A/C bc27-28 readings

[DAVES89]

The Red cycles from -2 to 10. It hangs for a long time on 0 and -1 before going to -2 and then back to 10. I still get plenty of cold air because I cleaned the evaporator, radiator, and condensor. I also adjusted the blend door [it was way wrong] for max A/C. I followed that up with taking the little rubber boots off each side of the dash and installing duct tape to block off air flow to the doors. 

 Also following a suggestion from kdirk, I installed new weatherstripping at all the vents when I was doing the dash swap to reduce/eliminate any "blow by"..

 The 'vert has always blown cool but I also cleaned all the things as stated above.

 The Black developed a leak in the A/C years ago and while it would work has no freon. This okay by me as the Black is a winter only car.

 I have 6-7 cans of R-12 in back stock and am probably a bit low on freon for the Red but I am reluctant to add a partial can and waste the rest. Besides until it trips for low refrigerent I don't want to start anything.

 

[Ronnie]

12 hours ago, 89RedDarkGrey said:

There must have been only a few Artists who drew the Main diagrams, used for car manuals- because both contain many of the same hand-sketched illustrations- but the FSM does not contain one photo of anything..

 

Would it have broke the bank- to hire someone in house or professional- to snap a few B&W 35mm of something?

 

It's possible that the drawings were done before there was a car in production to photograph. It probably took almost as much time to develop the service manuals as it did to produce the cars. I imagine GM wanted the factory service manuals to be the hands of the mechanics as soon as the cars hit the showroom.  Waiting until the final design of the car was approved before starting on the service manual would have prevented that from happening.. Haynes manuals do have some photographs but I find the drawings in the FSM to be more useful although they do have inconsistencies and errors sometimes.

[KDirk]

Keep in mind too that some illustrations were reused from similar car models. The E platform in which the Reatta and Riviera was built also encompassed the Eldorado and Toronado/Trofeo. Closely related was the C platform that the 85-93 Cadillac Deville was built on.

 

Having owned a 91 and a 95 Deville, and seeing the similarities both in the cars mechanical and suspension systems to the Reatta (less so in the electrical and body of course) as well as the factory manuals, I can tell you that GM recycled some illustrations from the other E and C platform cars. Remember too that the Reatta was largely derived from the Riviera that debuted in 1986, and save for the change from the older 3.8l engine to the first iteration called 3800 in 1988, most of the under hood setup was pretty close to the 86/87.  Point being that some of these illustrations were almost certainly 86/87 versions, and could have been retouched for expected changes. However, these may not reflect final production design depending on what point in development they were drafted.

 

Worth noting also that I'm fairly certain these illustrations were all hand drafted. While CAD was definitely available (and used by GM) in the mid 80's, this was still a transitional period where conventional drafting was heavily used. All in all, I think they did a passable job. There are always inconsistencies in a service manual of this scope, and running changes  are almost never added except by addendum (if even that). 30 years on, nearly nobody has a complete volume of the updates and bulletins that were issued.

 

All this gets back the point that a good seat of the pants troubleshooter will always be needed. Either you have that skill or you don't. I think it can be learned, but there is a certain innate skill needed to develop and exploit that ability fully. Some people just can't think that way from my observations.

 

 

Edited June 17, 2017 by KDirk (see edit history)

[89RedDarkGrey]

3 hours ago, Ronnie said:

It's possible that the drawings were done before there was a car in production to photograph. It probably took almost as much time to develop the service manuals as it did to produce the cars

 

https://en.wikipedia.org/wiki/Haynes_Manual

http://www.p6club.com/content/obituary-terry-davey

 

Very interesting reading.

 

[drtidmore]

17 hours ago, KDirk said:

FWIW, the 91 system is laid out differently with the accumulator mounted in front of the engine up near the coolant resovoir. Line set is different as a result, but functionally is identical.

 

Ok, so that adds to the mystery of exactly WHAT is going on with 2Seater's FSM as what he posted shows the accumulator in the same place as it resides on the 88/89.  I really wish 2seater would post what version of the FSM he owns.  This is one of those things like the errors in the FSM on the ABS diagnostics orientation where it would be great to chase this down to avoid future confusion.

[KDirk]

Ok, looking right now at both the 88 and 89 FSM  final edition, page 1B1-14 full page diagram, figure 17, both matches the page you posted from your 89 final edition. I do see that either the legend was type set again or the frame around the legend was enlarged slightly for 89 as in comparing the two, there is more margin at the bottom edge of the page on the 88 version than 89.

