My poor mans SMU - The Agilent 66311B

[FrankE]

It's back in the game again (well, at least to diagnose).  I replaced the 5A PCB fuses I blew late last night.  I replaced the Papst 12V fan which itself looks like a retrofit with a  San Ace 60 S Series 12V fan (from Farnell). Rated 0.4m³/min (14.1cu.ft/min) 16dBA. It is much quieter and I haven't yet fitted an isolation gasket.
It didn't come with a molex connector attached, (it's a 3 wire BTW) so I had to bodge it on with a connector block as I don't have a Molex crimper yet.
It soon ramps up full tilt despite no/low load i.e having 13.9V across it. Seriously, HP/Agilent is that fan speed really necessary
Input rails are fine.
No idea where to use as the common for the crowbar and reverse. My scope is broke so I can't see the gating. My 289 is away to Fluke, so it's getting a bit of a nuisanse swapping leads.

Now I can't get the output to shift from 8.000V.
I set it there after not previously being able to shift it from some other nominal voltage.
Perhaps it's on remote. "Address 0" is displayed in the boot sequence and I was fiddling in that menu.
I can't see the firmware option
Hmm, RTFM etc.

edit: oops, I forgot one of the 125mA  fuses had popped

I can't see any full schematics anywhere so there's a lot of guesswork
Once I get it fully working, would I get any better precision/ performance if I replaced the LM 78xx CGT and LM 79xxCGT the +/- 4% with LM 78xxACTG and LM79xxACTG the +/- 2% versions or are the subsequent stages taking care of the precision?

[FrankE]

Yay, repaired it. The crowbar circuit was gone (broken trace). Oh well, didn;t need those 2% LM 78xx LM79xx ACTG regulators. Might use them to build a 5680 10MHz Rb frequency standard.
It's running nice and quietly with a new fan
It has firmware A01.08.
What's to be gained from changing to a later ROM revision e.g. A 02.04 Any additional functionality?
I don't have a GPIB converter and they are horrendously expensive second-hand, so I'd need to buy an Atmel AT27C2048 and see if the local hackspace has an EEPROM programmer.
What can I do with the serial port? Is there any freeware one can use to control the 63311B or log?
I'm not sure how the calibration is but I get my 289 back from Fluke within a couple of days, which I assume will come with some sort of post repair/board change factory cal.

[TheSteve]

Nice to hear you've brought it back to life!

[FrankE]

Nice to hear you've brought it back to life!

Many thanks for the heads up on the deal.
It was my own fault, I might have broken it by allowing a scalpel to drop in it.
It's running really quietly 15.386V 2.6A powering a screenless laptop I found, so worth changing the fan.

This firmware upgrade. Is it worth it? I have the through hole version A01.08. I read SMT was introduced in 1999. Is 02.04 even compatible?
I got an Atmel IC today from RS, I don't have an EEPROM programmer and I couldn't justify buying even a TL866 this month but someone said they could do it with an Arduino. Whether 'could' is 'will' is yet to be seen.

[TheSteve]

No idea on the firmware update - I'd be tempted to leave it though, especially if you don't have a programmer handy. I love updating firmware on stuff but there is always the risk/reward factor. I'd also backup the calibration data before messing with it if you do attempt it.

[nidlaX]

Does anyone know how to remove the front panel board from the front panel chassis? The service manual suggests I pull the board back 1/8th of an inch near the encoder, but I can't feel any compliance and fear breaking the board if I do so.

[feedback.loop]

Does anyone know how to remove the front panel board from the front panel chassis? The service manual suggests I pull the board back 1/8th of an inch near the encoder, but I can't feel any compliance and fear breaking the board if I do so.

First you need to undo the nut on the encoder, then bend the tab on the board (see the pic) slightly so that it can go past the plastic stop, then you must be able to slide the board a few millimeters from under the hooks.

[nidlaX]

Does anyone know how to remove the front panel board from the front panel chassis? The service manual suggests I pull the board back 1/8th of an inch near the encoder, but I can't feel any compliance and fear breaking the board if I do so.

First you need to undo the nut on the encoder, then bend the tab on the board (see the pic) slightly so that it can go past the plastic stop, then you must be able to slide the board a few millimeters from under the hooks.

Thanks! Not sure how I missed the encoder being secured to the front panel chassis (although it's not mentioned in the disassembly procedure).

[ve7xen]

I decided that not having a knob was unacceptable, and have been wanting to try having small parts printed by a 3D printing house, so figured modelling and having a knob printed would be a good exercise.

