EEVblog Electronics Community Forum
Electronics => Metrology => Topic started by: TiN on May 15, 2018, 05:24:04 pm
-
One who did not build own calibrator from random crap cannot be called true volt-nut.
Back in 2014, when I was naive and overly optimistic, I thought it would be piece of cake to build a DIY MF-calibrator. There is still a thread (https://www.eevblog.com/forum/projects/project-kx-diy-calibrator-reference-sourcemeter/) as evidence of that. Fast forward to 2018, I've learned quite a bit, build over 30 LTZ1000-based references, accumulated lot of expensive not-necessary equipment and even got somewhat working 5.5-digit Time Electronics 9823 calibrator (https://xdevs.com/fix/te9823/).
However friend of mine introducted me to real deal, Fluke 5700A, and since we peeked inside of it (https://xdevs.com/fix/f5700a/) my soul was gone.
Only owning very personal 5700A can restore the status quo. As with usual volt-nut disease, first sign is suddle - I thought why not buy A11 REFDAC board from 5700A, that will give me finished and proven dual LTFLU-reference and DAC eh? Just add some RPI, power supply and level translation, some python code, and we can get module to generate some precision voltages, eh?
Well, shortly after I got ohm boards.... then I discovered that to power these things I will need about 6 different unipolar rails. Well, after XXXX $$$ spent over last 2 years I have now all 5700/5720A boards, some in multiple quantity. I even scored Fluke 5500A front panel, which I hope to modify to fit 5700A VFD display board.
Today I received last bits - A3 and A4 backplanes, so this project is to kick off from waste of moneys and dust collectors to hopefully something workable.
(https://xdevs.com/doc/Fluke/5700a/imgx/zf5700_1.jpg) (https://xdevs.com/doc/Fluke/5700a/imgx/zf5700.jpg)
Now the only "issue", I don't have a chassis. So far my best thinking is to make CNC plates from Aluminum and build something like a crate to hold all things together, in arrangement same as real-deal 5700A.
Who need chassis anyway, eh?
(https://xdevs.com/doc/Fluke/5700a/imgx/zf5700b_1.jpg) (https://xdevs.com/doc/Fluke/5700a/imgx/zf5700b.jpg)
To be continued.
First I need to restore cables with connectors on transformer and check baseplane boards and outguard power. Then we can continue to inguard power and then start adding boards as progress goes.
-
First smoke tests, just out-guard section. :-/O
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/smoke_assy_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/smoke_assy.jpg)
This is how it looks like. Rather tricky to position everything stable enough, because if heavy weight of transformer and it's steel frame. :phew:
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/smoke1_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/smoke1.jpg)
After few mishaps with front panel power and mainboard connector not matching the service manual schematics, I got it powered on
No communication error from outguard. I wonder why, heh. :-DD
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/smoke_err_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/smoke_err.jpg)
Now we go ahead for more involving tests.
-
RSVD for complete runs.
-
Looking forward to this epic build.
-
Crazy :-+ :popcorn:
-
I am always amazed what you show us, looking forward to see the final outcome
-
(https://media.makeameme.org/created/thats-some-dope.jpg)
-
Livestream for first smoke tests pls. :popcorn:
-
Inspired by your mentioning of the A11-REFDAC-stand-alone-possibility i also aquired a A11 a few years ago.
I have the necessary power supply about half way ready and already tested the Arduino-code to control the 82C55 and 82C54 on the A11. Assembling kinda slowed down after i received my Fluke 5440b Calibrator. :=\
-
Your work is amazingly impressive.
Thanks for sharing
-
:popcorn: :box:
It looks like nice new story 8)
-
TiN always beeing TiN.
One day Fluke will see it and throw some piece of gear on your house.
-
is Fluke's backyard in Taiwan? :D
-
(http://78.media.tumblr.com/ee78efeae0c0939dd45fe37739b258b0/tumblr_nblyh3ttjQ1sj67gjo1_500.gif)
A real "Frankenstein" :-DD :-+
-
Awesome thread! Thanks for sharing.
I am convinced that you will not stop when it is finished to original spec, but start looking for ways to make it ever better.
I am subscribed.
-
I decided not to lurk around and go right for the business.
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/testrig_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/testrig.jpg)
Now it's important to note, that I was working on my calibrator boards in background for last year or so, and all of these boards are already repaired and working except A18 regulator supply and high-voltage power supply board. I was lucky to have friend with 5700A that I have here in service now (A12 and A7 issues), so I got a chance to verify my boards one by one (obviously except power boards, not going to risk it).
So while it may look stupid plugging all 23 boards together for first tests, I actually have confidence that 91% of them working OK already. Question was only about A3/A4 and getting it all together.
Quick check on VFD brightness.
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/vfdleft_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/vfdleft.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/vfdright_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/vfdright.jpg)
VFD glass is dim and aged, but we have a future solution for that. It is good enough for current states, so no worries.
Shall we try bold attempt on running self-test? ;D
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/selftest_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/selftest.jpg)
It takes about 15 minutes to have calibrator self-verify itself, using internal LTFLU, 10K and 1 ohm and Fluke FTS references.
I don't have WBV option modules plugged in yet, it's too early for that.
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/fulltest_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/fulltest.jpg)
She passes! :-DD :popcorn:
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/testpass_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/testpass.jpg)
I also run the DCZ internal self-calibration. It did not spit any errors either.
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/dczcal_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/dczcal.jpg)
Based on today's progress, I think we have a progress :=\
Soon will need to start designing CAD drawing for chassis metalwork parts. First step would be to make each assembly rough 3D model, so I can make it all "fit".
-
If you have the 5700A of a friend in house it should be easy for you to copy the metal sheet parts into CAD and get them lasercutted and bend, right? I'm sure you find very cheap lasercutter at your location.
-branadic-
-
Happy Metrology Day 2018 everyone! :-+ :-DMM
(https://xdevs.com/doc/xDevs.com/Lab/xdevs_zombielab_1.jpg) (https://xdevs.com/doc/xDevs.com/Lab/xdevs_zombielab.jpg)
:popcorn:
-
Finest metrology porn at its very best.
Really gets me started :scared:
-
Random question TiN, what is that white thing in the upper left with blue tape on it? Is it just a light?
TonyG
-
Yes.
Troubleshooted A14 high-voltage control assembly today. Found few dead zeners which are not zenering anymore, and cooked high-voltage 10pA Ibias opamp, that drive the PA assembly, which in turn drives the high-voltage transformer.
Need to buy replacement for it, that is OPA445BM. $24 a pop, happily DK have them in stock.
-
Got some pretty PCBs to play with.
PCB A: Expanders for DIN backplane connectors, to electrocute myselftroubleshoot stuff during calibrator operation.
(https://xdevs.com/doc/Fluke/5720A/ztest/zflk_expa_1.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/zflk_expa.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/zflk_expb_1.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/zflk_expb.jpg)
PCB B: Thermal spreader PCB that sits on the back of A10 main resistance board on original Fluke 5700/5720A's. A10 which I got from ebay was stripped of this shield board, so I build my own version.
It is 4-layer, with solid copper plane on 3 outside layers. Side facing the board is isolated.
(https://xdevs.com/doc/Fluke/5720A/ztest/za10_tpcb_1.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/za10_tpcb.jpg)
I got plenty extras, so I can even modify few to use as front protection cover above for resistors. I don't want to ruin those ceramic hermetic networks by accident, those are unobtaniums.
(https://xdevs.com/doc/Fluke/5720A/ztest/za10_tpcbb_1.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/za10_tpcbb.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/za10_tpcbt_1.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/za10_tpcbt.jpg)
And compare to original Fluke one (tinned board on the right), which is 2-layer (metal only on outside).
I don't know about you, but I like mine better. :box:
(https://doc.xdevs.com/doc/Fluke/5720A/ztest/za10_compare_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/ztest/za10_compare.jpg)
There is also PCB C stuff, but that to be unveiled in due time.
(https://doc.xdevs.com/doc/xDevs.com/Lab/pcbc.jpg)
-
:D Looks like the VFD-replacement.
-
Echo88
Heck no :)
Extender in action:
(https://doc.xdevs.com/doc/Fluke/5720A/za9/fa_ext_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/fa_ext.jpg)
Anyone interested in each board overview? :-DD
Today I received spare A9. Just couldn't resist. :-//
(https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_top_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_top.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_bot_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_bot.jpg)
The Ohms Cal assembly (A9) contains the 1 Ohm, 1.9 Ohm, and zero short resistance values. It also contains a digital control stuff, and a 2-wire compensation circuit for of two-wire ohmmeters cal. A differential amplifier circuit and a 2V/5V/10V source from this board also are used during calibration of the ohms function. Also lot of latching DS2E/DS4E relays ;).
(https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_side_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_side.jpg)
Board made in 2004, sporting newer green better quality PCB. Shield removed for clarity.
(https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_shd_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_shd.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_isoc_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_isoc.jpg)
The floating supplies for U7 and U8 consist of a dc-to-dc converter made up of LT3526N PWM regulator (do not confuse it with modern LTC3526!) (https://xdevs.com/doc/Fluke/5720A/pdf/LT3526.pdf), transformer T1 and associated components. Switching-regulator control IC U6 switches +17S through the primary of T1 at about 30 kHz as determined by R27 and C21. The secondary of T1 is
rectified and filtered by CR1, CR2, CR9, CR10, C22, and C23. The voltage is then regulated by 8.2V zener diodes VR1 and VR2. Additional filtering is provided by L3, L4, R43, R44, C24, and C25. Two capacitors (C53 and C54) between ACOM and SCOM reduce switching noise that would otherwise appear across the calibration resistor.
You also can see primary 10KOhm (ceramic hermetical network), sneak in between relays. This is the resistor that gets calibration vs external 10 Kohm during ACAL. Four of these resistors are used in my voodoo prototype Fluke SL935 (https://xdevs.com/article/fsl935/#rstc) to form precise 10KOhm.
(https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_ref10_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_ref10.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_loz_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_loz.jpg)
The 1W value is made of four 4 Ohm wirewound resistors in parallel (R41). The 1.9 Ohm value is made of two 3.8 ohm wirewound resistors in parallel (R42). Relays K4 and K5 connect the 1 ohm and 1.9 Ohm values to OHMS OUT HI and OHMS SENSE HI. Relays K6 and K30 connect them to OHMS OUT LO and OHMS SENSE LO. Relays K7 and K8 select the short. There are low and high-side current cancellation circuitry too.
(https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_ctrl_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_ctrl.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_res_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720A/za9/za9_res.jpg)
-
Ok, ok, you've convinced me. Now i'm looking on ebay germany, if I can get these pcb's for a small amount of money too... :-DD
-
@Tin: Is it possible to run some linearity tests? Fluke 5700A against 3458A BOTH in fixed Range from 0V to 10V? That would be very interesting.
