TNP #43 - Rohde & Schwarz URE2 DC - 25MHz 300V RMS Voltmeter Teardown, Repair & Experiments

In this episode Shahriar takes a look at a fault Rohde & Schwarz URE2 which is a highly sought after RMS voltmeter. The ability of the RE2 to provide accurate RMS measurement up to 25MHz as well as provide DC measurement makes it an excellent instrument for characterizing DC/DC converter output noise characteristics.
The faulty instrument can measure AC, however no DC measurement can be made. The unit also occasionally generates an error code 0080 which points to a fault in the main analog board. Teardown of the instrument reveals a pair of front-end relays responsible for mode switching which are the likely cause of the DC measurement problem. The relays are replaced and the instrument returns to full operation. The performance and functionality of the unit is verified through several experiments.
Please consider supporting The Signal Path:
www.TheSignalP...
/ thesignalpath
Other links:
www.TheSignalP...
/ thesignalpath

[@michalzi]

I took a Brüel&Kjaer2636 to measure my HP3245A, because it is the best low-noise equipment I have (in this special application even better suited than my HP3458). Result: I actually see a tiny 47kHz square wave with an amplitude of ca. 220uV RMS! And as it disappears when unplugging the HP3245, it actually comes out the unit (no pickup-noise). 73 de Marc DL8ABE

[@Thesignalpath]

Thank you, looks like DAC clocking noise...

[@zacharysherf9397]

I thankful for a new signal path video to watch before bed!

[@Vintaginside]

I'm really glad you took into account my original comment when you uploaded the modified HP voltmeter. Thanks! On the other hand, UREs have some built-in switchable high and low pass filters which make your life easier when it comes to tricky measurements. Cheers,

[@ThouBeteraNay]

"We should see nothing" - we saw your cat, which is even better :D

[@GJackie24]

Shahriar excellent video as always! I love to see pieces of equipment I never knew existed. You have an amazing collection of new and old. I could easily spend a lifetime exploring this tech. Glad to see "Pooch" is always available to lend a paw at troubleshooting this stuff. How on earth do you find this stuff ? Love the tear downs and looking at all of the care that is taken in creating these instruments. Have you ever seen Marco Reps that guy is obsessed with obtaining the lowest PPM standards he can buy fix or design.

[@timlister177]

+1 for Pooch "help"...

[@jb5631]

2:24 - Would love to see see Pooch more often in the videos ❤

[@SeanBZA]

Sure a backlight is easy to retrofit, as the option likely uses the same display, and the actual backlight might already be there, just disabled in firmware, or there is a jumper you connect to enable it. Did that on a digital massmeter, as the backlight, because the firmware on the one I have is old, does not enable correctly, and I wanted the dim green light on all the time to show power on, so simply went and placed a small blob of solder across the CE junction of the transistor controlling it. Yes should replace the dim GaAs LEDs with brighter ones, or at least much more efficient ones, but they do the backlight function well enough still, and it is a nearly 30 year old instrument. The other one, which is a precision Oertling massmeter, uses a VFD display, which is, despite being close to half a century old, still bright. Did have the Scaeffner filter on the power supply give out that charactaristic smoke, so it was removed, and replaced with a salvaged set of chokes from a failed PC power supply instead. After all the transformer is potted, and impedance protected, and has internal one time thermal fuse as well, and no SMPS to be seen there, all linear power rails for low noise.

[@nikmilosevic1696]

Just checked my 3245A on a 3458A and got about 400mV AC at ~124KHz when DC set to 0V. Interesting.

[@RminusOR]

"I'd love to get my hand in one of these" 😅

[@EliasBakken]

Regarding the AC noise form the universal source, my HP 3245A has an ACV amplitude of 0.702 mV measured on an HP 3458A and 0.284 mV on a Siglent SDM3065X. They are not in agreement with each other, but perhaps that is a bandwidth issue? I see a tone at 8.59 MHz and lots of stuff in the 1 to 5 MHz range.

[@Thesignalpath]

Thank you. Looks like some noise is present in yours as well.

[@bblod4896]

Does the HP use a switching power supply? If yes, perhaps that's where the AC noise is coming from?

[@IanScottJohnston]

With my 3245A I get 0.377mV AC on my 3458A. This is via the 3245A BNC and also via the banana sockets I retro-fitted to the front panel. I also tried various cables, LMR200 coax, teflon/silver, gold banana's and all have about the same reading. With my 3458A open circuit I am getting around 0.434mV AC.........so I suspect this whole issue is noise pickup.

[@Thesignalpath]

I saw those videos you did on the retrofit, they were great. I think the source might be from the DAC. Illya did some additional experiments and it is consistent with some DAC clocking noise. Thank you for checking also!

[@USA_27]

Nice Content

[@xxportalxx.]