 

Otherwise, everything is identical including the illustration ID number H60400-1B-EK. Note that last bit, EK,  denotes the platforms to which the illustration  was applicable. Bohh E and K platofrms in this case. The Cadillac Seville was built on the K platform, FWIW, and the Deville moved to the K platform in 94. I do have a 90 manual (final) but is mislaid preseneltly. Will try to turn it up and compare to 88 and 89.

 

I'd say 2seater must have posted from a 90 manual, or from one of the preliminary editions given the different diagram. We will get this cleared up one way or another. The 91 FSM has a fully revised diagram that shows  the updated component locations and is on a different page, has a different figure number and illustration ID number.

 

Edit: found The 90 FSM. Diagram on page 1B1-5 matches what 2Seater posted. That said, 1B1-5 is a right hand page, and his photo looks like it is on the left. Illustration ID number is L63000-1B1-E-R1.

 

 

 

Edited June 17, 2017 by KDirk (see edit history)

[2seater]

On ‎6‎/‎16‎/‎2017 at 2:49 PM, Ronnie said:

 

 

On ‎6‎/‎16‎/‎2017 at 2:49 PM, Ronnie said:

 

 

The terminology you are using doesn't seem to be correct. There is a low side temperature sensor, who's resistance changes with temperature -  and there is a low pressure switch that only contains a set of contacts. Those contacts are either open or closed. There is no "low pressure sensor" that I'm aware of that would have a variable resistance reading. Perhaps your '90 is different from my '88???

 I agree completely. I used the term ASSuming my manual has simply reversed the labels, but it is certainly more than that. The sensor near the blower motor has to be the high side temperature. The one just after the orifice tube and before the evaporator should be the low side temperature. The low pressure switch is pretty much midway between the two temp. sensors in the larger return line.  All of my connectors are similar and unlike the ones shown via photo or illustration in the manual.

On ‎6‎/‎16‎/‎2017 at 2:49 PM, Ronnie said:

 

There is also a high side temperature sensor. That sensor and the low side temperature sensor are interchangeable parts if I remember correctly. It has been my experience that the high side sensor doesn't play a part in cycling the compressor on and off under normal operating conditions.

10-4. It explains the similarity in readings.

[2seater]

On ‎6‎/‎17‎/‎2017 at 8:45 AM, drtidmore said:

Ok, so that adds to the mystery of exactly WHAT is going on with 2Seater's FSM as what he posted shows the accumulator in the same place as it resides on the 88/89.  I really wish 2seater would post what version of the FSM he owns.  This is one of those things like the errors in the FSM on the ABS diagnostics orientation where it would be great to chase this down to avoid future confusion.

Sorry for the delay, I was at my cottage for the weekend. No internet and my phone is barely literate. Yes, it is the standard 1990 Reatta/Riviera FSM. Thanks for everyone's patience

[2seater]

On ‎6‎/‎15‎/‎2017 at 9:28 PM, Ronnie said:

 

I missed this on first reading. The low side temp sensor is on the passenger side.

 

 

Just to clarify, my layout is exactly as above, except spread out over a larger area. The pressure switch is located approx. where the high side temp. sensor is shown and the hi-side temp. sensor is to the right of the blower motor in the view above. The locations on my FSM drawing are in the correct locations but labeled incorrectly. I do not know if it because my '90 is early production, last four of vin 0145, or what it is? For example: my trunk light bulb doesn't match the FSM or any bulb chart I have ever seen. It is a single filament. non-polarized, two pin bulb, #1004. No case ground? The occasional discrepancy does make repairs a little more difficult. Like the bulb above, I ordered a few small led replacements, for interior and exterior use, and then find the listings are incorrect when making the change. It does get annoying and leads to delays.

 

My thanks to everyone that has helped me get this sorted out.

[Ronnie]

44 minutes ago, 2seater said:

Just to clarify, my layout is exactly as above, except spread out over a larger area. The pressure switch is located approx. where the high side temp. sensor is shown and the hi-side temp. sensor is to the right of the blower motor in the view above.

 

I took another look at my '88 sensors just to verify and they are basically laid out as shown in the illustration I posted. The high side temp sensor is just to the left of the blower. I'm surprised that they would have made changes to the layout of the AC lines but I'm sure they had good reasons.