-
Got tinted blue acrylic lens today for front panel mods. Looking much better with it. :-DD
Blue tint gives interesting looks too. Vendor did not have gray lens in stock.
(https://xdevs.com/doc/Fluke/5720A/ztest/z5720a_face_1.jpg) (https://xdevs.com/doc/Fluke/5720A/ztest/z5720a_face.jpg)
Now need to dremel some metal work to make 5700 display board fit the 5500A FP frame.
-
NICE!!! :-+
-
Looks
like better than a bought one!
-
Damm TiN you are NUT's - Not just Volt-nut's :clap:
Amazing assembly, and i'd say 99% impossible to get those boards for human payable prices in EU'land
/Bingo
-
Repaired today my A14 (High-voltage/High-current assembly).
Culprit part was $24 opamp:
(https://xdevs.com/doc/Fluke/5720A/za14/opa445_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/opa445.jpg)
Board overview:
(https://xdevs.com/doc/Fluke/5720A/za14/xa14_top_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_top.jpg)
There is I/O and relays control on the right side, A=-100 amp, and other analog circuitry. This board, together with A15 used to generate 1KV DC/AC and 2.2A ranges in calibrator. Big metal thing on the right - is custom transformer made by HIREL SYSTEMS for Fluke (P/N 775288). Transformer was checked to be OK, so that was a "phew" moment.
Original issue was that high-voltage ACV was working, but 2.2A and DCV not, tripping output to standby and failing self-test like:
A15 (HV +DC High voltage output fault) error 3111
A15 (HV +DC/Amp Ref/Error amplitude fault) error 3112
A15 (HV +DC preamplifier fault) error 3109
A15 (HV +DC series pass & current fault) error 3110
(https://xdevs.com/doc/Fluke/5720A/za14/xa14_bot_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_bot.jpg)
Board received damaged, thanks to "great" shipping :-- But nothing too hard to fix.
(https://xdevs.com/doc/Fluke/5720A/za14/xa14_dmg_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_dmg.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_fix_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_fix.jpg)
Also because transformer is not tightly mounted to board (it supposed to be fixed by inguard chassis frame, not PCB), it made a dent in board.
(https://xdevs.com/doc/Fluke/5720A/za14/xa14_dent_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_dent.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_rev_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za14/xa14_rev.jpg)
But inner traces still ok, so I'll just let that be.
After some troubleshooting and opamp replacement, she is fully functional now. :)
Now I only need to fix my A18, which had one of 220K 3W carbon-composite resistors go boo to 75 ohms and cooked PCB under it. :scared:
My strategy is typical - mill out the carbonized FR4 material, restore connections with copper jumpers and put new components (including replacement for all rest carbon comp mines)
(https://xdevs.com/doc/Fluke/5720A/za18/za18_cavea_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za18/za18_cavea.jpg)(https://xdevs.com/doc/Fluke/5720A/za18/za18_caveb_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za18/za18_caveb.jpg)
Isolating stuff on this area of the board is bit important, because this is +/-250V supply regulator for MFC's PA.
(https://xdevs.com/doc/Fluke/5720A/za18/za18_inner_1.jpg) (https://xdevs.com/doc/Fluke/5720A/za18/za18_inner.jpg)
Schematics of affected area with evil resistors marked out:
(https://xdevs.com/doc/Fluke/5700a/img/A18/f5700_a18pa_res.png)
To be continued...
-
Looks familiar. Those 220k resistors should be inspected annually or replaced based on parts used in the 5700A/20A II.
-
Hi Tin, what made the error “Guard crossing protocol failed to start”. I have the same issue with my fluke 5700.
-
That error happened when I power on only outguard boards. Essentially it means A20 cannot talk to A17 over the fiber.
If all power is OK, then maybe A17 died. I have repaired one recently.
-
I replaced the A17 with another good A17 board but i still get the same message. Could it be the a20?
-
Did you check power on both in guard and outguard domains? Check fiber cable too.
-
I will check it later, I am just waiting for another 5700 I borrowed from a friend. I will see later which one is the culprit. Thanks Tin
-
I'd be very careful with somebody else's 5700, as if one of your PSU boards cooked (old A18 can go bad easily due to CC resistors in PA supply section), then you can very much kill your friends 5700A. Also beware high voltages present in calibrator, even when output is not programmed for high voltage/operate mode. :-BROKE
-
Thanks for the advice Tin, now its already working. Found out that the optical connector needs soldering. Now its running good and passed all self test. Btw I got an R&S smiq03 maybe you are intersted in. I will check for the error messages so you can have another project. Thats for free because of the free advice.
-
I should rename this thread from Zombie to Hulk Edition.
Why? This is the case when few photos work better than hundred books... :wtf:
Key component - chassis, has arrived today. :-DD
Just few minor dents and scratches, no worries.. :-//
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_case_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_case.jpg)
Few, they say? ???
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_face_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_face.jpg)
5720A.. it was... :scared:
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_fpback_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_fpback.jpg)
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_open_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_open.jpg)
Of course I bought this case without any PCB. Can't afford it if you want to ask the price.. :popcorn:
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_fpfr_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_fpfr.jpg)
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_right_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_right.jpg)
Right side :D
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_nonflat_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_nonflat.jpg)
There are flatness "issue".
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_vslts_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_vslts.jpg)
Low Thermal binding post is gold-plated CuTe for reference. If I file 0.3mm off it for notch, it should fit in just fine.
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_keyaw_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_keyaw.jpg)
Film sheet that originally stick between VFD glass and front panel plastic is conductive. It's some sort of (ITO?) RFI/EMI shield mylar, to reduce radiated crap from high-voltage VFD spitting outside of calibrator.
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_emires_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_emires.jpg)
Put 5700A VFD board in...
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_vfdsp_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_vfdsp.jpg)
Like a bought one, eh?
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_fpne_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_fpne.jpg)
Some quick tests with reassembled panel
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_test1_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_test1.jpg)
"She will be alright"...
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_test6_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_test6.jpg)
More photos at xDevs.com post (https://xdevs.com/article/hulk5720a/).
-
Unbelievable, what kind of crap you can buy.
Anyhow, even if the proceeding steps are getting smaller and smaller, there's now light at the end of this long tunnel.
"First Light" expected in a few weeks, or so?
Frank
-
It physically hurts to see such a device tormented so brutally. But nurse TiN is already at work and caresses the patient. :popcorn:
-
It physically hurts to see such a device tormented so brutally. But nurse TiN is already at work and caresses the patient. :popcorn:
While wearing his oh, so fetching "naughty nurse" uniform! >:D
TIN ain't afraid of no crap! No how! No way! It's all good to him!
-
That case must have come the EEVblog's favourite eBay seller in Israel!
-
Lurking on this thread. One of my favourite on here at the moment :)
That Israeli seller is “interesting”. Did a bomb land on his warehouse? So much good stuff battered.
-
Lurking on this thread. One of my favourite on here at the moment :)
That Israeli seller is “interesting”. Did a bomb land on his warehouse? So much good stuff battered.
More likely he is digging through the dump bins of a military warehouse, and grabbing the stuff they ran over with the bulldozer that day.
-
Perhaps somebody who is familiar with mechanical field can help me a little?
Since this wonderful shaped chassis came without top inguard and outguard shield covers, I'm going to design and get manufactured suitable replacements.
CAD drawing with photo-texture over:
(https://xdevs.com/doc/Fluke/5720A/zcase/top_Shield.jpg)
Item I need help with is this:
(https://xdevs.com/doc/Fluke/5720A/zcase/top_nut.jpg)
It is a press-fit "washer" which allow bolt head to stay flush to the surface of the shield.
(https://xdevs.com/doc/Fluke/5720A/zcase/shield_nut_cut.jpg)
It does not have a thread. Press-fit diameter is 3/8". What is the proper name for it, so I can google it and source from some hardware store/supplier? I tried to google "press-fit nut, washer, blind screw press-fit" etc, but all I get are standoffs or threaded inserts.
Here is photo of these in the original cover:
(https://xdevs.com/doc/Fluke/5790A/img/av_shield_nut.jpg)
-
Try to look for knurl inserts
-
Insert or rivet nuts..
-
I think the fasteners used by Fluke were made for captive screws but Fluke inserts the screw in the other direction.
They would be similar to this https://bit.ly/2NSuWR9 (https://bit.ly/2NSuWR9)
The ones in the link are probably too long but it might help him narrow down his search.
-
https://en.wikipedia.org/wiki/Swage_nut
And a bored person with a lathe
-
https://eu.shop.bossard.com/group/en/9504/nuts-inserts/insert-nuts/bn-201_hex-self-clinching-nuts
You put it on the other side (like Fluke did) and you tighten it...
-
Those all are not the ones used. There is no thread on original insert but a lip for bolts head. Shield does not hold a bolt, it goes thru , just hiding the bolt head under surface.
-
Does Fluke still sell the shield as a spare part ? This might be easier than to get one made.
For the bolts, one might get away with just sink holes and mating screws. Not as pretty, but should be easier to get.
-
They did not even reply any of my parts quote requests before, so those were dead end wishes. Service manual implies that you can order parts, but you likely need to be NMI or big known cal lab for them to even spend time replying?
I could just bolt the shield using regular flathead screws, and it can work just fine. It is just an idea to make it look kinda like real deal that bothers me :). After all this time and funds inversted into this , skipping on little screws feeling like a betrayal :D.
-
Hi TiN, IMHO you want a set of these guys:
https://apexfasteners.com/fasteners/images/Flush-MountedPanelScrewComponents.pdf
It's true, they have thread on the insert, to retain the screw, but they are so designed that the screw head is flush with the insert, and the threads can be easyly drilled out.
Cheers,
DC1MC
-
but you likely need to be NMI or big known cal lab for them to even spend time replying?
My experience with Fluke Calibration is quite good. I had a few questions related to the uncertainties and their calculations of the recently bought 8508A and they answered in a very long detailed email and the contact person also created an excel sheet for me to play with the numbers. I also asked something about the specifications of the 5450A and the 5440B and they answered immediately.
Nevertheless, I bought a new 8508A a couple of weeks before and I asked with my company email.
-
They'd better answer questions that 22K$ DMM customer asks. :) However I didn't buy any 5700/5720A calibrators from Fluke, so my parts quote requests not justified I guess. :=\.
Attached better pics of the nut.
-
Hi,
How about https://www.mcmaster.com/#94350a130/=1dq264u (https://www.mcmaster.com/#94350a130/=1dq264u)
-
I think this is closest we can get, so I'll go with that.
