I just realized I was getting unusual measurements due to the circuit being tested. I was getting some type of ac interference until the circuit completely warmed up(30 minutes). I have to trace the problem down now but I feel much better about the meter. Thanks for the input everyone-It's good to know the fluke 87 is such a reliable meter.
Hi RocketScientist,
It's been said, but your Fluke is designed to arc (spark gaps) a little over 1 KV. That is a valid safety measure. If you are getting close to the limits and are unsure, use an HV probe. One that has been properly certified. The voltage coefficient that resistors have in those is very noticeable. They must actually either apply very high voltage (depending on the rating), or apply a known correction factor if the limit they have is only 1 KV.
Fluke meters also do have a controlled input impedance. Many other DVMs do not, so using an accurate HV probe with those is a waste of time. Calibration "as a system" is the best way to go in those cases.
The isolated inputs on the Scopemeter are great. I just never did like them, compared to the price. One of my least favorite calibration jobs were Scopemeters. Humidity is another I don't care for too much.
Man, you have some nice gear there. The MSO I am familiar with, and it's the easiest DSO to operate for old guys like me. They "feel" right, and they can display an eye pattern about the best - excepting the upper models. Your signal generator is also great. Getting the signals you need is now pretty easy. I'm using dedicated pulse generators for pulses, audio and RF oscillators for those signals. I think you can generate a 10.7 MHz signaland modulate it with that generator.
As for the Tektronix scopes, I've used both analog HP and Tek. I have to say I like the HP really well. I have hopes for an old 465 to restore, and I have an HP 1722A to restore as well. I think they both made excellent oscilloscopes. Advertising probably made the difference. I borrowed a Tek DPO 4000 series. The fan noise and excessive heat coming out bothered me. The DPO is also an odd shape that will not fit where a 'scope should go. The Agilent DSO/MSO 6000 product is longer, quieter, cooler and fits in a 'scope space. Operationally, they are both fine (the Agilent is easier for me to use) and I expect will last the 15 ~ 20 years we all expect from those product types.
As for low end products, well ... they had to do something to offer in the low end market. Rigol is a decent product. Obviously they will not compete with Agilent, they aren't supposed to. I tried a Rigol out and liked it more than the Tek 3000 series.
You know who else has a really nice scope (+ easy to use)? LeCroy. It performed well and was intuitive when using it. I found it easier to use than the Tek MSO 4000.
I had a week each with the Agilent MSO6000 and also the Tek MSO 4000. I tried all the analog things I normally do. That is difficult for digital 'scopes due to the sampling. For analog performance with a clear displayed image, you really do need a screaming high sampling rate. So it really became a test of what I could coax out of each instrument. The initial set up on the network was dead easy for most, especially for the Tek and Agilent. Retrieving information from these were easy except for units not network capable. Each would store setups on an external resource as well as internally. Personally, I love that feature! Getting to the nitty-gritty of proper triggering in order to capture certain wave forms was not a trivial task. It was easier on the Agilent, and I did get a passable (just usable) display. It was apparent that an analog 'scope is still required on the bench. Although I could suffer though using the Agilent 6000, a higher end model is really what you would need to align a CD player for example. The displays on the Tek 4000 were not good enough to use, but tantalizingly close.
In all fairness, I have had more exposure to Tektronix products, but I have a natural affinity for the Agilent product. They are both very good oscilloscopes that have advantages and weaknesses if you are coming from an analog oscilloscope. I think the one 'scope I have from Tektronix that I liked was a 350 MHz analog model. Of course, I can't remember the number. It went with the shop when I sold it.
-Chris
It's been said, but your Fluke is designed to arc (spark gaps) a little over 1 KV. That is a valid safety measure. If you are getting close to the limits and are unsure, use an HV probe. One that has been properly certified. The voltage coefficient that resistors have in those is very noticeable. They must actually either apply very high voltage (depending on the rating), or apply a known correction factor if the limit they have is only 1 KV.
Fluke meters also do have a controlled input impedance. Many other DVMs do not, so using an accurate HV probe with those is a waste of time. Calibration "as a system" is the best way to go in those cases.
The isolated inputs on the Scopemeter are great. I just never did like them, compared to the price. One of my least favorite calibration jobs were Scopemeters. Humidity is another I don't care for too much.
