Author Topic: Repair : HP 3245A precision DC/AC source (Read 58296 times)

TiN · « **on:** October 30, 2015, 11:14:57 am »

I feeling not truly exited about outcome of last HP gear repair, so I decided there is more to it.

Let's the community work this one out, will see how good our Hewlett-packard EEVBlog fanboys know their T&M lineup.

I have new piece of gear in-transit at this moment, and while I'm waiting for it, let's take a guess, what is it?

Almost sure that I'm hinting too much, but anyway, here's piece of photo of it:

Give me your ideas! Person who guess correctly will win a prize, mention in credits

P.S. Ones who know may remain silent

.

deadlylover · « **Reply #1 on:** October 30, 2015, 11:17:59 am »

3245A universal source?

I *think* it shared the same chassis as 3458A.

TiN · « **Reply #2 on:** October 30, 2015, 11:18:43 am »

Man, why you kill all the suspense?

crispy_tofu · « **Reply #3 on:** October 30, 2015, 11:24:40 am »

wOw, that was fast

deadlylover · « **Reply #4 on:** October 30, 2015, 11:28:09 am »

I only recognised it because I was previously 'tricked' by it when I was browsing 3458A listings on Taobao (Chinese Ebay).

lukier · « **Reply #5 on:** October 30, 2015, 11:35:37 am »

Awesome TiN!

I was looking for 3245A on eBay for a while, but they go for pretty penny.

There is very little info on these devices online. Service Manual is not very detailed, but I think CLIP can be bought (paper?) from equipment manual vendors.

I'm extremely curious how they pulled off 6.5 digit precise DC voltage output. Nowadays one would just buy AD5791B and Bob's your uncle. Jim Williams AN86 20 bit DAC would be probably too slow to settle both DACs and correct with ADC readings.

Another option is PWM. AFAIR this is the way DAC is implemented in some Fluke calibrators. It still surprises me that Fluke made PWM DAC precise, need to figure out how they did it at some point.

TiN · « **Reply #6 on:** October 30, 2015, 11:35:43 am »

Quote from: crispy_tofu on October 30, 2015, 11:24:40 am

wOw, that was fast

Yah, too fast... Moooods! Remove this guy

Ok. One I bought is sick and broken (surprised?), so it will be a full teardown and worklog on it too. Should be here somewhen next week, I hope. I could not resist to win it at $344 (+175 for shipping, yeah, it's big and heavy 15KG).

Vgkid · « **Reply #7 on:** October 30, 2015, 03:47:07 pm »

Looking forward to it.

TiN · « **Reply #8 on:** November 03, 2015, 04:27:58 am »

Box is here.
Should I make teardown tuesday video?

Vgkid · « **Reply #9 on:** November 03, 2015, 04:34:37 am »

twiddling my thumbs.

TiN · « **Reply #10 on:** November 03, 2015, 04:36:49 am »

I can even try livestreaming, if one know good online streaming engine, since Youtube borked theirs.

tec5c · « **Reply #11 on:** November 03, 2015, 05:07:07 am »

http://www.yawcam.com/ ??

TheSteve · « **Reply #12 on:** November 03, 2015, 06:13:43 am »

Yes please to a video tear down.

PTR_1275 · « **Reply #13 on:** November 03, 2015, 09:32:57 am »

I'm going to guess error 82 and has the high voltage option in it.

I was bidding on it, but good to see it go to a good home (even though it isn't on my bench)

dom0 · « **Reply #14 on:** November 03, 2015, 09:57:27 am »

Quote from: lukier on October 30, 2015, 11:35:37 am

Another option is PWM. AFAIR this is the way DAC is implemented in some Fluke calibrators. It still surprises me that Fluke made PWM DAC precise, need to figure out how they did it at some point.

PWM filtered by integrators is extremely linear. It is basically the counterpart to timing/integrating ADCs.

crispy_tofu · « **Reply #15 on:** November 03, 2015, 10:05:12 am »

Twitch.tv maybe? Looking forward to it!

