Agilent 34401a revisions

[dorkshoei]

The firmware version is just an academic excercise to determine when it was manufactured, right?

Wrong.

Does the firmware matters aside from understanding if it uses the new display board?

Read thru the service notes.  For example note 34401A-04 describes a (the?) difference between 01-01-01 and 02-01-01 firmware.

Serial Numbers: 3146A00100 / 3146A03335 (Firmware Revision 1 – REV 01-01-01)
Explanation of indicated calibration failure of AC current

Starting with Revision 2 firmware (REV 02-01-01) a calibration was added at the 10 mA
level on the 1 AAC current range. Units prior to this revision will meet specifications but
are not designed to accept a calibration at this point. However, using the STE9000 test
software (Application Part #5011-1013) they will indicate “fail” for this adjustment on the
SCAT II “adjustment report” if the AC current calibration is attempted.

[dorkshoei]

Also look at the "backdating" section of the 34401A service manual.

In reference to the above 01-01-01 -> 02-01-01 change you will see:

From 3146A03335 and Below Modify A1U502 34401-8881 Programmed EPROM

[NaxFM]

Just got my first 34401A!
It's made in the USA and it still has the HP branding on the back, the serial number is US36074928 and the firmware version is 10-05-02, so the most recent Agilent one.
Honesly I didn't expect such an updated firmware, the unit seems quite old...

[dorkshoei]

They are old :-)   I think Keysight made a 03 display board but other than this they didn't make that many changes before it was discontinued.

You can get the bumper and handles from China for cheap.    https://www.ebay.com/itm/184419854408

That's the new handle design.  You can get both but be sure to specify.  I can't recall if your unit had the old or new.  Maybe the new style was just for the oval blue button.  Someone will know, or it doesn't matter.

[HighVoltage]

Just got my first 34401A!
It's made in the USA and it still has the HP branding on the back, the serial number is US36074928 and the firmware version is 10-05-02, so the most recent Agilent one.
Honesly I didn't expect such an updated firmware, the unit seems quite old...

Nice, congratulations.
Someone exchanged the window already from a HP brand to Agilent brand, that is also nice.

Yes, I agree, get some bumpers and a handle and it will look even better !

[gerryc89]

Hi to all, i recentrly obtained two multimeters in bad condition, with different board revision, one with v05 and one with v11. I Have repaired all of these, with this occasion i have analyzed the differences between two models and i have extracted also the two firmware of the comm cpu v1 and v5. In the beckdating section of the service manual i have verified the difference and the result is this:

first change board revision
C111  100 pF ±5% 50 V  add
C402  220 pF ±5% 50 V  add
R327 21.5K ±1% .125W   add
R403 8.25K ±1% .125W   replace 12.1k
R422 46.4 ±1% .125W    add
R430 Zero Ohm          replace 14.7ohm
U102 custom resistor, i think is only a different production, i have measured the two different codes without desoldering and is very closest

second change only components
CR203 Zener 5.1V 5%    replace 4.7v
R207 2.37K ±1% .125W   replace 5.62k in manual, but mine is 1.78k, other in photos are 1k

other changes is only production variation, the last important change is the dipslay panel from version 07 and the firmwares of main and comm cpu

i have analyzed the differences and i have founded this:
The gate of mosfet q301 is connecet thrugh the r327 resistor, i think it work only in saturation and the resistor decraese the stress on the switch ic.
The power to U411 and relative cap C410 is delivered by u401 5v reference buffer thrugh the r422 resistor, instead from u553
Replaced the zener voltage that power the ohm current generator from 4.7v to 5.1v
The pin vee of the mux U411 is connected to gnd only with the inductor, insted the r430 in series
Other changes in the amplifiers chain of the final stages that go in the adc.
Someone think that all this changes can be do on the old pcb for update all hardware without concequences? Improve something? Obvius that a recalibration is necessary.
Here is the firmware of comm cpu u700 that i have dumped

http://www.ko4bb.com/getsimple/index.php?id=manuals&dir=HP_Agilent/HP_34401A_6.5_Digit_Multimeter

In the latest revision of the pcb is present a jumper that connect to ground the a4 address line of the ram, if the jumper is removed the gnd disappear from signal and remain connected to the cpu pin 27 that is an high speed output, on the old revision this pin is grounded and the a4 line of the ram is connected to vcc, the result is the same, a part of the ram is unused. Is interesting if on the firmware effectively is supported.

regards
gerry
 

[gerryc89]

other images

[Kleinstein]

Someone think that all this changes can be do on the old pcb for update all hardware without concequences? Obvius that a recalibration is necessary.

The old HW was working OK and it is not clear if the changes really improve things. Some of them could be because of minor changes to the custom chips  (e.g. c111 may have been inside on older hybrid version, or just needed because of other changes inside). 
I don't see any reason to change things - the changes don't look like they would fix a serious bug.