Man, you have some nice gear there. The MSO I am familiar with, and it's the easiest DSO to operate for old guys like me. They "feel" right, and they can display an eye pattern about the best - excepting the upper models. Your signal generator is also great. Getting the signals you need is now pretty easy. I'm using dedicated pulse generators for pulses, audio and RF oscillators for those signals. I think you can generate a 10.7 MHz signaland modulate it with that generator.
As for the Tektronix scopes, I've used both analog HP and Tek. I have to say I like the HP really well. I have hopes for an old 465 to restore, and I have an HP 1722A to restore as well. I think they both made excellent oscilloscopes. Advertising probably made the difference. I borrowed a Tek DPO 4000 series. The fan noise and excessive heat coming out bothered me. The DPO is also an odd shape that will not fit where a 'scope should go. The Agilent DSO/MSO 6000 product is longer, quieter, cooler and fits in a 'scope space. Operationally, they are both fine (the Agilent is easier for me to use) and I expect will last the 15 ~ 20 years we all expect from those product types.
As for low end products, well ... they had to do something to offer in the low end market. Rigol is a decent product. Obviously they will not compete with Agilent, they aren't supposed to. I tried a Rigol out and liked it more than the Tek 3000 series.
You know who else has a really nice scope (+ easy to use)? LeCroy. It performed well and was intuitive when using it. I found it easier to use than the Tek MSO 4000.
I had a week each with the Agilent MSO6000 and also the Tek MSO 4000. I tried all the analog things I normally do. That is difficult for digital 'scopes due to the sampling. For analog performance with a clear displayed image, you really do need a screaming high sampling rate. So it really became a test of what I could coax out of each instrument. The initial set up on the network was dead easy for most, especially for the Tek and Agilent. Retrieving information from these were easy except for units not network capable. Each would store setups on an external resource as well as internally. Personally, I love that feature! Getting to the nitty-gritty of proper triggering in order to capture certain wave forms was not a trivial task. It was easier on the Agilent, and I did get a passable (just usable) display. It was apparent that an analog 'scope is still required on the bench. Although I could suffer though using the Agilent 6000, a higher end model is really what you would need to align a CD player for example. The displays on the Tek 4000 were not good enough to use, but tantalizingly close.
In all fairness, I have had more exposure to Tektronix products, but I have a natural affinity for the Agilent product. They are both very good oscilloscopes that have advantages and weaknesses if you are coming from an analog oscilloscope. I think the one 'scope I have from Tektronix that I liked was a 350 MHz analog model. Of course, I can't remember the number. It went with the shop when I sold it.
-Chris
Hi imix500,
Watch out for Extech. My impressions from about 9 years ago were that they didn't hold cal very well, and the service rate for
Hi gto127,
Your '87 is an excellent meter. Take care of it. It will respond to HF noise as well, so that may have been your issue.
-Chris
You weren't the first and there is no way you will be the last. The one error we saw often was someone using the current scale and setting to measure the voltage on 550 V lines. That usually ends the life of the meter - permanently and spectacularly. I wish I had seen the incident with a few of these.
And there is the beauty of a Fluke DVM. They can be repaired for reasonable money normally. The newer models are closed case calibration and adjustment. Now that is the only way to live! Your optimizing requires no guesswork to get it close, it will be right on the money.
Watch out for Extech. My impressions from about 9 years ago were that they didn't hold cal very well, and the service rate for
only is 1/2 the dealer price or some such. It is overly expensive and only they could do the optimizing. Special program (like Fluke). Avoid.
Hi gto127,
Your '87 is an excellent meter. Take care of it. It will respond to HF noise as well, so that may have been your issue.
-Chris
Just for jollies -- there are a couple of real excellent application notes from Linear Technologies -- #86 "A Standards Lab Grade 20-bit DAC with 0.1ppm/C Drift", Ap Note #62 "Data Acquisition Circuit Collection", etc.
Here's a joint article written by an engineer from TI and Thaler: http://focus.ti.com/lit/an/slyt183/slyt183.pdf
I'm sure that there are a lot of vintage Hewlett Packard and Tektronix application notes, just a little bit harder to find. I have an early HP 16 bit DAC and each bit is individually adjustable!
