Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
- HP-Agilent-Keysight-Equipment
- Messages
Search
|
Re: HP-3456A - Question Regarding Resistance Measurement
What are the specs of the meter? Generally the divider has to be considerably more accurate than the overall meter specs. Its possible you are worrying about an error that doesn't really matter.
toggle quoted message
Show quoted text
?? If you have a precision source of voltage for calibration you can tell if the meter is within specs, perhaps more easily than if the resistor is in specs. On 10/12/2020 1:25 PM, n4buq wrote:
The 90M is the last resistor in the chain for the X10M OHMS range of a VTVM (a Technology Instrument Corporation (aka Acton) 800A). --
Richard Knoppow dickburk@... WB6KBL |
|
Re: HP-3456A - Question Regarding Resistance Measurement
The drift in carbon comp resistors can happen just sitting around. Depends on how much heat and moisture they are subject to, especially moisture. The AB/Ohmite resistors seem to be pretty stable with age, some others are not. When using old resistors it is always wise to measure them. There are a very few applications where comp resistors are superior to film resistors but only a few. The idea that comp resistors have less reactance is not true. I measured many of every kind on a Boonton RX meter, which is intended for this sort of measurement, and found film resistors had consistently lower reactance. Carbon and metal film are very stable. Metal are more stable and can be made to have any temperature coefficient including zero over a wide range of T depending on the mix of metals used. Carbon has the TC of carbon (what else) but is not permanent (I have lost the proper term for this, reciprocal maybe). Carbon film resistors will also run at red heat without changing value even though it boils off the conformal coating. We never tested metal film that way but all of the precision resistors were subject to a seven times overload test before being spiraled and finished. I am sure resistors are made differently these days, probably laser cutters for spiraling instead of the diamond or carbide saws we used. They are certainly cheaper.
toggle quoted message
Show quoted text
On 10/12/2020 12:57 PM, doug wrote:
--
Richard Knoppow dickburk@... WB6KBL |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Its possible to measure the two in series and in parallel but series puts it in the next higher range of the meter with lower precision. Parallel also results in values which are on the edge of the specs.
toggle quoted message
Show quoted text
?? I don't think cleaning will do anything. I am not enthusiastic about isopropyl alcohol because it can leave a residue. I prefer naphtha, AKA Ronsonol. Since you are connecting directly to the meter terminals I doubt this is a problem. I think possibly you are running into the combined error of the meter and the resistor. If both are a bit high the combined error will be out of spec even though both meter and resistor are in spec. I downloaded the operating and service manuals but did not notice what -hp- requires for a test standard. I worked long ago for a company who made established reliability resistors. These were metal film resistors made for NASA contractors. We made up to something like 0.01%. We had precision bridges, a room full of NIST (NBS then) certified standards and a couple of precision ratio bridges. Not trivial. These are extremely high values and I don't know any tricks for checking calibration other than precision standards. The whole basis of resistance measurement has changed since I worked at it. ??? One idea is simply to use the resistors and see if the instrument they are in behaves correctly. On 10/12/2020 12:25 PM, n4buq wrote:
I placed the resistors directly across the jacks so no extra leads involved. I'll try cleaning and remeasuring including series and parallel measurements. --
Richard Knoppow dickburk@... WB6KBL |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Bob Albert
When measuring such high resistance, alternate current/leakage paths come into play.? A guarded measurement is a good way to reduce such problems. However, since your readings are high, it makes me think the problem is elsewhere.? Before deciding what's causing the anomaly you ought to make some more tests, on other close tolerance resistors. Put the two resistors in parallel and see if the reading is what you expect.? Then put them in series.? Add in other components and make those calculations. Realize also that the accuracy specification for the instrument grows to a rather poor figure for high resistances. Voltage coefficient of resistance may be causing some errors.? Did you monitor the temperature?? All resistors change value with temperature.? Select 6 digit reading and watch the display while you stand back and allow air currents to stabilize. In other words, it's not as simple as a casual resistance measurement. Bob
On Monday, October 12, 2020, 11:54:59 AM PDT, n4buq <n4buq@...> wrote:
