40kV DC hv probe frequency response - UK Vintage Radio Repair and Restoration Discussion Forum

trobbins · 19th Oct 2020, 3:32 am

I recently got around to checking out an AVO RM215-L/2 AC/DC Breakdown Leakage and Isolation Tester using two vintage 40kV DC rated 1000:1 probes.

The HP 34111A 40kV probe is speced within +/-2% for 20-30C range when used with a 10M meter input resistance, and comprises a 999Meg with 1.11Meg divider. The RS 610-281 40kV probe doesn't have a spec I could find on-line, but appears to be of similar construction although the 1.11M is provided by a 910K + 470k trimpot, so perhaps trimmed for good initial accuracy, but a concern for aging and temperature and humidity.

Thanks to Bill and Ed (https://www.vintage-radio.net/forum/...ferrerid=13948) I was able to confirm all the functional and metering operation of the AVO was ok using 12VAC accessory supplies before firing up the HV output.

That gave me a HV supply to check the operation of the two 40kV DC probes I had, both on DC and 50Hz AC, and to cross-compare the AVO volt meter against the two probes. Although one HV probe requires 10Meg loading, I connected each probe with a cheapo Aneng 8009 DMM which I know are damn accurate at DC and up to 1kHz ACV, and have 11Meg DCV input resistance and possibly 10Meg for ACV input impedance. An advantage of the Aneng meters is that they are small battery operated units, so somewhat safer than using a mains powered DMM when probing HV.

Both probes and the AVO front panel meter measured the same within 1% resolution for ACV, and the probes both measured 8% high for DCV (which also measured about 5% ACV due to ripple from the doubler rectifier in the AVO). With 11 Meg loading on the HP 34111A HV probe, the meter input is nominally 1% high compared to a 10Meg rated load. The front panel AVO meter is likely to have some error as I was only comparing 1kV on a 4kVFS meter, so that cross-comparison all appeared fine for now.

But it got me thinking as to what the AC frequency response may be like, even though the probes aren't specified for ACV accuracy. So I used a 10Vrms sinewave generator (Wavetek 111) to energise the primary of a Parmeko power transformer with 1.45kV secondary and cross-compared ACV probe measurements with some lab meters that can cope with up to 750Vrms and have a wide frequency range. I was able to generate about 400-600Vrms for testing at frequencies up to 1kHz easily, so the frequency response of the Aneng meters was not going to be an influence.

The HP probe gave a slightly higher frequency of 90-100Hz for a 5% drop in measurement, compared to 80-90Hz for the RS probe, when using the Aneng 8009 meters. Those meters don't have a spec for input capacitance loading, which I presumed would have an impact on the frequency response, so I pulled out the only other meter I have for comparison purposes - a HP 3400A, which has a 10Meg and <50pF input spec (the spec is 20pF for 1V and above range).

It would appear that the input capacitance loading of the HP3400A was lower than the Aneng, as the -5% voltage reading drop occurred at about 120Hz for the HP 34111A probe, and about 200Hz for the RS probe.

So the HV probes appear to be fine for ACV measurement with lowish frequency, as per mains AC frequency, or single phase rectified ripple, but perhaps a concern for accurate 3ph rectified ripple on HVDC.

Keith956 · 19th Oct 2020, 9:07 am

Scope probes normally have a capacitive divider that sets the AC response, as well as the resistive divider for DC response. If these probes are only intended for DC - which they probably are - then what you are seeing is quite believable.

THere's a good article on scope probes here - http://www.dfad.com.au/links/THE%20S...ES%20OCt09.pdf

trobbins · 19th Oct 2020, 10:44 am

Yes I believe the high frequency HV probes embed a metal tube/sleeve in the external insulator to provide distributed capacitance to the central high resistance core extending to the tip. Some one x-rayed a HP probe I recall, and I believe a few have tried to 3d print some probes although the concern has been that the available plastic is not really appropriate for higher voltages.

The RS probe handle was removeable, and sort of showed a long central foil extending in to the insulated tube section, but the unit was otherwise not designed to be dismembered. The HP probe also wasn't obviously going to pull apart. I don't really have a reason to try and make up an insulated metal sleeve to investigate how much capacitive balance could be achieved, so just happy to see what upper frequency accuracy could be achieved and what the Aneng 8009's could achieve compared to a known low capacitance input meter.

