S-Prihadi

Guys,

I am sad and pissed today....... .... , I should have done the refrigerant fill level verification 2 years ago and my compressor maximum life would be longer.
Short of 35% or by average 208 grams from a 590 grams capacity R134A , may have made my compressor oil circulation poor enough to an extend wear and tear at the compressor could be higher. This a bad condition for HVAC.
Compressor oil tag/mix along the R134A for circulation.



I measured the recovered R134A.







Deficit is leak. Based on a study and even EPA uses this, 4-7 grams per year is the norm. Newer cars supposedly is a 4 grams per year leak at the compressor shaft dynamic seal.


I honestly I do not think my HVAC is leaking, as a 9+ years if indeed a leak exist, I would be very very low on R134A by today if not empty.
I think MB Indonesia during assembly and recharging did a poor job. They ALREADY screwed up my water pump from bad quality water and also the coolant % mix.
My HVAC system is kinda "wet". I am not able to pull better than 500 microns using 1/4" hose at HP and micron gauge is here at HP side and another 3/8" hose at LP.
I am getting only at best 566 microns in 95 minutes.
I am still pulling vacuum now and will keep it sucking till 10-12 hours if I need to.
This means MB Indonesia never vacuum down dry enough the day the HVAC was commissioned and PAG compressor oil its by nature hygroscopic too, so double wammy.




=================================


Using TARE or zero-ded with hose weight included ( empty hose ), I can see total gross recovered R134A being 326 grams of R134A and not 300 grams, but the 26 grams are at the hose, so I can't recover it into the recovery tank.



Yes, the 1/4" short 4-5 feet hose can contain easy 20-30 grams of R134A in liquid state. My portable recovery machine pumps and condensed the R134A to be liquid.
I recovered from LP side at first, mainly vapor, as to not suck any compressor oil along with the R134A. Later I recovered from both LP and HP.

Above : 45 grams is theoretical maximum.



So, I measured recovery tank weight before and after without the hose. Easier.



==========================

The recovery machine can suck down to -9.3 PSI below atmosphere or equal to -19 inch Hg, not bad.




==========================

I have replaced both my LP and HP valve core today.
I use 3-4 psi constant flowing/pressurized nitrogen inside the HVAC system while I do the replacement of the valve core, so no ambient air can enter the HVAC system.


===========================


NITROGEN PRESSURE - LEAK TEST

Having digital gauge accuracy to 0.1 psi is a blessing and a curse.
Temperature change will effect pressure reading. It was raining very heavy and the night got cooler by like 3 degrees.

Nitrogen at 110 PSI as test pressure is used because at 30C, R134A stabilized in my car will read about 103 PSI.
I was tempted to test to 150PSI but me don't want to risk my evap.







Let see how low vacuum microns I can get in another 6 hours.


Will update the work progress....................

juanmor40

@S-Prihadi I think the missing refrigerant is likely unfilled, . However, you must also look at the brighter side, that your system may cool a lot better once you put it back to specifications. Finger crossed

S-Prihadi

Yes, will be cooler but how much extra wear and tear have occured ?

See how wet the HVAC system is, easy to see from my vacuum pump oil color. I forgot to buy spare vacuum pump oil !!!

Today is the last business day for Indonesia.
The Moslem Eid al-Adha is on 21st April and shops won't open at least 7 days from tomorrow.
So me can't buy vacuum pump oil. I lost 26 microns due to this wet oil. My pump can only pull to 35 microns instead of 9 microns, read at pump own fitting.


I screwed up today, because I was so pissed my system is wet.
I use heat gun to assist R134A boiling OFF, works decent. yeah !!!

See the micron reading on the tablet, the R134A gasses off and easier to suck out.


I then needed to wee-wee/urinate and I "stuck" my heat gun to blow hot air to the radiator, so I can leave it for 90 seconds for the toilet.
I forgot to push aside the cooling fan blade from direct heat and it melted in a mere 90 seconds......... DUGGHHH !!!!