TiN · « **Reply #16 on:** November 03, 2015, 11:15:16 am »

Twitch needed to install 3rd party garbageware (which not even working right on my Win2008).

Ustream it is.

I'll do proper video with camera as well, so don't expect much from cheapie webcam.

TiN · « **Reply #17 on:** November 03, 2015, 10:17:19 pm »

I'll post details later (6am here, not slept yet!), but in two words : It was easiest repair in the world, did not even need to heat up soldering iron.

Here are some internal goodies:

Cleaning front panel:

Vishay hermetic foils on HV board (4pcs 40K000 and one 4K4):

lukier · « **Reply #18 on:** November 03, 2015, 10:47:21 pm »

Quote from: TiN on November 03, 2015, 10:17:19 pm

I'll post details later (6am here, not slept yet!), but in two words : It was easiest repair in the world, did not even need to heat up soldering iron.

Great job!

Get some sleep and let us know what was the issue. I remember the eBay listing (even asked the seller for the shipping cost but in the end it was too much) and it showed some errors.

Thanks for sharing the CLIP - PWM DAC indeed! (and 12 bit high speed DAC in low resolution mode). Interesting to see so many switching power supplies in a precision instrument (even for the inguard section).

Nice to see mostly off the shelf parts, except few hybrid resistor networks in the analog section and gate arrays (PWM DAC controller etc).

I see you added a bit of foam around LM399 to keep it nicely warm and cozy

Strangely enough, in the CLIP LM399 zener current is sourced from 5V. Heater uses GND and +18V. I remember other Agilent gear using +-15V for the heater and there was some discussion that this is a more stable arrangement

lowimpedance · « **Reply #19 on:** November 03, 2015, 11:00:59 pm »

Nice score , whats the bet it was a dirty or loose connector/socket.

A welcome change from the prior HP experience Im sure.

TheSteve · « **Reply #20 on:** November 03, 2015, 11:20:59 pm »

Front panel connector loose? This rivals my last ebay score - HP network analyzer that was sold as dead, was a dirty 110/220 VAC switch.

Vgkid · « **Reply #21 on:** November 04, 2015, 03:40:14 am »

Thanks for the teardown. I assumed that it would bear some resemblance to the 34401A, but nope.

TiN · « **Reply #22 on:** November 04, 2015, 04:12:46 am »

Why 34401? This beast does not even have ADC

Article with more photos and details

As you can get hints from article, I plan to do some modifications and tweaks, such as replacing voltage reference with LTZ1000 module and adding lower gain output boards, so I could generate accurate low-noise millivolts/microvolts. There are two open BNC openings in front panel face, so I can install triax connectors for extra outputs.

Vgkid · « **Reply #23 on:** November 04, 2015, 05:08:37 am »

Prices(1988 catalog pg 430) $4200 base, 2500 2nd channel, 400 waveform software option, option 002 is not in this edition.

lukier · « **Reply #24 on:** November 04, 2015, 03:37:15 pm »

Hi TiN,

if in the future you'll be playing more with 3245A internals, would you mind probing XMSP/XMSN and XSLP/XLSN signals? I'm curious what is the scheme employed there, PWM or pulse density, what PWM frequencies (as we have 200 Hz and 194 Hz low pass filters), center aligned and complementary PWMs or what.

I'm analyzing their PWM DAC circuit and it is a bit cryptic due to the sheer number of analog switches involved in all the modes (and probably selftests). In HiRes scenario LSB PWM is connected with the MSB via RP100A resistors ?, then through low pass filter, but the LSB has its own path and LPF as well and both of them can be switched to the common DAC output. In these scenario the HS DAC seems to provide DC offset. In high speed scenario the LSB+MSB DAC seems to provide DC offset (called vernier)?

Very interesting device indeed. Maybe with HV option and LTZ1000 it could as well replace these bulky EDC voltage/current standards.

What INL can one expect from such PWM DAC? Linearity is not specified in the 3245A datasheet AFAIR, but if they say it is 6.5 digit I would assume 1 ppm.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Repair : HP 3245A precision DC/AC source (Read 58296 times)

Share me