Prema 6048 teardown

[Rax]

Repeated my experiment as above, and I'm getting pretty good (sub-ppm) consistency on the FX measurements from one nulling cycle to another.

[Rax]

Please don't adjust any of your precision devices, but do everything on paper. Without notes it's easy to get lost. Some people use a lab notebook to keep important information like that in a safe place.

You could just measure the Fluke reference as good as possible and note down the result, maybe together with ambient temperature. After the P6048 mod you could repeat the exact same procedure and determine the new value. This way you can always calculate "unmodded", calibrated results by subtracting the shift caused by the mod. It's easier to revert the mod as long as you don't adjust the Prema calibration.

Regards, Dieter

Thank you for this, Dieter. I'd really like to save the current constants from the meter, and I may do that before doing anything else.

But after having a calibrated 3458A at hand, it seems the Prema - which was about 4ppm off in its original hardware configuration, at least per the 3458A - is now at what seems to be 6.64ppm high at 10V. So, essentially, the difference between the different op amps is these two ppms. I can double check this with my 5440A - which I deem verified by the 3458A - but I'm also considering having the Prema cal/adjusted to the 5440A. I'd bring it to sub-ppm levels on 10V from the over 6ppms it's at now, while I'm very happy for figuring a working combo of op amps curing the INL issue (which I should double check is cured, surely...).

Just something I'm considering. Of course, I can always put back the original op amps and restore calibration to factory constants.

[dietert1]

I think a deviation of 4 ppm is completely normal for equipment that hasn't been calibrated for years. The extra 2 or 3 ppm shift from upgrading the opamps are very little, too.
If you are able to read the actual calibration info of your Prema 6048 and save it digitally, would you provide some hints how to do it? I tried to find something in the user manual but in vain. Is it by pulling the calibration memory IC?

Regards, Dieter

[Rax]

I think a deviation of 4 ppm is completely normal for equipment that hasn't been calibrated for years.

Just about compliant with the 1yr spec, if I read it correctly (5.5ppm of 10V?).

If you are able to read the actual calibration info of your Prema 6048 and save it digitally, would you provide some hints how to do it? I tried to find something in the user manual but in vain. Is it by pulling the calibration memory IC?
Regards, Dieter

What I'm thinking is pulling U12, ripping out the data on it and saving it with my xgecu reader so I'd be able to rewrite it on some other IC (such as FM16W08s which I got for this purpose). I think this should transplant the factory cal constants into the replacement contemporary IC. Non-volatile memory not relying on a battery.

[perdrix]

You may find the attached useful (it includes calibration instructions) as well as schematics.

[Rax]

You may find the attached useful (it includes calibration instructions) as well as schematics.

Thank you, David. I think at least the three of us in this thread doing work on their 6048s have the manual at hand (if you haven't already, please do check the op amps mod!), but I took a quick, cursory look back in the thread, and it looks like the manual PDF hasn't been attached anywhere. Thanks for filling that gap!

[Rax]

Is it by pulling the calibration memory IC?

Regards, Dieter

Dieter, I think the "floating" calibration info is saved on U12, a 48Z02 which in my unit is a Dallas DS1220Y. I saved the memory out of it with my xGecu T48 to a BIN file (which I can provide should you want it, though I think it only carries my calibrations constants).

I'm currently evaluating how I can transplant it on an FM18W08, which seems to be a pin-compatible replacement with non-volatile memory but without a battery. My plan is to then have a full hardware backup of U12 including the current cal of the instrument. I'd then be able to recalibrate it with the capacity to go back.

Now, I hit a bit of a roadblock - I'm not 100% sure how to proceed. I programmed the FM16W08 with the DS1220Y BIN, but there seems to be more to it than I initially thought: https://xdevs.com/fix/hp3458a_u4/#fram. Rewiring and placing of the data on the lowest bank?

If anyone reading this has completed or is familiar with the procedure, I'd appreciate some input.

[Rax]

Some pics of the work today. I had to remove the side panel to be able to pry U12 out of its socket. These f-ers stick like glue after a decade or three. I am using a Brown Dog (Cimarron) adapter to make the FM16W08 (SOIC) into DIP.

[Rax]

Well, for one, the FM16W08 is a 28 pin IC, the DS1220 is 24. Shame on me thinking this is a "drop in."

[branadic]

Attached images of a proper adapter.

-branadic-

[Rax]

Last night, I got as far as finding some adapters (such as those you point out, branadic) available both for direct purchase and gerber files (https://github.com/tebo1000/DS1220-replacement-PCB), which is when it downed on me I blew up my "extended weekend" project. I could probably source some "80s date codes" DS1220s in my neck of the woods during the weekend, but that's about all I can do.

I'm planning to simply program the BIN onto the FM16W08 and test in the meter. If I get some Error 8 or whatnot, reverse it (put back the DS1220, though I'll still try building a few with different calibration constants).

In terms of programming the FM16W08s with the DS1220 BINs, are there some known specific precautions to take? Given the former are four times the size of the latter. The other thing I'm not quite sure of is if the programmer needs to be set to "see" a DS1220 once the FRAM is on the adapter, but that's what makes sense to me, so that's what I'll do.

[Rax]

gerber files (https://github.com/tebo1000/DS1220-replacement-PCB)

I wonder though if anyone can confirm these adapters (those above) are working as intended. I ordered a set from OSH Park, but it'd be great to have them vetted before I go too far with them.

[Rax]

Is it by pulling the calibration memory IC?

To answer this question specifically, Dieter (and as discussed offline), that is correct. I'm not trying to approach this through GPIB. It would be nice to be able to pull this data that way, but I don't think I need that given my specific needs.

