-
This thing sucking my life out.
No problem... just ship it to me and I will dispose of it properly. I have the technology to do so... I am a certified Zombie Slayer - Class 1. Hurry, before it's too late! You really don't want to wake up some morning and find out that your brains have been munched by the beastie! -
Hooked 3458A to the output and 11 hours after got my HTML report with measured check points. Only relative specs with K=1 used.
Generated PDF-report. 100Kohm point is not fail, there is procedure error cause excessive ppms
Comparison of DCL 11V. Right scale is deviation in ppm. Correction offsets applied (-1.0 ppm for 5700A, +0.2 ppm for HULK).
Also there is little problem on my A11 REFDAC board, but it looks okay for first tries (and there are no inguard/external shields on HULK whatsoever now).
Green is HULK calibrator vs 3458A
Brown is 5700A calibrator vs another 3458A.
I'll need get my 720A in order to test better.
A good start for now, to be improved later on, hopefully.
Running second more refined calkit sequence now, with better resistance data.
-
What is the difference between 3458-B and 3458-REF ?
-
Less then 1ppm. They both calibrated off same references.
-
I heard that voodoo Fluke meter is the ducks guts for precision measurements and metrology.
Well, wonder no more, we are about to find out.
Our Hulk calibrator deserve only the best treatment, right?
Too bad it's last calibrated in 2014, but it once I establish it's gain errors vs my calibrated 1ohm/10K/10V standards (thanks to Hulk ACAL transfer!), we can then correct all data accordingly. Most interesting will be resistance testing of course, because this box have option 001, meaning it's able to do ratio measurements directly to my resistance standards like FSL935 or ESI SR104. -
Very nice
PS: I see a 4w connection but no 4w symbol on the 8508A display? -
TiN,
you're, you're a, you're a slummetrology millionaire
e61_phil, 4W is not used in DCV mode, isn't it?
If you want to check a DMM on a calibrator in all modes, it makes sense to directly connect the full cable set -
e61_phil, 4W is not used in DCV mode, isn't it?
If you want to check a DMM on a calibrator in all modes, it makes sense to directly connect the full cable set
On the 8508A it is even in Volt mode possible to use 4w. This is done to avoid a stack of connectors. However, the stack of connectors is a greater problem on a 3458A than on a 8508A (temperature gradients).
-
Both 3458A and 8508 can do "4W" DCV, as ratio mode. Since both meters are measuring , not sourcing , there is no much point to do 4W on voltage just for one signal input.
If anyone have idea what to run as comparison with 8508A, I'm all ears. Need to get my 720A fixed first and calibrated to test things like linearity.
First impression of Fluke - it is super slow and UI is quite awkward. Not that I care about UI much, it will be used with python snakes anyway. Definitely see Datron's design roots, even GPIB command set is nearly identical. I expect 3458A beat Fluke on DCV and ACV but fall short for resistance/current. -
Both 3458A and 8508 can do "4W" DCV, as ratio mode
. Since both meters are measuring , not sourcing , there is no much point to do 4W on voltage just for one signal input.
4wV has nothing to do with ratio. 4wV is to force the voltage on the "right" position for measuring. The other problem are stacked connectors. Especially on the 3458A the input jacks are warmer than the envirement. This causes a temperature drop over the stacked connectors and some EMF. Mr. Zirpel, a retired guy from the PTB "Gleichstromlabor", can talk hours and hours over this topic and the problems related to this.
He also told me that they use the 8508A as a transfer system between AC and DC, because it is that stable. Therefore, I'm not sure if the 3458A really beats the 8508A in AC. In DC, the 3458A might be more linear in the 10V range. But at least the 1000V range of the 8508A is much better. One have to keep in mind, that the 8508A don't rely on ACAL. Everything which needs to be stable in this meter is in a way stable, that it has (at least) the same specifications which the 3458A only reaches which ACAL.
If your 720A is still broken: Why not running some tests like the ones Dr. Frank showed here? Hook all meters on your calibrator, trigger all at the same time and run through the ranges. -
Im not sure why you need the 720A, since the 3458A should be at least as linear as the 720A apart from the 1000V-region, where a 752A really shines. I eagerly await the 8508A-teardown.
-
Now TiN, it's time for conquering the world.
-
e61_phil is right, the 1281 / 8508A are fully staffed with Vishay ultra stable resistors, where 3458A uses TaN arrays only.