Here's a joint article written by an engineer from TI and Thaler: http://focus.ti.com/lit/an/slyt183/slyt183.pdf
I'm sure that there are a lot of vintage Hewlett Packard and Tektronix application notes, just a little bit harder to find. I have an early HP 16 bit DAC and each bit is individually adjustable!
Fluke Meters
Anatech, thanks for the reply and tips. We're in agreement.
And back to the thread, if there's good reason to suspect the Fluke has a problem, I'd try the voltage reference trick. And if that gives close to the expected reading, I'd probably leave it alone. Of course simple DC voltage accuracy is different, as Anatech pointed out, from some of the other calibration issues like AC, frequency response, current, etc. But it would give you a valid "sanity check" for the meter.
My ancient 87 still does very well. Despite all the years and abuse, the only problem is a few display segments have different contrast. The display in my old 30+ year old Fluke 8022B is worse off but it still works just fine--especially for a very early digital meter from the 1970's! The 8000 series put Fluke on the map.
Just for fun, and confirmation of accuracy, here's a pic of the Fluke 8022B, 189 and 89 (left to right) in front of the Agilent 34410A all displaying the same 10 volt source. The 1970's fluke even has the "BT" low battery indicator on but is still fairly accurate after 30+ years without any calibration.
10.00 (8022B)
10.004 (189)
10.01 (89)
10.0069 (Agilent)
Anatech, thanks for the reply and tips. We're in agreement.
And back to the thread, if there's good reason to suspect the Fluke has a problem, I'd try the voltage reference trick. And if that gives close to the expected reading, I'd probably leave it alone. Of course simple DC voltage accuracy is different, as Anatech pointed out, from some of the other calibration issues like AC, frequency response, current, etc. But it would give you a valid "sanity check" for the meter.
My ancient 87 still does very well. Despite all the years and abuse, the only problem is a few display segments have different contrast. The display in my old 30+ year old Fluke 8022B is worse off but it still works just fine--especially for a very early digital meter from the 1970's! The 8000 series put Fluke on the map.
Just for fun, and confirmation of accuracy, here's a pic of the Fluke 8022B, 189 and 89 (left to right) in front of the Agilent 34410A all displaying the same 10 volt source. The 1970's fluke even has the "BT" low battery indicator on but is still fairly accurate after 30+ years without any calibration.
10.00 (8022B)
10.004 (189)
10.01 (89)
10.0069 (Agilent)
Attachments
Hi RocketScientist,
That's about what I would expect to see. You have a group of the better meters there. (I do hate the side buttons on the 802x series).
The displays only need the PCB, display and elastomers cleaned. That's about a 15 minute job for each, okay - maybe another 5 minutes for the 8022B. The display is a pain compared to the 87. How are the 189 models put together? About the same as the 87?
I'm liking the new 34401A! I'm jealous but can't complain about the service I have had with mine. Your point about the accuracy for these meters are what I've been trying to say for years. Stay out of them, don't drop them and they will reward you with great accuracy over the years.
I have seen Agilent's new hand-helds. I would definitely risk trying one out. They have great built in features as well, and a great display. Very easy to read.
-Chris
That's about what I would expect to see. You have a group of the better meters there. (I do hate the side buttons on the 802x series).
The displays only need the PCB, display and elastomers cleaned. That's about a 15 minute job for each, okay - maybe another 5 minutes for the 8022B. The display is a pain compared to the 87. How are the 189 models put together? About the same as the 87?
I'm liking the new 34401A! I'm jealous but can't complain about the service I have had with mine. Your point about the accuracy for these meters are what I've been trying to say for years. Stay out of them, don't drop them and they will reward you with great accuracy over the years.
I have seen Agilent's new hand-helds. I would definitely risk trying one out. They have great built in features as well, and a great display. Very easy to read.
-Chris
anatech said:
Thanks. The 189 is a mixed bag. It has tons of features the 87 lacks like much higher count/more accuracy, dB, nS, temp, frequency, memory/logging, multiple peak hold modes, wider frequency response, lower/higher ranges, etc. But, despite using 4 AA's instead of a 9v, the battery life isn't as good, and the user interface is a bit obtuse. It doesn't have a graphical display so you get some really cryptic messages on the 5 digit main display trying to set up some of the more advanced functions.
