HIOKI 3522 LCR Meter teardown

[Michel95200]

Hello
Hioki 3522 100kHz LCR meter uses traditionnal autobalancing technique with undersampling ADC. Main parts are:

* OPA2134 + BUF634 for transimpedance amplifier
* 5 range resistors selectable with relays TQ2-5V
* voltage and current range amplifiers use OPA2134 and NJU201 analog switches
* test signal generator uses AD9831A DDS
* test voltage level: multiplying DAC AD7945
* output buffer : BUF634
* measurements are performed using 16-bits ADC ADS7809U with undersampling technique
* undersampling clock uses another DDS IC AD9831A

[ogden]

Nice job! What was motivation to reverse engineer schematics of this module? Any further plans? BTW you could increase contrast of pictures.

[_Wim_]

Wow, very nice job. That must have been quite a bit of work! 

[Michel95200]

I am independant and for a part of my job I am an electronic designer for a well known european instrumentation company. For this work I make sometimes this kind of reverse engineering to see the technique used by competitors and the parts they use. I did the same for other LCR meters if some are interrested. It took me half a day to draw the complete schematic of this one. I would like to find a not working Keysight E4980A or the analog board only do to the same.

[dr.diesel]

I did the same for other LCR meters if some are interrested.

Very interested!  Hameg?

[ogden]

I did the same for other LCR meters if some are interrested.

Many would be interested. Please create new thread for each instrument in question.

It took me half a day to draw the complete schematic of this one.

Impressive skills. Do you use some specific "reversing know how" or just useful tips you can share to us?

[DaJMasta]

I've got a 3532-50 meter and while it uses a lot of the same parts, it's got more of everything!  Three DACs, a dual channel audio ADC, another few range switches (two different kinds of relays, most like the ones on your board), and long, long chains of filtering opamps.  Haven't done any proper reverse engineering, but despite the differences it's neat to see a lot of similar parts.

[Michel95200]

Generally modern LCR meters use dual ADC to sample voltage and current at the same time. In the 3522 voltage and current are sampled sequentially.
In the Hioki 3531 I have also reverse-engineered the ADC is a dual audio ADC AK5389 from Asahi Kasei. This instrument is very interresting, it uses also undersampling technique with external fast sample and hold. For high frequency band it uses a reconstruction of the current using demodulation of the error and reconstruction of the current using sine/cosine signals coming from two DDS. Schematic is more complicated than the 3522. Maybe the 3532-50 is similar, do you have a picture of the main board ?

[DaJMasta]

I have some mediocre quality diagnostic pictures 

A similar technique may also be used in this one, if I remember right, the signal on either input of the ADC was a 1kHz sine of some description, regardless of the frequencies I tried... though I don't 100% remember.  I believe the SO16 chips around the ADC are all analog muxes, and the lines of 4 SO8 packages are single opamps, I believe fairly high spec but not too much GBP given the operating frequency of the device.  There's also a bit more on the board under the digital board which isn't pictured - at least one extra trail of filtering amps, which would be under the bottom left portion of the full pic of the board.  You can see the input/output terminals on the left side of the two images that show the frontend stuff.

[Michel95200]

Hioki 3532 seems technically closed to the 3531. It uses a reconstruction technique using demodulators/modulators. I cannot see the reference of these parts, in the 3531 classic MC1496D are used. Three DDS are present, one for sine/cosine generator and another one for undersampling clock generation. The sample and hold circuits seems identical (the area with blue through hole capacitors). I have delimited the different functional blocks.

[DaJMasta]

Very interesting!  Those suspected MC1496Ds are actually just more DG411DYs, but there very well could be that section under the digital board in an area not pictured by these pics.  I remember being a little baffled when looking over it, so it's sort of nice to hear that it uses a fairly complex measurement scheme 

I remember being real surprised to find local regulation in the form of 78XX and 79XX regulators in those larger all-plastic TO-220 style cases.  I suspect they use some more modern variant with better stability and PSRR, but I hadn't really considered them as good regulator choices for high sensitivity analog sections.

[ogden]

Those suspected MC1496Ds are actually just more DG411DYs,

Two pairs of DDS+LM361 are strong hint that indeed it is IQ (sin/cos) mixer system and they use DG411DY as mixer (demodulator).

Hioki 3532 seems technically closed to the 3531.

Could you share it's schematics?

Three DDS are present, one for sine/cosine generator and another one for undersampling clock generation.

Third DDS (ic202) looks like test signal going to Hcur.

[edit] I don't know how legit that seller and his offer is, but look at those piles of LCR meters!

[hoangduongo]

Hello,
It might not be totally related. but does anyone here have the schematic of the Hioki 9268/9269 DC bias modules that go with the 3522 LCR meter?
At the very least, the basic principle and feasible working plan will also be much appreciated, as I am fancy to imitate one for my 3522.
Thanks a lot!
Duong

[coromonadalix]

you have Hioki 3522 on Ebay

Even low as 4xx$ usd dollars for parts


This seller has many of them ??
https://www.ebay.com/itm/Used-Hioki-HIOKI-3522-50LCR-1Pcs/173558354050?hash=item2868e25082:g:0QQAAOSwtM1brCao

[hoangduongo]

Coromonadalix,
I don't think I fully understand your question.
I already have the 3522 meter (purchased from a local factory liquidation), now I'm looking to copy/build the DC bias modules (like the 9268/9269), as the ones appear on ebay are even more expensive than what I paid for my 3522.
Thanks, Duong

[coromonadalix]

Ill correct my post,  you have many 3522 on Ebay, and one seller link seems to have ton's of them ??  they sell around 1k$$ usd

The lcd seems expensive ?

Is it a good lcr

I tried to build an dc biais board for an Fluke 6304 i had,  i had the chance to copy the 6306 dc biais board from an eevblog member, but the same parts used where very difficult to get,  150$ in parts only,  no pcb made    

thks

[hoangduongo]

Hi,
Imho, the 3522 is a decent meter, especially for me who works with audio most of the time. But also because of that I frequently need to measure chokes/transformer with DC current passing through, which then leads to the demand of the DC bias modules (the 9269 DC current module is preferred for this reason).
I looked up and found your project with the 6304 meter and 9565 module, looks terrific but I'm not sure if they use the same principle with the Hioki.
Thanks

[xu59]

朋友们好！
    哪位能告诉我如何进入日置3532-50数字电桥的的校准界面。不胜感谢！

[KungFuJosh]

朋友们好！
    哪位能告诉我如何进入日置3532-50数字电桥的的校准界面。不胜感谢！

Me too! Has anybody figured out how to get into calibration for the 3532-50 or similar?

Thanks,
Josh

[KungFuJosh]

Hioki 3532 seems technically closed to the 3531. It uses a reconstruction technique using demodulators/modulators. I cannot see the reference of these parts, in the 3531 classic MC1496D are used. Three DDS are present, one for sine/cosine generator and another one for undersampling clock generation. The sample and hold circuits seems identical (the area with blue through hole capacitors). I have delimited the different functional blocks.

Do you still have your 3532? I'm curious what in-circuit resistance readings you get for the group of resistors around R361.

Thanks,
Josh