Help on testing a HP 355C Step Attenuator

[Tony_G]

Hi All,

I just recently acquired a set of HP Step Attenuators (355C & 355D). I wanted to check them out so I downloaded the manual from Keysight and I'm trying to follow the quick check procedure for the 355C:



The problem that I'm running into is that with the 415E range set to 10dB and 0dB expand the meter is pegged all the way over to the right - even with the gain & vernier turned fully counterclockwise.

I don't have a 651B Test Oscillator so I'm trying to duplicate it with my DG1032Z Arb. I have that set to an output load of 50 ohms, connected to an inline 50 ohm terminator, set to 1 KHz sine wave at 0.3 Vrms (verified on my DMM) and connected to the 355C.

It looks like it's a typo but if it is then they committed to it on table 3 (next page, where you write down the values).

So my question(s) is, what am I doing wrong? Is it really just a typo or am I just not duplicating the 651B properly?

Any help appreciated.

Thanks,

TonyG

[cncjerry]

Approx. .223v RMS over 50 ohms is 1mw or 0dBm.  I would get the swr meter reading correctly first assuming your meter goes that low.

Instead of using the swr meter just loop it back thru the attenuator to your scope.  Then check the RMS value on the scope calculating the dBm value as you switch the attenuator.   Your scope might have dBm as a measurement value, probably does.

[Tony_G]

Thanks CNC for the suggestion - I just gave it a try.

Unfortunately my scope only goes down to 1mV and my Signal Generator only goes to +20dBm so the best I can read down to is a -80dBm with any "accuracy". I think this is the reason they're using the 415E as it has 60dB of amplification on it.

Actually I might just try that and see what happens - Thanks again for the suggestion.

TonyG

[Tony_G]

I just tried this and it looks like I'm running into the same problem I had originally (before I put this post up). The problem is that I can get accurate numbers down to 90dB attenuation but then it just flattens out.

With the way the 355D is built by the time I get to 90db I have tested each individual attenuation stage - I'm either doing something wrong or I'm just running into the noise floor limit of my setup here (which I wouldn't be surprised on).

TonyG

[Tony_G]

To step back a bit, I recently acquired a set of 355C & D step attenuators. I tried doing the test procedure for both and the 355C passed fine (with changes to the test procedure) but I've been unable to get the 355D to be close to in spec for the 100-120dB ranges.

I tried a lot of stuff and eventually I thought I should try to get the exact procedure in the manual to work with the 355C as that will validate my setup.

I now have a working, in spec 651B Test Oscillator so I set up the procedure as described in the image I included in the original post. I have exactly the same problem. So it must be a typo and not specific to test equipment. If I set the range to 0dB and use the Expand know to go from 0 - 4dB it gives me values that I would expect for the 355C.

I'm just taking another look at the 415E and I'll run through the tests again.

[Tony_G]

Using CNC's suggestion I just measured the step attenuator directly against the output of the 651B using my DM3058.

The spec says that it is +/- 0.1dB at 1KHz and the largest variance I measured was just under an order of magnitude less than that. So I can say that the 355C is bang on.



Next step is to try and look at the 355D but I suspect without using the test setup they have in the manual the 90dB+ ranges will be too small for my gear to measure.

TonyG

[nctnico]

Don't measure using an oscilloscope but an RMS voltmeter and a frequency of 1kHz. 0.1dB flatness is around 1% error so an osciloscope just isn't accurate enough by itself.

[TimFox]

If you believe the attenuator is working properly for AC, you can verify the attenuation values using a DC source and an accurate voltmeter.  Remember to terminate the attenuator with an accurate 50 ohms (which need not be in an RF package for the DC measurement) and do not exceed a DC voltage input that is larger than the RMS maximum input specification.

[nctnico]

Using AC is easier especially if you have a DMM which can show the difference between 2 signals and show that in dB directly.

[Tony_G]

Thanks guys for the suggestions - I can read accurately to 80dB and I get a 90dB reading but I don't have the resolution on my DMM to go lower (it can only read 10's of uV). Unfortunately I don't have anything that can get the input signal up above +23 odd dBm.

The DC spec maxes out at 5V which wouldn't be enough extra to overcome my resolution issue.

[ywara]

Remember that there are only 4 attenuator elements in these devices.

If all 4 elements check out good (i.e. with that element switched in, it performs correctly), then there is no need to check every step.

[TimFox]

Thanks guys for the suggestions - I can read accurately to 80dB and I get a 90dB reading but I don't have the resolution on my DMM to go lower (it can only read 10's of uV). Unfortunately I don't have anything that can get the input signal up above +23 odd dBm.

The DC spec maxes out at 5V which wouldn't be enough extra to overcome my resolution issue.

90 dB below 5 V DC is 158 uV.  10 uV resolution is about 6% of that, or 24 dB below that value.
Note:  if you use audio-frequency AC to measure the attenuator, you must be very careful about grounding.  If both the AC source and the AC voltmeter are grounded, you will see too much output at high attenuation.  Probably best to use a battery-powered voltmeter.