[gerryc89]

Someone think that all this changes can be do on the old pcb for update all hardware without concequences? Obvius that a recalibration is necessary.

The old HW was working OK and it is not clear if the changes really improve things. Some of them could be because of minor changes to the custom chips  (e.g. c111 may have been inside on older hybrid version, or just needed because of other changes inside). 
I don't see any reason to change things - the changes don't look like they would fix a serious bug.

I have the same opinion, but is interesting if enchange a crosstalk noise on the adc switch U411, the current of the ohm gen i think is the same only change the open voltage.
You have some ideas regard the last opamp stage of the comparator input for the asic? The asic is the same, only change the custom resistor array but the ohm values seems to be the same.
Gerry

[Kleinstein]

The comparator input to the asci would effect how well the integrator tracks zero. So a change here can effect how large the excursion of the intergrator can be.
For some reason HP claims that this would help with the INL due to dielectric absorbtion, as there is less average voltage in the intergrator. However as far as I see it, the extra small contribution via R406 adds a linear part and thus mainly a small effect on the ADC gain, but hardly the linearity. With the rather fast modulation there is anyway not much INL effect due to dielectric absorption expected, unless the integration capacitor is a high DA type.
The part still makes sense, as less maximum voltage also means that a smaller integration capacitor can be used and for the MS3 ADC the auxiliary ADCs sees a smaller voltage range. So The R403 / R406 part can improve the resolution, which is one of the weaker points of the MS3-adc.

I don't know why the value or R403 changed. It could be a BOM simplification or from looking at the worst case instead of average integrator output voltage. The ideal value for the R403/R406 part also depends on things like the delay of the comparator and the size of the integration capacitor (may include tolerances). The change may be just to allow for more scattering in the capacitors (lower reject rate from capacitors on the small side).

Using U411 to also supply the HC4053 switch is a balance: the supply of the 4053 has some small, but visible effect on the ADC ( My ADC design is similar in using a 4053 type switch).
Using the voltage from U411 makes the supply more stable (U553 can show quite some temperature effect).  With the ADC constantly running I would not expect that much variations in the supply current of the HC4053. So the crosstalk back to the 5 V ref may not be that bad.  The change likely improves the ADC a little and adds a little noise to the ohm- part current sources and input amplifier.
The 100 nF decoupling capacitor seems to be sufficient. I don't understand the odd 46 Ohm value. If critical they could have used more decoupling.

[gerryc89]

Other two dmm back to life, full recap and i have also upgraded the main eprom and comm cpu. I have utilized a 2Mbit flash with bootblock, it work because the reset pin is already connected to vcc(plan to install another otp and remove the socket for evitate possible data corruptions because the write pin is connected to the cpu), i have also adapted the calibration eeprom inserting from the v11 cal the 0x14-0x4F zone to address 0x14 of the original, rewrite the zone 0x0c-0x13 with v11 data, and cut the file to 0x1FF, it work for v2 to v6 i think. The display v1 appear to work on v11 but i have scanned the pcb for reproduction of the v2 panel. No error messages for calibration, i have partially tested and appear to work correctly, i have to try the loopback of the rs232 and gpib but the software seems to start and run correctly on the 87C51FA from NXP. I have also exchanged the red black rear connectors of the new version to the front of the old.
Gerry

[Miti]

Other two dmm back to life, full recap and i have also upgraded the main eprom and comm cpu.

Why did you have to upgrade the comm cpu? Is there any writing about this?

Cheers,
Miti

[gerryc89]

Other two dmm back to life, full recap and i have also upgraded the main eprom and comm cpu.

Why did you have to upgrade the comm cpu? Is there any writing about this?

Cheers,
Miti

I have upgraded for evitate compatibilty problems, i don't know if is effectively necessary, and for try to extract the firmware from the controller that isn't avaiable online until now. In the change history the u502 and u700 are changed together more of two times. For programming the 87c51fa the tl866 or tl866ii work very well, only the tl866II have a bug in the software that if you close the xgpro application the chip can't be verified anymore, but is a problem to verify the encryption array, it program the chip correctly, i have signaled the bug to xgecu. The rs232 comunication work ok after update.
Gerry

[gerryc89]

I have recreated the layout of the v2 front panel and populated it with all components. The mcu is an AT89C51, it is the eeprom version of otp 87C51. It work very well and fit exactly like the original, the spare code of only the pcb is 34401-26512 , assembled with components is 34401-66512. I have others 9 pcb's if somewone is interested, are gold plated with mask, the only thing that is difficult to get is the resonator, the only that i have found is too little and i have mountend upside-down with wires.
Gerry

[coromonadalix]

Woah  you really did a new front pcb    kudos   

[NoisyBoy]

Excellent work on the PCB.  I was able to order a replacement front board through Keysight a couple years ago to replace my dimmed VFD, but the part is no longer orderable now, so your work will really help out those who needs front board repair beyond the VFD.