Right photo is after I tried to fix it.....LOL.
I hope this damage will not produce any significant imbalance.


.

juanmor40

I have been in those situations where everything seems to be closed when I need a part or a tool. Not fun, even more, if you have a good chunk of parts out or disconnected.

On the heat gun, what can I say? Rule 1: do not leave active tools unattended or w/o BIG notes for others to NO TOUCH until I am back. Been there with: "I just collected a few things that looked in the wrong place"

Hopefully, that small missing piece will not cause any issues. Doubt it will about 50% load, but you have heard it in "take-off mode", it makes give you a surprise, fingers crossed.

CaliBenzDriver

Surya I am sure your AC's going to cool much better after intensive care.

I was thinking... while your system is empty would be a good time to swap the A/C pressure sensor near condenser... it may have abilities to leak besides supersonic fan.

Do a quick visual inspection for obvious traces of PAG.

If you're missing a good amount of refrigerant, it has escaped the sealed circuit.
🤞

S-Prihadi

Yes Cali, aging pressure sensor replacement is mighty good idea.
I will do that within the next 12 months and I want to replace the desiccant drier cartridge too.
Something like this : https://www.ebay.com/itm/301964125970
Since this is a molecular sieve only, I can go for a good 3rd party one, as long as they are sold vacuum sealed and inside an aluminum bag and not plastic.

================

Since this is my first time doing HVAC work, my work-flow is still bad and un-organized.
I now realized why Tom Lech will only use a 4 port and 4 valves manifold gauge, it save a lot of time and less risk of ambient air introduction.
He uses this one : https://www.hvacandtoolsdirect.com/p...-micron-gauge/
Switching task between vacuuming & injecting nitrogen & injecting/recovering R134A on a 2 valves 3 ports manifold is a BIG pain when we want to maintain deep vacuum within, at all time.
The holy grail of HVAC work is, ambient air is a CONTAMINATION, so we do not want any and hose cleanliness must be 1st priority.
Price wise the SMAN 480V at US$725 is not expensive actually because of the many sensors it has build-in if purchased separately is already US$600 ish and many more capabilities its like free. I drool every time I see it.

==================


So, progress report.

I managed by 9 hours later, to get the microns to 378.


It is not a straight forward straight 9 hours vacuuming. In between I stopped twice to wait for 40 minutes to get moisture to off-gas within the HVAC system and let my vacuum pump cool down too.
The vacuum pump has a standing room fan to assist its cooling, I do not want to burn it.

I then recharge R134A to gross 612 grams.
I took my time practicing the designated R134A injection valve. I am a virgin at this. I want super slow release.







I use gross 612 gram, where NET target is 590 grams because I allow 20-30 grams loss at hose.
So at least I probably get a TRUE FILL of 570 to 580 grams as worse case miscalculation.




Since I do not use the gauge manifold to do the recharge, but direct hose from R134A bottle, and I do not operate my engine to suck out what is left in the hose by using LP port,
I got the chance to see how much a 5 feet 1/4" hose actually can contain liquid R134A.

R134A in liquid form is to be charged to HP port while engine if OFF. Bottle is to be upside down.
The left over R134A if in the charging hose can be charged/injected to the car HVAC, but engine + HVAC need to run and hence create suction at LP port. This must use gauge manifold where its middle zone can share both the LP and HP zone.
Slowly one can adjust the valve to meter the left over R134A at the charging hose to be in vapor state, as compressor can't compress liquid.
As I said, I did not use gauge manifold set, so I dumped to atmosphere the left over R134A remaining in the charging hose.
Here I wished I have a 4 port 4 valve gauge manifold set.


Indeed, the liquid to gas phase of R134A removed so much heat, frozen fitting we get. Approx 35 grams of liquid R134A, seems enough to loose a finger from frost bite...Euugghhhh...scarry stuff.




R134A, approx 590 grams or even 300 grams, it will show same pressure.
Below is after the R134A charging to estimated 590 grams.


