Gossen) it would be:

1. AC Volt position, all meters connected in parallel, measuring mains 240volt.       

2. DC Amp position, all meters in series, measuring roughly 4.5A DC current load.

This could be good tests to see if any other slowness quirks is lurking in the 121GW FW.

[GameProgrammer79]

totally agree, good to taste volts and amps as well. 

Thanks for the video Dave, yes the 1.05 FW is noticeably faster than 1.04 and at least as fast as the U1282A. And as expected the Brymen 869 is quite faster but then it also only goes down to 10 milliohms resolution where the 121GW goes down to 1 milliohms.

It would be great if you could do two additional auto-ranging comparisons between 121GW, U1282A and U1272A (and perhaps the Gossen) it would be:

1. AC Volt position, all meters connected in parallel, measuring mains 240volt.       

2. DC Amp position, all meters in series, measuring roughly 4.5A DC current load.

This could be good tests to see if any other slowness quirks is lurking in the 121GW FW.

[beanflying]

Please no more SOOKING about slow auto ranging it is until a release of official new firmware fairly pointless and the issue more than well known. Design of firmware by committee with no code in front of you and zero direct contact to the manufacturer even more pointless. Please look up the term of WOFTAM, which is exactly what it would be called in business this behavior would additionally earn you an ass kicking. Let Dave/UEI sort it and then let us know.

More importantly is my meter is reading 0.1 mV low in 29.4 degrees and took less than 5 seconds to stabilise unlike the 50+ year old 740B that took 30 minutes+ . Reference against the Agilent measures 5.00021 at 25C

[dcac]

I’m very convinced it’s up to Dave if he wants to make any further auto-raging comparisons. By his own admission - he just didn’t notice how slow the production model was, possibly his focus was on other problems like ie. the bluetooth interface. But if he now feels confident the 121GW really is on par with comparable DMM’s,  well then - that’s great!

[GameProgrammer79]

I’m very convinced it’s up to Dave if he wants to make any further auto-raging comparisons. By his own admission - he just didn’t notice how slow the production model was, possibly his focus was on other problems like ie. the bluetooth interface. But if he now feels confident the 121GW really is on par with comparable DMM’s,  well then - that’s great!

Really appreciate the amount of attention it has received and number of issues reported so far. It will definitely get better with each iteration of the firmware. We should not get fussy about bugs, it is every where including large firms like Intel (spectre and meltdown) are not immune to it. There are some 3 billion dollar exploding projects as well (http://www.techradar.com/news/this-is-how-much-samsung-says-itll-lose-on-the-galaxy-note-7-recall)

121GW is future proof, simply because one can change the firmware with ease, firmware bugs are technically fixable. I am not sure if it runs a kernel or simple RTOS, probably community with catch up with alternative firmware before long.

[Towger]

Dave,

The natives on Kickstarter are starting to get restless.  It would not go a miss to post an update, with links to the video etc.  It should be the primary source of updates, but as they are spread over the forum, Amp Hour, Twitter and YouTube are being missed.

[Diosol]

I also made a quickie of 121GW. Dave's second video got me interested inspecting what happens while measuring resistance with multiple meters.
Fluke 322 clamp meter works a little bit differently and other meters went haywire. Switching lead polarity solved that issue.
Some day I'll do inspection of the voltages while measuring resistance and take a look does the Fluke 322 have reverse polarity.

Insulation tester had the highest current of 1,00 mA and maybe that's why it showed 1,4 MOhms in baseline setup.

[joeqsmith]

I saw your test fixture and it looks pretty good.   I assume printed.  If you wanted to replicate the tests I have been running,  the switch is fully rotated both directions to complete one cycle.  I cycle them 50,000 times in 60 hours while measuring the contact resistance.  If the resistance gets about 1K for 1000 cycles in a row, I abort the test.  Other than that, I just let the run.   Pictures before and after and document if the use any sort of lubrication on the contacts.  That's about it. 

If you don't plan to record the contact resistance, you may want to consider inspecting the contacts every 10,000 cycles like Brymen.

Looking forward to seeing what you decide to do and the end results. 

[EEVblog]

[joeqsmith]

First, thanks for taking the time to run this test.  Second, thanks for posting how you are running it and the results.  I have been curious if the pads had changed at all.   

Five minutes in, I switched to 1080.   Your 50K cycles would be a half for Brymen and what I am doing but still, what is up with this meter?   There is no way I would have expected to see these results after sending you those original pictures.    Impressive indeed. 

I could try and repeat it assuming the contacts are the same but it's not really going to prove anything.   Thanks again and looking forward to seeing the production parts ran.   

[EEVblog]

Five minutes in, I switched to 1080.   Your 50K cycles would be a half for Brymen and what I am doing but still, what is up with this meter?   There is no way I would have expected to see these results after sending you those original pictures.    Impressive indeed. 

I expected worse too, but that's what we got.
I don't know what serial number I sent you, but it would have been around a similar date, so likely exactly the same switch and contacts.

[Scottjd]

Five minutes in, I switched to 1080.   Your 50K cycles would be a half for Brymen and what I am doing but still, what is up with this meter?   There is no way I would have expected to see these results after sending you those original pictures.    Impressive indeed. 