Since outer covers aren't quite fixable, I also made few CADs for those. Like a bought one, eh?
Let's see how hard would it be to get these manufactured in one off qty.
-
Solid Works?
-
They are solid :)
Little update today.
Took apart frame parts. Inguard frame is easily fixable, but outguard right wall need some work.
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_framea_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_framea.jpg)
Even plastic guide frames are nearly intact, which is good news. I don't want all those boards floating inside willy nilly.
(https://xdevs.com/doc/Fluke/5720A/zcase/hulk_frameb_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/hulk_frameb.jpg)
Also chopped off posts area from old ruined front panel plastic. Compare 57XX post construction :-+ vs 5500A one with shrouded banana receptacles :--.
(https://xdevs.com/doc/Fluke/5720A/zcase/posts_top_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/posts_top.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/posts_bot_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/posts_bot.jpg)
I'll trim excess plastic on sides, and make it fit just like glove on my 5500A front panel frame. Should all fit seamlessly.
(https://xdevs.com/doc/Fluke/5720A/zcase/posts_merge_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/posts_merge.jpg)
-
Today's update, work on posts a bit to get the panel fit.
Condition before:
(https://xdevs.com/doc/Fluke/5720A/zcase/ltpost_before.jpg)
Condition after on my Fluke 5500A frame.
(https://xdevs.com/doc/Fluke/5720A/zcase/ltposts_top_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/ltposts_top.jpg)
I had to file 0.3mm for key on the rods to make them fit the original Fluke base. Other the cap plastic color, rest of the LowThermal connector is identical.
(https://xdevs.com/doc/Fluke/5720A/zcase/ltposts_out_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/ltposts_out.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/ltpost_back_1.jpg) (https://xdevs.com/doc/Fluke/5720A/zcase/ltpost_back.jpg)
USD $175 worth of connectors. Done. :)
-
Wow, that panel looked like junk but whatever juice you use in your magic wand turned it into practically a new item. Impressive work.
BTW, super jealous of your stuff, I want it all! :D
I'm gonna have to make friends with the local test equipment outlet/reseller here in Akihabara to see what gems I can score from their junk pile..
-
Midas touch
-
After few days of bang-bang smash-smash, I got main chassis parts somewhat straight. Today decided to do a test fit and start assembly of the frame.
Here are few shots from today's timelapse footage. ;)
(https://xdevs.com/doc/Fluke/5720A/hulk/step0_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/step0.jpg)
The cage didn't suffer much damage, and its geometry is acceptable. A3 and A4 motherboards align well.
(https://xdevs.com/doc/Fluke/5720A/hulk/step1_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/step1.jpg)
Couldn't resist, and plugged all (almost) my inguard PCBAs inside to make sure all slides nice. No issue here, yay :)
(https://xdevs.com/doc/Fluke/5720A/hulk/step2_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/step2.jpg)
Now time for heavy mains transformer frame and cooling fans. You can see transformer repaired wiring (previous owner chopped wiring right off :)).
(https://xdevs.com/doc/Fluke/5720A/hulk/step3_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/step3.jpg)
Now test fit with front panel assembly.
(https://xdevs.com/doc/Fluke/5720A/hulk/step4_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/step4.jpg)
Outguard PSU and CPU cards install nicely too.
(https://xdevs.com/doc/Fluke/5720A/hulk/step5_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/step5.jpg)
Like some said - there is distant light in the end of this XXXXX$ USD tunnel :D.
-
Congratulation to an almost finished investment. At the end, how much money did you safe compared to a fully functional but used unit?
-branadic-
-
I paid more than few recent sold ones on eBay, but I also get extra goodies too and few multiple boards to tinker with. It is also not one big payment but spendings over two years since I started collecting bits&pieces. If you want a unit, go buy used one, obviously. I knew from the start that it would be even or more $ than used box.
Anyhow, it's a learning process and fun that matters for me, and I'm happy to be able to do it. If you factor in cost of time , then most of "good deals" of broken T&M are more expensive than complete working units, but we still do it, right? :)
-
I paid more than few recent sold ones on eBay, but I also get extra goodies too and few multiple boards to tinker with. It is also not one big payment but spendings over two years since I started collecting bits&pieces. If you want a unit, go buy used one, obviously. I knew from the start that it would be even or more $ than used box.
Anyhow, it's a learning process and fun that matters for me, and I'm happy to be able to do it. If you factor in cost of time , then most of "good deals" of broken T&M are more expensive than complete working units, but we still do it, right? :)
Illya,
yep, the end of the tunnel comes closer..congratulations, and again - well done.
In the old days, it was quite easy to get old and damaged equipment for a bargain, with an easy fix.
Even from the US, including tax and VAT, it was possible to get top notch test equipment for a very low price.
Today, it's not that easy any more, up to impossible, maybe also due to the revival of DIY electronics and metrology instruments, by the eevblog effect..
Anyhow, I really appreciate very much all your re-engineering work you've done and that you're still doing.. and probably spending all your income on that stuff..
It's fun for you, for sure, but also a big knowledge benefit for everyone in this community.
And a big fun for me, also, to follow your progress.
At last, I just want to point to your unique final conclusions and categorization of the 3458As minor and major faults (U180!) .. that has helped a lot many people..
Ok, I think that's enough of ultimative adulation .. get this thing running, finally!
Frank
-
old A18 can go bad easily due to CC resistors in PA supply section
just my experience, if A18 is too dusty it would cause output tripped if set value above 20V, and we get errors from A14~A16.
-
Initial chassis works done, so I decided it's time to test fit stuff and give it a smoke test.
(https://xdevs.com/doc/Fluke/5720A/hulk/hz_left_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/hz_left.jpg)
(https://xdevs.com/doc/Fluke/5720A/hulk/hz_right_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/hz_right.jpg)
(https://xdevs.com/doc/Fluke/5720A/hulk/hz_back_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/hz_back.jpg)
(https://xdevs.com/doc/Fluke/5720A/hulk/hz_front_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/hz_front.jpg)
(https://xdevs.com/doc/Fluke/5720A/hulk/hz_top_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/hz_top.jpg)
There is still some bang-bang smash-smash work need to be done on the rear right corner, to get sh`t align properly, but I don't have patience to deal with that now.
So minimal set of boards installed, to allow me repair and do some checks on inguard A18 regulator board (yes, one that got carbonized by carbon resistor and have drilled holes now :-/O).
After few hours of jumpwiring and checking, I've got it working. No smoke and blown parts so far, yay!
Positive PA+ 375VDC output: :-+
(https://xdevs.com/doc/Fluke/5720A/hulk/hza18_pap_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/hza18_pap.jpg)
Negative PA- -190VDC output. :-+
(https://xdevs.com/doc/Fluke/5720A/hulk/hza18_pan_1.jpg) (https://xdevs.com/doc/Fluke/5720A/hulk/hza18_pan.jpg)
Tomorrow last piece arrives, so I'll continue to check regulator rails and voltages, and once all is good, will fit all boards in and see if the Hulk chuch. :scared:
-
TiN,
Great job so far. I know you can't wait to run its first calibration. :popcorn:
-
Received my bad A8 today (good A8 I had before had to go as part of chassis trade :P) .
Hybrid ceramic assembly looks good, rest I can repair easily. :-DD
So excited, my hands are shaking! (better not when probing 565 volts on A18!) :scared:
-
saw the light.
such an epic.
-
:popcorn: Yes, all boards are in (except 30MHz WB option, that is next step).
(https://xdevs.com/doc/Fluke/5720A/hulk/test_hulk.jpg)
Sorry for ugly photo, just a quick shot, as I went to bed after 4am tonight. This thing sucking my life out.
Who can sleep anyway when there is 5720A in the room? :-DD
And after few days of mangling with random non-documented diagnostics errors and swapping my spare cards, I finally have Hulk running.
Finished ACAL last (this) night, and now it is running thru paces with HP3458B.
Next step is to fix my wideband option board set (it trips output to standby in 12kHz-1.2MHz range, but works fine in 1.2MHz-30MHz range), and wait for some parts to arrive, like metal covers and random bits.
(https://xdevs.com/doc/Fluke/5720A/hulk/hulk_calkit1.png)
Then I'll have it fully assembled and be on 30-day calibration interval versus my 10K/10V/1ohm/4920M standards, to check if there are any hidden rabbit holes awaiting.
-
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 (https://xdevs.com/doc/Fluke/5720A/cal/HLK_3458B_TEST1_AUG_1_2018_no_covers_5720A11.pdf). 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.
(https://xdevs.com/doc/Fluke/5720A/cal/hulk_dcl10v.png)
I'll need get my 720A in order to test better.
A good start for now, to be improved later on, hopefully. 8)
Running second more refined calkit sequence now, with better resistance data. :-DMM
(https://xdevs.com/doc/Fluke/5720A/cal/hulk_rmeas2_aug_2_2018.png)
-
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? ;D
(https://xdevs.com/doc/Fluke/8508A/img/f8508a_init_1.jpg) (https://xdevs.com/doc/Fluke/8508A/img/f8508a_init.jpg)
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 (https://xdevs.com/article/fsl935) or ESI SR104 (https://xdevs.com/review/sr104).
-
Very nice :-+
PS: I see a 4w connection but no 4w symbol on the 8508A display?
-
TiN, |O you're, you're a, you're a slummetrology millionaire |O
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. ;D :-+
-
Now TiN, it's time for conquering the world.
(https://images-mm.s3.amazonaws.com/Pinky_Brain_To_Do_List_Gray_Shirt.jpg)
-
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.
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.
For ACV we have more powerful contender - Wavetek 4920M. I'd like to see 8508A try to beat that one.
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.
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 :D), now that I have decent source.
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.
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. :D
I eagerly await the 8508A-teardown.
I'll await for that too, but it's not happening, xDevs labs ain't rich enough :D
1281 / 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":
(https://xdevs.com/doc/_Passives/vpg2018/vpgres2.jpg)
Initial datalog (https://xdevs.com/f8508a_comp1/).
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 (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. :scared: ;D
-
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 (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.
-
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 ;)
Philipp,
I still could not figure out, why and whom in particular you are indirectly calling a FLUKE or HP fanboy, especially as in this section, where metrology, not brands is in the key interest of people.
If you have me in mind, being an hp fanboy, I understand it even less.. also why you don't want to discuss that comparison further..
Or did you simply not understand, what I tried to demonstrate, that the hp 3458A due to its low tests currents (surely also due to its mediocre parallel measurement of high ohm resistance) by far can't be as good as the 8508A in its worst of three Ohm mode?
A pity that you skipped the discussion at that point.