I watched a webminar by an automotive HVAC expert, he said a mere 15% fill R134A will show enough pressure to deceive someone thinking their car is NOT UNDERCHARGED.
It must be the saturated state thingy, where liquid and gas form R134A co-exist, until either it cross to superheated/100% gas or subcooled/100% liquid region.

Our TXV is a hard shut off type, hence LP shows only 82.3 PSI while HP shows 100.8 PSI when recently charged.
It will take like an hour or more for these two pressures to equalize and HP side will always be approx 1 PSI higher.


===========================


Before I share the test after recharge, I would like to show the intermittent erratic behavior of N70B1 ( Interior semperature sensor with integrated fan ) caught in video. Dang extra work soon



I wonder, how VIP is N70b1 in terms of operational input it provides for a decision ?
The B10/4 I labeled A basically is the same data as N70b1 I labeled B for its intended data, it is only their vertical position is different.







I found a TSB on N70b1. https://static.nhtsa.gov/odi/tsbs/20...67780-9999.pdf
Since MB uses same design on their cars, more so if the year of production is close, it seems N70b1 is a VIP sensor.
When a B10/4 value is 25C and then B70b1 is 5C, which will the HVAC computer choose ?



SO HOW DID THE STATIONARY TEST GOES, WITH PROPER R134A FILL CAPACITY ?
Well, not much different during stationary , because my HVAC is programmed to save energy. To be fair, it cools a bit faster from start to 4.0 - 5.0 Celsius evap temp.
The compressor is allowed to do full power ( 1.00 to 0.95A ) only at the very start of 30C ambient temp, and then it throttled down to stingy bastar-dz 0.50 ish Amp as soon as 4.0 to 5.0 Celsius evap temperature
achieved. This should change when I set 2.0 C as evap temperature priority and the cooling fan set to HVAC dominant and not low noise dominant.

Also the heater core DIY by-pass valve is still on my list

CaliBenzDriver

OCP remote temp sensor

Both our cars share this A/C sensor issue with the remote temperature telling goofy lies.
It's almost certain, a lot more people are unknowingly affected just as well.

The temp reading is bad already at the OCP level even before going anywhere.

The hardware looks ok.
I've tried another OCP with identical results.

To me this appears to be like an OCP software bug or some missing supply voltage to drive that sensor... I don't know what's behind that amazing game

The sensing element looks like a simple thermistor more so than a diode junction.

We can try to test the OCP Module :
pock around for voltage and
scop around with OCP diagram...
the usual GND, PWR, CAN...

Last time I delt with OCP I recall it is belongs to two distinct CANS and a LIN network. It's a busy box built around a nice microcontroller but with buggy software development.

CaliBenzDriver

I think we want to double check that LIN network transfers. Make sure the pins are live. We're not go8ng to decode the datastream... just do minimum sanity checking.

The video shows very interesting temp data instabilities in a way I could see classified as poor LIN transfers out of OCP ???


> On a limb:
Are we going to learn something helpful here that in turn will help with the amazing story with the poor voltage control below 12.6V using best Korean knockoff LIN sensor by HYundai ?? They go helped getting started !

Q: Who knows ... ?


Let's record the Release Numbers information for our current FIRMWARE - You have the Ninja updates database to boost that to last... hopefully just a cheap software bug or code mismatch between modules...
Stable codetree is pretty much the opposite of the bleeding edge development - We are the paying Beta testers 👏

S-Prihadi

For my car, by wiring diagram N70b1 is called N70/3b1 , because a Panoramic car OCP is called N70/3, not N70....but the module for Pano is called A98 and is using CAN B.
On Xentry, for my car there is no N70 of any kind

So N70/3 is like shared *****-house and only using LIN 2 network, not CAN-BUS.
If N70 , this one is using CAN B.





N70/3b1 interior temp sensor in green circle.

