Hi guys,
I mainly use an old Fluke 87 on the bench, which is a very good meter. Until recently when I needed extra precision I just used a Keysight 34461A, which was very, very nice.
I’ve just bit the bullet and bought an LCR meter (BK precision 878), so hopefully will have something better than the 87 to test resistors and capacitors.
Alas I’ve lost access to the 34461A, so I’m thinking of buying something for home. I see a reasonable number of HP3478A meters around. They appear to tick all the boxes; 5.5 digit, reasonably cheap, HP...
Anyway, what are people’s thoughts of them?
I mainly use an old Fluke 87 on the bench, which is a very good meter. Until recently when I needed extra precision I just used a Keysight 34461A, which was very, very nice.
I’ve just bit the bullet and bought an LCR meter (BK precision 878), so hopefully will have something better than the 87 to test resistors and capacitors.
Alas I’ve lost access to the 34461A, so I’m thinking of buying something for home. I see a reasonable number of HP3478A meters around. They appear to tick all the boxes; 5.5 digit, reasonably cheap, HP...
Anyway, what are people’s thoughts of them?
I've had my 3478A for about 3 years now and it works great. Be sure to change the CMOS battery out in it if you buy one. If the battery goes flat you'll lose the cal data. You also have to be careful when changing the battery. There are some threads online that detail how to change the battery without losing the cal data.
There's also a fellow online that has published a free program to control the 3478A via GPIB. It's here if you're interested.
There's also a fellow online that has published a free program to control the 3478A via GPIB. It's here if you're interested.
I got one and my gripes other than cal data vanishing with the cell backup dying is that its upper DC voltage range is only 300V and has no math/memory functions. Math function is very handy for resistor matching.
I also have the 3403C, 34401A and 3455A. The 3455A can do math, 3403C has the highest crest factor for its true RMS detector. I have a couple of Flukes, a 76 and 87.
Six DVMs is a lot but one meter can't do all the job that I need measuring. A lot of my test gear has this situation of one is simply not enough for the job at hand.
I also have the 3403C, 34401A and 3455A. The 3455A can do math, 3403C has the highest crest factor for its true RMS detector. I have a couple of Flukes, a 76 and 87.
Six DVMs is a lot but one meter can't do all the job that I need measuring. A lot of my test gear has this situation of one is simply not enough for the job at hand.
The 8846A goes for ~5-10 times what 3478A’s do. I’d be single pretty quick if I spent that much on a multimeter.
Warrjon, I don’t understand the issue. The 3458A I can use is 8.5 digits and in cal. Any 3478A I’d be calibrating is only 5.5 digits. From reading the tech manual for the 3458A, I would simply measure a reasonable source with the 3458A, write down the value, then connect to the 3478A, enable calibration, and push the range up or range down buttons until the reading is the same as the first 5.5 digits of the 3458A.
It’s only 5.5 digits, so I can cobble up sources that’ll stay put to that level for the few minutes needed. Accuracy is unimportant, as I’ll measure the actual value. Only stability matters.
Warrjon, I don’t understand the issue. The 3458A I can use is 8.5 digits and in cal. Any 3478A I’d be calibrating is only 5.5 digits. From reading the tech manual for the 3458A, I would simply measure a reasonable source with the 3458A, write down the value, then connect to the 3478A, enable calibration, and push the range up or range down buttons until the reading is the same as the first 5.5 digits of the 3458A.
It’s only 5.5 digits, so I can cobble up sources that’ll stay put to that level for the few minutes needed. Accuracy is unimportant, as I’ll measure the actual value. Only stability matters.
Last edited:
The number of digits is not relevant to the specification of the meter. For example the spec (4 wire Ω) on the 3458 for 100Ω is ±0.00231Ω and the the 3478 is ±0.00175Ω for the same resistance reading. So the 3458 spec is not even 2:1, and this is the easy bit along with DC volts. When you get to AC voltage and current how are you going to verify frequency response.
Other things you will need to take into account are
Corrections for the inaccuracies in the 3458 - this info will be in the calibration package for the 3458A.
Thermal drift caused drafts and or variations in temperature in the room if you are not in a temp controlled Lab. This can be significant is the low ranges.
AC source for voltage and current between 20Hz and 300kHz This is where we used AC/DC transfer to calibrate standards equipment.
I worked in a calibration Lab until I retired so have repaired and calibrated lots of 3458's and other Fluke/Agilent standards equipment. You have a good quality multimeter and adjusting it without the correct equipment and environment you may as well use a Fluke 187.
Last edited:
Dunno where you’re getting your specs from. The spec for the 300Ω 4 wire resistance range on the 3458A (from the operators manual table 4.3) is +/- 0.0045% + 4 counts (within 1 deg of cal temp and within 24 hours of cal. Given each count in this range is 1mΩ, for 100Ω I make measurement uncertainty +/- 8.5mΩ.
A more realistic spec is the +/-5 deg from cal, 1 year one, of +/- 0.017% + 5 counts, or +/- 23mΩ for our 100Ω resistor.