Frank
-
Hi Frank,
we moved the discussion to the new thread quarks opened.
I already said there, that I read something between the lines which isn't there. Sorry for that!
And yes, I simply misunderstood you. For me it seems that you want to say something like "hey, this 3.5 digit meter is more or less like the 3458A, because you have to switch the 3458A to NDIG 3 to compare it." But that was my fault and not yours.
-
On my quest to make it real 5720A-level unit, I got some candy from RFparts.com (pleasure to deal with, I only wish their site would be less buggy :) )
(https://xdevs.com/doc/Fluke/5720A/za16new/rfparts_1.jpg)
Bought this new revision A16 some time ago, with missing and bodged out parts, including one of RF transistors, hence the photo above.
This is power amplifier assembly, that makes 220V ranges and related functions. I have older version A16 which works fine, but we not going to settle to second best.
(https://xdevs.com/doc/Fluke/5720A/za16new/hulk_a16n_1.jpg)
Somebody stole capacitors from it too. :)
But magical hybrid is all nice and intact, covered under plastic oven hood:
(https://xdevs.com/doc/Fluke/5720A/za16new/hulk_a16n_chopf_1.jpg)
So I went ahead, put all brand new missing parts back in and plug it into the Hulk with fingers crossed. (it was already 4am or so, so no patience to check voltage rails/etc. NOT recommended way of repair!)
(https://xdevs.com/doc/Fluke/5720A/za16new/hulk_a16_upgr_1.jpg)
No magic sparks or orange smoke escape, start self-test..... PASS.... start CalCheck procedure.... all good, yay?
Connected output to my beloved 4920M, set 30VAC...
(https://xdevs.com/doc/Fluke/5720A/za16new/hulk_30vace_1.jpg)
And no problem with output
(https://xdevs.com/doc/Fluke/5720A/za16new/hulk_30vac_1.jpg)
Did full calibration procedure afterwards, no problem. Max change was 1610% for 220VDC range (no wonder, assembly is different).
(https://xdevs.com/doc/Fluke/5720A/za16new/mfc_na16z_1.jpg)
One sleep-less night, and bingo, board is back to life. :)
After A11 REFDAC, this module is the largest revision change between legacy 5700A and updated 5720A units.
(https://xdevs.com/doc/Fluke/5720A/za16new/a16_updated_sch.png)
And here's calibration change report, for curious minds:
------------------------------------------------------------------------
FLUKE CORPORATION S/N :) 5720A OUTPUT CALIBRATION
PRINTED ON 8/16/18 AT 20:33:55 xDevs.com HULK, Initial CAL
------------------------------------------------------------------------
------------------------------------------------------------------------
MODULES PRESENT
------------------------------------------------------------------------
Software Revision 1.4+B+*
Switching Matrix
DC Volt Module
AC Volt Module
220V Module
1100V/2A Module
Current Module
Ohms Module
Hires Osc Module
Rear Panel
------------------------------------------------------------------------
ARTIFACT CALIBRATION DATES AND TEMPERATURES
------------------------------------------------------------------------
Specification Interval: 24hour Confidence: 99%`
MAIN OUTPUT: (Most recent) 8/15/18, 22.4 Degrees C
(Previous) 8/01/18, 27.0 Degrees C
(Days since last cal) 1
ZERO: (Most recent) 8/16/18
(Days since last cal) 0
------------------------------------------------------------------------
STANDARDS USED
------------------------------------------------------------------------
STANDARD NUMBER CERT. DATE
VOLTAGE _FX792X______ ____07/04_______ 9.999999 V
10 kOhm _FSL935_______ ___05/30_______ 9.999976 kOhm
1 Ohm __FSL935_______ ___05/30_______ 1.0000594 Ohm
------------------------------------------------------------------------
EXTERNALLY CALIBRATED INTERNAL REFERENCES
------------------------------------------------------------------------
REFERENCE Most Recent Previous Shift (ppm)
6.5 V DC Reference 6.957491 6.957486 0.66
13 V DC Reference 13.855310 13.855310 0.00
Resistance Reference 0.9999856 0.9999836 1.97
1.0 Ohm Resistor 0.9998467 0.9998039 42.80
1.9 Ohm Resistor 1.8999940 1.8998703 65.12
------------------------------------------------------------------------
DC VOLTAGE OUTPUT SHIFTS
------------------------------------------------------------------------
RANGE Point Zero Shift Full Scale Shift Spec Shift
(+/-) (% spec)
220 mV +FS 0.00000 mV -0.00014 mV -0.65 ppm 7.27 ppm -8.98
-FS 0.00000 mV 0.00019 mV 0.87 ppm 7.27 ppm 11.98
2.2V +FS 0.0000000 V 0.0000018 V 0.80 ppm 3.86 ppm 20.75
-FS -0.0000000 V -0.0000013 V -0.58 ppm 3.86 ppm -15.09
11V +FS 0.000000 V 0.000004 V 0.40 ppm 2.77 ppm 14.31
-FS 0.000000 V -0.000002 V -0.18 ppm 2.77 ppm -6.42
22V +FS -0.000002 V -0.000002 V -0.08 ppm 2.73 ppm -2.92
-FS 0.000002 V -0.000003 V -0.13 ppm 2.73 ppm -4.73
220V +FS 0.00007 V -0.01312 V -59.63 ppm 3.73 ppm-1599.83
-FS 0.00007 V 0.01320 V 60.02 ppm 3.73 ppm 1610.18
1100V +FS 0.0001 V -0.0004 V -0.36 ppm 5.45 ppm -6.52
-FS 0.0001 V 0.0003 V 0.31 ppm 5.45 ppm 5.70
------------------------------------------------------------------------
AC VOLTAGE OUTPUT SHIFTS
------------------------------------------------------------------------
RANGE Point Full Scale Shift Spec Shift
(+/-) (% spec)
2.2 mV FS -0.000029 mV -0.0013 % 0.24 % -0.57
20.00 kHz -0.000025 mV -0.0011 % 0.24 % -0.48
50.00 kHz -0.000032 mV -0.0014 % 0.25 % -0.58
100.00 kHz -0.000029 mV -0.0013 % 0.32 % -0.40
119.99 kHz -0.000024 mV -0.0011 % 0.65 % -0.17
0.1200 MHz -0.000013 mV -0.0006 % 0.65 % -0.09
0.2000 MHz -0.000003 mV -0.0002 % 0.65 % -0.02
0.3000 MHz 0.000029 mV 0.0013 % 0.65 % 0.21
0.4000 MHz 0.000066 mV 0.0030 % 1.28 % 0.23
0.5000 MHz 0.000113 mV 0.0051 % 1.28 % 0.40
0.6000 MHz 0.000180 mV 0.0082 % 1.43 % 0.57
0.7000 MHz 0.000253 mV 0.0115 % 1.43 % 0.81
0.8000 MHz 0.000348 mV 0.0158 % 1.43 % 1.11
0.9000 MHz 0.000442 mV 0.0201 % 1.43 % 1.41
1.0000 MHz 0.000587 mV 0.0267 % 1.43 % 1.87
1.1000 MHz 0.000734 mV 0.0334 % NO SPEC 0.00
1.1999 MHz 0.000845 mV 0.0384 % NO SPEC 0.00
22 mV FS -0.00025 mV -11.48 ppm 312.27 ppm -3.68
20.00 kHz -0.00029 mV -13.14 ppm 312.27 ppm -4.21
50.00 kHz -0.00023 mV -10.24 ppm 447.27 ppm -2.29
100.00 kHz -0.00018 mV -8.41 ppm 772.73 ppm -1.09
119.99 kHz -0.00014 mV -0.0006 % 0.15 % -0.41
0.1200 MHz -0.00019 mV -0.0009 % 0.15 % -0.57
0.2000 MHz 0.00011 mV 0.0005 % 0.15 % 0.32
0.3000 MHz 0.00104 mV 0.0047 % 0.15 % 3.07
0.4000 MHz 0.00082 mV 0.0037 % 0.25 % 1.47
0.5000 MHz 0.00136 mV 0.0062 % 0.25 % 2.44
0.6000 MHz 0.00216 mV 0.0098 % 0.40 % 2.44
0.7000 MHz 0.00301 mV 0.0137 % 0.40 % 3.38
0.8000 MHz 0.00410 mV 0.0186 % 0.40 % 4.62
0.9000 MHz 0.00544 mV 0.0247 % 0.40 % 6.12
1.0000 MHz 0.00678 mV 0.0308 % 0.40 % 7.63
1.1000 MHz 0.00956 mV 0.0435 % NO SPEC 0.00
1.1999 MHz 0.01088 mV 0.0494 % NO SPEC 0.00
220 mV FS -0.0009 mV -4.07 ppm 121.36 ppm -3.35
20.00 kHz -0.0004 mV -2.03 ppm 121.36 ppm -1.67
50.00 kHz -0.0011 mV -5.14 ppm 256.36 ppm -2.00
100.00 kHz -0.0008 mV -3.71 ppm 590.91 ppm -0.63
119.99 kHz -0.0004 mV -1.65 ppm 963.64 ppm -0.17
0.1200 MHz 0.0007 mV 3.36 ppm 963.64 ppm 0.35
0.2000 MHz 0.0017 mV 7.77 ppm 963.64 ppm 0.81
0.3000 MHz 0.0050 mV 22.61 ppm 963.64 ppm 2.35
0.4000 MHz 0.0086 mV 0.0039 % 0.15 % 2.55
0.5000 MHz 0.0133 mV 0.0061 % 0.15 % 3.94
0.6000 MHz 0.0200 mV 0.0091 % 0.30 % 3.06
0.7000 MHz 0.0274 mV 0.0124 % 0.30 % 4.18
0.8000 MHz 0.0368 mV 0.0167 % 0.30 % 5.63
0.9000 MHz 0.0462 mV 0.0210 % 0.30 % 7.07
1.0000 MHz 0.0607 mV 0.0276 % 0.30 % 9.28
1.1000 MHz 0.0755 mV 0.0343 % NO SPEC 0.00
1.1999 MHz 0.0866 mV 0.0394 % NO SPEC 0.00
2.2V FS -0.000006 V -2.62 ppm 49.55 ppm -5.29
20.00 kHz -0.000001 V -0.58 ppm 49.55 ppm -1.17
50.00 kHz -0.000008 V -3.69 ppm 85.45 ppm -4.32
100.00 kHz -0.000005 V -2.27 ppm 138.18 ppm -1.64
119.99 kHz -0.000000 V -0.20 ppm 425.45 ppm -0.05
0.1200 MHz 0.000011 V 4.80 ppm 425.45 ppm 1.13
0.2000 MHz 0.000020 V 9.22 ppm 425.45 ppm 2.17
0.3000 MHz 0.000053 V 24.06 ppm 425.45 ppm 5.65
0.4000 MHz 0.000089 V 0.0041 % 0.11 % 3.64
0.5000 MHz 0.000136 V 0.0062 % 0.11 % 5.57
0.6000 MHz 0.000204 V 0.0093 % 0.18 % 5.19
0.7000 MHz 0.000277 V 0.0126 % 0.18 % 7.06
0.8000 MHz 0.000372 V 0.0169 % 0.18 % 9.48
0.9000 MHz 0.000466 V 0.0212 % 0.18 % 11.88
1.0000 MHz 0.000610 V 0.0277 % 0.18 % 15.57
1.1000 MHz 0.000758 V 0.0345 % NO SPEC 0.00
1.1999 MHz 0.000869 V 0.0395 % NO SPEC 0.00
22V FS -0.00002 V -0.72 ppm 48.18 ppm -1.49
20.00 kHz -0.00005 V -2.38 ppm 48.18 ppm -4.94
50.00 kHz 0.00001 V 0.52 ppm 85.45 ppm 0.61
100.00 kHz 0.00005 V 2.36 ppm 121.36 ppm 1.94
119.99 kHz 0.00010 V 4.43 ppm 336.36 ppm 1.32
0.1200 MHz 0.00004 V 1.96 ppm 336.36 ppm 0.58
0.2000 MHz 0.00034 V 15.65 ppm 336.36 ppm 4.65
0.3000 MHz 0.00128 V 58.17 ppm 336.36 ppm 17.29
0.4000 MHz 0.00106 V 0.0048 % 0.11 % 4.32
0.5000 MHz 0.00160 V 0.0073 % 0.11 % 6.53
0.6000 MHz 0.00240 V 0.0109 % 0.17 % 6.49
0.7000 MHz 0.00324 V 0.0147 % 0.17 % 8.76
0.8000 MHz 0.00434 V 0.0197 % 0.17 % 11.73
0.9000 MHz 0.00567 V 0.0258 % 0.17 % 15.33
1.0000 MHz 0.00701 V 0.0319 % 0.17 % 18.96
1.1000 MHz 0.00980 V 0.0445 % NO SPEC 0.00
1.1999 MHz 0.01112 V 0.0505 % NO SPEC 0.00
220V FS -0.0054 V -24.47 ppm 60.18 ppm -40.66
5.000 kHz -0.0045 V -20.32 ppm 60.18 ppm -33.77
10.000 kHz -0.0059 V -27.01 ppm 60.18 ppm -44.88
20.00 kHz -0.0068 V -30.93 ppm 60.18 ppm -51.39
50.00 kHz -0.0207 V -94.10 ppm 95.45 ppm -98.58
100.00 kHz -0.0831 V -377.95 ppm 173.64 ppm -217.67
119.99 kHz -0.1258 V -571.69 ppm 990.91 ppm -57.69
0.1200 MHz -0.1243 V -565.11 ppm 990.91 ppm -57.03
0.2000 MHz -0.3558 V -1617.40 ppm 990.91 ppm -163.22
0.3000 MHz -0.7557 V -3434.82 ppm 990.91 ppm -346.63
0.4000 MHz -1.1857 V -0.5389 % 0.52 % -103.10
0.5000 MHz -1.4540 V -0.6609 % 0.52 % -126.44
0.6000 MHz -0.9597 V -0.4362 % 0.85 % -51.60
0.7000 MHz -1.6055 V -0.7298 % 0.85 % -86.32
0.8000 MHz -2.3775 V -1.0807 % 0.85 % -127.82
0.9000 MHz -1.9068 V -0.8667 % 0.85 % -102.51
1.0000 MHz -1.3834 V -0.6288 % 0.85 % -74.37
1100V FS -0.003 V -2.46 ppm 73.64 ppm -3.34
0.3000 kHz 0.003 V 3.06 ppm 73.64 ppm 4.16
0.5000 kHz 0.000 V 0.27 ppm 73.64 ppm 0.37
0.7000 kHz 0.001 V 1.11 ppm 73.64 ppm 1.51
1.0000 kHz -0.001 V -0.68 ppm 73.64 ppm -0.92
------------------------------------------------------------------------
DC CURRENT OUTPUT SHIFTS
------------------------------------------------------------------------
RANGE Point Zero Shift Full Scale Shift Spec Shift
(+/-) (% spec)
220 uA +FS 0.0001 uA 0.0003 uA 1.18 ppm 71.82 ppm 1.64
-FS 0.0001 uA -0.0001 uA -0.60 ppm 71.82 ppm -0.84
2.2 mA +FS -0.000003 mA 0.000001 mA 0.35 ppm 33.64 ppm 1.03
-FS -0.000002 mA -0.000006 mA -2.92 ppm 33.64 ppm -8.69
22 mA +FS 0.00000 mA 0.00003 mA 1.56 ppm 32.27 ppm 4.83
-FS 0.00000 mA -0.00003 mA -1.57 ppm 32.27 ppm -4.86
220 mA +FS 0.0000 mA -0.0005 mA -2.07 ppm 53.32 ppm -3.88
-FS 0.0001 mA 0.0005 mA 2.44 ppm 53.32 ppm 4.58
2.2A +FS 0.000017 A -0.000311 A-141.40 ppm 115.22 ppm -122.73
-FS 0.000018 A 0.000346 A 157.26 ppm 115.22 ppm 136.49
------------------------------------------------------------------------
AC CURRENT OUTPUT SHIFTS
------------------------------------------------------------------------
RANGE Point Full Scale Shift Spec Shift
(+/-) (% spec)
220 uA FS 0.000 uA 0.15 ppm 165.45 ppm 0.09
0.5000 kHz -0.001 uA -4.00 ppm 165.45 ppm -2.42
1.0000 kHz -0.000 uA -0.55 ppm 165.45 ppm -0.33
2.000 kHz 0.000 uA 1.23 ppm 368.18 ppm 0.33
5.000 kHz -0.001 uA -3.23 ppm 368.18 ppm -0.88
7.000 kHz -0.001 uA -0.0006 % 0.14 % -0.47
10.000 kHz -0.002 uA -0.0010 % 0.14 % -0.70
2.2 mA FS 0.00000 mA 0.89 ppm 138.18 ppm 0.64
0.5000 kHz 0.00000 mA 1.80 ppm 138.18 ppm 1.30
1.0000 kHz 0.00000 mA 0.92 ppm 138.18 ppm 0.66
2.000 kHz 0.00000 mA 1.21 ppm 269.09 ppm 0.45
5.000 kHz 0.00000 mA 0.43 ppm 269.09 ppm 0.16
7.000 kHz -0.00000 mA -0.0001 % 0.14 % -0.09
10.000 kHz -0.00000 mA -0.0002 % 0.14 % -0.13
22 mA FS 0.0000 mA 0.82 ppm 138.18 ppm 0.59
0.5000 kHz -0.0000 mA -1.49 ppm 138.18 ppm -1.08
1.0000 kHz 0.0001 mA 2.57 ppm 138.18 ppm 1.86
2.000 kHz 0.0000 mA 0.83 ppm 241.82 ppm 0.34
5.000 kHz 0.0000 mA 0.81 ppm 241.82 ppm 0.33
7.000 kHz -0.0000 mA -0.0002 % 0.13 % -0.13
10.000 kHz -0.0001 mA -0.0003 % 0.13 % -0.24
220 mA FS -0.001 mA -3.00 ppm 133.64 ppm -2.25
0.5000 kHz -0.000 mA -1.74 ppm 133.64 ppm -1.30
1.0000 kHz -0.000 mA -0.52 ppm 133.64 ppm -0.39
2.000 kHz 0.000 mA 1.67 ppm 228.18 ppm 0.73
5.000 kHz -0.001 mA -5.36 ppm 228.18 ppm -2.35
7.000 kHz -0.000 mA -0.0001 % 0.11 % -0.13
10.000 kHz -0.002 mA -0.0008 % 0.11 % -0.79
2.2A FS -0.00033 A -150.08 ppm 308.18 ppm -48.70
0.5000 kHz -0.00034 A -152.39 ppm 308.18 ppm -49.45
1.0000 kHz -0.00033 A -148.33 ppm 308.18 ppm -48.13
2.000 kHz -0.00033 A -150.07 ppm 485.45 ppm -30.91
5.000 kHz -0.00033 A -150.09 ppm 485.45 ppm -30.92
7.000 kHz -0.00034 A -0.0153 % 0.61 % -2.51
10.000 kHz -0.00034 A -0.0154 % 0.61 % -2.53
------------------------------------------------------------------------
RESISTANCE OUTPUT SHIFTS
------------------------------------------------------------------------
Point Full Scale Shift Spec Shift
(+/-) (% spec)
0.0000000 Ohm 0.0000000 Ohm 0.00 Ohm 0.00 Ohm 0.00
0.9998039 Ohm 0.0000428 Ohm 42.80 ppm 85.00 ppm 50.36
1.8998703 Ohm 0.0001237 Ohm 65.12 ppm 85.00 ppm 76.61
10.000290 Ohm 0.000037 Ohm 3.73 ppm 23.00 ppm 16.24
18.999925 Ohm 0.000051 Ohm 2.71 ppm 23.00 ppm 11.77
100.00301 Ohm 0.00016 Ohm 1.62 ppm 10.00 ppm 16.18
189.99756 Ohm 0.00049 Ohm 2.59 ppm 10.00 ppm 25.91
1.0000085 kOhm 0.0000019 kOhm 1.86 ppm 8.00 ppm 23.19
1.9000219 kOhm 0.0000035 kOhm 1.82 ppm 8.00 ppm 22.70
9.999776 kOhm 0.000015 kOhm 1.54 ppm 8.00 ppm 19.31
18.999374 kOhm 0.000030 kOhm 1.58 ppm 9.00 ppm 17.57
99.99470 kOhm 0.00012 kOhm 1.16 ppm 9.00 ppm 12.92
189.98859 kOhm 0.00068 kOhm 3.59 ppm 9.00 ppm 39.93
0.9999808 MOhm 0.0000017 MOhm 1.73 ppm 16.00 ppm 10.84
1.8999653 MOhm 0.0000062 MOhm 3.28 ppm 17.00 ppm 19.31
9.998990 MOhm 0.000051 MOhm 5.08 ppm 33.00 ppm 15.39
18.998386 MOhm 0.000109 MOhm 5.76 ppm 43.00 ppm 13.38
100.00693 MOhm 0.00184 MOhm 18.41 ppm 100.00 ppm 18.41
-
Looking good.