Above is pe54.21-p-2107-97dab





Above is pe54.21-p-2107-97dab



Above left side is : pe77.20-p-2200-97daa, Electric wiring diagram for panoramic sliding roof, Model 212 with code 413 (Automatic panoramic sliding sunroof) Control unit A98


So the data flow as follows :
N70/3b1 interior temp sensor physically located at OCP N70/3 , communicate via LIN2 to N10/1 Front SAM.
N10/1 Front SAM uses X30/32 CAN B distributor to communicate to AAC Air Cond Controller N22/7
No wonder by slow naughty Xentry Pass Thru hanged when I use it pass 20+ minutes monitoring AAC Air Cond Controller N22/7 during last pressure test, here : https://mbworld.org/forums/e-class-w...ml#post8740291

Once upon a time on 17th March 2023.





MB did not indicate LIN2, damn !!! .............as the data channel for N70/3

From : pe83.40-p-2052-97daa, Electrical function schematic for automatic air conditioning, input signals MODEL 212





Now, non pano-roof car, using N70 which is using CAN B





The only way I can test my N70/3b1 thermistor is to measure its resistance directly and apply COLD/HEAT.
To apply HEAT is easy, to apply COLD is more challenging.

CaliBenzDriver

the thermistor component is all right. It's what's done to Xfer that data that introduces chaos.

Where do the LIN network route to... from F-SAM to 10 slow devices - how many of those are acting goofy?

S-Prihadi

Yes, very probable its the software translation thingy ....

juanmor40

In case you want to test it. You can also put it in a plastic bag so it is never wet.

https://academy.fredsappliance.com/r...-of-ice-water/

S-Prihadi

I need to test it as installed. Easier and more accurate as I can see what Xentry is showing.

I found something which can be useful. I will avoid over cooling for fear of cracking

I can get this locally, someone imported this. I hope it does not contain propane like some cooling spray.


Do you know that CRC has freeze spray using R134A 100% ... damn.
http://images.salsify.com/image/uplo...vmnwm3wgy6.pdf
https://www.crcindustries.com/produc...-10-wt-oz.html


=============================

Final test.
HOW IS MY COMPRESSOR OIL CONDITION ?







Well, this is an aging oil. The amber color is from heat for sure and some water moisture, hence not super clear amber.








Probably I lost at HP hose and at that Inspection Sight Glass approx 3-5cc of compressor oil.


============


Return R134A back to car LP port. Slowly, so that R134A is vapor form only sucked by LP port/compresor.



======================


FROM DENSO - COMPRESSOR OIL ONLY

Above : Virgin new oil.












===========================

If a very new car HVAC system or recently overhauled, where its oil has not aged yet and no UV dye, it would look this good.I am soooooo jelous.



==========================


How in the hell can my compressor be cooler than 90C ? No way, not at bumper to bumper traffic.
Ambient temp in engine bay alone can be 96C when engine already loaded and slowed down to crawl speed, and the work of compression of the compressor will give more heat...Dugghhh !!!





Stationary test alone from cold start 30C, compressor body is already at 85C 12 minutes into HVAC operational. Data from May 2021.





Stationary test outside my shaded garage, with Under Hood or engine bay ambient temperature installed. 28th Jan 2023. Typically I do 20 ish minutes test.

90C hot air blowing to a running compressor = VERY HOT COMPRESSOR.


=================


Dang, in this case I might as well do compressor oil "balancing" with new oil, when I do the desiccant drier cartridge replacement. Maybe I can get 60cc new compressor oil swapped, out of the supposedly 120cc total brand new system fill capacity.


Above is assumed "dry" compressor. Oil balancing is when one replaced exact quantity of remaining oil of an old compressor , for the new compressor.
I can then replaced the 2 O-rings on the compressor hose end too.





If I use HP port to recover refrigerant 100% , on a hot engine, I may get 20cc of compressor oil tagging along the R134A.
So total 60cc of new compressor oil to be injected is possible.


.

S-Prihadi

I want to share a tip.