Based on the photos here and at the Agilent Find-A-Part, plus some measurements of my actual unit, I did my best at a model in OnShape (shown is not exactly what was printed, I made some refinements):


I had the design printed by Sculpteo, using their SLS process, in 'polished beige plastic'. With express shipping it cost about $15 and arrived in a week.

Overall I am happy with the result. The only flaw was that the printing process left about 0.1mm extra clearance around the knob shaft. I had expected 'expansion' and not 'shrinkage' so had modeled to exactly the shaft dimensions, so now it's a bit sloppy. The STL I exported was also somewhat 'low res', resulting in the stepped appearance. I have changed the model to make the knob about 1mm taller, since it seems a bit shorter than the real one, added 0.1mm total interference on the shaft dimensions to make it tighter, and exported at a higher fidelity. I don't intend to reprint, as I think what I've got will be perfectly functional.




I have made the model public on OnShape, so you can copy it to your own workspace and modify as you see fit:

https://cad.onshape.com/documents/2770017399f21096171bcc75/w/82d82831cbaa64cec2e5b5a7/e/d79452ab36c21b057637b9b5

You can also download an STL file for direct printing here (note this is a refined model that I have not myself printed):

http://gotroot.ca/media/agilent_knob_model.stl

[pigrew]

I started trying to learn about Windows Presentation Framework over the weekend. I mostly can do the GUI part now, but data binding is still a mystery. Maybe next week I'll figure that stuff out.

Anyway, the software is able to talk to my 66309D pretty well.

<https://github.com/pigrew/HP663xxCtrl>

I'm not yet posting any builds, as there are a few small issues that need to be fixed (like adding a Save button). But, it's in a somewhat usable state. It'll probably work with any of the 663xx series (but I forgot to tell it about the 66312A and 66332A). If you find issues with its ability to talk to other models, let me know or send me a pull request on GitHub.

I tried to replicate Agilent's "data log" feature, but I'm not able to see most of my advertisement packets, so I'm not sure how useful it would be for me to estimate energy usage. Their software restricted it to the 66319/66321, but I'm not sure why. The user manual didn't describe any commands supported by the newer hardware or firmware v3 that would seem to be required.

The attached is the current draw of a TI BLE microprocessor. I've triggered it on an advertisement packet. It's transmitting on three different channels.

[pigrew]

One other request:

My SMU's front panel lets me select "AUTO" current ranging. What's the SCPI command to enter and leave auto-range mode? I couldn't find it in the manual.

[samofab]

I can report that the software compiles and connects to my 66309D.
Few bugs to iron out... in mine the decimal points for the display of current voltage and current on the left are missing (displaying 100000 instead of 1.0V)
Otherwise it looks totally great! Finally an easy way to use our 66309's & Co to their full potential!

Info for somebody else who'll try to compile it: VisaComLib is in file
GlobMgr.dll found in C:\Program Files (x86)\IVI Foundation\VISA\VisaCom
at least on my computer.

thumbs up!

[samofab]

My SMU's front panel lets me select "AUTO" current ranging. What's the SCPI command to enter and leave auto-range mode? I couldn't find it in the manual.

Perhaps you can look in strings section of the firmware if there's an undocumented SCPI command.

[pigrew]

My SMU's front panel lets me select "AUTO" current ranging. What's the SCPI command to enter and leave auto-range mode? I couldn't find it in the manual.

Perhaps you can look in strings section of the firmware if there's an undocumented SCPI command.

I tried downloading the firmware image posted in this thread, but couldn't find any strings in it. Is the image obfuscated or compressed?

Sent from my LG-D850 using Tapatalk

[pigrew]

Few bugs to iron out... in mine the decimal points for the display of current voltage and current on the left are missing (displaying 100000 instead of 1.0V)

I think this must have to do with my assumption that everywhere is like the America, and uses periods as decimal points.

I just tweaked a few things, and more things will now show up as commas for you...

But, for storing data files (like CSV), I assume that you would want to use periods as decimal points?

Were there issues with parsing numerical responses from the instrument? I probably should tell it to use the InvariantCulture more often....

(And yes, it still has quite a few small bugs that I can think of, but it nearly working well enough for my needs)

[nctnico]

But, for storing data files (like CSV), I assume that you would want to use periods as decimal points?

Getting that right is a study in it's own. The quickest solution is to make sure you use the local language setting. This may require extra initialisation in your software.

[pigrew]

But, for storing data files (like CSV), I assume that you would want to use periods as decimal points?

Getting that right is a study in it's own. The quickest solution is to make sure you use the local language setting. This may require extra initialisation in your software.

I've made the GUI and the communication with the instrument better respect the system's culture.