Im not sure how good your shielding is, but if you can try measuring the powersupply again from a different location. I remember once I was in an elctronics lab that was accross the street from a radio tower, we'd pick up all kinds of noise on our instruments while testing unshielded circuits!

[@xDevscom_EE]

Happy holidays to TSP! I've just checked 3245A (no options) and my own 3245A (option 001) with 3458A and Fluke 5790A and don't see 0.8mV. Sensitivity limit of 5790A is 0.6mV. 3458A reads some noise under 0.3mV but it does jump around around there even without 3245A connected. Perhaps what you saw is just some noise pickup?

[@Thesignalpath]

Thank you! Happy holidays to you too. It looks like there is some small noise from your unit, could of course be pickup…

[@ryebis]

@@Thesignalpath my 34401a sees some noise ~0.02mv from 3245A in the mV range but nothing significant in the 1V or 10V range which is surprising.

[@SeanBZA]

@@Thesignalpath Could also be environmental noise coming from the power rails. Time to put a search on your auction sites for a 2kVA ferroresonant line conditioner, to run all the lab equipment off of. It does make a big difference, though you do get a lot higher third harmonic distortion on the mains waveform from the core going into and out of saturation, but linear power supply, and SMPS input, diodes do run cooler with the flatter waveform, as the ci=onduction time per cycle is increased. High frequency noise, along with all the voltage variations, does almost disappear. Just that the 2kVA units will mean you do not need a home gym if you move it around at all, they are very heavy. Older ones you might need to replace the capacitors from them having drifted down in value, note that you do need to get the same voltage rating, as that tank can run up to 600VAC or more across the capacitors under certain conditions. Plus they are not exactly efficient at low load, I call mine the foot heater, as it lived under the table, though there was the one time it lived up on a high shelf, which was very firmly attached to the brick wall, and I could stand on the shelf no problem.

[@xDevscom_EE]

After more testing I can actually see small tone at 45.27kHz from one unit and 70kHz in second. Only when DCV 0V mode used and in high resolution. Reasonable suspect would be PWMDAC circuitry used to generate high resolution output in 3245A.

[@ryebis]

@@xDevscom_EE Interesting, I would've assumed the low pass filter stages in the 3245A output path to remove such a noise.

[@Slesar.]

Могу ошибаться, но мне видится что сейчас любой осциллограф заменяет этот прибор, так как имеет измеритель RMS, в зависимости от разрядности осциллографа до нескольких знаков после точки рассчитывает RMS сигнала отображаемого на экране и конечно же предоставляет лучшее представление о сигнале в сравнении с этим прибором.

[@PlasmaHH]

It seems to me that if you pluck 0Vrms into the accuracy formula of the specs, the instrument is still within it

[@itsBlueCave]

You said earlier that the relays were working fine , why did you changed then thanks

[@Thesignalpath]

I didn't say the relays were good.

[@karlbesser1696]

And what about analog and digital modulated signals in the upper frequency range?

[@jimomertz]

Seems strange to me that they designed it with a combination of DIP and surface mount. Why wouldn’t they just upgrade everything to surface mount? Probably could have fit everything on one side of the board.

[@oliverer3]

The only reason I can think of is that it would make mounting components on both sides of the board easier if one side is through hole.

[@steveoffer2917]

Should that display have a backlight? I think they have an incandescent that always fails. Maybe an LED replacement could be made. The meter is helpful when used in conjunction with a scope and low lighting.

[@Thesignalpath]

Only the URE3 model has the backlight. URE2 unfortunately does not.

[@simontay4851]

Well then why not fit your own backlight to your URE2? It would be easy for you.

[@Thesignalpath]

I will definitely look into it. It would make the instrument a lot better for sure.

[@steveoffer2917]

@@Thesignalpath Awesome, much appreciated.

[@simsys_outdoor]

Better replace also this few silver electrolytic caps (simens type). They are always faulty after this long time..

[@R2AUK]

9:30 Are these -48.9 dBm at 26 kHz still there if you turn off the HP instrument?

[@Thesignalpath]

No, they are gone.

[@bambumbambu]

do you also have the schematic for it?

[@aliemlek]

👍👍👌👌

[@foobarables]

How is the RMS conversion done in this instrument?

[@douro20]

I'm not sure since I currently can't find a schematic but I think it may be using a FET-based analog converter like the original URE. I don't see any dedicated RMS converter IC or thermal converter, and I highly doubt that digital conversion is being done since this is a true RMS instrument.

[@NiHaoMike64]

@@douro20 I can see a hybrid analog/digital approach making sense. Square and average in analog (these require a lot of bandwidth), take the root in digital (this only needs to be done as often as you need a new reading). While you're at it, you can also compensate for analog nonidealities in digital and/or split up the averaging into an analog and digital stage to eliminate the need for a high value, very low leakage capacitor in the filter.

[@mp1454]

🐈♥️👍🏻👍🏻👍🏻