Therefore, 8508A beats 3458A also concerning T.C. stability in all modes, especially in varying room temperature...like in TiNs cave...
And they probably have used a much more stable LTFLU reference module.. currently even me have problems with lab temperature.
Due to current heat wave over Europe, i.e. 35..39°C over the last 3..4 weeks, temperature has risen in the basement from usual 21.5°C to 24.5°C, so I really see the mediocre 0.15 ppm/°C of the LTZ1000A standard module
Good, that I have 5 LTZ1000 modules, and the 5442A, which are all temperature compensated / stabilized
Frank -
Very nice to finally see a 8508A in your lab.
I am sure you will soon find out why I prefer it over 3458A. -
4wV has nothing to do with ratio. 4wV is to force the voltage on the "right" position for measuring.
Que? Not sure what you're saying there.
Naiive me expected the second pair to be some sort of guards, to be used in case one wants to measure sources with very high impedance. A glance in the (detailed and freely on-line available handbook https://us.flukecal.com/literature/product-manuals/8508a-reference-multimeter-users-manual) tells that those are however explicitely for calibration purposes:
--8<--
4wV Allows operation with calibrators which provide a remote 4 wire sensing
capability by providing connections between INPUT HI and SENSE HI and
between INPUT LO and SENSE LO.
The 4wV annunciator in the left-hand display indicates that 4-wire sensing
is active.
-->8--
It seems to be as Dr. Frank stated earlier, just a convenience feature:
p. 5-4
--8<--
In 4wV mode Sense Hi is internally connected toHi and Sense Lo is internally connected to Lo.
-->8-- -
tells that those are however explicitely for calibration purposes:
Where is the part which explicitely states it is only for calibration? It is for every source with 4w output (sense and force).
That's what I tried to say. It allows the forcing and sensing cables to properly connect to the instrument. On the 3458A you have to hook four cables to two connectors. I will try to be more precise next time, sorry.It seems to be as Dr. Frank stated earlier, just a convenience feature:
If one follows the explanation of the PTB guy, it is much more than only convenience. -
4wV has nothing to do with ratio. 4wV is to force the voltage on the "right" position for measuring.
Okay, I see what you mean now after reading a bit of manual. Enabled it now, restarted datalog. Don't see a difference yet.QuoteHe also told me that they use the 8508A as a transfer system between AC and DC, because it is that stable. Therefore, I'm not sure if the 3458A really beats the 8508A in AC.
For ACV we have more powerful contender - Wavetek 4920M. I'd like to see 8508A try to beat that one.QuoteIn DC, the 3458A might be more linear in the 10V range. But at least the 1000V range of the 8508A is much better. One have to keep in mind, that the 8508A don't rely on ACAL. Everything which needs to be stable in this meter is in a way stable, that it has (at least) the same specifications which the 3458A only reaches which ACAL.
3458A specs are very conservative. There are lot of speculations in sentences above, so I will surely be running comparison tests with all that to see actual data of this 8508 and four of my 8.5d meters (6581T does not deserve to be mentioned
), now that I have decent source. QuoteIf your 720A is still broken: Why not running some tests like the ones Dr. Frank showed here? Hook all meters on your calibrator, trigger all at the same time and run through the ranges.
I need 720A anyway for other purposes. Unlike 3458A, linearity of 720A I can at least calibrate out. I don't have friends (yet?) to get access for PJVS system to test linearity of 3458A directly.QuoteI eagerly await the 8508A-teardown.
I'll await for that too, but it's not happening, xDevs labs ain't rich enoughQuote from: Dr.Frank1281 / 8508A are fully staffed with Vishay ultra stable resistors, where 3458A uses TaN arrays only.
After waiting gazillion weeks , I might just got some solution for that "problem":
Initial datalog.
3 hours from power on, voltage still climbing, +0.68ppm so far. Data before 9:50 was saved with lost precision due to snake bug, hence the steps.
Gap between 9:50 and 10.05pm is ACAL ALL on 3458B.
-
I'm not sure if I already posted this. Attached is a comparison of measuring a 20Meg USF371 Caddock Resistor. The resistos sits in a thermal stabilized chamber and has guarded connections.
The comparision isn't fair to the 3458A, because it is on 20% FS and on 100% at the 8508A. Anyhow, it is impressive.