[2seater]

I decided to try an experiment to see what my particular car sees as indicated low side temperature. I disconnected the plug at the sensor, made two small copper pins to insert in the connector and connected a 10k trimmer pot across the pins. I adjusted the pot. to various resistance levels and read the bd28 indicated. I would then disconnect one side of the pot. and check the indicated resistance value. This by no means complete as I was doing this alone and a lot of back and forth had to be done. The reason I say this is there is definitely a high and low value for each temperature reading, so a helper would perhaps provide an easier way to find the boundaries but I hope the result below will provide basic usable information:

 

Resistance     indicated temperature

Open                -39

9.96k                -2

9.67                  -1

8.99/9.17           0

8.64                    1

8.18                    2

7.46                    4

6.97                    5

6.72                    6

6.57                    7

6.17                    8

5.63                    9

5.57                    10

2.62                    26 this matches the high side indicated

[89RedDarkGrey]

Would a BRIGHT light shined on the Sunload sensor change any of these values?

[2seater]

1 hour ago, 89RedDarkGrey said:

Would a BRIGHT light shined on the Sunload sensor change any of these values?

I have no idea but I would doubt it. I think the sun load is likely a separate input the climate system uses to adjust the perceived interior temperature? This test should have isolated the sensor input to the BCM to simulate the variable resistance of the thermistor in the line? I should add this was with key on and engine off. I imagine this could spoof the system into running as long as the resistance value was set above -1*C as long as everything else remained in range but didn't seem like a good idea.

[drtidmore]

On 6/19/2017 at 2:24 PM, 2seater said:

I decided to try an experiment to see what my particular car sees as indicated low side temperature. I disconnected the plug at the sensor, made two small copper pins to insert in the connector and connected a 10k trimmer pot across the pins. I adjusted the pot. to various resistance levels and read the bd28 indicated. I would then disconnect one side of the pot. and check the indicated resistance value. This by no means complete as I was doing this alone and a lot of back and forth had to be done. The reason I say this is there is definitely a high and low value for each temperature reading, so a helper would perhaps provide an easier way to find the boundaries but I hope the result below will provide basic usable information:

 

Resistance     indicated temperature

Open                -39

9.96k                -2

9.67                  -1

8.99/9.17           0

8.64                    1

8.18                    2

7.46                    4

6.97                    5

6.72                    6

6.57                    7

6.17                    8

5.63                    9

5.57                    10

2.62                    26 this matches the high side indicated

2seater,

The fact that your trim pot test came up with the same resistance as what my sensor measured at roughly the same temp (i.e. 26C), validates the sensor on my Reatta.  Thanks for taking the time to run this experiment.  

 

To solve the puzzle of what the low or high temp sensor should read at ANY temp, we can approach it a tad backwards.  We KNOW that the BCM was programmed with the appropriate equation and coefficients to resolve the temp based on the resistance, so we can plug in 3 points from the data above and create the curve based on NTP thermistor norms.  Once the 3 coefficients for the Steinhart-Hart equation have been found, then we can plug in ANY resistance back into the equation and see how close it tracks what 2Seater found the BD028 reading registered.  

 

Using the link below, you need to first pick 3 points, about 10 degree apart (I chose 26, 10, 0) and allow it to solve the 3 simultaneous equations.  Then on the right hand side, at the bottom, you can either enter a resistance OR a temp and it will resolve the other.  

http://www.thinksrs.com/downloads/programs/Therm Calc/NTCCalibrator/NTCcalculator.htm

 

So, exactly what does this gain us?  Well, for any suspect high or low temp sensor on the Reatta, as long as we know the temp of the sensor (i.e. infrared thermometer) and the resistance, we can determine IF it fits the curve.  If not, we can then use Ohms Law for parallel resistance (product of resistors divided by sum of resistors) to calculate a trim resistor to move the curve back in line with what the BCM is expecting.  Longer term, this data also means that we can find a suitable replacement thermistor to repair our sensors and we can keep our A/C systems running.

 

I will try and run down a suitable thermistor and post what I find

 

Edited June 21, 2017 by drtidmore (see edit history)

[drtidmore]

On 6/19/2017 at 2:57 PM, 89RedDarkGrey said:

Would a BRIGHT light shined on the Sunload sensor change any of these values?

no, the sunload sensor is used by the A/C programmer to manage the air mix and blower speed and has nothing to do with the BD028 reading

Edited June 20, 2017 by drtidmore (see edit history)

[2seater]

41 minutes ago, drtidmore said:

2seater,

The fact that your trim pot test came up with the same resistance as what my sensor measured at roughly the same temp (i.e. 26C), validates the sensor on my Reatta.  Thanks for taking the time to run this experiment.  

 

To solve the puzzle of what the low or high temp sensor should read at ANY temp, we can approach it a tad backwards.  We KNOW that the BCM was programmed with the appropriate equation and coefficients to resolve the temp based on the resistance, so we can plug in 3 points from the data above and create the curve based on NTP thermistor norms.  Once the 3 coefficients for the Steinhart-Hart equation have been found, then we can plug in ANY resistance back into the equation and see how close it tracks what 2Seater found the BD028 reading registered.  

