Author Topic: EEVblog 121GW Multimeter Issues  (Read 670144 times)

0 Members and 4 Guests are viewing this topic.

Offline gr8tfly

  • Contributor
  • Posts: 10
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #850 on: June 04, 2019, 04:40:31 am »
I had been logging some weird mains fluctuations and found the log files corrupted too. Duplicate data, and it seems to start after ending the first log (stop, then start, which creates a new file).

Anxious to try out v2.0. Saw Daves brief demo and the differences on resistance and capacitance look terrific! Maybe they tackled logging... I will give that a go and post back.
 
The following users thanked this post: AlanS, JayHattler

Offline kf4hzu

  • Newbie
  • Posts: 9
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #851 on: June 05, 2019, 01:17:23 am »
@hyvamiesh Have you tried updating to firmware 2.00? I just ordered a 121GW myself and would be interested to hear if this is a hardware or software issue.
 

Offline hyvamiesh

  • Newbie
  • Posts: 4
  • Country: fi
Re: EEVblog 121GW Multimeter Issues
« Reply #852 on: June 05, 2019, 08:42:37 am »
This is a hardware issue as far as I know. I've updated mine to 2.0 from 1.57 and the problem persists.
 

Offline CDaniel

  • Frequent Contributor
  • **
  • Posts: 411
  • Country: ro
Re: EEVblog 121GW Multimeter Issues
« Reply #853 on: June 05, 2019, 08:49:31 am »
This is surely an issue with your multimeter , the beeper circuit is very simple and the schematic is available
 

Offline Brumby

  • Supporter
  • ****
  • Posts: 12288
  • Country: au
Re: EEVblog 121GW Multimeter Issues
« Reply #854 on: June 05, 2019, 10:25:12 am »
Has anyone else had the beeper fail on them intermittently?
... ...
Anyone have suggestions where to concentrate my search for a bad solder-joint?
Around the beeper?
 

Offline dcac

  • Frequent Contributor
  • **
  • Posts: 336
Re: EEVblog 121GW Multimeter Issues
« Reply #855 on: June 05, 2019, 08:56:27 pm »
Has anyone else had the beeper fail on them intermittently?
... ...
Anyone have suggestions where to concentrate my search for a bad solder-joint?
Around the beeper?

If you have another dmm, measure the dc voltage over the beeper when it's supposed to beep i.e. in continuity mode, you should measure something like 3 volts.
 

Offline afedorov

  • Newbie
  • Posts: 5
  • Country: ru
Re: EEVblog 121GW Multimeter Issues
« Reply #856 on: June 05, 2019, 08:59:37 pm »
Tested four and a half hour logging at 1s on v2.00 and everything seems fine so far. Thanks a lot, Dave and David.
 
The following users thanked this post: gr8tfly

Offline gr8tfly

  • Contributor
  • Posts: 10
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #857 on: June 05, 2019, 09:19:44 pm »
Good news! Thanks. I'm going to get mine updated and get some logs running on our mains power again.

I forgot to mention I was running at max sample rate. I did a rough calculation of that and came out with 206.97 ms/sample. Maybe as a future feature, a calibrated rate of something slower, but still more than 1 Hz, that would it would be guaranteed to meet – say 500 ms/sample (250 would be better, but my gut says that would be hard to meet). My use case was measuring power line fluctuations of up to ~2 Hz, so I really needed to run at max rate and see what I see. Even if I run at the uncalibrated max rate, I can back calculate an average sample rate and get close to my fluctuation event time-of-day notes (to correlate samples to real-time events).
 

Offline kf4hzu

  • Newbie
  • Posts: 9
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #858 on: June 08, 2019, 05:27:56 pm »
I got in my 121GW and so far really like it! I also would see value in having a consistent sample rate like 250ms instead of pushing the hardware to its max but having several milliseconds of jitter in the readings.
 

Offline dadler

  • Supporter
  • ****
  • Posts: 851
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #859 on: June 11, 2019, 05:15:53 am »
I just upgraded to firmware 2.0, and I am experiencing an issue with auto-ranging using the resistance function and nulling out the leads via REL.

In auto-ranging mode, if I connect the leads together to null out the lead resistance, and press REL, the meter leaves auto-ranging mode and locks into the range it is in (milliohms). This results in essentially all readings being OFL, as I seldom want to measure milliohms at this point. If I press RANGE to manually change ranges, REL gets reset -- which sort of defeats the purpose of REL.