-
e61_phil, 4W is not used in DCV mode, isn't it?
If you want to check a DMM on a calibrator in all modes, it makes sense to directly connect the full cable set
On the 8508A it is even in Volt mode possible to use 4w. This is done to avoid a stack of connectors. However, the stack of connectors is a greater problem on a 3458A than on a 8508A (temperature gradients).
What it "really" is : Double "isothermal" banana jacks. The two upper red jacks are connected together with a relay - and so are the two black ones. Electrically (assuming zero current from infinite input impedance of the voltmeter) it is the same as banana 2-to-1 Y-adapters or stacks of two banana plugs.
However - thermally there is a difference (to a stack of banana plugs). Assuming that the meter is hotter than the environment, there is a temperature 'gradient' from inside the front panel / instrument chassis to the outside. So the 4wV moves the Y-adapter or (the vertical) 2-plug-stack to a (horisontal) "thermal reference plane" inside the instrument.
Interesting "nV thinking". It has no special use or relation to "force and sense"- it can be used for anything that benefits from a double (parallel connected) input jack. So essentially a voltmeter with double paralleled input jacks.
Copy from manual http://assets.fluke.com/manuals/8508A___umeng0300.pdf p. (pdf) 43 of 198:
"Allows operation with calibrators which provide a remote 4 wire sensing
capability by providing, at an internal reference plane within the DMM,
connections between INPUT HI and SENSE HI and between INPUT LO and SENSE LO.
Use of an internal reference plane can be helpful in 4wire AC, AC transfer, or 4wire AC – DC transfer applications. "
-
I'm not sure if I already posted this. Attached is a comparison of measuring a 20Meg USF371 Caddock Resistor. The resistos sits in a thermal stabilized chamber and has guarded connections.
The comparision isn't fair to the 3458A, because it is on 20% FS and on 100% at the 8508A. Anyhow, it is impressive.
And 8508A in 200V mode, compared to 3458A with its mediocre 1V, or so? -
What it "really" is : Double "isothermal" banana jacks. The two upper red jacks are connected together with a relay - and so are the two black ones. Electrically (assuming zero current from infinite input impedance of the voltmeter) it is the same as banana 2-to-1 Y-adapters or stacks of two banana plugs.
Thanks, to make it much clearer. That was what I was trying to say.I'm not sure if I already posted this. Attached is a comparison of measuring a 20Meg USF371 Caddock Resistor. The resistos sits in a thermal stabilized chamber and has guarded connections.
The comparision isn't fair to the 3458A, because it is on 20% FS and on 100% at the 8508A. Anyhow, it is impressive.
And 8508A in 200V mode, compared to 3458A with its mediocre 1V, or so?
Yes, but the design of the current source and the possibility of a proper guarding on the 8508A is much more important than the high voltage. For the measurement I used two coaxial connections. Both shields connected to the Ohms Guard. The driving signal is low impedant and more or less immune to leakage. The current input is held on the ohms guard potential, which reduces the leakage a lot.
-
A 'fair' comparison can be done by the 8508A specification already.
Normal and high voltage mode beat the low power mode by far, which is the region, where the 3458A operates. 5, or 25 times more stable.
-
I don't get all this double banana stack talk. Who uses banana on precision measurements, when spades are available?
Also I noticed back of the meter is really warm, +45ish C.
There is large temperature difference as result between rear posts and normal cable temperature.
Actively cooled 3458A is never this warm anywhere at binding posts, including rear ones.
Perhaps all this should move into 8508 thread, or in few pages more we'll have to rename dear Hulk into DMM fight.
-
Perhaps all this should move into 8508 thread, or in few pages more we'll have to rename dear Hulk into DMM fight.
that’s a good idea, maybe it makes sense to move my old 8.5 digit DMM post into metrology
https://www.eevblog.com/forum/testgear/8-5-digit-dmm/msg258880/#msg2to58880
or I can open new and insert relevant data -
Great idea, but please stay away from fighting one DMM against another. We're all not married with Fluke or HP. I think both meters have pros and cons. But I don't agree with your comparision Dr. Frank, but I don't want to get deeper in this discussion
@Tin: I will try to get our thermal camera and make some pictures of binding posts and so on on both meters. btw the PTB guy said Transmille is gone one step further and they build a special cage around the binding posts to keep them all at the same temperature. The absolute temperature is much less important than the difference between the inputs.