What is the IET Labs reference resistor in the back?
-
Awesome project, thanks for sharing!
Max
-
Got some news from a bird. Missing chassis parts are manufactured and will be on their way to me soon. So excited!
(https://xdevs.com/doc/Fluke/5720A/hulk/chassis_new_parts.jpg).
These are internal shields for both inguard and outguard sections and exterior top and bottom covers. Yes, I ordered exterior covers with black powder coating. I think black will look much better on my 5720 Hulk edition and match my SL935 and 792X standards :-DD.
I also could not resist getting second set of resistance boards (pun intended).
(https://xdevs.com/doc/Fluke/5720A/a9a10_second/sohms_a9c_1.jpg) (https://xdevs.com/doc/Fluke/5720A/a9a10_second/sohms_a9c.jpg) (https://xdevs.com/doc/Fluke/5720A/a9a10_second/sohms_a10c_1.jpg) (https://xdevs.com/doc/Fluke/5720A/a9a10_second/sohms_a10c.jpg)
Second A9 OHM CAL (that contains calibration circuitry, reference 10K resistor hermetic network and 1.0, 1.9 ohm resistors and 4-wire short) is also 5720A version. Second A10 (working ohms PCBA) is older 5700A version (on photo it is the one with white COTO relays), which have little bit different networks and older CADDOCK 90Meg resistor.
Will be interesting to compare both boards stability, as one set will be just sitting on shelf, while second set to operate in calibrator 24/7/365. :)
-
Is it worth buying a Fluke 5720A instead of a cheaper (and easier to get) 5700A at work? I'm especially interested in DC Volts up to 1000V.
Do you have some experience about the real differences and not only data sheet values?
-
I'd go for 5720A, it have multiple circuit updates for better reliability. Also DC linearity of 5720A is better.
High frequency AC is major improvement as well, but I did not get to testing that part yet.
Also recent 5700A are likely have most of hardware at 5720A level anyway, it's not that you can get replacement boards for 5700A anyway (by official service).
There are also 5700A series II, and 5700A/EP units, and 5700A/AF units, and few more obscure ones :)
-
Hulk covers need to be green... otherwise Hulk gets angry, and you don't want to see Hulk get angry. :-DD
-
My favourite delivery express guy made a visit and left me some goodies. :-+
From CAD drawing to reality, just like that...
(https://xdevs.com/doc/Fluke/5720A/armor/armor_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/armor.jpg)
Let's see how many dimensions I got wrong :D.
(https://xdevs.com/doc/Fluke/5720A/armor/inguard_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/inguard.jpg)
I had to make inner shields for both inguard floating cage and for outguard controller side.
(https://xdevs.com/doc/Fluke/5720A/armor/topcb_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/topcb.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/topcvr_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/topcvr.jpg)
Almost "Before" and "after" photo, with only difference that after is not a same part, but a DIY replacement :). Despite the desire of texaspyro and others, it is not green :-DD
(https://xdevs.com/doc/Fluke/5720A/armor/hulkbolt_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/hulkbolt.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/hulkins_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/hulkins.jpg)
I also ended up ordering custom CNC brass inserts, for captive screw mounts for inguard shield, so it would mimic original Fluke design closely. :) Those ended up a nice fit!
(https://xdevs.com/doc/Fluke/5720A/armor/hulkin_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/hulkin.jpg)
But not so much for A14 PCB notch cutouts! They are off by few mm, I guess somebody in xDevs chat got me distracted (https://xdevs.com/chat).
Dremel to the rescue, nothing can stop some shhrrr-shhrrr with cutting wheel. Not even angry neighbors next-door. :scared:
(https://xdevs.com/doc/Fluke/5720A/armor/hulkalive_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/hulkalive.jpg)
She'll be alright, like we used to say here at EEVBlog! :)
(https://xdevs.com/doc/Fluke/5720A/armor/hulkoff_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/hulkoff.jpg)
There are still some little tweaks to do, but so far I'm very happy with the outcome.
(https://xdevs.com/doc/Fluke/5720A/armor/armored_1.jpg) (https://xdevs.com/doc/Fluke/5720A/armor/armored.jpg)
In fact, calibrator already doing it's job, I just recently used it to calibrate all functions of freshly repaired 8842A, but that is a story of the other thread. :popcorn:
-
Looks great. :-+
You can still get some decals for it. https://www.etsy.com/market/hulk_decal (https://www.etsy.com/market/hulk_decal)
-
Definitely a 5720B ...
-
Respect, and you know exactely what parts are inside. Knowing what is inside the device is much more important than the "external" stuff.
-
I need to upgrade just one high voltage board to be 100% 5720-level on hardware. :)
Maybe I'll do something with top plastic frame to make it more "black" :-DD
Quick experiment with cloth tape:
(https://xdevs.com/doc/Fluke/5720A/armor/hulkblack_1.jpg)
-
How is terminal shield fixed onto front panel, by screw?
通过我的 PRA-AL00 上的 Tapatalk发言
-
No, it's insert into GUARD binding post. You can see screw that holds banana terminal inside.
-
Thanks to lukier for kick in the butt, I got around to run some python snakes to measure (or should I better say, estimate?) linearity errors of the above mentioned Hulk calibrator.
So here are some pretty charts, impressed even me... Can't be that good. :-DD
Method is rather simple, python application programs DC voltage of the Hulk output from -10.9 to +10.9, with range locked to 11V. Steps in programming are 0.1V, for good resolution. Calibrator ran internal zero before test.
As detector we use ADC with best linearity possible : 3458A. Actually two of them, connected to Hulk output, using Belden cable and bare copper spade lugs.
(https://xdevs.com/doc/Fluke/5720A/cal/test_hulk_3458ab_10v_noac.png)
Then collected data points are calculated for deviation from best-fit polynom, and error calculated in ppms from 10VDC scale. Numpy code to do that:
p = np.polyfit(real,ideal,1) # Fit polynom to data
pv = np.polyval(p,real) # Calculate deviation from polynom
diff = ideal-pv # Determine delta
diff_ppm = (diff/10)*1000000.0 # Calculate scale in ppms
Some comments for less verse readers:
* Thin blue line is meter 3458A (GPIB 3)
* Thin pink line is second meter 3458B (GPIB 2)
* Thick brown line is INL data of 3458B minus INL data of 3458A. This essentially to remove INL of the calibrator source itself, to get difference between two 3458As.
* Orangle line is bonus, Fluke 8508A versus Hulk calibrator. Settings RESL8, FAST_OFF, equivalent to about NPLC1040. :blah:
Differential INL between 3458A is quite cool +0.09 to -0.06 ppm peak to peak. Matching the 0.1 ppm specification of the 3458A well, for what it's worth. :o
Then I turned off aircon in calroom, and let ambient temperature to climb +29C. Idea is that it may help INL data, by removing drafts and temperature variations?
(https://xdevs.com/doc/Fluke/5720A/cal/test_hulk_3458ab_2002_10v_noac_1.png) (https://xdevs.com/doc/Fluke/5720A/cal/test_hulk_3458ab_2002_10v_noac.png)
This test still running (Beware, full sweep takes multiple hours, so not 100% correct way, as best INL specified only for minutes/tens minutes periods, not hours), that's why data not finished.
* Thin blue line is meter 3458A (GPIB 3)
* Thin pink line is second meter 3458B (GPIB 2)
* Thick brown line is INL data of 3458B minus INL data of 3458A. This essentially to remove INL of the calibrator source itself, to get difference between two 3458As.
* Orangle line is bonus, Fluke 8508A versus Hulk calibrator. Settings RESL7 this time, FAST_OFF, equivalent to about NPLC104. :blah:
* This graph also have Keithley 2002 meter added for giggles. NPLC50, DFILT OFF.
* Bold magenta and violet lines are difference of K2002 from either of 3458s.
As sanity check, I got remote access to <90day calibrated Fluke 5730 in far far away galaxy. So I ran very same code on 5730A and 3458A (another one, not mine) so we can glance over the data from this setup as well.
10V , NPLC50. Same configuration of the calibrator as my 5720A Hulk.
(https://xdevs.com/doc/Fluke/5730A/test_5730_3458_10v_nplc50_1.png) (https://xdevs.com/doc/Fluke/5730A/test_5730_3458_10v_nplc50.png)
1V range, NPLC 100. Rangelock at 2.2V range.. Otherwise same configuration.
(https://xdevs.com/doc/Fluke/5730A/test_5730_3458_1v_nplc100_1.png) (https://xdevs.com/doc/Fluke/5730A/test_5730_3458_1v_nplc100.png)
Comments? :)
-
Comments? :)
Mmm ... , well ..., it seems OK, Right?
-
Lookin good!
You should compare tests with Fluke 720A, as well
-
It still looks like there is a small jump at around 0 V, when the 5720 reverses polarity. AFAIK this is done with relays, so that except for a possible jump at around zero the 5720 should be to a large part be symmetric. A possible additional test could be with the voltmeter(s) in revers connection.
So the calibrator looks really good - except for that small jump around 0. With a careful measurement, one might be able to correct some of this in software.
As the run seems to take quite some time, there might be a temperature effect too. To see this a second run under similar conditions could help. Its kind of finding a compromise between speed and noise.
-
Yes, polarity reversal in 57xx done by relay on the output of the DAC.
I don't sweat much about that jump, it's 0.2ppm after all. Own noise of the DMM's reference is that much anyway. It is not that often that you need transfer negative voltage to positive voltage.
A possible additional test could be with the voltmeter(s) in revers connection.
Interesting idea, i'll try. Also will add 3458C into mix (third meter that have 2ppm/hour drifty ADC), curious to see if drifty ADC also have bad linearity.
As the run seems to take quite some time, there might be a temperature effect too. To see this a second run under similar conditions could help. Its kind of finding a compromise between speed and noise.
I did fast test with same rig but using NPLC10 instead of NPLC100. 8508A in RESL6 mode, and its linearity goes to toilet in this mode, unlike 3458A@NPLC10.
(https://xdevs.com/doc/Fluke/5720A/cal/test_hulk11rng_3458ab_2002_10v_nplc10_1.png) (https://xdevs.com/doc/Fluke/5720A/cal/test_hulk11rng_3458ab_2002_10v_nplc10.png)
-
I did fast test with same rig but using NPLC10 instead of NPLC100. 8508A in RESL6 mode, and its linearity goes to toilet in this mode, unlike 3458A@NPLC10.
I think that is a quantization problem. If you select 6 digits on the 8508A you will get only 6 digits instead of the 3458A which gives always a billion digits over GPIB no matter how it is configured.
-
Quantization or not, as result you can't get good linearity out of 8508A without making it slowest meter on the planet ;).
