UPDATE: Here is one of the photos from the previously mentioned thread for people who want something higher resolution to look at.
Seems a reasonable design to me. input protection looks ok. i see a ptc. the input resistor is a ladder of smd resistors to get adequate standoff voltage without flashover.
The relays are dual coil with mechanical latch , as should be ( agilent uses single coil but mechanically latched relays )
Does anyone know why this is important ? Yes ? No ? let me enlighen those of you that don't know.
An energized relay coil creates a magnetic field around the contacts. Any signal traveling through the relay and thus the magnetic field sees this as a back EMF . In essence this magnetic field acts as an impedance and depending on the strength of the field, and the frequency of ths signal traveling you get strange rsults. For dc you don't care but when measuring ac voltage or currents you don't want that ! You get errors. as frequency increases it gets worse.. So you either us a magnetically shield relay ( Like the little red Coto-wabash , or teledyne relays ) but those can t handle current , or cannot switch current ( they 'stick' as long as current is running ). or you use relays with mechanical latching. so the coil does not need to be kept energized. in signal generators you will see that the output attenuator relays are always mechanical latch.
right, back on track...
i see two grinded off partnumbers .. thats stupid. it doesn't protect your design. why bother...
there is no real reference. they use a so-8 type reference , most likely an AD part .... for a 6 1/2 digit i would have expected a thermally controlled buried zener like the lm399 (yes you can still get those .. )
No real hybrid , laser trimmed dividers which worries me a bit for stemperature stability and long term drift...
I see a nice wide gap between inguard and outguard logic. The brains are obscured by the top board but no doubt we will find the blackfin DSP since a lot of other rigol stuff has that.
A couple of CPLD's contain the glue logic . the GPIb controler is not an classic 9914 as that part is obsoloete but a copy in a CPLD. At leas tthey use real GPIb buffer chips (75161) and don't muck around with simply using the io pins of a ttl chip...
the inguard logic is controlled by a cpld and it looks like they ar not using optocouplers but an analog devcies digital signal coupler that works magnetically ( obscured by the white and black leads to the fuse
I wonder what is under the metal shield. could be the true-rms convertor as that one is sensitive ...
a common mode line filter sits between the power plug and transformer . there dosn't seem to be a true on/off button but rather a standby switch. so the transformer is always draining power ...
Someone complained about long boot times... be happy it's not a Fluke /Tek benc meter ... those things run loonix. Whoever came up with the idea that an operating system was needed to make a digital multimeter should be stapled to the wall... I hope this rigol doesn't fall in the same category... if i find out it runs loonix it is immediately disqualified.
I have one on the way to test as well ( but i'm after other things... things like how fast can it take accurate readings, what's its throughput over gpib , longterm stability , temperature stability. etc ) flashy user interface mean nothing. i need a meter i can trust to give a correct result.
One of the tests i like to do is turn off the autocal , short the inputs , launch autozero , apply a known voltage and then take 10000 readings without autozero and see how bad the machine drifts... stick it in a tempchamber and fluctuate ambient frmo 18 degree c to 40 degree c and see what it does.... after 10000 readings , short the input and take the residual. to see how far we are off zero. the agilent 34401 does very good on that test ...
I can't have a machine that starts drifting if someone opens a door .... or if the airco kicks on and out.. ( i've seen that ! measuremensts with a cyclic up and down shift to their baseline. perfectly in sync with the airco kicking in and out .... in the new lab we have a different airco system that is always on it is a dual stage system with a heater, a chiller and a humidity control. Temperature is always fixed and changes maybe 0.5 degrees c . But in the old lab stuff like that would influence precision measurements...)
as for the 'growlers' on the forum... let them 'growl'
-edit- apparently that picture is a 5-1/2 digit. for tat it may be ok with a lesse reference. the 61/2 digit needs something something better.
i wonder what converter they are using ...