 

Using the link below, you need to first pick 3 points, about 10 degree apart (I chose 26, 10, 0) and allow it to solve the 3 simultaneous equations.  Then on the right hand side, at the bottom, you can either enter a resistance OR a temp and it will resolve the other.  

http://www.thinksrs.com/downloads/programs/Therm Calc/NTCCalibrator/NTCcalculator.htm

 

So, exactly how does this gain us?  Well, for any suspect high or low temp sensor on the Reatta, as long as we know the temp of the sensor (i.e. infrared thermometer) and the resistance, we can determine IF it fits the curve.  If not, we can then use Ohms Law for parallel resistance (product of resistors divided by sum of resistors) to calculate a trim resistor to move the curve back in line with what the BCM is expecting.  Longer term, this data also means that we can find a suitable replacement thermistor to repair our sensors and we can keep our A/C systems running.

 

I will try and run down a suitable thermistor and post what I find

 

Yikes! I think I got lost around the second line, but the calculator is pretty cool. I must admit, the two different curves shown are a bit confusing, at least to me. It looks like the calculator sort of idiot proofs it though giving hard numbers rather than trying to interpret the curve. Would it be helpful to widen the range of the readings or hone in on the areas we are interested in? I am guessing that temps. around normal ambient would be what users might first encounter, engine off and after sitting for a while? The two readings that both indicate zero were discovered by accident and not by plan. The trimmer pot is not the easiest to finally adjust with Homer Simpson fingers :)) If we have enough data, I am good with that too.

[drtidmore]

6 minutes ago, 2seater said:

Yikes! I think I got lost around the second line, but the calculator is pretty cool. I must admit, the two different curves shown are a bit confusing, at least to me. It looks like the calculator sort of idiot proofs it though giving hard numbers rather than trying to interpret the curve. Would it be helpful to widen the range of the readings or hone in on the areas we are interested in? I am guessing that temps. around normal ambient would be what users might first encounter, engine off and after sitting for a while? The two readings that both indicate zero were discovered by accident and not by plan. The trimmer pot is not the easiest to finally adjust with Homer Simpson fingers :)) If we have enough data, I am good with that too.

Getting a few more BD028 readings from lower resistance, say up to 35C would be nice so as to provide additional verification that the curve still fits, BUT I suspect we already have a good fit.  

Edited June 20, 2017 by drtidmore (see edit history)

[2seater]

Well, it's in the fifties and raining right now, so a little time in the garage would be a good diversion. Welcome to the first day of summer.

[89RedDarkGrey]

56 minutes ago, 2seater said:

I will try and run down a suitable thermistor and post what I find

 

Would these websites and it's product help here?

 

http://www.omega.com/pptst/55000.html

and

http://www.resistorguide.com/thermistor/

Edited June 20, 2017 by 89RedDarkGrey (see edit history)

[2seater]

44 minutes ago, drtidmore said:

Getting a few more BD028 readings from lower resistance, say up to 35C would be nice so as to provide additional verification that the curve still fits, BUT I suspect we already have a good fit.  

I should have switched to a lower resistance pot as the 10k gets really fiddly down low, but it worked out okay. This what I found:

 

k ohms           indicated temp. *C

2.918                23

2.660                26

2.390                28

2.110                31

1.885                33

1.784                35

1.675                37

1.555                39

 

This my plug for the sensor with pins inserted. Maybe not the best choice for background but you can see it is different than any illustration or photo I have seen?

 

 

[drtidmore]

11 minutes ago, 2seater said:

I should have switched to a lower resistance pot as the 10k gets really fiddly down low, but it worked out okay. This what I found:

 

k ohms           indicated temp. *C

2.918                23

2.660                26

2.390                28

2.110                31

1.885                33

1.784                35

1.675                37

1.555                39

 

 

 

 

Don't know if you have plug these latest #s into the calculator yet, but they show the problem with most ANY thermistor curve and that is no curve is a perfect fit.  As the low temp side is the one that is of the most interest, I stayed with the temps from 31C and down.  At those temps, the curve pretty well predicts what you found the BD028 to read vs resistance.  

 

The only reason I am even interested in the higher temp values is to allow us to come up with a simple measurement using an ohm meter that will determine how badly any specific low side temp sensor has drifted (i.e. easier to measure at ambient and installed rather then pulling it to measure at 0C).  Since we now know what resistance we need to get the BCM to calculate 0C, we could also test the sensor in an ice water bath if accuracy was desired.  It is really the resistance of the thermistor at 0C that determines how well the BCM controls the compressor cycling as well as getting the evaporator as close to a constant 0C state (i.e. slight overcharging) as possible.  