I expected worse too, but that's what we got.
I don't know what serial number I sent you, but it would have been around a similar date, so likely exactly the same switch and contacts.

If these were the production parts, I would like to try and understand why such a dramatic difference.  I wouldn't have guessed it had more than a thousand cycles on it but for all I know, UEI may have cycle tested it.   Looking back over the pictures, it appears the contacts are what was wearing.  Even looking at the last few that they sent, it looks like something was still going on.  Strange.   

Looking at the video and the large gap in the fingers that rotate the knob, are you sure you are getting the full stroke of the switch?  It looks like it may only rotate 90 degrees.  Still even it that was the case, I would have still expected that area to have a fair amount of wear.   

Just scratching my head.

I’ll add to the head scratching. Or let’s just say thinking out loud.
I agree with Joes counting as this is only a half cycles since a user will turn on, then off and that’s a full cycle. Just like they charging cycles on a battery, charge and discharge for a full cycle.

My questions thinking out loud would be around the way the PLA adapter is mounted to the stepper motor. Seems to me and knowing PLA (it looks like PLA) the part will wear some and if a set screw was used the adapter on the motor will start to slip. I’m not a big fan of PLA for mechanical use from personal experience when I first started doing 3D printing.

Maybe something a little stronger on the adapter with something that also has a little flex like a nylon mix filament. The nylon will allow a little flex without breaking the part, and you can over turn the switch just a little past the off positions. My thinking is most people always turn past the off point, they only let go of the switch because the switch stopped them. I’m not talking crazy over shoot on the switch with massive torque, just a little to replicate real human usage. If you set the controller to over shoot with a PLA part even if it’s made with 100% fill, it will start to stress crack, break, or give away on the tolerances of the whole for the set screw and slip.

But I guess that’s why I’m wondering how the part is held into the stepper motor?
It’s a nice rig, but just might benefit from a few little tweaks?
Scott

[1anX]

I agree with Joes counting as this is only a half cycles since a user will turn on, then off and that’s a full cycle. Just like they cling cycles on a battery, charge and discharge for a full cycle.

Hmmm, I disagree with your observation in regards to the operation of this meter!

There is an OFF position at each end of the selector, so in effect a user could easily select a range and then switch OFF by continuing in the same direction. Hence one cycle of the switch is achieved by rotating in one direction.

Regardless of how you view the definition of one cycle, at the very least we now have a measured number of cycles to interpret contact wear!

Its interesting, but of course it needs to be repeated with a production unit which I'm sure will happen just as soon as UEi provide one.

[EEVblog]

My questions thinking out loud would be around the way the PLA adapter is mounted to the stepper motor. Seems to me and knowing PLA (it looks like PLA) the part will wear some and if a set screw was used the adapter on the motor will start to slip.

It's not slipping.
That meter did 50,000 genuine rotations.

[Scottjd]

I agree with Joes counting as this is only a half cycles since a user will turn on, then off and that’s a full cycle. Just like they cling cycles on a battery, charge and discharge for a full cycle.

Hmmm, I disagree with your observation in regards to the operation of this meter!
There is an OFF position at each end of the selector, so in effect a user could easily select a range and then switch OFF by continuing in the same direction. Hence one cycle of the switch is achieved by rotating in one direction.

I see your point. But at the same time I think most users will always go for the same “off” position. Maybe based on if they use the meter mostly for current readings, or voltage readings. With mode selected aside I’m still thinking most people will always go to the left off position if they own more then one meter just based on muscle memory and the fact the most other meters they own also go off to the left.

For me personally I will always turn it off to the left because I’m used to turning my meters off to this direction. I do own one meter that has a dual off just like the 121GW (Digitek DT-4000ZC) and it always goes off to the left. Even if I use a mode on the right side of the range switch it still gets turned off to the left. If I had bought a meter that only turned off to the right and didn’t have a left off option, this would probably annoy me to the point of returning it.

But as you said, either way 25K, or 50K is still a lot of switching, and at least it’s getting tested now before being put back into manufacturing and sold to more people.
So what is that saying, “Better late then never”.

[Brumby]

For me personally I will always turn it off to the left because I’m used to turning my meters off to this direction. I do own one meter that has a dual off just like the 121GW (Digitek DT-4000ZC) and it always goes off to the left. Even if I use a mode on the right side of the range switch it still gets turned off to the left. If I had bought a meter that only turned off to the right and didn’t have a left off option, this would probably annoy me to the point of returning it.

Same.

[EEVblog]

Now I need to know if there is or has ever been such a meter?

Near enough

[HKJ]

For me personally I will always turn it off to the left because I’m used to turning my meters off to this direction. I do own one meter that has a dual off just like the 121GW (Digitek DT-4000ZC) and it always goes off to the left. Even if I use a mode on the right side of the range switch it still gets turned off to the left. If I had bought a meter that only turned off to the right and didn’t have a left off option, this would probably annoy me to the point of returning it.

Same.

Now I need to know if there is or has ever been such a meter?

I have not seen that, but some meters has off at the center.