However, my question was regarding data files. I'm thinking that the data files should be independent of what computer wrote them? Also, I usually expect commas to separate fields in CSV files, but this would conflict with using a comma as a decimal point.

[nctnico]

What works for me when dealing with CSV, configuration, XML, etc files is using a semi-colon to seperate fields and a period as a decimal seperator. In order for this to work I wrote a some C/C++ routines which override the locale setting and process/print numbers with periods as decimal seperators. Displaying is another thing. If all goes well the programming language should detect the selected locale and format numbers accordingly.

[samofab]

Your last commit has a bug in ApplyProgram_Click.

you replaced
if (!Double.TryParse(CH1VTextBox.Text, out details.V1)) {
with
if (Validation.GetHasError(CH1VTextBox) || Validation.GetHasError(CH1ITextBox)) {

so you don't actually populate the details with values in textbox fields.

If you give me commit rights, I can fix small errors like this.

[pigrew]

Your last commit has a bug in ApplyProgram_Click.

......

so you don't actually populate the details with values in textbox fields.

Too late at night. Committed untested code. I think that it's fixed now, and I found a number of other bugs. I was trying to switch the fields over to using a ViewModel (I've never used one before), as the easiest fix for handling number formatting issues. It also let me add range validation checks somewhat more easily.

[samofab]

yea, it works now.
found another bug.. if you try to acquire much more than 3000 points the read operation timeouts at ReadIEEEBlock

[nctnico]

@pigrew: your software looks interesting. When I have some time I'll try it.

[pigrew]

I've put the software on the back-burner for the last few weeks. I ended up buying a similar Keithley supply (2304A), and will write a driver to also control that instrument with the same software. In the mean time, I'm fixing a few minor bugs I've found.

I also found that there is a native .Net driver interface for visa (ivi.net). I had been using IVI VISA COM. It looks like IVI VISA.Net has been supported for the last year or so, and makes it much more clear how to write drivers that don't have memory leaks and other issues. So, unless there are objections, I'll transition to using Ivi.net. It looks like it is supported on all recent versions of Windows (NI Visa 15+, Keysight IO 17.2+).

I promise to post a build of the software in the next week, once known issues are fixed and Keithley support is added.

Since it'll support other instruments, I'll make a new thread for the software (and post a link here).

[kaz911]

I decided that not having a knob was unacceptable, and have been wanting to try having small parts printed by a 3D printing house, so figured modelling and having a knob printed would be a good exercise.

Based on the photos here and at the Agilent Find-A-Part, plus some measurements of my actual unit, I did my best at a model in OnShape (shown is not exactly what was printed, I made some refinements):


I had the design printed by Sculpteo, using their SLS process, in 'polished beige plastic'. With express shipping it cost about $15 and arrived in a week.

Overall I am happy with the result. The only flaw was that the printing process left about 0.1mm extra clearance around the knob shaft. I had expected 'expansion' and not 'shrinkage' so had modeled to exactly the shaft dimensions, so now it's a bit sloppy. The STL I exported was also somewhat 'low res', resulting in the stepped appearance. I have changed the model to make the knob about 1mm taller, since it seems a bit shorter than the real one, added 0.1mm total interference on the shaft dimensions to make it tighter, and exported at a higher fidelity. I don't intend to reprint, as I think what I've got will be perfectly functional.




I have made the model public on OnShape, so you can copy it to your own workspace and modify as you see fit:

https://cad.onshape.com/documents/2770017399f21096171bcc75/w/82d82831cbaa64cec2e5b5a7/e/d79452ab36c21b057637b9b5

You can also download an STL file for direct printing here (note this is a refined model that I have not myself printed):

http://gotroot.ca/media/agilent_knob_model.stl

Hmm - I think Keysight spares quotes about $3.86 for the dial when you look for spares using the part number. I got 66311b from Ian for GBP 40,- without Dial knob - so I looked it up. But I then remembered i had an old E3610 panel in the scrap pile - and that has the same dial/knob so I solved it. But nice work making one - I'll try and print it on my Zortrax next time I turn it on to see how it does on ABS.

The 66311b I got from Ian was non functional - and it was a 120v unit - someone had probably plugged into 230/240 V uk supply. 1 main fuse and one onboard through hole 5A fuse later - the thing is now spinning but noisy as h.. But very precise and works well.

[HighVoltage]

- the thing is now spinning but noisy as h.. But very precise and works well.

If you add a 100 Ohm resistor in line with the fan supply, it gets much quieter and still has plenty of air flow.
Or you use an ultra quiet fan with built in temperature control.
I have two of these and they are both very quiet now.