I have two brand new 1% tolerance, metal-film resistors that I am looking to use in a project.? One is 90M and the other is 50M and I'm using my HP-3456A to measure them. The 50M tests within 1% (around 49.8M). The 90M tests at just over 92M which is about 2% over nominal. 2-Wire and 4-Wire show about the same differences. My meter is not recently calibrated so I know I cannot really trust either measurement; however, does it make sense that if both resistors are indeed within 1% (as they should be given that they're new), would one test correctly and one test incorrectly? I realize that new parts may not be in spec and at least one of these may be an example of that but these were bought from reputable suppliers (e.g. not eBay purchases) so I suspect they are within tolerance and my meter may be giving me false readings. If both tested with the same relative difference (e.g. both showed +2%), then that might make better sense but I'm curious as to whether it's possible/probable that if both are in tolerance that the meter would show that kind of variance. Yeah, I know.? Weird question and maybe too many variables but thought I'd see what the experts say. Thanks, Barry - N4BUQ |
|
Re: HP-3456A - Question Regarding Resistance Measurement
I can tell you from experience that while the 3456A is great, it can give false readings if the power supplies aren't spot on.? Back when I only had one of them, I was working on a precision voltage standard (2.5V & 5V), and while the 5V was dead on, the 2.5V was off by .25V (I'm guessing now, because this was years ago).? My multimeters all said it was 2.5V, but I trusted the 3456A...? The manufacturer of the standard couldn't figure out why I was getting the strange value because (I think) one voltage was used as a reference for the other one.? I finally determined, with help from the group, that one of the caps in the power supply was bad, but it only effected a very small range (like 1.8 to 3.2V).? Before I did the repair, I stepped through the entire range with a voltage calibrator, and the 1.8-3.2V was the only place that was off.? I just happened to find it. -Dave
On Monday, October 12, 2020, 11:54:55 AM PDT, n4buq <n4buq@...> wrote:
I have two brand new 1% tolerance, metal-film resistors that I am looking to use in a project.? One is 90M and the other is 50M and I'm using my HP-3456A to measure them. The 50M tests within 1% (around 49.8M). The 90M tests at just over 92M which is about 2% over nominal. 2-Wire and 4-Wire show about the same differences. My meter is not recently calibrated so I know I cannot really trust either measurement; however, does it make sense that if both resistors are indeed within 1% (as they should be given that they're new), would one test correctly and one test incorrectly? I realize that new parts may not be in spec and at least one of these may be an example of that but these were bought from reputable suppliers (e.g. not eBay purchases) so I suspect they are within tolerance and my meter may be giving me false readings. If both tested with the same relative difference (e.g. both showed +2%), then that might make better sense but I'm curious as to whether it's possible/probable that if both are in tolerance that the meter would show that kind of variance. Yeah, I know.? Weird question and maybe too many variables but thought I'd see what the experts say. Thanks, Barry - N4BUQ |
|
Re: HP-3456A - Question Regarding Resistance Measurement
The 90M is the last resistor in the chain for the X10M OHMS range of a VTVM (a Technology Instrument Corporation (aka Acton) 800A).
toggle quoted message
Show quoted text
The 50M is part of the input resistive divider for ACV and DCV functions. Thanks, Barry - N4BUQ ----- Original Message -----
From: "ArtekManuals" <manuals@...> |
|
Re: HP435 measuring head? kW's?
I'm not selling anything. Consider that the 11683A option H01 has a modified 8481A with its sensor replaced with a 196 ohm resistor.? This is fed by a 196 K resistor that is either connected to the switch OR to an external voltage source.? The modified 8481A is merely acting as a chopper, switching back and forth between the voltage across the 196 ohm resistor and ground. Of course, it's possible that an overload blew the FET chopper, but if your meter zeros and you can calibrate to 1 mW with 0.14493 V at the 196 K resistor, I'd consider it highly unlikely.? It's also possible that the FET balance was upset when modifying the 8481A, but there are procedures to check/fix that. With my homebrew 11683A-H01, my 436A gave surprisingly close readings to the theoretical ones for the voltage applied.? I think the largest deviation?was 0.02 dB.? I got similar results with a 438A.? I cross-checked with a real 8482A and some attenuators using the 436A's 1 mW output.? Results again were within the tolerance of the attenuators. I have no reason to believe that this modification of an 8481A with a blown sensor doesn't work. As far as vendors with blown 8481As are concerned, they ask way too much for them.? They are practically worthless given that 'parkertest' charge pretty much the same to repair one ($465) as to sell a refurbished one ($475?obo).? I paid about $50 shipped for the dead 8481A that I converted.? FWIW, I've returned two 'working' 8481As in the past - one just plain dead and the other that didn't meet SWR specs. Maybe one of those converted 8481H sensors will work for $150... Orin. On Mon, Oct 12, 2020 at 10:47 AM Roy Thistle <roy.thistle@...> wrote: On Sat, Oct 10, 2020 at 10:46 AM, Dave Smith W6TE wrote: |
|
Re: HP-3456A - Question Regarding Resistance Measurement
On 10/12/20 3:13 PM, george edmonds via groups.io wrote:
Hi BarryMy only experience with this is of a LARGE change in value, of a screen-grid resistor. Does this actually happen with a resistor that is not in use? --doug, WA2SAY
|
|
Re: HP-3456A - Question Regarding Resistance Measurement
I think you need to watch out on the high resistance ranges of the 3456A, I believe the quoted accuracy is noticeably worse above 10M, for instance, if you look at p1-5 of the operation manual you'll see that the 24hr accuracy jumps from <.042% of reading for R=<10M to 1.3% for 10-100M and 1.8% 100-1G