If you want to test oil for very minor water contamination, it is easy. The oil is NOT to be engine oil which can be dirty from by-product of combustion.
Oil for HVAC compressor, air/dive compressor or vacuum pump compressor is what we can test.

Below is from my dive compressor after 30 hours operation. Brand new it has a mild amber tint and not 100% clear like vacuum pump or Denso D8 oil.
Dive compressor has no choice, it will be exposed to water moisture during compression blow-by from compression chamber into the crankcase.
The oil itself is very cool overall, only 60C as highest when in crankcase, but at piston rings it can be 150C plus.

Oil with moisture will turn cloudy when cooled down, just like air, the cooler it is, the less water moisture air or oil can hold.
The hotter it is, more water moisture they can hold.

Same oil below, one at approx 30C and the other cooled down to like 5C.

S-Prihadi

Ha ha ha, I found the nitrogen pressure calculator..........yipeeee !!!

So.........




I'll be dammed, a mere 1F down from 86F = 0.23 psi pressure drop , if starting at 110.5 PSI










Wow, 1C cooler 0.42 PSI lower if from 27C to 26C, starting at 110.5 PSI.



100% Nitrogen or dry air, same calculation

S-Prihadi

TODAY, 23rd April : Seems the N70/3b1 is fixed already. Probably need some attention only.
No problem with it. Perhaps "cycling" the connectors was the cure
I tested HOT and COOL, I vibrate it yada yada ........... its OK, that is after I disconnected all the connectors and test PCB on my table and install back PCB.
Before I remove this Overhead Control Panel ( N703b1 ), for a few seconds I thought vibration was the cause, it was not, it was the glitch I been seeing.

Yesterday I took car for 30 minutes run, bloody N70/3b1 was going COLD/Glitch so often and one lasting 8+ minutes and up to like 2 Celcius.... LOL.







Today evening , HVAC STATIONARY test again and all good , N70/3b1 stable for the full near 30 minutest test after the "love" I given it


===================

The tear down............


INSTALLED AS-IS , RESISTANCE VALUE, at 30.1C. 7.7K Ohms.
I do not know what is the actual stand alone value of this NTC thermistor.


.


HOT AIR TEST







COLD TEST, at best 20C.




.






What is the difference between PTC and NTC thermistor?
For NTC, the resistance decreases when the temperature increases. For PTC, the resistance increases when the temperature increases.


















======




Got the chance to use my Tesa tape. The anti squeak one




I hope the N70/3b1 of mine will not go grumpy again on me .

S-Prihadi

Sorry.......... the back side of the PCB is not attached yet.

CaliBenzDriver

How long would you guess this OCP reset will last? (1, 2, 4,... Wks).

I've seen temp. numbers pretty much random both positves/negatives like -70°F in OCP but shown as blanks in HVAC.

That makes OCP temp unreliable. HVAC doesn't seem to be bothered by this issue. It's an extra factor helping HVAC deliver poor performances.

Are we positive what bus the Temp sensing works with: LIN2 (vs. CAN-B for OCP) ?

I think what we're observing maybe a side effect of fixing painted GND's... but hell, MD don't care much about side effects
​​​​​
The way I look at this is a call to troubleshoot the LIN networks made unriable by unstable voltage reference, here mostly GND to begin with


Now the exciting part:
It's usually hard to be excited about things...here we may have the opportunity to understand how to fix a whole new class of Big Benz Bug:
The UNRELIABLE LIN XFERS that CORRUPTS DATA.


Are any old timers folks of "Pre-CAN" era familiar with this sort of LIN issue ????

Proven Root cause... TBD.

S-Prihadi

Well, I have no idea

The problem with N70b1 or N70/3b1 is, without a scanner one will never know its value.
Without a scanner and recording the scanner page/screen for the entire drive, one would not know its exact occurrence time, duration and bogus temperature value unless one has a friend to keep a constant watch.
I think there could be different algo for N70/3b1 LIN based vs N70b1 CAN BUS based, as I do not see any negative effect ( yet ) of the N70/3b1 showing crazy low under 5 Celcisus for that 8 minutes.
I think the TSB or issue of N70b1 bogus value making HVAC run hotter is for the CAN BUS ones.