 

The code in our BCMs was set to work sufficiently well across the range of thermistor curves GM expected from the sensors, so we should not get overly concerned with absolute accuracy.  

 

This information does give us the information to find a suitable replacement thermistor that we can use to salvage existing housings.

[2seater]

Unfortunately it is unknown how close the readings are to the borders between cells, so that may be part of the issue in getting a precisely predictable curve. I am sure it is reasonably close, and of course more data points in the calculator would probably help too. In a perfect world, a series of trimmers in the desired ranges, or maybe a ten turn pot, plus a way to watch as the dashboard indicator changes would better establish the center of each cell. The ECM is decoded pretty well. Too bad the BCM isn't.

 

On a related subject, if this drift is normal, how well can we trust the plethora of other similar ones, all through the systems, coolant, air temp, even in the MAF? I'm sure GM expected some of that, but maybe not for going on 30 years

[drtidmore]

22 minutes ago, 2seater said:

On a related subject, if this drift is normal, how well can we trust the plethora of other similar ones, all through the systems, coolant, air temp, even in the MAF? I'm sure GM expected some of that, but maybe not for going on 30 years

Thermistors typically don't exhibit a drift issue except under high temperature conditions (well outside the range they are subjected to in our systems).  What I actually believe is happening is NOT thermistor drift but the weld/solder joints between the thermistor leads and the support leads are becoming brittle and introducing added resistance (i.e. making the temp seem lower than actual).  I have not had one of these out where I could investigate if the leads were spot welded or soldered.  If soldered, a simple reheating with flux followed by a good cleaning with denatured alcohol likely could cure the problem.  If they are spot welded, I would still attempt to solder the joints as that might well restore the proper conductivity.  Cleaning thoroughly after soldering is an absolute must!

 

We are going to need to figure out a good solution regardless as even though other GM cars of the same vintage used these sensors, this is another one of those items that is approaching unobtainium status! 

 

[Ronnie]

I know this isn't scientific but if your goal is to determine if the sensors are reading correctly isn't this test sufficient for doing that?  A/C Temp Sensor Troubleshooting  I thought the goal of the sensor testing was to determine if a sensor was keeping the AC system from cooling properly. It would seem that if you have a half dozen different sensors agreeing on the ambient temperature (after all sensors have equalized overnight) that it could be assumed that they are all reading correctly. If you have one reading, (BD28 for example), that is not in step with the others I think you could assume it is probably bad.

[drtidmore]

9 minutes ago, Ronnie said:

I know this isn't scientific but if your goal is to determine if the sensors are reading correctly isn't this test sufficient for doing that?  A/C Temp Sensor Troubleshooting  I thought the goal of the sensor testing was to determine if a sensor was keeping the AC system from cooling properly. It would seem that if you have a half dozen different sensors agreeing on the ambient temperature (after all sensors have equalized overnight) that it could be assumed that they are all reading correctly. If you have one reading, (BD28 for example), that is not in step with the others I think you could assume it is probably bad.

Ronnie,

In order to correct a drifting sensor, even if short lived, you have to know its resistance at a specific temp. Getting a good thermistor curve allows us to predict the proper resistance at ANY temp and therefore determine how much is out of range.  This is much quicker and more accurate than hit or miss on the trimming resistor value.  

 

That said, the comparison against the other temp registering sensors on the Reatta after sitting overnight is an easy way to see IF the sensor is out in the weeds but again, we need the data 2seater provided to create the thermistor curve that the BCM expects in order to accurately trim the sensor OR find a suitable replacement. 

 

Of course none of this would be important IF we could buy new sensors!

Edited June 20, 2017 by drtidmore (see edit history)

[Ronnie]

10 minutes ago, drtidmore said:

This is much quicker and more accurate than hit or miss on the trimming resistor value.  

 

I agree. It did take me a lot of trial and error to get the correct resistor to put my sensor back in range where the AC cooled properly. I thought I ended up with a 25 ohm resistor but it must have been a .25 ohm after seeing the numbers 2seater posted.

 

I think your first long post was a little over my head so I missed the point.of what you and 2seater were trying to accomplish. Thanks for posting.

[2seater]

I know this has sort of taken on a life of its own but it appeals to my curiosity. I did a quick hand plot of the readings I posted.and I noticed a couple of obvious outliers. I think I will try this in reverse based on the predicted values from the online calculator to see how close it is. The high temperature area seems pretty smooth, even done by hand, so I will try just the lower temperature area.

