KHG DIY or Alternative Reagent

[DanyL]

If someone could pop the back off a unit and take a photo of any leds/sensors in the reagent path, and any markings on them, the sensor type could be determined.

From that information a patch to the firmware that removes the test should not be all that difficult.

Or it could simply be a matter of replacing the sensor with components that send psuedo sensor data back to the MCU to make it think the reagent is within spec.

edit:
The second is probably the easiest, as it is super easy to interrogate the data between a sensor and the MCU no matter what type of device it is.

For the first part - It’ll likely help in reversing, but if anything my first goal would be to target the locking mechanism rather than the detection method.

They may change things in a future update in ways I wouldn’t describe here to still recognize it, and that would require further work to mask the differences in other ways - but I think that software wise it’ll be the easier path going forward.

As for a hardware patch - that’s actually a really good idea, and could be something as easy as adding the right resistor to mask the differences.

The question is wether the change is that obvious and not some sort of a long term sampling to detect small variations that could otherwise be interpreted as an acceptable noise.

Another point I would like to add is that from further looking at the firmware, it seems update files do not necessarily have to contain all parts of the firmware, and it may actually reserve some memory to exchange payloads on the fly or load them separately from another source, some of which looks to be stored on the SD card - so it’ll be helpful to have a full image of the SD card as well.

Edit: Also, if anyone of the owners is able to solder 2 wires to the board and has an FTDI on hand - a serial access to the device would be a gold mine in terms of useful information we can extract from it for reversing purposes.

 

[Randy Holmes-Farley]

If it is absorption of light, adding an appropriate dye seems easier to me than the firmware rewriting, but I expect that just mirrors our areas of expertise.

My first step would be to measure the absorbance of the OEM reagent as a function of wavelength to see what dye is used.

 

[Jposch]

I have done all the upgrades my self over the years. I contacted Alex and he said he just charges for the parts that were needed so I figured what the heck. I dont know about everyones khg but I test my alk with my red sea kit every other week or so and most of the time I need to reset the reference and I sure would never trust mine to dose alk. I mainly use mine to watch trends for something going wrong.

I do NOT trust mine to dose alk, but I manually test alk with salifert monthly and it's always good. I even use a cheap $25 pH probe from Amazon because I guess the open glass bulb probes are allergic to me. I broke each one, 3 in a row when calibrating. I wasn't even rough, so I gave up on them.

 

[Jposch]

What is the consensus on how the unit knows if you diy the reagent?

Fixing the diy reagent to match may be a better long term choice than downgrading the unit, although I understand it carries risks testing diy reagents.

I posted a link to an optical sensor technology that measures droplet size and composition, and the form factor closely matches the khg sensors. This helps to increase accuracy, but they also added some indicator dye that the sensor can detect. It'll catch oem reagent dilution being off as well, so it's quite sensitive.

 

[Jposch]

Some folks mentioned an optical sensor early in the thread?

something like this:

https://cdn.intechopen.com/pdfs/26343/InTech-Optical_chemical_sensors_design_and_applications.pdf

 

[coralbeauties]

Has the optical sensor always detected what all the video shows? Is it just the newer firmware where we are having problems running diy reagent. My monitor is approximately 6-7 years old and Ive run diy through it for at least half its life or more.

 

[Randy Holmes-Farley]

For those who have recently bought a KHG, where in the purchase and setup and use does one first find out that one has to use their reagent or it will lock up?

IMO, it is unacceptable to only find out after you have bought it. I read through the current web site and saw no such info, and could load it into my shopping cart at multiple sites without seeing any such info.

 

[Jposch]

For those who have recently bought a KHG, where in the purchase and setup and use does one first find out that one has to use their reagent or it will lock up?

IMO, it is unacceptable to only find out after you have bought it. I read through the current web site and saw no such info, and could load it into my shopping cart at multiple sites without seeing any such info.

Afaik, it's just a post in their Facebook group.

 

[Jposch]

Has the optical sensor always detected what all the video shows? Is it just the newer firmware where we are having problems running diy reagent. My monitor is approximately 6-7 years old and Ive run diy through it for at least half its life or more.

I don't believe the sensor was added until 2021. When it's testing. Do you see a red light flash when reagent drips?

 

[Jposch]

Video in slow motion

 

[coralbeauties]

I don't believe the sensor was added until 2021. When it's testing. Do you see a red light flash when reagent drips?