Build quality and mechanical design of the 189 seems every bit as good as the 87 and in some ways better. You don't have to remove screws or the entire back just to change the batteries. And the stand works much better than the rubbery one on the 87.
I haven't tried the new Fluke 289. It's expensive and probably has even less battery life. For me, the combination of the 189 and 192 Scopemeter will do most anything (and more) that the 289 can do. But I bet it's probably easier to use the more advanced functions on the 289 as it (finally) has a graphic LCD that can do multiple lines of text, prompts, etc.
The Agilent 34401A is great. My only gripe, like so much other new test equipment these days, is it has a fan. It's fairly quiet as fans go but it's still there. These days, with several instruments and power supplies running simultaneously, the test bench can get kinda noisy. We audiophiles don't like noise.
The quietest fan award goes to the Scopemeter. You have to put your ear to the case to hear it. It probably helps it doesn't have any vents anywhere. So, near as I can tell, the fan must just circulate the internal air--probably over the FPGA or whatever is moving samples at 500 Mhz.
Hi RocketScientist,
Agilent is normally good about considering noise from the instrument. Gee, mine doesn't have a fan at all. But then again, it's not a vacuum cleaner either. Do you put a bit of filter foam over your fan inlet? Might be a good idea if there isn't a filter in place already.
I think I'd be using a power supply for the 189 if available. Mind you, it may add enough noise to make the last digit meaningless. Have to check that.
The Scopemeters are a pain to get in and out of, not unless they made things easier. You probably don't want any vents either! An air temperature exchanger would work though.
Fluke is one company that knows what they are doing with hand-help products. These days I would want a Fluke or Agilent, and Escort would probably be good also. They also OEM for the big boys.
I don't think those 87'a will ever really die. It's a landmark instrument that all others have to measure up to. Just treat it well.
-Chris
Agilent is normally good about considering noise from the instrument. Gee, mine doesn't have a fan at all. But then again, it's not a vacuum cleaner either. Do you put a bit of filter foam over your fan inlet? Might be a good idea if there isn't a filter in place already.
I think I'd be using a power supply for the 189 if available. Mind you, it may add enough noise to make the last digit meaningless. Have to check that.
The Scopemeters are a pain to get in and out of, not unless they made things easier. You probably don't want any vents either! An air temperature exchanger would work though.
Fluke is one company that knows what they are doing with hand-help products. These days I would want a Fluke or Agilent, and Escort would probably be good also. They also OEM for the big boys.
I don't think those 87'a will ever really die. It's a landmark instrument that all others have to measure up to. Just treat it well.
-Chris
anatech said:
I *believe* the Scopemeters are actually a Philips design. I don't know if that was just the first generation, or if they're all designed by Philips, but I know it's *not* the same team in the northwest that designs the meters. I met an engineer who worked in the Fluke DMM group. And while paying him some compliments he told me about Philips and the Scopemeters. But I don't remember the exact story.
And, no, the 189 cannot be AC powered. I think you're correct it's hard to do that and maintain the specs in a low cost device. You also have to deal with the 1000 volt isolation requirement.
And virtually every instrument I have with a fan has it configured to exhaust. The air typically enters in several places. So filters are not really an option.
Anyway, yeah. It's nice still be able to buy products that are over engineered. They're getting more and more rare as everyone turns to (mostly) disposable produce from China Inc.
Hi Walter,
Should be an easy fix. Clean the elastomers with 99% Alcohol, and the PCB contacts. Was it dropped? It looks like the contacts are not aligned properly.
Clean the carbon connections in the elastomer very gently using a Q-tip or cotton swab. You should be able to buy replacements from Fluke in case you mess them up. Just work carefully, there is no need to scrub.
Hi umteman,
Not even close these days. Buy a really accurate voltage reference chip and use that instead. I shutter to think how old that mercury cell is! It was good for a VTVM calibration though.
-Chris
Should be an easy fix. Clean the elastomers with 99% Alcohol, and the PCB contacts. Was it dropped? It looks like the contacts are not aligned properly.
Clean the carbon connections in the elastomer very gently using a Q-tip or cotton swab. You should be able to buy replacements from Fluke in case you mess them up. Just work carefully, there is no need to scrub.
Hi umteman,
Not even close these days. Buy a really accurate voltage reference chip and use that instead. I shutter to think how old that mercury cell is! It was good for a VTVM calibration though.
-Chris
Hi Walter,
No problem at all.
You have a really nice meter there, please let us know how everything turns out. These are newer than the ones I worked on, but should be similar. The elastomer should be light gray in colour and fairly flexible. Contact is made by slight pressure from the display frame that lightly compresses the display glass (with conductive traces that are difficult to make out) towards the pads on the PCB. I'd like to know if this has changed at all. The total job should take less than 20 minutes from case off to case together, less time with practice. I hope you don't get that practice of course!
-Chris
No problem at all.
You have a really nice meter there, please let us know how everything turns out. These are newer than the ones I worked on, but should be similar. The elastomer should be light gray in colour and fairly flexible. Contact is made by slight pressure from the display frame that lightly compresses the display glass (with conductive traces that are difficult to make out) towards the pads on the PCB. I'd like to know if this has changed at all. The total job should take less than 20 minutes from case off to case together, less time with practice. I hope you don't get that practice of course!
-Chris
No, his best bet would be to leave it alone. He's more likely to screw it up than 'fix' it. Fluke is good stuff and I'd be more comfortable with a 20 year old Fluke that hasn't been 'diddled' than one that was 'fixed' by someone not knowing what he's doing.
When I was doing HD video design I was using the Analog Devices AD588 voltage reference. On initial test of a batch of boards I was reading 4.998 on 3 consecutive cards. VERY strange as they always read 5.000 so what was different? The Fluke 8060A I picked up on eBay so I opened it up and set it to read 5.000 but that is the only time I've ever 'tweaked' a meter in 35 years. Do NOT take this a criticism of eBay. They had nothing at all to do with the meter calibration. More likely the previous owner felt the need to 'fix' it.
When you get to the 500 Volt areas you have to be careful that the 11 megohm load of the meter isn't changing the real reading. If that's the case, high voltage probes change to loading to 1 gigohm at the expense of some accuracy. An old 50,000/volt analog meter loads 50 megs on the 1000 volt range.
I think you're referring to Weston cells as the reference.
G²
Hi stratus46,
Absolutely agree.
Hi Walter, All,
Fluke meters hold their calibration extremely well, as do HP / Agilent meters (they make hand held meters too). So, unless there is a particular need and proof that you need to calibrate even a 3.5 digit meter, don't touch it.
A couple points I've mentioned before but I'll bring them up again. With a higher end Fluke (or any other good meter), the high frequency calibration will vary depending on the distances between various parts on the PCB and the case. A professional calibration facility can not calibrate a Fluke 87 (for example) out of its case unless the technician throws the reading out during the calibration towards the direction of drift. Then it's just adjust, reassemble and check. Repeat if not in tolerance. What I did was to take a Fluke 87 top case and drill holes exactly over the adjustment locations. Then I could assemble the meter in it's own rear case and my adjustment jig. This allowed a much faster calibration that was also much closer to specification. Still the reading varied with the proper top cover, even if you remove it and reassemble without doing anything. So I still had to fudge the adjustment in the direction of drift.
Some of the newer meters are completely software adjustable, or "closed case calibration". This is a major improvement for calibration accuracy. Very clearly then, if you aren't using the proper calibration standards and sources, you'll only mess things up. This is true even if you have the right equipment but no calibration jig. Closed case calibration with stored data points for correction is definitely the way to go. These require special software and link cable.
Also realize that lower quality meters often fail their own calibration limits - brand new in the box! A major difference here is in the input divider. The better products put this on one thick film assembly that lowers capacitance and allows the elements to track thermally. Lower quality meters use individual resistors. Some of these are normal 1/8 watt, 5% (!) resistors. Even 1 % resistors can create large deviations between ranges since worst case for in tolerance parts is 2%. What this means is that often the less expensive meters are "Beyond Economic Repair", or out of tolerance and not worth correcting the problem. Can the home hobbyist do this? Yes. Service time is your enemy for charges. Just order really accurate resistors, non-inductive metal film suggested, and rebuild the divider. Then either use an accurate voltage reference chip, or send it in for calibration (about $100 USD for something basic).
I think it pays to buy a minimum of one really good meter. You can compare other, cheaper meters against one you know is correct. Don't expect good accuracy on the AC ranges at frequencies well above 120 Hz. A reading at 1 KHz could be pure fiction unless the meter is rated for that range. It's sometimes just the nature of the beast that some meters will not ever really be accurate. They are still useful as long as you remember that 11.23 VDC really means "about 11 and a quarter volts". Don't get trapped into the thought that because there are digits on the display that they are therefore accurate. Some analog meters are more accurate than many digital meters. How sad is that? Many older HP meters are like this. Also, the analog meter may use a DC to RMS converter chip (HP and others), whereas most digital meters only peak detect the waveform and report a lower figure that represents what the RMS reading for a low distortion sine wave would have read.
An example I can offer is an HP 974A meter I bought recently. It's rated to a frequency of 100 KHz, and a DC accuracy of 0.05% (look it up!) with a 4 1/2 digit display. This is an old meter I bought at auction. Checking it against a bench meter of better accuracy shows that the calibration is still good. That's great because I would have to send it in for a good calibration.
It's difficult to avoid turning controls in a meter for some people. Restraint here is helpful.
-Chris
Absolutely agree.
Hi Walter, All,
Fluke meters hold their calibration extremely well, as do HP / Agilent meters (they make hand held meters too). So, unless there is a particular need and proof that you need to calibrate even a 3.5 digit meter, don't touch it.
A couple points I've mentioned before but I'll bring them up again. With a higher end Fluke (or any other good meter), the high frequency calibration will vary depending on the distances between various parts on the PCB and the case. A professional calibration facility can not calibrate a Fluke 87 (for example) out of its case unless the technician throws the reading out during the calibration towards the direction of drift. Then it's just adjust, reassemble and check. Repeat if not in tolerance. What I did was to take a Fluke 87 top case and drill holes exactly over the adjustment locations. Then I could assemble the meter in it's own rear case and my adjustment jig. This allowed a much faster calibration that was also much closer to specification. Still the reading varied with the proper top cover, even if you remove it and reassemble without doing anything. So I still had to fudge the adjustment in the direction of drift.
Some of the newer meters are completely software adjustable, or "closed case calibration". This is a major improvement for calibration accuracy. Very clearly then, if you aren't using the proper calibration standards and sources, you'll only mess things up. This is true even if you have the right equipment but no calibration jig. Closed case calibration with stored data points for correction is definitely the way to go. These require special software and link cable.
Also realize that lower quality meters often fail their own calibration limits - brand new in the box! A major difference here is in the input divider. The better products put this on one thick film assembly that lowers capacitance and allows the elements to track thermally. Lower quality meters use individual resistors. Some of these are normal 1/8 watt, 5% (!) resistors. Even 1 % resistors can create large deviations between ranges since worst case for in tolerance parts is 2%. What this means is that often the less expensive meters are "Beyond Economic Repair", or out of tolerance and not worth correcting the problem. Can the home hobbyist do this? Yes. Service time is your enemy for charges. Just order really accurate resistors, non-inductive metal film suggested, and rebuild the divider. Then either use an accurate voltage reference chip, or send it in for calibration (about $100 USD for something basic).
I think it pays to buy a minimum of one really good meter. You can compare other, cheaper meters against one you know is correct. Don't expect good accuracy on the AC ranges at frequencies well above 120 Hz. A reading at 1 KHz could be pure fiction unless the meter is rated for that range. It's sometimes just the nature of the beast that some meters will not ever really be accurate. They are still useful as long as you remember that 11.23 VDC really means "about 11 and a quarter volts". Don't get trapped into the thought that because there are digits on the display that they are therefore accurate. Some analog meters are more accurate than many digital meters. How sad is that? Many older HP meters are like this. Also, the analog meter may use a DC to RMS converter chip (HP and others), whereas most digital meters only peak detect the waveform and report a lower figure that represents what the RMS reading for a low distortion sine wave would have read.
An example I can offer is an HP 974A meter I bought recently. It's rated to a frequency of 100 KHz, and a DC accuracy of 0.05% (look it up!) with a 4 1/2 digit display. This is an old meter I bought at auction. Checking it against a bench meter of better accuracy shows that the calibration is still good. That's great because I would have to send it in for a good calibration.
It's difficult to avoid turning controls in a meter for some people. Restraint here is helpful.
-Chris
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.