[2seater]

I used the calculator to get predicted values based on my data points of 23*C, 10*C and 0*C. I added a second column of predicted resistance values to the right of the original numbers. The predicted values worked perfectly and showed the correct value in diagnostic bd28. This also provides a little more definition of the range of acceptable resistance for each temperature. The only place the prediction was off was at temperatures above the 23*C I used for the calculator. It predicted 2.501k would be 26*C, but bd28 showed 27*C. I found from the previous test that 2.62-2.66kOhms works for 26*C so perhaps I should have used a higher value in the calculator to cover the entire range? I will try that to see if the predictions fall more line.

EDIT: I ran the numbers again using 37*C as the top number and everything falls in line nicely and the low temperature stays within the needed range. I guess that calculator is smarter than I am

original-predicted     indicated temperature

Open                               -39

9.96k  9.893                    -2

9.67    9.436                    -1

8.99/9.17 9.00                  0

8.64    8.583                     1

8.18    8.184                     2

7.46    7.439                     4

6.97    7.091                     5

6.72    6.758                     6

6.57    6.441                     7

6.17    6.137                     8

5.63    5.847                     9

5.57    5.570                    10

2.62/2.66                         26 this matches the high side indicated

 

 

Edited June 20, 2017 by 2seater (see edit history)

[drtidmore]

It is pretty apparent that the thermistor used by GM in our Reatta's have a value of around 2.6Kohms@26C based on the data 2Seater has gleaned by spoofing the BCM with a trim pot to allow a range of resistances vs CRT BD028 readings to be logged.  This would confirm what I read on my installed sensor (2.624K ohms@25.56C).  I have been searching for a suitable replacement thermistor that we might be able to use to "repair" an failed low temp sensor, by soldering in a new thermistor.  While at the time that GM designed the Reatta's A/C system the thermistor chosen might have been commonplace, but similar value thermistors are not on the radar presently.  I can find all manner of 2.2Kohm@25C and that would work by simply adding a 400 ohm resistor in SERIES with the thermistor (i.e. placed in the lead connecting to the low temp sensor housing).  Note that this is NOT the same as putting a resistor in parallel with the sensor in order to trim its readings DOWNWARD slightly.  

 

 

Edited June 21, 2017 by drtidmore (see edit history)

[2seater]

I lost my long winded reply :((

 

I will stop looking at tweaking the curve as it seems we are pretty close now, unless it would be helpful to verify the numbers in the ambient temperature area where the average user would test, say 15*C-35*C. If we are good enough now, great. My math skills seem to be having difficulty. This is just an exercise to get the concept down but this is my example: Per Ronnie's link I checked bd27 = 21*C, bd28 = 19*C against my temp. gun = 20.9*C. So I would need to reduce my resistance approx. 200 Ohms, in this fictional example, to correct bd28.

 

I know nothing about the performance of thermistors, but if 25*C is a common rating point, it would seem 2.75k-2.8k would be in line with what we need. I defer to drtidmore on all things related to this subject.

 

Edited June 21, 2017 by 2seater (see edit history)

[drtidmore]

1 hour ago, 2seater said:

I will stop looking at tweaking the curve as it seems we are pretty close now, unless it would be helpful to verify the numbers in the ambient temperature area where the average user would test, say 15*C-35*C. If we are good enough now, great. My math skills seem to be having difficulty. This is just an exercise to get the concept down but this is my example: Per Ronnie's link I checked bd27 = 21*C, bd28 = 19*C against my temp. gun = 20.9*C. So I would need to reduce my resistance approx. 200 Ohms, in this fictional example, to correct bd28.

 

I know nothing about the performance of thermistors, but if 25*C is a common rating point, it would seem 2.75k-2.8k would be in line with what we need. I defer to drtidmore on all things related to this subject.

 

I think we are good with the readings at this point.  I am not finding a thermistor in the proper resistance range, but as I mentioned a 2.2Kohm@25C is common these days which we could use with a series installed 400ohm resistor (the curve for the NTP thermistor itself would be ok just offset which the 400ohm resistor would correct).

 

As for the fact that your BD028 seems to indicate a lower temp than the BD027 which is confirmed by your infrared thermometer, you ARE giving up a couple of degrees of cooling.  As the resistance of the thermistor rises as the temp falls, the curve predicts that your low temp sensor should read about 3.6Kohms.  Did you think to get a direct resistance measurement on the sensor at the same time?  Assuming that it does read around 3.6K@21C, then we need to trim it back to around 3.28K to restore accuracy, which would be a 37Kohm resistor in PARALLEL across the thermistor.  