toggle quoted message
Show quoted text
The resistance cal procedure has no adjustment above 10M as I recall, I have a couple of them and made a DIY resistance calibration box from some high stab metal foil resistors a while ago but I seem to remember that the two highest ranges (100M and 1G ohm) were just a reading check? Adrian On 12/10/2020 19:54, n4buq wrote:
I have two brand new 1% tolerance, metal-film resistors that I am looking to use in a project. One is 90M and the other is 50M and I'm using my HP-3456A to measure them. |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Sounds like since you cant trust the calibration since it is not current that this is all within the range of uncertainty
toggle quoted message
Show quoted text
I am pretty sure that Mouser and Digikey do not actually measure the value of 1% resistors going out the door. The resistors are likely made in China ( need I say more) This sounds like an analogy of the man with two watches "A man with one resistor always know exactly what the value is..a man with two resistors is never quite sure" What are you trying to do that requires 1% accuracy at these high values? Dave NR1DX On 10/12/2020 3:13 PM, Richard Knoppow wrote:
By 90M do you mean Megohms?? Do you have any other resistors of similar value and close tolerance, they can be used to check the meter. I assume they are expensive otherwise I suggest getting a second resistor. If megohms these are very high values. If the resistor had read low I would have suspected some leakage resistance but that would not make it read high.-- Dave Manuals@... www.ArtekManuals.com --
This email has been checked for viruses by Avast antivirus software. |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Not if they are metal film resistors, those are very stable both for age and temperature. The error is almost certainly not from either. I agree that contamination will make the value go down, assuming it is conductive along the body of the resistor.
toggle quoted message
Show quoted text
?? Almost all modern resistors are either carbon or metal film resistors. They are both quite stable although metal film are more stable with temp. The resistors that drift are carbon composition resistors. The resistance element is composed of particles of carbon suspended in a binder. With time the area of contact between the particles is reduced making the value increase. In general, the higher the value of the resistor the more it drifts but even low value resistors can drift a lot if subjected to some heat. The reason carbon comp resistors were never offered in tolerances of less than 5% is that the heat of soldering them can move them more than that.? Because moisture also causes them to drift the quality of the casing is important. Some brands, notably Allen-Bradley and Ohmite (the are actually all AB) are about the most stable. ?? I wonder if the OP has access to another meter or bridge to test the resistor. ??? Note the specs for the meter on its 100Meg range is 1.3%+1 count. So, its getting close to the meter's limit. On 10/12/2020 12:13 PM, george edmonds via groups.io wrote:
Hi Barry --
Richard Knoppow dickburk@... WB6KBL |
|
Re: HP435 measuring head? kW's?
¿ªÔÆÌåÓýI've converted a number of HP 11683As to 11683A-H01s over the last 15 years which KT charges a lot for but it amounts to almost nothing.? All they do is to put a bypass switch on the rear panel so that you can either use the internal voltage divider or
an external DC source.? Anyone can do it.? The DCV accuracy that you need is only something that is roughly .05%.? The problem with the switch isn't the switch itself but the linearity of the divider resistors.? Even a brand-new switch assembly from KT doesn't
always meet that spec and neither do a lot of 11683As.? It's easier and more accurate if you use an external voltage source & a 848x sensor assembly although the 11683A is more convenient.
Richard
?
From: [email protected] <[email protected]> on behalf of Roy Thistle via groups.io <roy.thistle@...>
Sent: Monday, October 12, 2020 12:47 PM To: [email protected] <[email protected]> Subject: Re: [HP-Agilent-Keysight-equipment] HP435 measuring head? kW's? ?
On Sat, Oct 10, 2020 at 10:46 AM, Dave Smith W6TE wrote:
Building a 11683A isn¡¯t as simple as using an old 8481A and a few precision resistors.You know...I'm glad you said this... and I enjoy that guy's blog... but, there are only not very loud crickets, when it comes to posts/reports of whether home-brew 11683As actually function as intended. (I'd compare this project to many of the old Radio Electronics magazine projects that... allegedly... didn't live up to the specs... and often needed a rework, or an editorial insertion, in subsequent issues, of the mag. These "projects" are consumer entertainment for perusers. And, they're candidates for future projects: projects fun to imagine, but never actually constructed. ) That conjectured... I'll take a different tack... and would like to pose... allegedly... how many of the posters here (and elsewhere... i.e. YouTube)... are making claims considering the validity of used, and cheap, "heads" (i.e. H.P. 8080 coaxial power sensors)... how many are making claims because they are vendors up selling junk? |
|
Re: HP-3456A - Question Regarding Resistance Measurement
I placed the resistors directly across the jacks so no extra leads involved. I'll try cleaning and remeasuring including series and parallel measurements.