I am able to see the data off of the DS1220Y (though the programmer interface) but it's not directly legible. Just enough to allow me to assume there's something on there, and make some basic comparisons (for instance, after transplanting this code into the FM16W08, which seems to check out).

I'm not sure if there's a way to interpret it, but I assume it would need to rely on the Prema software in some manner, which may preclude that path.

[Rax]

I've adjusted the meter tonight per my 5440A (10V range only). I trust it to about .5ppm/10V. I'm about 1C below everything the 5440A has seen in December when it was running alongside the calibrated 3458A, so I exercise a bit of judgement in the software calibration adjustment.

At this point, I have a DS1220Y with these calibration constants, one with the last calibration calibration constants, and will next program one with the factory constants.

I'm not sure how easy it's been for others to extract U12 out of its socket, but I'm having a heck of a time doing it. There's an "cardboard HV shield" that's in the way. The FM16W08 on adapter will be so much easier to plug in.

My end goal is to have three FRAMs with these three sets of calibration constants. Such that I'd have these three sets in an easy to swap in manner and "battery-agnostic," so they'd last 150 years with no stress.

[dietert1]

It may be easier to take out the whole processor board. There are the flat cables, two nuts that hold down the board near the front and two GPIB screws in the back. As always, avoid ESD damage while handling the board. When inserting the board back again, assure the calibration switch remains off.

Regards, Dieter

[Rax]

Continuing to examine the shifts I get after switching those op amps:

• 2V - a 1V excitation from the 5440A returns .99999595V vs. .99999535V (the latter being the reading on the 3458A from a couple of weeks ago)
• .2V - a 100mV excitation returns 99.99431mV vs. 99.99616mV (which certainly gets into the noise)

The above is with 80s integration.

I think this is much lower than I expected - or is there maybe a scaling aspect to the shift? In the sense of being proportionally lower at lower readings. I didn't think it would. If I'm reading the specs correctly, this is actually compliant with spec accuracy on the 2V scale but not on the .2V (off by 1.25uV, which is well down in the weeds...).

[Kleinstein]

For replacing U9 , I would expect an effect on the readings near zero, effecting the INL problem there as an extra offset for the positive / neg side. The error near zero also slightly effects the calibration of the scale factor, but this may be in the uncertainty of the usual calibrations.
For replacing U10 one should have a small effect on the scale factor: the offset of U10 adds to the 7 V reference voltage.
In addition there can be a change in the zero point, but this is usually corrected with the zero function.

The changes should scale with the ranges and this less visible in the 2 V range and even less for the 0.2 V range. The gain between the ranges should not be effected by a change in the OP-amps.

[Rax]

The changes should scale with the ranges and this less visible in the 2 V range and even less for the 0.2 V range.

I think that is what I'm seeing. The shift at a 10V excitation/measurement (20V range) was not small - all said and done about 8-9ppm [including a likely 4ppm drift by the meter due to aging]. In the 2V range at 1V, if my math is correct, just about 1ppm of 1V [much less, actually, judging by values measured after multiple 80s integrations (all the way to 24hrs later). But then, who has time for an integration so long it can be considered "aging?..."]. In the .2V at 100mV excitation I see more shift, or lack of accuracy, but I blame the low level of the excitation which gets things close to noise levels.

[...] = ulterior notes

[Rax]

Finally got "Bonetti"-based adapters from OSH Park and soldered some FM16W08s, but I'm getting errors when trying to write the BIN onto it as a "pseudo-DS1220." Both the "program" and the "verify" seem to return errors.

When trying to write the 6048 BIN onto the FM16W08 installed on a straight-out DIP adapter instead, the programmer didn't complain. Both the "program" and the "verify" procedure went through without a hitch.

Any thoughts as yo why I'm unable to write this successfully? The only thing I can think of is I didn't install the SMD cap the "Bonetti" board has provisions for. But I can't imagine this to be the issue here.

Pics enclosed.

[Kleinstein]

The Decoupling capacitor could make a difference, as decoupling should be close to the chip and the adapter / ZIF socket add quite some path.
An obvious point is that with the adatper the programmer should be set to a DS1220 or similar chip, no longer the FM16W08.

[Rax]

Maybe I'm impacted by this: https://www.eevblog.com/forum/manufacture/fyi-xgecu-tl866ii-and-t48-cant-program-fm16w08-fram-chips/.

[Rax]

The Decoupling capacitor could make a difference, as decoupling should be close to the chip and the adapter / ZIF socket add quite some path.

I didn't have an 0805 at hand, so I used a very small leaded MLCC to no avail.

An obvious point is that with the adatper the programmer should be set to a DS1220 or similar chip, no longer the FM16W08.

Yes, indeed, and that's how I did it.

[branadic]

Why do you use "pseudo-DS1220" instead of choosing the FW16W08?

-branadic-

[Rax]

Why do you use "pseudo-DS1220" instead of choosing the FW16W08?

-branadic-

Because, once the FM16W08 is installed on that adapter, it effectively is a "non-battery" DS1220 for all intents and purposes. The programmer should see a DS1220 to be able to program it (pins assignments, etc.). And hopefully the Prema also "sees" it as DS1220 once installed in its U12 socket.

Am I missing something?...

[Rax]

Why do you use "pseudo-DS1220" instead of choosing the FW16W08?

-branadic-

On a side note, as I was able to program the BIN on an FM16W08 placed on a BrownDog 28 SOIC to DIP adapter, and verify it successfully, I guess I could transplant that already programmed IC onto the "Bonetti" adapter. But it feels I shouldn't need to do that. I also am not sure the programming would become a successful emulation of a DS1220 so the Prema would like it.