 

Before you do anything however, I am following a hunch on what is happening with the drifting thermistors and Ronnie may have just confirmed it.  My hypothesis is that the thermistor, which don't typically have a drift failure mode except when repeated exposed to HIGH temps well above anything in the A/C system), is FINE!  The problem may well be the solder joints on either side are becoming brittle as solder DOES tend to do this when constantly temp cycled.  Ronnie located his old failed one and found that the thermistor could be MOVED between the two support/connecting legs of the housing.  He is shipping it to me for further investigation.  If this turns out to JUST be a solder issue, the solution is simple, reflow the solder on both sides.  If this is the case, we have a simple fix for at least some of the failing sensors.  This is not to say that ALL low temp sensor failure modes will be solder joint related, but my hunch is the majority WILL fall into that category. 

Edited June 21, 2017 by drtidmore (see edit history)

[2seater]

1 hour ago, drtidmore said:

I think we are good with the readings at this point.  I am not finding a thermistor in the proper resistance range, but as I mentioned a 2.2Kohm@25C is common these days which we could use with a series installed 400ohm resistor (the curve for the NTP thermistor itself would be ok just offset which the 400ohm resistor would correct).

 

As for the fact that your BD028 seems to indicate a lower temp than the BD027 which is confirmed by your infrared thermometer, you ARE giving up a couple of degrees of cooling.  As the resistance of the thermistor rises as the temp falls, the curve predicts that your low temp sensor should read about 3.6Kohms.  Did you think to get a direct resistance measurement on the sensor at the same time?  Assuming that it does read around 3.6K@21C, then we need to trim it back to around 3.28K to restore accuracy, which would be a 37Kohm resistor in PARALLEL across the thermistor.  

 

Before you do anything however, I am following a hunch on what is happening with the drifting thermistors and Ronnie may have just confirmed it.  My hypothesis is that the thermistor, which don't typically have a drift failure mode except when repeated exposed to HIGH temps well above anything in the A/C system), is FINE!  The problem may well be the solder joints on either side are becoming brittle as solder DOES tend to do this when constantly temp cycled.  Ronnie located his old failed one and found that the thermistor could be MOVED between the two support/connecting legs of the housing.  He is shipping it to me for further investigation.  If this turns out to JUST be a solder issue, the solution is simple, reflow the solder on both sides.  If this is the case, we have a simple fix for at least some of the failing sensors.  This is not to say that ALL low temp sensor failure modes will be solder joint related, but my hunch is the majority WILL fall into that category. 

I need to find a mating plug for the sensors to make a better way to measure the sensor itself. It is easy to bend that fine pin in the sensor with a mini-alligator clip. I thought I had some of the little Weatherpack/Metripak pins/sockets but I must be out? I am sure you are correct about the cooling not being optimized, but that will likely wait until I drain the system and replace the compressor. I might just switch the two temp. sensors as the high side one seems to be pretty accurate. Maybe a little junkyard searching will yield a few sensors?

 

That Ronnie is on the ball with just the right part. Unless I am wrong, it is unfortunate the system has to be drained to pull the sensor. I cannot find any of my local shops I use have equipment, or will do, R12 systems. Maybe they just don't want to mess with it? Thank you for the information on the sensor shift.

[Ronnie]

6 minutes ago, 2seater said:

Unless I am wrong, it is unfortunate the system has to be drained to pull the sensor.

 

I won't ever admit doing this but it is possible to swap the sensors without loosing all the freon or needing to pull a vacuum on the system. If you do it on a cold day Freon pressure in the system drops dramatically. You could probably do it on a warm day to but a cold day is better.

 

Be prepared to do the swap by having everything you need at hand - including gloves to protect your hands from cold escaping gases. Swap out the sensors while freon is coming out. Moisture can't go into the system while freon is coming out. Once the swap is done recharge and you're ready to go.

 

A guy who lives in East Tennessee did this and it worked well.

[drtidmore]

6 minutes ago, 2seater said:

I need to find a mating plug for the sensors to make a better way to measure the sensor itself. It is easy to bend that fine pin in the sensor with a mini-alligator clip. I thought I had some of the little Weatherpack/Metripak pins/sockets but I must be out? I am sure you are correct about the cooling not being optimized, but that will likely wait until I drain the system and replace the compressor. I might just switch the two temp. sensors as the high side one seems to be pretty accurate. Maybe a little junkyard searching will yield a few sensors?