toggle quoted message
Show quoted text
Thanks, Barry - N4BUQ ----- Original Message -----
From: "Froggie the Gremlin" <jonpaul@...> |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Yes - I wasn't thinking correctly but contaminants would cause lower readings.
toggle quoted message
Show quoted text
They're supposed to be new resistors (from Mouser and Digikey). Thanks, Barry - N4BUQ ----- Original Message -----
From: "george edmonds via groups.io" <G6HIG@...> |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Yes - 90 Megohms and 50 Megohms. I do have some other new, 1% resistors of slightly lower value (e.g. 9M and 900k) and I do have an older meter calibrator (Fluke 760A) capable of 10M but, like my other gear, is not calibrated.
toggle quoted message
Show quoted text
You're correct, though, that if there were contaminants on the 90M, it would have read lower, not higher. I should clean both with IPA and retest. These are the resistive dividers for the OHMS function an old VTVM I'm restoring and so hope to have them as close as possible to their nominal values. Thanks, Barry - N4BUQ ----- Original Message -----
From: "Richard Knoppow" <dickburk@...> |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Hi Barry Contamination will make resistors go down in value not up. How old are these resistors, are they NOS. Resistors invariably go up in value with time, especially if over 10M. 73 George G6HIG Dover UK On Monday, 12 October 2020, 20:06:34 BST, n4buq <n4buq@...> wrote: Not in a strict sense.? Both were at relatively the same ambient temperature (they'd been in the same environment before testing).? I mostly held them by their leads but did notice that I could change the resistance slightly by touching them which I sort-of expected given the high resistance of either one. This wasn't exactly a laboratory-controlled experiment and there could have been some temperature differences between the two but I'd suspect it wasn't that great. I wonder if I need to clean them a bit better before testing them again.? I presume hand oils might affect these with the 90M seeing more of that than the other. Thanks, Barry - N4BUQ ----- Original Message ----- > From: "Bruce" <bruce@...> > To: [email protected] > Sent: Monday, October 12, 2020 1:58:59 PM > Subject: Re: [HP-Agilent-Keysight-equipment] HP-3456A - Question Regarding Resistance Measurement > > Did you considet temperature coefficient > Quoting n4buq <n4buq@...>: > > > I have two brand new 1% tolerance, metal-film resistors that I am > > looking to use in a project.? One is 90M and the other is 50M and > > I'm using my HP-3456A to measure them. > > > > The 50M tests within 1% (around 49.8M). > > > > The 90M tests at just over 92M which is about 2% over nominal. > > > > 2-Wire and 4-Wire show about the same differences. > > > > My meter is not recently calibrated so I know I cannot really trust > > either measurement; however, does it make sense that if both > > resistors are indeed within 1% (as they should be given that they're > > new), would one test correctly and one test incorrectly? > > > > I realize that new parts may not be in spec and at least one of > > these may be an example of that but these were bought from reputable > > suppliers (e.g. not eBay purchases) so I suspect they are within > > tolerance and my meter may be giving me false readings. > > > > If both tested with the same relative difference (e.g. both showed > > +2%), then that might make better sense but I'm curious as to > > whether it's possible/probable that if both are in tolerance that > > the meter would show that kind of variance. > > > > Yeah, I know.? Weird question and maybe too many variables but > > thought I'd see what the experts say. > > > > Thanks, > > Barry - N4BUQ > > > > > > > > > > > > > > > |
|
Re: HP-3456A - Question Regarding Resistance Measurement
By 90M do you mean Megohms?? Do you have any other resistors of similar value and close tolerance, they can be used to check the meter. I assume they are expensive otherwise I suggest getting a second resistor. If megohms these are very high values. If the resistor had read low I would have suspected some leakage resistance but that would not make it read high.
toggle quoted message
Show quoted text
? It IS possible for a new resistor to be off value. On 10/12/2020 11:54 AM, n4buq wrote:
I have two brand new 1% tolerance, metal-film resistors that I am looking to use in a project. One is 90M and the other is 50M and I'm using my HP-3456A to measure them. --
Richard Knoppow dickburk@... WB6KBL |
|
Re: HP-3456A - Question Regarding Resistance Measurement
Not in a strict sense. Both were at relatively the same ambient temperature (they'd been in the same environment before testing). I mostly held them by their leads but did notice that I could change the resistance slightly by touching them which I sort-of expected given the high resistance of either one.
toggle quoted message
Show quoted text
This wasn't exactly a laboratory-controlled experiment and there could have been some temperature differences between the two but I'd suspect it wasn't that great. I wonder if I need to clean them a bit better before testing them again. I presume hand oils might affect these with the 90M seeing more of that than the other. Thanks, Barry - N4BUQ ----- Original Message -----
From: "Bruce" <bruce@...> |
Froggie the Gremlin
to navigate to use esc to dismiss