[gerryc89]

Thanks to all the informations on the forum for the calibration EEPROM and checksum i can do It, this Is my little contribuire I think that is possibile to create, without much effort, a led version with the same CPU and some shift registrers led drivers. Kingbright make a type of 14 segments and dot in smd format that Is to little to integrate in the panel.
Gerry

[Hydron]

Is there a reliable source for the VFDs (by themselves) at the moment? I've seen some comments about the aliexpress/ebay ones being a bit hit and miss, and my 34401A VFD is usable but a bit tired, so it would be nice to replace it if there is a source of good new ones. That replacement board is tempting though - would let you build a new display without risking breaking a dim-but-working one in the process!

[NoisyBoy]

I have seen the same feedback from replacement VFD, lots of hits and misses, depending on the marketplace sellers, and even from the same seller.  I have also seen some where the product does not match the description, or the pictures posted, and I hate to wait to get something from China, find out it doesn't work and try to get an refund or replacement.  But still, that's the cheapest route, I just never went down that path.  Switching out the VFD is not that hard, I have done a couple with donor units with original parts, a desoldering gun would help greatly, so the risk for ruining the PCB is low as long as there is a good match in the replacement parts with the original (which can cause trouble with the metal shroud with shorting pins out). 

Back then, I was able to work with someone I know at keysight to obtain a parts authorization for order, so I was able to place an order through sales for the entire PCA front panel for $149.  But currently, the only way to get the parts is through their repair service, so that option is no longer available.

For measurement that requires high accuracy, I tend to leave my DMM on.  On my 34401As, I now would leave the display off until I need to do a series of measurements to avoid any VFD aging issues, this is one practice I would suggest for other 34401A owners.  It is a bit of a hassle, but not too bad once it becomes part of your workflow.

I think all these made the work from gerryc89 so much more important, so we can keep our 34401A in service.  I love it because it is fanless and very stable long term, I am not sure if there are other 6.5 digit meter that can do this.  I would love to see what he or others would do with alternate display technologies.

[coromonadalix]

 As we wrote on other threads Keysight where about to  statuate  a no longer available  front pcb with the original vfd soldered on it

it was about 195$ usd at the time, 2 years ago,  good luck if you find one

i bought 2x clone vfd on Aliexpress when they came out  35$ usd each,   had no problems with them,  maybe i was lucky

Bough recently a clone for the 34410a  version,   no problems,  but a very tight fit,  have to be very careful, the bent pins are done too far away from the vfd glass,   works very well    200$ CAD delivered

And finally  you have an oled made display version  here from Qu1ck, i helped him with probing signals from a more recent model ....  Fw 10-x-x


and yes  some meters have a display off command sent thru gpib  too

[NoisyBoy]

Yes, you can turn the display off via GPIB, but I typically do not want to power up my PC just to do that, so I just do it via the front panel.  Both my 34401A are later built with firmware 11, are you saying some earlier firmware does not have the ability to turn off the display through the front panel? 

I will have to look up the OLED upgrade, is OLED burn-in an issue?  That's why the LED option sound appealing, but clearly the items other than digits would be a problem to replicate without something like an OLED.

[coromonadalix]

yes oleds have some issues

there is a project here too have an external character lcd connected to a 34401a,    maybe with some code fiddling  you can adapt  it to your needs

or  the qu1ck  oled project could be adapted in a way with an bigger graphical / characters lcd too ??  since the protocol decoding was shared ....  the fun is yours 

or unless you want to take the   https://www.eevblog.com/forum/buysellwanted/ebay-led-display-module-fluke-8840a42a-(diy-kit)-replace-broken-damaged-vfd/

But i dont think you have enough supply power in the 34401a, i would not advice to take power from any source inside it ...


or finally   whack you own controller sending  vfd shut down commands on the gpib connector  loll           

[gerryc89]

One power supply that have a great current capacity Is present, Is the filament secondary, if the central winding of the trafo Is disconnected (Blue wire, that create the chatode referente voltage with a zener), It can be full wave rectified and used to power some led displays without touch balanced 18v rails, or only for the mcu.
Gerry

[coromonadalix]

yeah could try that filament supply,  but you have other supply for digital and analog sections, if you touch them you could create instabilities

you have enough place to add an small psu on the ac main line and feed a substitue display mcu  etc ....

[alan.bain]

On the 01-01-01 version it is worth noting that the command to query the firmware version on the 34401 constructs it from two RAM locations (on the 7-5-2 firmware there are 0x1708 for the release (-2) and 0x1A3E for the minor version (location is 50, /10=>5). So there are ways in which this could go wrong! Only the first digit major version (7) is a direct read from ROM. 
And of course the SYST:VERS? command gives the SCPI set supported (this is from ROM) so on 7-5-2 this is 1991.0