-
Quantization or not, as result you can't get good linearity out of 8508A without making it slowest meter on the planet ;).
Yes, both meters have pros and cons. And speed is definitely a point for the 3458A :)
-
I'd like to point to similar linearity measurements, I've done, 3458A vs. 5442A:
(https://www.eevblog.com/forum/index.php?action=dlattach;topic=43282.0;attach=206836;image)
I made 8 runs, always one direction on the 5442A, say from -11V to +11V, but also with reversed leads, so to measure the positive voltages of the 5442A as negative inputs on the 3458A, and vv., and afterwards mirroring the results in xls.
You see, due to this delicate measurement, you'll always get different results, mostly shifted values, but sometimes also qualitatively different.
So to really make good statements, maybe also quantitatively, it is required to make several runs, best would be circular ones (-11.. +11 .. -11 ...).
These singular measurements with 3458A, 5720 and 8508 may not reveal the whole story, I expect all instruments to perform a bit better in reality.
Especially the 3458A vs. 3458A should improve.
Frank
-
Xmas is early this year at xDevs.com labs...
Thanks to international support and USPS services, I am busy again, building something dangerous... Hulk deserves some more juice.... :scared: :bullshit:
(https://xdevs.com/doc/Fluke/5725A/img/frame.jpg)
-
DCC?
-
Transconductance Amplifier :scared:
-
Yes, it is. I had thing shipped in bits and pieces with 3 boxes, due to international weight and shipping limitations. Don't ask how much that cost..
(https://xdevs.com/doc/Fluke/5725A/img/goodies_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/goodies.jpg)
Pile of parts, eh?
(https://xdevs.com/doc/Fluke/5725A/img/pcbs_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/pcbs.jpg)
PCBs set in a way how they are in the unit.
Key metrology component in the 5725A booster is High voltage sense assembly:
(https://xdevs.com/doc/Fluke/5725A/img/hvsense_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/hvsense.jpg)
I was lucky to get one of later versions of this module in the unit, perhaps replaced during service sometime around year 2002 in the life of the box, as rest of the boards are dated 1993. This board have Fluke Thermal Sensor chip for AC/DC transfers, just like AC function boards in 57xx/792A and 5790A. Attenuator for 1100V AC and DC range is custom Fluke hermetic resistor network hiding under PCB block on top middle section.
How is all together in one piece again, so Digikey parts ordered to proceed with further repairs/troubleshooting as required. All electrolytic (except four huge 50000uF 20V caps, don't have those in stock) and tantalum capacitors, all carbon composition resistors will be replaced without questioning first, as preventive measure.
(https://xdevs.com/doc/Fluke/5725A/img/gtfo_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/gtfo.jpg)
Quick power on check didn't cause any fires or smokes, and Hulk happily reported fan problem and locked down all controls. Good start I guess.
In case you wonder what's wrong with the fan:
(https://xdevs.com/doc/Fluke/5725A/img/deadfanb_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/deadfanb.jpg)
-
Is that a mini-version of the Guildline MightyMux-scanner on the ip-cam?
-
No, it is not. :popcorn:
-
1529-R Chub-E4 Thermometer Readout......
-
The fan looks like a low air restriction model!
-
Bob's your uncle! :-DD
Repair concluded that....there was no repair needed... :wtf:
My intuition didn't fail me this time too, and building DIY interconnect cable between Hulk and Booster was a right move.
(https://xdevs.com/doc/Fluke/5725A/img/diycable_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/diycable.jpg)
It may not look much, but I did my best educated guess, provide guaded low-noise Keithley PTFE coax for analog signals, low triboelectric noise Gore coax for calibration signals, guaded PTFE twinax for current lines (MFC can move all current ranges to booster output terminals, not just boosted 11A range) and multiwire digital shielded cable for data/comms/isolated power.
(https://xdevs.com/doc/Fluke/5725A/img/booster_current_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/booster_current.jpg) (https://xdevs.com/doc/Fluke/5725A/img/booster_voltage_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/booster_voltage.jpg)
Now my little homelab capability extended to 11ADC/11AAC 10kHz and 1100VAC 30kHz/750VAC 100kHz.
Time to cook some DUT's :-DMM :-BROKE :scared: :bullshit:
P.S. another progress is that I managed not to kill myself working on this. These boxes are very well capable of lethal shocks, with high-power 500V present on heatsinks and exposed connectors! :bullshit: Time to go for some sleep, 3:30 am here..
-
Bob's your uncle! :-DD
Repair concluded that....there was no repair needed... :wtf:
My intuition didn't fail me this time too, and building DIY interconnect cable between Hulk and Booster was a right move.
(https://xdevs.com/doc/Fluke/5725A/img/diycable_1.jpg) (https://xdevs.com/doc/Fluke/5725A/img/diycable.jpg)
That bright green power cable is remarkable
-
Calibration check of 5725A before ACAL:
+ ------------------------------------------------------------------------
+ ARTIFACT CALIBRATION DATES AND TEMPERATURES
+ ------------------------------------------------------------------------
+ Specification Interval: 24hour Confidence: 99%`
+
+ MAIN OUTPUT: (Most recent) 10/19/18, 22.6 Degrees C
+ (Previous) 10/17/18, 24.0 Degrees C
+ (Days since last cal) 48
+
+ ZERO: (Most recent) 12/06/18
+ (Days since last cal) 0
+
+ 5725: (Most recent) 5/03/16, 23.0 Degrees C
+ (Previous) 4/28/16, 23.0 Degrees C
+ (Days since last cal) 946
It was last calibrated in 2016, some 946 days ago.
------------------------------------------------------------------------
5725 AC VOLTAGE OUTPUT SHIFTS
------------------------------------------------------------------------
RANGE Point Full Scale Shift Spec Shift
(+/-) (% spec)
5725 FS 0.004 V 3.19 ppm 78.64 ppm 4.06
1100V 1.0000 kHz 0.007 V 6.63 ppm 78.64 ppm 8.44
1100V 5.000 kHz 0.003 V 2.67 ppm 110.45 ppm 2.42
1100V 10.000 kHz 0.003 V 2.47 ppm 110.45 ppm 2.23
1100V 12.00 kHz 0.001 V 0.54 ppm 110.45 ppm 0.48
1100V 15.00 kHz 0.005 V 4.39 ppm 110.45 ppm 3.98
1100V 17.00 kHz 0.005 V 4.21 ppm 110.45 ppm 3.81
1100V 20.00 kHz 0.022 V 19.86 ppm 110.45 ppm 17.98
1100V 30.00 kHz 0.036 V 32.43 ppm 240.00 ppm 13.51
750V 30.01 kHz 0.011 V 14.80 ppm 244.67 ppm 6.05
750V 40.00 kHz 0.032 V 42.14 ppm 244.67 ppm 17.22
750V 50.00 kHz 0.027 V 35.63 ppm 244.67 ppm 14.56
750V 60.00 kHz 0.030 V 39.44 ppm 660.00 ppm 5.98
750V 70.00 kHz 0.049 V 65.86 ppm 660.00 ppm 9.98
750V 80.00 kHz 0.065 V 86.42 ppm 660.00 ppm 13.09
750V 90.00 kHz 0.100 V 133.59 ppm 660.00 ppm 20.24
750V 100.00 kHz 0.133 V 177.48 ppm 660.00 ppm 26.89
------------------------------------------------------------------------
5725 DC CURRENT OUTPUT SHIFTS
------------------------------------------------------------------------
RANGE Point Zero Shift Full Scale Shift Spec Shift
(+/-) (% spec)
5725 +FS 0.00003 A -0.00055 A -50.26 ppm 372.73 ppm -13.48
-FS 0.00003 A 0.00061 A 55.25 ppm 372.73 ppm 14.82
------------------------------------------------------------------------
5725 AC CURRENT OUTPUT SHIFTS
------------------------------------------------------------------------
RANGE Point Full Scale Shift Spec Shift
(+/-) (% spec)
5725 1.0000 kHz -0.0005 A -46.97 ppm 385.45 ppm -12.19
Worst deviation is 27% of the spec at 750V 100kHz versus 24 hour spec? I'll take that as a win. :-+
-
Call me nuts, but I think the M6 earthquake that happened yesterday noon (April 18, 2019) killed resistor in my 5720A. :wtf: :-BROKE
(https://xdevs.com/doc/Fluke/5720A/a10/wtfz1a_2.jpg) (https://xdevs.com/doc/Fluke/5720A/a10/wtfz1a.jpg) (https://xdevs.com/doc/Fluke/5720A/a10/wtfz1b_2.jpg) (https://xdevs.com/doc/Fluke/5720A/a10/wtfz1b.jpg)
I have no other reasonable explanation to this. Nobody touched the unit, it was just sitting, happily running tests since full assembly and calibration 68 days ago.
This network have 10,90,900,19,171,1710 ohms and provide main resistance and current output functionality of the calibrator, among two other hermetical film Fluke resistors next to it.
Enough to say, there is no off-the-shelf replacement for this part :'(
-
That looks awful, you have my deepest sympathy.
Probably the array had been soldered under some tension, which now unloaded due to the strong vibrations.
Hope, that you find replacement part.
Frank
-
Funny enough, I have actually 10 of these networks in my ovenized resistance standard Fluke SL935 prototype.
Their 10 ohm elements are all connected in parallel form 1 ohm.
(https://xdevs.com/doc/Fluke/SL935/flukerc_1.jpg) (https://xdevs.com/article/fsl935/)
But I'm not going to vandalize calibrated standard for the resistor. :-\
-
Well hopefully you can find another, that really is a shame after all the work you've put into it.
-
Good thing the earthquake didn't break bigger things, like your house! :scared:
-
Sorry to hear that. It could be also your recent lab rearrangment. (or maybe both: earthquake and lab rearrangment).
I was only at first floor of some mall when the earthquake happened. Just on first floor I knew it was strong, so much more at higher floors like your place. Some earthquakes make building to swing (horizontally), other earthquakes make building to shake both horizontally and vertically (those are really scary and could make stuff jump). I felt one of those with jumps 4 years ago, and I was only at a 4th floor. If calibrator had another device stacked then the effect is much worse.
While there are some 'seismic rack' manufacturers in taiwan, I don't think it will be easy to buy, nor cheap. I have found Ikea 'Metod' kitchen cabinets to fit my needs. The metod cabinets also have pull out shelves, so it is easy to take device out and work on rear side of device.
But of course, I don't have a heavyweight Hulk monster at home, so don't know it will be helpful to you at all.
-
Can you buy it as a spare part from Fluke? I live down the road from them so i'm happy to swing up to Everett, pick it up and post to you if needed.