None of my other meters behave in this fashion -- and I'm pretty confident I know how to null out meter leads. Is this a bug in firmware 2.0, or does REL only work when manually ranging?
« Last Edit: June 11, 2019, 05:20:35 am by dadler »
 

Offline newbrain

  • Super Contributor
  • ***
  • Posts: 1714
  • Country: se
Re: EEVblog 121GW Multimeter Issues
« Reply #860 on: June 11, 2019, 08:05:55 am »
None of my other meters behave in this fashion
That's interesting, as all the autoranging DMMs I have/had did lock the range with relative measurements (e.g. Fluke 87V, page 34, same for Fluke 289).
At the same time I've always wondered whether that was an easy to overcome limitation...
Care to share which brads/models don't do that?
Nandemo wa shiranai wa yo, shitteru koto dake.
 

Offline dadler

  • Supporter
  • ****
  • Posts: 851
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #861 on: June 11, 2019, 08:53:31 am »
None of my other meters behave in this fashion
That's interesting, as all the autoranging DMMs I have/had did lock the range with relative measurements (e.g. Fluke 87V, page 34, same for Fluke 289).
At the same time I've always wondered whether that was an easy to overcome limitation...
Care to share which brads/models don't do that?

Thanks for making me check. I actually discovered that the couple of Fluke meters I own switch to manual once REL is pushed, although they lock to a much more useful range. I seldom use these meters and had to get them out of storage, hehe.

The 121GW shows OFL if I try to test a 100 ohm resistor after null-ing the leads in auto-range mode.

Just went through my easy to access meters, others are tucked away or sans batteries:
  • Brymen BM869s: Auto-ranging works with REL
  • Brymen BM257s: Auto-ranging works with REL
  • DER EE DE-5000: Auto-ranging works with REL
  • Keysight 34465A: Auto-ranging works with REL
  • Keithley 2015-P: Auto-ranging works with REL
  • Fluke 17B+: Auto-ranging switches to manual, but the range is <= 430 ohms
  • Fluke 297: Auto-ranging switches to manual, but the range is <= 540 ohms
  • 121GW: Auto-ranging switches to manual, but the range is <= 54 ohms

I am so used to the first 5 meters that I honestly believed the 121GW contained a bug (maybe it does?). Certainly seems like something that would be trivial to implement, if this isn't a bug or oversight. If one doesn't know the resistance, but still desires to zero the leads, the lack of REL auto-ranging support requires an extra measurement. Per resistance measured. Not exactly fun when checking a board or a bunch of microscopic SMT resistors.

Heck, I'd be happier if at least I could manually cycle through the ranges on the 121GW and not lose my REL measurement.
« Last Edit: June 11, 2019, 09:10:23 am by dadler »
 
The following users thanked this post: newbrain, Marco1971

Offline Hydron

  • Frequent Contributor
  • **
  • Posts: 978
  • Country: gb
Re: EEVblog 121GW Multimeter Issues
« Reply #862 on: June 11, 2019, 09:38:09 am »
Just checked and a Fluke 87V goes into manual 600 ohm range when REL is used (in both normal and high-res modes). More useful than ~60 ohms!

Note that on both the Fluke and the 121GW you can manually range _first_ and then REL out the leads if needed, trading off resolution for a larger upper limit (given the extra milli-ohm range available, this is more useful for the 121GW than the Fluke).

I will admit that I hadn't yet run into this particular annoyance in real use - too many meters and not enough time!
 
The following users thanked this post: dadler, Marco1971

Offline dadler

  • Supporter
  • ****
  • Posts: 851
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #863 on: June 11, 2019, 10:58:02 am »
It's more of a psychological issue than a practical one, honestly. For true milliohm offsets near zero, this probably gets lost in the noise once the meter changes ranges. But when you've built a habit of nulling out your leads all the time, it can be sort of annoying  :rant:
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 5795
  • Country: ca
Re: EEVblog 121GW Multimeter Issues
« Reply #864 on: June 15, 2019, 05:45:00 pm »
fluke 189   REL  give :  manual range <500 ohms ?
 
The following users thanked this post: Marco1971

Offline kf4hzu

  • Newbie
  • Posts: 9
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #865 on: June 17, 2019, 04:09:10 am »
I noticed today when sampling a ~125KHz square wave from an FPGA that the duty cycle reading on the 121GW is fairly slow to stabilize. I compared it with another meter I have and it stabilizes to the correct reading in under half a second where the 121GW takes over 10 seconds to get to the same reading. The frequency and period measurement are very quick to stabilize. I'm guessing the duty cycle slowness can be fixed with a future firmware update but I am not sure where to report this. Hopefully this is the right place!