I have done the Workshop Coding.

1st one is, set EVAP to be at 2C instead of Humidity Dependent.

One thing I am curious to know, the definition of EVAP temperature is by a sensor 5mm away from evap metal skin as used by MB, or a "calculated" EVAP skin temperature ?



2nd thing I changed is the HVAC priority for radiator fan speed.



NOTE : Workshop Coding need password. I can get from BenzNinja for a 24 hour password for the Ninja-Xentry, but I instead used my Baby Launch Creader Elite


===============

22nd April Test, before 2 new parameters coded. On the road test.

B35T Plus meter is showing middle left vent temperature.


23rd April Test. New parameters coded. Stationary test at outside my garage. This is also where N70/3b1 been "fixed".



The only difference I can see from Xentry for the 2Celsius EVAP set temperature is that the compressor duty cycle seems to be very very slightly more aggressive by 0.05 amps or 5% from maximum or 10% before new parameter coded.
Usually 0.50amp as the common normal duty cycle when evap has reached 4 - 5C, this time I am seeing often up to 0.56 amps as common.
As for radiator fan speed I can't confirm anything "faster" yet, as there is duty cycle meter I use this time.


This is the car moving approx at 70 ish KM/H.



I think the HVAC software is being very conservative.
I also must compare moving car test of 22nd April to another moving car test soon for the new coded parameter.
Stationary test relies heavily on the cooling fan speed logic and to do a mere 60 KM/H equivalent is a big ask from the electric fan.

CaliBenzDriver

Surya, I think the unlimited fan speed at stationary speed is a feature made for your tropical climate... it's legit to maximize cooling of your turbo engine.

AC FAN settings (cools high pressure liquid side)

I am sure it would be a nice HVAC study. I don't think we can discover better cooling is detrimental

We know your alternator is going to make the 1,200.W fan fly with low-impedance 14.4V supply on-board...
> Amp-Clamp meter says what for the fan PWM load ?

Whatch out for world-class glitches near 100A pulsed - This could potentially toast ECU very short distance to the prefuse box fan feed. It's going to be either a proper Benz class drama (large spikes in prefuse) - who knows !
Thank Valeo for their ALT regulators... smarter than Bosch ECU


Meaning your OCP Temp uses LIN to reach AAC ⚡

Your OCP TEMPERATURE is now able to reach your AAC module with coherent numbers.. NICE!!
I have vested interest in understanding why that amazing OCP sensor is ghosting and how to retire this particular issue...
(¿after fixing GND?)

S-Prihadi

Cali,
Your OCP is the CAN BUS type N70 or LIN like mine N70/3?

CaliBenzDriver

I think my OCP is using LIN bus but you are saying regular roof OCP is N70 and that means its using CAN Bus. (Full size roof uses N70/3 over LIN.

I can easily understand how the HVAC AAC module could receive corrupted data. The odd bit is how OCP also have goofy data for an internal data.

S-Prihadi

If a scanner shows N70 during complete module list read out, that means it is CAN BUS type.
For mine, scanner can not read N70/3 existence , but DTC will be issued by N70/3 via its upstream master which is Front SAM or AAC module.

Same like my tranny pedal shift baby PCB board called N135 , the reporting upstream master is the steering column N80.
No scanner or even Xentry can see N135 stand alone, until I go into N80 module.

S-Prihadi

N70b1 interior sensor so far so good.

25 minutes video logging, speed up 12.5 times to become 2 minutes. Tested today, Sunday.




I wonder, why the hell Autel call N70b1 as TM61 ??

CaliBenzDriver

that is really good karma... your N70 is working right now. It could last 🤞

The whole CAN/LIN options are confusing (i believe the crazy temps are over LIN).

Nice ride, miss the cursing