So, it turns out modern silicon chip design is more software than hardware type work. Who would have thunk? (But it makes sense).
I'm way out of my depth here with my new job...... I am absolutely
not a software person......
So, it turns out modern silicon chip design is more software than hardware type work. Who would have thunk? (But it makes sense).
I'm way out of my depth here with my new job...... I am absolutely not a software person......
I disagree, except in a boring literal ways.
Hardware is inherently massively parallel (duh!). That doesn't change when describing/constraining/testing a design using schematics or text. It is necessary that the text explicitly expresses the parallelism (standard computer languages fail to manage that).
The way the design is expressed
looks like the
syntax of a classic software language, but the
semantics are completely different. That causes never-ending problems for softies that think it is easy to convert from C/etc to VHDL/Verilog.
OTOH, yes, you do need to understand text editors/IDEs and compilation. Big deal
OT3H, the compilation tools are massive, and there is a significant learning curve. That's because hardware is complex
I much prefer working with schematics and PCB's than code. To my eyes, VHDL may as well be C or Python or etc etc... Just give me a schematic and some resistors.
Could not help myself. Gotta stop doing this...
OK, drumroll please!
See pic.
TomG.
For a moment I thought I was having a heart attack, but it was actually just my wallet cramping up and hurting when I looked at that picture.
McBryce.
Funny, I had that same feeling, now I know what it is....
TomG.
Congratulations on your new job Terra !
A whole new set of skill to be acquired... then once you have become an expert you can help us out on the forum !
I never played with that stuff, all I can remember from 30 years ago is the VHDL was an absolute pain when trying to make logic hips out of an FPGA because you just couldn't control the propagation delays so the chip could have erratic behaviour even though on paper it was perfectly functional... but to be fair I also seem to remember, many years later, that somehow that problem was largely overcome and now designers didn't have to care all that much about it, and could draw whatever they wanted and somehow the FPGA would actually do what it was intended to do.
I guess the compilers were much improved ?!
Yes I know absolutely squat... and I don't see that changing since I don't quite see what use I could have for that kind of stuff as a hobbyist...
Anyway, it's cool that your new employer is giving you a chance at this, even though he knows you know squat about it.
I don't see that kind of miracle happening over here. They only employ people that have a good Eng. degree from a reputable expensive private school, and are already masters in their field, beginners ignored.
Have fun...
I never played with that stuff, all I can remember from 30 years ago is the VHDL was an absolute pain when trying to make logic hips out of an FPGA because you just couldn't control the propagation delays so the chip could have erratic behaviour even though on paper it was perfectly functional...
You can't do that with conventional digital logic designs, either.
You have to structure the design so min/typ propagation delays are unimportant.
You have to state the clock period and external timing requirements, in the form of constraints.
It is then up to the compiler, and place and route tools to fiddle around with the internal implementation to meet those constraints.
If that can't be achieved, then you have to restructure your design.
That should remove "erratic" behaviour. Whether the (non-erratic) behaviour is correct is something your tests ought to reveal
3585A - searching the problem: input section
Wow thanks for all the hard work you put into little guy, he must love you so much by now !
Incredible for the probe tip ! I thought it missing would be a show stopper and condemn little guy to the trash... and in the end it was a standard part, who would have thought !!
The tolerance on anything with a glass envelope is rather..wild. Like that Nik-c tube on my Rochar DVM that just won't fit at all inside the tube bracket...... or like you here, light bulbs that won't !
Luckily small light bulbs you can buy lots of, for cheap... so you can just pick one that fits.
Thanks for your work !
And thanks for sending him to me, to repair, will put him to good use next time I'm fixing one of the IC based counters.
The irony with the 3406A probe accessories, is the isolator probe accessory that these screw on tips are for
is missing from the set, as you can see there is a choice, but only the sharp pointy tip would have been supplied with the 10525A.
The witches hat & two sleeves at the top of the picture are not part of the set either.
I know what you mean with glass parts that don't fit, seem to remember trouble with some cheap flanged neons for the power switches, that also don't fit well.
David
Absolutely agree, the 3458As are a rare case where time works in your favour.
The drift of the LTZ1000 does significantly decrease over time, so the new ones need some years to stabilize ...
I guess I'll see about the drift. I have an array of 732s to check that out. My reference 3458a drifted 1.6ppm in 20 months. Non 002 option, BTW.
TomG.
002 Option just selects the initial error. I guess those have been pre-aged a little bit
. Any, after a few years the drift should decrease significantly.
My box is a first or second generation one with a complete Keysight refurbishment (new caps, line filters, fan, front panel, VFD, nonvolatile memory, ...)
So I guess the LTZ1000 is about 20 years old. Nice !
My box is a first or second generation one with a complete Keysight refurbishment (new caps, line filters, fan, front panel, VFD, nonvolatile memory, ...)
So I guess the LTZ1000 is about 20 years old. Nice !
Unless... They swapped the LTZ1000 as part of the refurbishment too?
McBryce.
(An addendum for the shaded device)
First happened so that I ordered a set of six AA to C adapters from ePay and got one, next try from T emu was a fast success.
But the place is too tight, maybe correct but too tight anyway.
So sub-C order is placed, but then what.
I remembered that I've saved a 18V battery from lawn trimmer, for identification purposes.
So no problem, ripping that two decades old battery apart and dang, six cells have life, sane also, around 1V.
Now four of them are charging, 120mA and 6.1V, and keeping their cool.
Just some 4-wire shorting plugs, and 4-wire banana to BNC adapters to enable me to connect the kelvin resistance standards I am making for my LCR meters directly to my multimeter.
If no volunteers appear from DE I guess I can order them for you from NL?
@TERRA Operative
I'm a Bürklin customer, so I think, I can help here.
20 pieces of the order number 12F170 are EUR 37.60 plus 19% VAT and shipping.
What about Robert's suggestion?
Those have the order number 12F160 and 20 of them cost 25.40 plus VAT and shipping.
Let me now if I should order them for you.
The part I am looking at seems to be actually discontinued (unless I'm looking at things wrong?) so I was thinking that availability was down to what suppliers have in stock.
I'll send off an email to Staubli Japan and see what they say. If no-go, I'll ping back here to work something out.
I'm looking at using the stud version as I need to space it from the board for longer reach to get around the deep-ish sockets and the rubber bumper around the front panel on my 34461A (some of my boards are a bit larger than the edge of the front panel).
I'll check out the Multicomp ones too, thanks for the link, although Farnell etc don't operate in Japan it seems..
20 range is silent.
It can't make its mind between 9.99 - 10.1 so generally it's sort of like spot on.
After many hours its disagreement with its self has widened, maybe it's cooking maize.
Leftover cells could do a difference in EY503, if I could find out how to open the battery case without a chainsaw.
Last first charge of new pack long ago didn't go well, the pack missed a temp sensor.
Now after next recharge it's back to its more steady state.
And is more steady than first time, after who knows how long hibernation.
It's actually quite difficult to charge those cells without a proper power supply.
I left charging on over night, but current was very safe, maybe under C/20.
Finally stopped next day when 6V was giving 30mA.
If memory serves the original charge time was 3h.