19th Oct 2020, 3:32 am	#1
trobbins Heptode Join Date: Oct 2009 Location: Melbourne Australia Posts: 901	40kV DC hv probe frequency response I recently got around to checking out an AVO RM215-L/2 AC/DC Breakdown Leakage and Isolation Tester using two vintage 40kV DC rated 1000:1 probes. The HP 34111A 40kV probe is speced within +/-2% for 20-30C range when used with a 10M meter input resistance, and comprises a 999Meg with 1.11Meg divider. The RS 610-281 40kV probe doesn't have a spec I could find on-line, but appears to be of similar construction although the 1.11M is provided by a 910K + 470k trimpot, so perhaps trimmed for good initial accuracy, but a concern for aging and temperature and humidity. Thanks to Bill and Ed (https://www.vintage-radio.net/forum/...ferrerid=13948) I was able to confirm all the functional and metering operation of the AVO was ok using 12VAC accessory supplies before firing up the HV output. That gave me a HV supply to check the operation of the two 40kV DC probes I had, both on DC and 50Hz AC, and to cross-compare the AVO volt meter against the two probes. Although one HV probe requires 10Meg loading, I connected each probe with a cheapo Aneng 8009 DMM which I know are damn accurate at DC and up to 1kHz ACV, and have 11Meg DCV input resistance and possibly 10Meg for ACV input impedance. An advantage of the Aneng meters is that they are small battery operated units, so somewhat safer than using a mains powered DMM when probing HV. Both probes and the AVO front panel meter measured the same within 1% resolution for ACV, and the probes both measured 8% high for DCV (which also measured about 5% ACV due to ripple from the doubler rectifier in the AVO). With 11 Meg loading on the HP 34111A HV probe, the meter input is nominally 1% high compared to a 10Meg rated load. The front panel AVO meter is likely to have some error as I was only comparing 1kV on a 4kVFS meter, so that cross-comparison all appeared fine for now. But it got me thinking as to what the AC frequency response may be like, even though the probes aren't specified for ACV accuracy. So I used a 10Vrms sinewave generator (Wavetek 111) to energise the primary of a Parmeko power transformer with 1.45kV secondary and cross-compared ACV probe measurements with some lab meters that can cope with up to 750Vrms and have a wide frequency range. I was able to generate about 400-600Vrms for testing at frequencies up to 1kHz easily, so the frequency response of the Aneng meters was not going to be an influence. The HP probe gave a slightly higher frequency of 90-100Hz for a 5% drop in measurement, compared to 80-90Hz for the RS probe, when using the Aneng 8009 meters. Those meters don't have a spec for input capacitance loading, which I presumed would have an impact on the frequency response, so I pulled out the only other meter I have for comparison purposes - a HP 3400A, which has a 10Meg and <50pF input spec (the spec is 20pF for 1V and above range). It would appear that the input capacitance loading of the HP3400A was lower than the Aneng, as the -5% voltage reading drop occurred at about 120Hz for the HP 34111A probe, and about 200Hz for the RS probe. So the HV probes appear to be fine for ACV measurement with lowish frequency, as per mains AC frequency, or single phase rectified ripple, but perhaps a concern for accurate 3ph rectified ripple on HVDC. Attached Thumbnails

19th Oct 2020, 9:07 am	#2
Keith956 Heptode Join Date: Jun 2019 Location: Oxfordshire Posts: 737	Re: 40kV DC hv probe frequency response Scope probes normally have a capacitive divider that sets the AC response, as well as the resistive divider for DC response. If these probes are only intended for DC - which they probably are - then what you are seeing is quite believable. THere's a good article on scope probes here - http://www.dfad.com.au/links/THE%20S...ES%20OCt09.pdf

19th Oct 2020, 10:44 am	#3
trobbins Heptode Join Date: Oct 2009 Location: Melbourne Australia Posts: 901	Re: 40kV DC hv probe frequency response Yes I believe the high frequency HV probes embed a metal tube/sleeve in the external insulator to provide distributed capacitance to the central high resistance core extending to the tip. Some one x-rayed a HP probe I recall, and I believe a few have tried to 3d print some probes although the concern has been that the available plastic is not really appropriate for higher voltages. The RS probe handle was removeable, and sort of showed a long central foil extending in to the insulated tube section, but the unit was otherwise not designed to be dismembered. The HP probe also wasn't obviously going to pull apart. I don't really have a reason to try and make up an insulated metal sleeve to investigate how much capacitive balance could be achieved, so just happy to see what upper frequency accuracy could be achieved and what the Aneng 8009's could achieve compared to a known low capacitance input meter.