EEVblog #1366 - Mystery Bunker Item Teardown

[EEVblog]

Another mystery item from the storage bunker.

[SiliconWizard]

Yeah looks like this: https://www.geminibv.com/labware/vitatron-filter-spectrophotometer-mcp/

[Kleinstein]

The filters are likely something like 5-10 nm wide. They look like interference type filters and thus may have some extra passage way out - thus the extra range filter set.
I vaguely remember a price of some $50 for a similar size 670 nm filter. The more odd wavelegths may be more expensive.

The detector is likely just a UV sensitive silicon photdiode - still expensive. No need to have it specially characterized - there would be a kind of empty measurment with just water or similar to get a reference. This would also include the sensor, filter and lamp. Kind of odd that there is only one input and not an extra tank for water for those reference and cleaning.

The "heater for the vial actually looks like 2 small TEC elements. So not just a heater, but temperature control with both heat and cool.

The small board would be the temperature control and also "voltage" regulation for the lamp. There may be an additional photodiode to measure the light from the lamp instead of simple voltage control.

The off color on the small board is just from the light.  To get good intensity also at 380 nm chances are the lamp rans relatively high power and a few voltage above nominal.


Still an odd design with so much cable to the sensor  , usually one wants to keep the capacitance low.
The system may be actually measuring DC and not the more normal AC way with a chopper and phase senstive detector.
This may explain the high grade OP in TO99 - still a good one. 

[bitter_mike]

As far as the application, this is the sort of instrument which would be used to monitor the progress of a chemical reaction. Either the reactant or the product of the reaction would absorb light at a particular wavelength. You start the reaction going and stick the sampling tube in there. Then the pump continuously pulls some of the reaction mixture out of the reaction vessel (a beaker or flask or whatever) and measures the absorbance at the wavelength you set it to. This way, you can monitor the appearance of the reaction product or the disappearance of the reactant over time. In this sort of application, it makes sense for the little sample vial to be small. If it were large, you would be averaging together the reaction mixture from a wider range of time points and would lose resolution.

A modern instrument like this would cost in the high 5 figures/low 6 figures depending on what features you get with it.

[ace1903]

Does anyone know if pictures of the PCBs are available?
If I remember well there was some site where Dave uploads pictures of the PCBs from teardown videos.
I like videos like this but I would like to see more overview of the signal chain and to learn something from it.
I am sure that this instrument uses some interesting input amplifier with low bias current, then probably some logarithmic converter ic, and finally a discrete ADC circuit.
Usually, I look at the IC used then it triggers my curiosity and learn something new from the datasheets and application notes related to the ICs used in the instruments.

[Kleinstein]

The low bias amplifier is likely one of the OPs in TO99, the AD545. AD549 - still quite good by today's standards.
I somewhat doubt that there would be a logarithmic amplifier. It would be more like a little range switching if more dynamic range is needed than the ADC can offer. The critical case is relatively low absorbance and there it is important to have really good stability of the light source and detector and also amplifier gain. A log amplifier would only cause trouble in this case. Very strong absorption would more like want a stable zero. Only than a log amplifier would make sense.
 
I would have normally have expects some mechanical chopper for the light - but I have not seen it, and also not a second light detector.

[ace1903]

That was my point. If we can see the PCB we can have an idea if there are two sensors. You are right about the LOG amplifier.
I am attaching two pictures from one of my old PCBs. It is UVvis unit with a sensor with a quartz window.
The circuit is built around 2N5908 dual N JFET and I believe paper integrating capacitor.
It is the same age, manufactured ~1985.
Most of the time I do embedded software development but in my free time, I like to exercise my brain with interesting circuits and measurement methods.

[Rdx]

Interesting how the pump motor coupling is actually just a short piece of that same tube they use for pumping the liquid. And the alignment of the two shafts seems less than ideal. Even more impressive that the tube still seems to be fine after all those years.

[ace1903]

I think that pump is of peristaltic type and drives fluid in and out in equal amounts. Dave misidentified it as a valve.
The last part in the chain with switches and gear I can not identify. Maybe a preparation unit for washing fluid?
Those two tiny Peltier elements can not warm up the cell up to 90 degrees for efficient cleaning with fluid flowing at a reasonable rate.
@Rdx Where are you seeing pump motor coupling?

[thm_w]

Does anyone know if pictures of the PCBs are available?
If I remember well there was some site where Dave uploads pictures of the PCBs from teardown videos.
I like videos like this but I would like to see more overview of the signal chain and to learn something from it.
I am sure that this instrument uses some interesting input amplifier with low bias current, then probably some logarithmic converter ic, and finally a discrete ADC circuit.
Usually, I look at the IC used then it triggers my curiosity and learn something new from the datasheets and application notes related to the ICs used in the instruments.

https://www.flickr.com/photos/eevblog/ didnt see it

Interesting how the pump motor coupling is actually just a short piece of that same tube they use for pumping the liquid. And the alignment of the two shafts seems less than ideal. Even more impressive that the tube still seems to be fine after all those years.

Might be PTFE, that stuff lasts forever and is heavil used in chemistry.

[EEVblog]

Does anyone know if pictures of the PCBs are available?
If I remember well there was some site where Dave uploads pictures of the PCBs from teardown videos.
I like videos like this but I would like to see more overview of the signal chain and to learn something from it.
I am sure that this instrument uses some interesting input amplifier with low bias current, then probably some logarithmic converter ic, and finally a discrete ADC circuit.
Usually, I look at the IC used then it triggers my curiosity and learn something new from the datasheets and application notes related to the ICs used in the instruments.

https://www.flickr.com/photos/eevblog/ didnt see it

I only photograph stuff that is particually interesting, didn't in this case.

[Kleinstein]

The video footage is good enough to see the analog part and the main board.

It very much looks like a TIA with large 15 M resistors at the input (would be the normal configuration for a photo-diode). There is not that much what follow: 2 x 4 switches  1 good OP (AD547) and one good comparator. This may be some kind of discrete build ADC controlled by the CPU. It is still a little odd: for a dual slope ADC I am missing 1 more medium grade OP as a buffer - there is a LM324 near the digital part, but this would be a bit odd combined with the higher grade OPs used for the rest.  Maybe the have separate resistors just before separate resistors as kind of odd variant of the dual slope with variable gain.
 

[Rdx]

@ace1903 You can see it in the video at around 16:16 -  the motor that is driving the worm gear of the pump mechanically coupled

[ace1903]

I am wrong. Valve is clearly marked: 1 x NO & 1 x NC Pinch Valve -- 225P091-11.
@Rdx  Now I see coupling at 16:16. It is clear. The rest of the tubing is darker. Maybe some nasty chemicals were involved besides heat from the lamp.
Interesting idea to use microswitches to detect the top and bottom position of the piston.
This pinch valve was quite popular in scientific society. Found a bunch of interesting papers related to it.

[ace1903]

Can anyone see where is the second port on the pump?

[Fryguy]

The pistonpump itself doesn't need a second port - just a pair of check-valves to automatically switch the between the in/out ports . You'll find them inside that black switchingblock connected to the pump , the pinch-valve and the output-port on the rear side . You can see it at 15:05 top center of the image . Another possibility will be that the black block is just for connecting everything together and the pinch-valve does the switching .  The microswitches below the pump for the pistons upper and lower end position are connected to the pinch-valve . . .