 

That Ronnie is on the ball with just the right part. Unless I am wrong, it is unfortunate the system has to be drained to pull the sensor. I cannot find any of my local shops I use have equipment, or will do, R12 systems. Maybe they just don't want to mess with it? Thank you for the information on the sensor shift.

I have read on the forum that if you are quick on the swap, you can pull the old low temp sensor and screw on a new one without loosing all the freon, although you WILL loose some but at least you don't have to pull a vacuum if you are quick on the swap.  R-12 just got to be such a legal nightmare that shops stopped dealing with it.  

 

As for an APB on salvaged low/high sensors, I am going to see what I discover with Ronnie's failed sensor first, but if you run across ANY of them (GM used them on almost all their high end cars of that period), then I think it is good idea to pull them, regardless of condition, now that we have a good way of testing them and potentially a way to repair them.

 

 

[89RedDarkGrey]

What I have read- there is a difference between vapor and liquid regarding "saturation" temperatures and pressures. The temp sensor is downstream from the Orifice Tube, which acts like a venturi, dramatically lowering the pressure, correct?

 

I found some charts- if you can use them, and a company that makes all stuff to order if needed. I'm just trying to help, not run anything or post stupid stuff.

 

http://durexindustries.com/

 

[drtidmore]

40 minutes ago, 89RedDarkGrey said:

What I have read- there is a difference between vapor and liquid regarding "saturation" temperatures and pressures. The temp sensor is downstream from the Orifice Tube, which acts like a venturi, dramatically lowering the pressure, correct?

 

It is a tad more complex than just a simple venturi.  It has to do with expansion of the refrigerant liquid as it exits the orifice.  No expansion=no lowering of the energy density=no ability to absorb energy in the evaporator.  The expansion is due to both a large area into which the high pressure liquid exiting the orifice can expand, but in addition, the compressor is creating a dramatic pressure differential aft of the orifice.  If you take ANY compressed gas and allow it to rapidly expand, what we perceive is the it gets REALLY cold, but in reality all that has happened is that the energy density has been vastly reduced and less energy density=lower pressure=lower temp.  Now expanded and much less energy dense, the gas can travel thru the evaporator absorbing energy (heat) from the interior of the car which then after passing thru the suction side of the compressor which compresses the lower density gas into a much higher density gas then is passed thru the condenser where energy is removed turning the high density gas into a high density liquid and back it goes to the orifice. 

[89RedDarkGrey]

Like PC duster (an excellent AC refrigerant) when sprayed as a gas it is basically room temperature- but invert the can and spray- and super cold liquid shoots out, and the can gets icy too. The duster could be substituted in experiments for thermistors, too? It's way cheaper than R12 or 134a, but just as good as R12 and way better than 134a.

 

Thank you for the explanation.

[drtidmore]

And back to the beginning of this thread, we want to "manage" the expansion amount such that it does NOT overexpand, dropping the temp at the low temp sensor, below the -2C trip point, which we do by ever so slightly overcharging the system to create a tad higher suction side pressure as this is a closed system which can only hold so much refrigerant (again, more refrigerant=higher pressure=higher density=higher energy density and LESS room to expand).  In the good old days of GM's suction throttling system, this was done dynamically and the compressor ran all time with the suction throttling valve controlling the suction side pressure precisely at the point to achieve 0C in the evaporator at all times (there is a direct relationship between suction side pressure and evaporator temp).  

Edited June 21, 2017 by drtidmore (see edit history)

[drtidmore]

8 minutes ago, 89RedDarkGrey said:

Like PC duster (an excellent AC refrigerant) when sprayed as a gas it is basically room temperature- but invert the can and spray- and super cold liquid shoots out, and the can gets icy too. The duster could be substituted in experiments for thermistors, too? It's way cheaper than R12 or 134a, but just as good as R12 and way better than 134a.

The problem with using PC duster for evaluating a thermistor is that we need to know the temperature of the exposed thermistor and PC Duster is a pretty uncontrollable environment coming out the can!  However, it certainly would allow you to watch the resistance of a thermistor rise as the temp fell.

[89RedDarkGrey]

1 hour ago, drtidmore said:

The problem with using PC duster for evaluating a thermistor is that we need to know the temperature of the exposed thermistor and PC Duster is a pretty uncontrollable environment coming out the can

 

If 2seater can make a Flow Bench, TEVES accumulator tester- why not a "Thermistor tester"? Tubing, fittings, an evaporator, etc.- maybe a gutted junkyard minifridge? A NEW one is $60 at Wal*Mart, used even cheaper. All the needed components- just smaller.

 

It'd be cool to tinker with