TonyG
-
Can you buy it as a spare part from Fluke?
I can't even buy innocent parts like plastic handle for 57xxA from Fluke, they are not willing to even reply with a quote a DIY enthusiast like me. Leave alone parts for component-level repair, even though this resistor network is also used unchanged on modern 5730A. But thank you for support :-+
ramon
Rearrangement happen postmortem, after I discovered calibrator is dead and needed access to guts, so remove all parts. Before quake everything was working perfectly fine, it passed my weekly calibration check on all ranges/functions just that day morning.
-
Did you try to contact the local partner? In germany they work togehter with CalPlus GmbH and many of my contacts to Fluke Calibration are going through CalPlus.
-
I did a livestream on A3 board troubleshooting on Hulk-2 project. Beware, 4+ hours long of boring stuff. :-/O
https://youtu.be/LTO8yIvzz9M (https://youtu.be/LTO8yIvzz9M)
-
Does the broken resistor belong to A11 module? If so, there is one A11 module on ebay (https://www.ebay.de/itm/Fluke-5700A-5720A-A11-Dac-Module/264247032272).
-branadic-
-
No, it's ohms assembly, A10. Replacement board on it's way. :scared:
Meanwhile I'm fixing A4 PCB for second calibrator project.
Board have internal short on layer 3 and layer 4 (6 layer PCB) between 900VAC and 44VAC winding circuits due to PCB burndown.
(https://xdevs.com/doc/Fluke/5720B/a4/a4cave_1.jpg)
Had to carve a well to get access and cut conductive FR4 charr out.
(https://xdevs.com/doc/Fluke/5720B/a4/a4cave_2.jpg)
Should I create new thread about this 5700 "B" build or keep posting in this one?
-
Keep it here. When someone comes back years from now they don't have to search for part two.
-
Okay, we can reference first unit as H1 and second as H2 then.
Some work on A4, restoring layer 3 and layer 4 damages.
Mains switch wiring after cutting out charred board patch:
(https://xdevs.com/doc/Fluke/5720B/a4/a4_mains_done_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/a4_mains_done.jpg)
And cave work in two spots where traces and charred PCB were milled out out:
(https://xdevs.com/doc/Fluke/5720B/a4/a4_cave1b_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/a4_cave1b.jpg)
(https://xdevs.com/doc/Fluke/5720B/a4/a4_cave2b_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/a4_cave2b.jpg)
Once done will fill all with epoxy and solder copper foil on top to restore shielding.
-
Now that is some serious dedication! :-+
-
Nice surgery! Be sure to check the insulation resistance
-
Digikey parts arrived today. Among others I've ordered single-component heat-cure epoxy to fill the carved parts of PCB material. It worked quite nicely for first ever try. :)
(https://xdevs.com/doc/Fluke/5720B/a4/epoxy_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/epoxy.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/a4_epox1_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/a4_epox1.jpg)
Measured resistance between affected windings >1GOhm, no issue there.
Also some other random parts, including brand new Sunon Maglev 92mm fans. I want to make this second calibrator quieter than first one, since it will not have high voltage A14/A15 boards I guess it could run lower speed fan as well.
(https://xdevs.com/doc/Fluke/5720B/a4/a4_bp_conn_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/a4_bp_conn.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/h2_fans_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a4/h2_fans.jpg)
-
Never tried one part epoxy, how was it to work with (thick or runny)? Was there any hardening before applying of heat to cure?
-
It's like icecream. I squeezed a bit of it to the spot and used toothpick to fill all the gaps and voids. It's not runny, rather easy to apply without going all over the place. Since it specified to have 2 year shelf life at room temp (22C) there is no hardening without heat. I've cooked board area at about 90-100°C for 30 minutes. It got semitransparent after cure, and hardness of soft type plastic, like ABS. Photo above shows state before cure.
I've tested board after epoxy at 1000VDC, leakage about few tens nanoamps between affected signals. There was no guarding or shielding during test, so I'd say no problems for this particular application.
-
I want to make this second calibrator quieter than first one, since it will not have high voltage A14/A15 boards I guess it could run lower speed fan as well.
It looks the 2nd one is 732C + 1 \$\Omega\$ & 10k \$\Omega\$ standard :popcorn:
-
Sudden saturday morning livestream got some progress on "Hulk-2" 5700A calibrator.
https://www.youtube.com/watch?v=49nSJwYZszc (https://www.youtube.com/watch?v=49nSJwYZszc)
I've got it all together to do first calibration attempts. Of course, there should be many hidden issues in this unit, so I don't expect that calibration to stay very long.
Also for more fun, I've used first 5720A calibrator as calibration source for today.
(https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_cal_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_cal.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_caldates_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_caldates.jpg)
Also upgraded firmware to 5700EP version, which brings GPIB command level to 5720A. This is helpful for me, so I can keep using same python app, developed mostly for 5720A to do more tests in future.
Had to format calibration EEPROM, which is not a problem. Original data in EEPROM is for some other 5700A anyway, not this random mix of board.
(https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_caldone_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_caldone.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_dcv10_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_dcv10.jpg)
Calibration value shift is only -2190598% off the 24-hour calibrator spec.
No worries, she'll be alright. :-DD
Keithley 2002 will be used as a check standard here.
And we got long desired 10V out of the box. Only +0.2ppm off K2002 reading (which is calibrated last year July).
(https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res10_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res10.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res100_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res100.jpg)
Resistance is nice and happy too. 10 Ohm is +0.2ppm, 100 Ohm is +6.5 ppm.
(https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res10k_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res10k.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res100k_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res100k.jpg)
10 kOhm difference is -0.51 ppm. Calibrator specification for this resistance is +/-6.5 ppm.
Also 100 kOhm is +0.2ppm.
(https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res1meg_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_res1meg.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_dci1ma_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_dci1ma.jpg)
Even 1 Meg is mere -3.2 ppm, very happy with results so far on resistance. Not so much on DCI, there is 330 ppm error (which I expected, A7 current board known to have issue)
(https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_dcv20_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_dcv20.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_acv2v_1.jpg) (https://doc.xdevs.com/doc/Fluke/5720B/tests/h2_acv2v.jpg)
Look at that 20 DCV. Isn't it wonderful? ACV quick check also confirmed operation. Will use Wavetek 4920M AC Voltage Standard (https://xdevs.com/fix/d4920m/) for more detailed tests on ACV when DCI issues fixed.
-
Impressive as always !
-
Time to look at A7 assembly, which I already confirmed drifty from last year tests. This assembly perform two functions - generate frequency for ACV/ACI function, as it have high-resolution oscillator circuit. Second function is to voltage to current conversion, to generate 220uA/2.2mA/22mA and 220mA DCI/ACI current ranges. And this is where the problem is - all current ranges have drift about -400 or more ppm per 24 hours. I've checked relays, opamps and input signals - those are all good.
(https://xdevs.com/doc/Fluke/5720B/tests/fail_dci.png)
Same on 220uA range:
(https://xdevs.com/doc/Fluke/5720B/a7m/100uA_dci_fail.png)
So the suspect is magic Fluke resistor network which sits in front of a complimentary drive hybrid assembly. Network have two isolated resistor 4K/40Kohm networks.
(https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_rnet_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_rnet.jpg)
I've measured one side as 4.000K/40.000K but other side was 3.998K/40.248K. That made me thinking this other side have a drift, as this network is hermetic and laser trimmed.
Close look:
(https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_top_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_top.jpg) (https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_bot_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_bot.jpg)
To test the theory, I've removed network and replaced one side with two BMF resistors for a test (actually three, 2x2K in series + 40K VHP102L from dead 3458A :)
(https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_moda_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_moda.jpg) (https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_modb_1.jpg) (https://xdevs.com/doc/Fluke/5720B/a7m/a7z1_modb.jpg)
Logging started, will see if current drift goes away. In that case I will need to find permanent solution for Z1 network replacement.
-
Congrats
look like Hulk2 is all fixed now. :-+
-
What could be the cause for this partial fail of resistor network?
Degradation of junction between resistive material and lead?
-
They keep coming, I couldn't help it :(
(https://xdevs.com/doc/Fluke/5720A/a11_u4/a11_u4_1.jpg) (https://xdevs.com/doc/Fluke/5720A/a11_u4/a11_u4.jpg)
:palm:
-
What could be the cause for this partial fail of resistor network?
Degradation of junction between resistive material and lead?
From the picture the network looks rather bad - like flaws from the production: It looks like there are parts of the TaN resistive layer missing, especially for the right divider) and in the area between the gold traces there seem to be still some blue resistor layer left. This network should have not passed QC.
-
I can't see parts of the TaN resistive layer missing, but a fine insulation groove instead, to form the resistive element and some dust on the top of the glass.
Edit: Looking more carefully it seems there is some small crack in the glass (upper right side) which leads to some humidity inside the package and which also changed the color of the golden traces.
-branadic-
-
The confusing thing is that all this visible flaws are on the good right half.
I could not find any visible flaws on the drifting left half...
The background why I asked was more if there is a common failure scheme with this type of resistors and if drifting DACs - specifically U180 - from 3458A suffers from the same.
-
The background why I asked was more if there is a common failure scheme with this type of resistors and if drifting DACs - specifically U180 - from 3458A suffers from the same.
Propably not, different resistor technology, different package, so the cause for drifty U180 might be different.
-branadic-
-
Hulks are getting their new Pelican travel cases :-+
No expense spared on precious Fluke ppms.
(https://xdevs.com/doc/Fluke/5720A/travel_case/IMG_20190814_104404_1.jpg) (https://xdevs.com/doc/Fluke/5720A/travel_case/IMG_20190814_104404_1.jpg)
Models are 1690 and iM3075. I wanted two 1660's which are the perfect size for 5720/5725A, but local Peli rep did not have 1660 in stock.
-
Hi,
Did you replace electrolytic caps on A18 card?
My 5700A II output tripped above 22V range, its A18 has not enough voltage such as unregulated +480V is down to 300V.
Fluke repair center is no longer fully support 5700A |O
-
Yes, of course. All electrolytic capacitors on every board and all CC resistors were replaced, because my units run 24/7, often unattended for many days..
-
dl1640,
sounds like one of the larger CC resistors may be shorted. I have found this to be true on a couple of A18 now. These are the ones connected to the high voltage zeners.
-
I took the board out of the box, checked R208 R210 R235 R236 (220k 3W), all good. :-//
Maybe I'll have to desolder these big caps and test them...
-
Hello, i wish know anymore on this shiel please. How is connect (on Guard, ok, but is it connect on other plug? or cable?)? And, it's here for what expectation? on DC Voltage, for emf? or AC voltage, and reduce noise?
Thaks a lot for the information ...