[Marco] Looked at a lot of meters. However, he thinks HP3458A is the best, even though they were introduced more than 30 years ago in 1989. Someone donated one to [Marco], but it showed some error messages and showed unstable behavior when it started, so he needed some repairs.
According to [Marco], the error code indicates a problem with the multi-slope analog-to-digital converter, which is what makes the meter unique. The meter has 8.5 digits, so the normal conversion phase will not cut it.
The good news about this issue is that it provides us with an excuse to look inside the box. Every motherboard inside looks as complicated as a modern PC motherboard. Within this accuracy range, the circuit board is covered in a customized high-performance resistor network.
The standard method of converting a voltage to a number uses the time required to charge and discharge a capacitor, and the required time represents the voltage. The meter uses multiple possible slope resistors, [Marco] explains how the meter uses a fast and less accurate slope to get a rough reading, and then uses a slow and accurate slope to refine the lower numbers.
The custom chip has an IC and a custom resistor network. If it fails, the meter is almost impossible to repair without going to the factory service center to buy a new circuit board for about $3,000. The custom chip seems to be working properly, and replacing the comparator that is known to fail does not help.
What’s next? Purchase all the parts you can find for the circuit board (approximately $100), and then replace all the parts. We like his way of removing redundant component leads during the reconstruction process. At first, this seemed feasible, but the self-calibration failed. It seems that the custom IC may have been broken, so he eventually replaced the entire converter board.
This cleared the major error, but some measurements still had problems, causing another board to be repaired. The circuit in question performs RMS conversion on AC signals. The meter has a variety of methods to measure RMS.
This video is a great detective story, and you will learn a lot about high-resolution meters. When everything is normal, we will see some strange things, such as cables acting as capacitors and noisy fans.
I once worked with an engineer who designed the analog part. He said that this is a huge effort, and they have done more work than they expected. He believes this is part of the reason why HP/Agilent/Keysight has started but never completed the upgrade version. Only Fluke has a comparable DMM, and it can be said that the 3458 is still the best. It is too difficult to mass produce better products.
Someone told me that the AVO8 is the best multimeter that money can buy. It is carved on a stone, which Moses took down from the mountain during the triumph. I was obviously misled.
Since AVO8 is not common on this side of the pond, I found this to be an interesting reading… http://www.richardsradios.co.uk/avo8.html
I crave AVO 8 when I was a teenager, but their prices were beyond my ability. 40 years later, I have a Mk II on my bench. In the weird situation where I work on a valve radio, I am very happy to use a meter with the correct cycle.
All these better sophistry about other multimeters stem from a misunderstanding of the expected application of the HP3458A. It is not used for general fault finding, but for semiconductor characterization, and its accuracy in the uA and uV range is indeed excellent. The 4-wire measurement function (see 6 binding posts) and HPIB control are additional evidence that it is mainly used to characterize semiconductor devices.
I bought an old 5.5 Keithley and calibrated it by a friend. In the past year, it was really convenient. From matching transistors to measuring the input impedance of audio amplifiers.
Fluke 77 may be a good general-purpose instrument, but it is not the “best” instrument in any environment. No matter what your requirements are, Fluke sells better: cars? 88V. Explosive environment? 87V explosion-proof harsh environment? 28 Two. General Industry? 87V. data record? 287 / 289. Industrial process control? 789.
In addition to other tasks that the 77 can’t perform at all, any of these instruments can handle any task that Fluke 77 can complete, with higher accuracy and wider bandwidth. temperature? Conductivity? PWM duty cycle/pulse width? frequency? microampere? Rotating speed? True RMS voltage? good luck.
When it sells for $300 on Amazon, we can’t even say that Fluke 77 is a budget option for amateurs. Of course, it is cheaper than the other meters listed, but that doesn’t tell much. (289 is currently being sold to interested parties for $570). The reality is that if you use meters to make money, then the correct Fluke will quickly pay for itself. Maybe you only need 77 functions. Okay, buy a 77.
The thing is like this. Perhaps business users can strictly define their daily needs. Perhaps a company sent 77s technicians out, and the supervisor was holding something more capable (such as 87s with thermocouples) for rare situations that require temperature measurement. This seems to be a wise thing to reduce the upfront cost, the risk due to theft or loss, etc., but you can start the upgrade every hour you waste on the meter.
Hobbyists rarely have strictly defined requirements, nor do they have a depreciation plan that can be used to amortize costs over several years. If we have to buy two meters, it is usually better to buy the right one for the first time.
Patiently, I finally found my used Fluke 189 (the predecessor of 289) on craigslist at a discounted price. It seems that it has never left its box and is completely unmarked. My advice to other hobbyists is to buy the most powerful used Fluke you can afford. That might even be 77.
I will never understand the inner workings of that type of gear. Obviously, he did, and it was very interesting to watch him fix something that other people could understandably give up.
My daily carry meter is Fluke 8060A, which I bought back in 1983. When the Simpson 260 ruled the technician toolkit, it was a game-changing instrument, and the 8060A was still good. Around 1990, I had to send my 8060A back to Fluke because the display driver chip was broken, but after that repair, I have been using the 8060A regularly. I recently calibrated the Keysight 34461A 6.5 digit benchtop meter. During the temporary voltage measurement, the deviation of the Fluke 8060 from the 34461A within its rated bandwidth was within 1%. This is not bad for a meter that has been dangling in the kit for 30 years since the last calibration.
I have an old Fluke 80sumthinsumpthinA. About 20 years ago, I bought the last replacement LCD that Fluke had in stock for it!
By using our website and services, you explicitly agree to the placement of our performance, functionality and advertising cookies. learn more
Post time: Oct-21-2021