Edit: Below is the verilog code that generates the square wave. It is run on a 64MHz clock (posedge).
Code: [Select]
pin4 <= tickms;

if(tickcounter3 == 8'd128) begin
tickms <= 1'b0;
end

if(tickcounter3 == 8'd255) begin
tickcounter3 <= 8'b0;
tickms <= 1'b1;
end
else begin
tickcounter3 <= tickcounter3+8'b1;
end
« Last Edit: June 17, 2019, 04:16:16 am by kf4hzu »
 

Offline newbrain

  • Super Contributor
  • ***
  • Posts: 1714
  • Country: se
Re: EEVblog 121GW Multimeter Issues
« Reply #866 on: June 18, 2019, 02:02:48 pm »
I noticed today when sampling a ~125KHz square wave
Given the Verilog, those look more as 250 kHz (64 MHz / 256), slightly above the specified frequency for Duty Cycle (200 kHz).
Also, the minimum pulse width is 5 µs, so the 200 kHz limit is a bit of a best case.
Moreover, the stated sensitivity is 2.5 V RMS: what's the FPGA output level?

That said, it might be slow even with in range values, but being my 121GW 3000 km away I'm not able to do any test...
« Last Edit: June 18, 2019, 02:06:52 pm by newbrain »
Nandemo wa shiranai wa yo, shitteru koto dake.
 

Offline CDaniel

  • Frequent Contributor
  • **
  • Posts: 411
  • Country: ro
Re: EEVblog 121GW Multimeter Issues
« Reply #867 on: June 19, 2019, 05:50:19 am »
I checked with a signal generator , square wave at 100KHz and 50KHz at 15-20%, duty cycle seems to be is fast and accurate around 5Vpp , so what you would expect from a 5V digital circuit . Higher or lower than this level is slow to stabilize and increasingly inaccurate . Anyway , bellow 5% is not accurate to be trusted at all , maybe at low frequency .
« Last Edit: June 19, 2019, 06:23:59 am by CDaniel »
 
The following users thanked this post: newbrain, Marco1971

Offline kf4hzu

  • Newbie
  • Posts: 9
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #868 on: June 19, 2019, 06:51:56 am »
Given the Verilog, those look more as 250 kHz (64 MHz / 256), slightly above the specified frequency for Duty Cycle (200 kHz).
Also, the minimum pulse width is 5 µs, so the 200 kHz limit is a bit of a best case.
Moreover, the stated sensitivity is 2.5 V RMS: what's the FPGA output level?

I think your math may be incorrect for the frequency. That code sets the output to 0 when the counter is at 128, then sets the output to 1 when the counter is 255. This would mean a full peak-to-peak cycle would be 512 clock ticks. Luckily measuring the period using the meter is fast and accurate which can confirm the frequency measurement. It shows 0.0080ms which would be 125KHz at a 50% duty cycle.

.....a 5V digital circuit . Higher or lower than this level is slow to stabilize and increasingly inaccurate .

Ok both you @CDaniel and @newbrain mentioned voltage. The pins nominally output 3.3V but the DC measurement of the 125KHz output pin is 1.6713VDC (I would expect this, right? PWM?) or using DC+AC it shows as 2.4143V. Since other readings work fine with the lower voltage I suspect only duty cycle is being impacted by the 3.3V output.

Since Dave designed this meter primarily for electronics work I guess I expected all functions to work well at the voltages that you would find on MCUs and FPGAs. I really like the meter so don't take this as a complaint. I just think this may be able to be fixed in software since another much cheaper meter (Aneng AN8008, which Dave reviewed) is able to measure the duty cycle quickly and accurately.

I just wanted this to be documented in case it can be fixed in a future firmware revision.
 

Offline newbrain

  • Super Contributor
  • ***
  • Posts: 1714
  • Country: se
Re: EEVblog 121GW Multimeter Issues
« Reply #869 on: June 19, 2019, 07:37:19 am »
Anyway , bellow 5% is not accurate to be trusted at all , maybe at low frequency .
$$Vrms=Vp\sqrt{D}$$

So, in our case that would be about 1.12 Vrms, about half of the specified sensitivity for duty cycle measurement.
Bur yes, it seems the usable ranges of Vp  and frequency are a bit limited.
Nandemo wa shiranai wa yo, shitteru koto dake.
 