Mind you, I’ll be perfectly happy to measure 100Ω resistors to within 23mΩ. That’s a whole lot better than I can manage with my 87.
I confess I’ve never worked in a calibration lab, and only very recently have encountered the requirement for calibration for equipment at my workplace (including the 3478A referenced above). I’m not particularly interested in absolute accuracy, as long as I can get repeatable results and reasonable measurement precision.
A more realistic spec is the +/-5 deg from cal, 1 year one, of +/- 0.017% + 5 counts, or +/- 23mΩ for our 100Ω resistor.
Mind you, I’ll be perfectly happy to measure 100Ω resistors to within 23mΩ. That’s a whole lot better than I can manage with my 87.
I confess I’ve never worked in a calibration lab, and only very recently have encountered the requirement for calibration for equipment at my workplace (including the 3478A referenced above). I’m not particularly interested in absolute accuracy, as long as I can get repeatable results and reasonable measurement precision.
Last edited:
hi Suzy,
I have 2 of them, one I keep as a spare. if you can get one for around <100 i think a good deal. I got a printed service manual too of the bay for about 25 the LCD display ain't the best but they keep running. you can only cal 2 wire or 4 wire ohms. It has a wide BW true rms AC voltmeter, (AD536?) good for audio There was an update of the reference in later S/N's. iirc, the bench brief said to solder in a isolation diode to a second cell then remove the old one.. i have the file stored away, if you need it I'll post it. you here the relays chattering away in auto DCV but they seem to be very reliable. that's it for my thumbs up It does nicely beside the 5316A freq counter
I worked in the Motorola cal lab as a young tech, I learnt so much, funny they never had a 3478A, a few 3458's though, they go way back only a few had HPIB. We used the 3458A as a transferable standard/reference until they got a fluke 5101B calibrator.
Rick
I have 2 of them, one I keep as a spare. if you can get one for around <100 i think a good deal. I got a printed service manual too of the bay for about 25 the LCD display ain't the best but they keep running. you can only cal 2 wire or 4 wire ohms. It has a wide BW true rms AC voltmeter, (AD536?) good for audio There was an update of the reference in later S/N's. iirc, the bench brief said to solder in a isolation diode to a second cell then remove the old one.. i have the file stored away, if you need it I'll post it. you here the relays chattering away in auto DCV but they seem to be very reliable. that's it for my thumbs up It does nicely beside the 5316A freq counter
I worked in the Motorola cal lab as a young tech, I learnt so much, funny they never had a 3478A, a few 3458's though, they go way back only a few had HPIB. We used the 3458A as a transferable standard/reference until they got a fluke 5101B calibrator.Rick
Last edited:
Its the stable source/resistor/current thats the challenge but I'm confident you can resolve that. Do you have artifact standards for the 3458? That would improve the reference accuracy. And its not like you are doing things at home where traceability is a big issue.
I have 3 6+ digit DMMs Fluke 8506A, Prema 6001 and Keithley 2015. They all agree to the first 5 digits and no more. The Keithley's distortion measurement feature is neat but did not replace anything else and is not used. Its also notorious for flickering digits which is why I'm not happy with it on the bench. The Prema is an oddball here in the US but has potential however doesn't see much use except on the cal bench. The Fluke is pretty nice but huge and not for troubleshooting.
The best bench meter for troubleshooting I have used is the Fluke 8860A but mine died so I got the Keithley. I'll look closely at the HP3478A as an alternative. They are pretty reasonable here at around $150 US. However I really like the big LED's on the older Fluke.
I have 3 6+ digit DMMs Fluke 8506A, Prema 6001 and Keithley 2015. They all agree to the first 5 digits and no more. The Keithley's distortion measurement feature is neat but did not replace anything else and is not used. Its also notorious for flickering digits which is why I'm not happy with it on the bench. The Prema is an oddball here in the US but has potential however doesn't see much use except on the cal bench. The Fluke is pretty nice but huge and not for troubleshooting.
The best bench meter for troubleshooting I have used is the Fluke 8860A but mine died so I got the Keithley. I'll look closely at the HP3478A as an alternative. They are pretty reasonable here at around $150 US. However I really like the big LED's on the older Fluke.
Your spec is 8.5mΩ and what I listed is 2.31mΩ I took the spec directly from the calibration specification in the Agilent manual for the 3458A.
This was only an example.
What are you going to use to adjust the frequency response of AC voltage and current, you will need access to a FLuke 5700/20 and 5725 amplifier to have any chance of getting near adjusting AC. I did this for a living and can tell you it is not as simple as using the 3458A as a transfer standard. IMO it would be way easier to replace the battery.
I'm less interested in calibrating AC voltage and current, as I don't generally have much need for accuracy for these. I'll have an initial go using my 3325B, but like many things I get into I'll probably disappear down the rabbit hole for a while and come up with things.
Again, I can measure sources with the better calibrated meter and compare.
Again, I can measure sources with the better calibrated meter and compare.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.