My original unit that is 6-7 years old has always had the blinking red led. That is one reason why I have to wonder if all the authentication is all scare tactics. I have a newer unit running 9.3r31 firmware. I have run oem reagent in it since I bought it with a years supply of reagent. I will make up a batch of diy and risk trying it and see what happens. Im starting to think like the others if I have to buy the oem reagent I might as well use the trident with the generic reagent that is half the price of neptunes.
Jeff

 

[Randy Holmes-Farley]

Afaik, it's just a post in their Facebook group.

If that is true, its appalling.

 

[Jposch]

If that is true, its appalling.

The entirety of authentication of reagents is, IMO. even moreso that's its not obvious BEFORE purchase.
Right to repair laws come to mind.
That's like a car manufacturer disabling your vehicle for using an non-oem approved brand of fuel, or engine oil, etc.

My original unit that is 6-7 years old has always had the blinking red led. That is one reason why I have to wonder if all the authentication is all scare tactics. I have a newer unit running 9.3r31 firmware. I have run oem reagent in it since I bought it with a years supply of reagent. I will make up a batch of diy and risk trying it and see what happens. Im starting to think like the others if I have to buy the oem reagent I might as well use the trident with the generic reagent that is half the price of neptunes.
Jeff

I've seen posts of people having it occur. It's possible the older production models have the senor as well.
I really like the unit otherwise. Very stable and so many redundant sensors and functions to nearly eliminate overflowing or giving a false reading. You can set alerts for deviation between results, etc.
The reagent thing is concerning as well as the cloud connection. At any point, the company could push an update (currently all manual user operation) that would brick the unit. How far will they go to risk reputation over milking devote customers for monthly reagent income?
At the end of the day, if the cost was half or less for reagents, I wouldn't even have bothered.
Another thing that really erks me, is the newer firmwares burn through reagent to "keep probe working accurately." I call BS.

 

[Quatrc]

HYDROCHLORIC ACID 35% diy reagent from KH kepper? i need get 0.1N ?
400mL acid and 600ml osmose water? or 200 ml acid- 800ml osmose water?

 

[Randy Holmes-Farley]

HYDROCHLORIC ACID 35% diy reagent from KH kepper? i need get 0.1N ?
400mL acid and 600ml osmose water? or 200 ml acid- 800ml osmose water?

I've not calculated the exact amount, but those concentrations are way, way too high.

35% is a bit over 11 M (11 N). 9 mL into 991 mL of RO/DI water is fairly close to 0.1 N.

 

[Jposch]

I've not calculated the exact amount, but those concentrations are way, way too high.

35% is a bit over 11 M (11 N). 9 mL into 991 mL of RO/DI water is fairly close to 0.1 N.

That's approximately what i use for hardware store acid, which can vary a bit.
For those who've sewn rhe khg Facebook post about clogged nozzles, mine is still clean after a year of hardware store HCL. My guess is in that case, they used less than pure water, or perhaps some acid that wasn't pure hcl and h20.

 

[pav306]

Does anyone know the reagent strength for the D-D kh manager?

Is it the same as most of the other machines 0.02n?

 

[Randy Holmes-Farley]

Does anyone know the reagent strength for the D-D kh manager?

Is it the same as most of the other machines 0.02n?

I do not know.

 

[gbroadbridge]

Does anyone know the reagent strength for the D-D kh manager?

Is it the same as most of the other machines 0.02n?

The instruction manual indicates that the unit checks the concentration of the titration solution by measuring its pH and they state it must be at 25 degrees C for checking.

I'm not sure how much the pH of HCl pH varies by concentration, but perhaps the easiest way to check if the unit will work with DIY is to simply mix up some 0.02N HCl and see if the unit says it is okay.

The device uses a 50ml water sample from the aquarium, which is the same as other devices, so it is probably close.

The manual also says the test water can be returned to the aquarium so i't not harmful so again likely to be HCl

https://www.theaquariumsolution.com/sites/default/files/downloads/KH%20MANAGER%20INSTRUCTION%20MANUAL%20V1.2.pdf

 

[Randy Holmes-Farley]

I'm not sure how much the pH of HCl pH varies by concentration, but perhaps the easiest way to check if the unit will work with DIY is to simply mix up some 0.02N HCl and see if the unit says it is okay.

Not sure how tight that pH control is, but a 25% change in potency is only a 0.1 pH unit change.