Offline newbrain

  • Super Contributor
  • ***
  • Posts: 1714
  • Country: se
Re: EEVblog 121GW Multimeter Issues
« Reply #870 on: June 19, 2019, 07:49:53 am »
I think your math may be incorrect for the frequency. That code sets the output to 0 when the counter is at 128, then sets the output to 1 when the counter is 255. This would mean a full peak-to-peak cycle would be 512 clock ticks.

The fragment is not complete, but you stated that it's triggered by a posedge of a 64 MHz clock.
I see the output is 1 for 128 clock ticks and then 0 for the remaining 128, that gives 256 clock ticks for a full cycle (128 1, 128 0, then repeats).

Am I missing something obvious?
Nandemo wa shiranai wa yo, shitteru koto dake.
 

Offline CDaniel

  • Frequent Contributor
  • **
  • Posts: 411
  • Country: ro
Re: EEVblog 121GW Multimeter Issues
« Reply #871 on: June 19, 2019, 08:36:28 am »
I checked again and it's not the sensitivity , it's the  pulse duration , bellow 5% is accurate only up to 10KHz .

But as I said at 3,3Vpp it is not accurate at all at high frequency , again only to about 10KHz , at 100KHz the error is very big if the duty cycle is <50% or >50% .
At 5V logic it is very accurate at 100KHz , maybe 0,2% off , like in specs for 5%-100%.

I don't want to say that is bad or out of specs , it is how it is , personally I don't care too much . For 5V logic circuits it is very OK . But people should know its limitations .

« Last Edit: June 19, 2019, 02:35:55 pm by CDaniel »
 

Offline kf4hzu

  • Newbie
  • Posts: 9
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #872 on: June 19, 2019, 10:29:13 am »
Am I missing something obvious?

Hmm. No, I think I'm missing coffee :) Here is all the related code. I'm going to try a lower output frequency (1khz?) tonight and see how that goes. I also should be able to use something else to generate a 125KHz square wave with 5V.

Code: [Select]
TinyFPGA_BX_pll pll(.REFERENCECLK(clkGB),
  .PLLOUTGLOBALA(clk128),
  .PLLOUTGLOBALB(clk64),
  .RESET(1'b1));

led ledmain (
    .ledclk(clk64),
    .led2clk(clk),
    .led(led),
    .usb_pu(usb_pu),
    .pin1(pin1),
    .pin4(pin4)
    );

Code: [Select]
module led(
    input ledclk,
    input led2clk,
    output reg led,
    output usb_pu,
    output reg pin1,
    output reg pin4
    );

reg [7:0] tickcounter3;

reg tickms;

always @ (posedge ledclk) begin
pin4 <= tickms;

if(tickcounter3 == 8'd128) begin
tickms <= 1'b0;
end

if(tickcounter3 == 8'd255) begin
tickcounter3 <= 8'b0;
tickms <= 1'b1;
end
else begin
tickcounter3 <= tickcounter3+8'b1;
end

end
 

Offline kf4hzu

  • Newbie
  • Posts: 9
  • Country: us
Re: EEVblog 121GW Multimeter Issues
« Reply #873 on: June 20, 2019, 10:29:47 am »
OK, using the below code on an Arduno Uno (5V pins) does greatly decrease the time to a stabilized duty cycle reading at ~125KHz. Does this mean improving this at 3.3V is not possible due to hardware limitations of the meter? And yes I plan to test at lower frequencies on both the FPGA and MCU.

Code: [Select]
#include <util/delay.h>

void setup()
{
  pinMode(9, OUTPUT);
}

void loop()
{
  noInterrupts();
  while(1) {
    PORTB |= _BV(PORTB1); // turn on
    _delay_us(3.81);
    PORTB &= ~_BV(PORTB1); // turn off
    _delay_us(3.81);
  }
}
 

Offline CDaniel

  • Frequent Contributor
  • **
  • Posts: 411
  • Country: ro
Re: EEVblog 121GW Multimeter Issues
« Reply #874 on: June 21, 2019, 07:58:59 pm »
I don't think is possible , my Fluke 867B is better because Hz ( duty cycle ) is a soft button , you are still in volts mode and you can choose manually a voltage range where works the best . This is how it should be done according to manual .

But 121GW use the rotary switch and has fixed sensitivity , whatever that might be when switched to Hz .
« Last Edit: June 21, 2019, 08:07:49 pm by CDaniel »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf