| Products > Test Equipment |
| 40,000 Count Multimeter under $50 - HoldPeak HP-770D |
| << < (7/9) > >> |
| geb:
The thread over at Kazus discussing HP-770D capacitors... This has the capacitor replacement and also the EEPROM tweak for 55000 counts. https://kazus.ru/forums/showthread.php?t=120510 --- Quote ---I never actually thought/knew about this before. Should I then essentially just ignore the last digit of my multimeter if I'm going for more-precise measurements? --- End quote --- I've seen people suggest that we should ignore the last digit. On some cheap meters, this is a safe strategy. But it isn't automatically true. Many of the lowest-cost meters were made with Hycon HY12P65. That chip is rated for 5000 counts but its ADC is 19-bit (2 ^ 19 = 524288). It's found in 2000, 4000, 6000, and 9999 count meters. HY12P65 is even used in Zotek ZT-209 and ZT-303* with firmware extending it to 19999 counts. My ZT-303 meter is quite accurate after mods and calibration, so I think kindly of the HY12P65. With a rating of 50000 counts, HP-770D's HY3131 ADC is better than HY12P65. But my sample of HP-770D is not impressing me. My modded ZT-303 (paralleled caps and ICL8069A voltage ref), and minimally modded ZT-102 (EEPROM 9999 limit based on AN8008), and even my newly arrived ZT-702S are pretty reliable on their last digit. If I had time, I would continue to study my HP-770D and try to figure out what's making it read so differently at one power-up versus another. I'd also like to know if other HP-770D owners have the same issue. * - The PCB's silk screen says it's an HY12P65, but who knows for sure? |
| iMo:
--- Quote from: LooseJunkHater on September 19, 2023, 07:22:26 pm --- --- Quote from: iMo on September 19, 2023, 10:11:50 am ---The cheapo 20000/30000/40000 meters are basically with the "older" chips usually used as 4 digits ones (with the last 5th digit used as a "guard digit" and not showing it on the display). That allows to show stable readings and small none-linearity. Two my 30-40y old meters work that way, the last digit is not shown while the chips itself can do one digit more.. --- End quote --- I never actually thought/knew about this before. Should I then essentially just ignore the last digit of my multimeter if I'm going for more-precise measurements? --- End quote --- For example the HY3131 is a "24bit sigma delta ADC" according to the DS. That is theoretically ~16million resolution (8.5 digits). You will not get such resolution in reality because of noise, drift, pcb design, etc. etc. (many other params are involved). The actual parameters (ie number of significant digits) are therefore mostly given by the external factors to the ADC. The cheaper parts and poorer design, the less digits you actually get. Therefore the cheap DMM vendors rather show less digits, as to have let the last digits jump all over the place. If you had access to the firmware of the MCU (which reads the ADC's values) you could introduce some averaging/smoothing to get a better output. |
| iMo:
--- Quote from: geb on September 19, 2023, 02:18:31 am ---..It calibrates by using basically the same method as ZT-102, AN8002, and other cheap meters using a Hycon MCU. Close the jumper, turn on the unit and switch its dial quickly to ohms. Wait for its self-calibration to finish; it'll beep in mid-calibration and again at the end. Use HOLD and RANGE to decrease and increase, and SEL to advance. One word of warning: If you are displaying 40000 and you increase, the calibration routine will fail and you'll need to reload the EEPROM from a backup.. --- End quote --- Could you point us on a detailed guide how to adjust the meter, plz? |
| LooseJunkHater:
--- Quote from: geb on September 20, 2023, 03:20:57 am ---The thread over at Kazus discussing HP-770D capacitors... This has the capacitor replacement and also the EEPROM tweak for 55000 counts. https://kazus.ru/forums/showthread.php?t=120510 --- End quote --- I translated much of that thread and the associated thread. https://kazus.ru/forums/showthread.php?t=120130&page=3 Translated from Russian to English: " I replaced all the stock ceramics 1206/0805 (X5R - 10/4.7uf) in the power supply connections of both chips with decent branded lowesr 100uF polymers and 47uF ceramics. The total capacitance for the +3V supply was about 250 µF instead of 40 µF. I am quite pleased with the result of the modification, the glitches have disappeared, and it works a little faster than the new one. The reference voltage measurements have not changed. Measurements of resistors <0.3 Ohm have become more adequate, for example, instead of 0.1 Ohm it was 0 Ohm, now it seems stable 0.07-0.08 Ohm. " Associated image: https://kazus.ru/forums/attachment.php?attachmentid=160835&d=1614161531 _________________________________________________________________________ From this thread: https://kazus.ru/forums/showthread.php?t=120510 " What was done: first I replaced the stock capacitors, 1206 turned out to suck Y5V, and 0805/0603 - X5R/X7R https://kazus.ru/forums/showthread.php?t=120130&page=3# using 3V power supply I replaced the ME6208 LDO with a low noise LP2985-3.0 " Ohm's was calibrated simply by using "closed probes" (I guess a near-zero ohm resistance connection?) "the diode measurement remains (it cannot be successfully calibrated), the direct voltage drop differs from brands like Appa607.71 by 1-2%." "For power supply, I replaced all the stock capacitor junk (all 0.1 uF were Y5V); LDO HE233 was replaced with low noise LP5907-3.3V; 4.7uf was added to the Icl8069 ion." " Replaced it according to the diagram, without recalibration: C1, C3, C17, C21, C23, C11, C13 The next step was to increase the capacitance (add it in parallel) to the reference pins of the chips; recalibration was required, but the effect was noticeable - the device began to measure electrolytes 2-3 times faster. Add. Photos can be viewed here http://ixbt.photo/?id=album:66908 " [I'm unsure what they mean by "diagram" (schematic), and I also don't see what values they changed for C1, C3, C17, C21, C23, C11, C13] There was also some other content about modifying the firmware values to increase the counts, but it's not very useful due to the poor translation. This thread appears to have the stock firmware (in case you modify the firmware and mess it up; I of course can't verify the legitimacy of the firmware): https://kazus.ru/forums/showthread.php?t=116225 Direct link to firmware download: https://kazus.ru/forums/attachment.php?attachmentid=125109&d=1518199910 |
| geb:
iMo, I've been asked to create a detailed guide for HY12P65 and derivatives. It's a fair bit of work to document it with all the side-topics and variations, and it might have to cover topics like "But I just want to calibrate this one favorite mode and be done with it -- what can I skip?" If my work deadlines would ease up, I could consider it. In the meantime, I don't think I have one favorite guide that you could apply to HP-770D. A translated DTM0660 datasheet might be worth reading as a starting point. Section 12 covers calibration. But change all of the 5's to 2 or 3 or 4. For example, in the first step while the selector is still pointing to the ohms scale, whereas the datasheet calls for 500mV and 50K sources, your first known-good reference sources for HP-770D could be 400mV and 20K, or 300mV and 40mA, and so on. Press SEL to skip ahead, HOLD to decrease, and RANGE to increase. And never hit RANGE when the count is above 39999!) LooseJunkHater, I've attached the screenshot from the same Kazus thread about the increase in counts. It requires editing the data EEPROM's contents, which means attaching to the EEPROM and reading/writing it. Change the bytes outlined in orange to the values in the picture. But please note that "firmware" in this case is merely the contents of the data EEPROM. It would be wonderful to get a copy of the firmware (I mean the program code), and/or a guide to the contents of the data EEPROM. |
| Navigation |
| Message Index |
| Next page |
| Previous page |