Frequency of 5V signal from EV fan clutch to the PCM

santon

Original poster
Member
Jun 3, 2020
94
Israel
I am trying to develop a small controller to emulate the electro-viscous (EV) fan clutch that was used on early models of Trailblazers (prior to 2008, I think). With the EV fan clutch, the speed of the cooling fan is controlled by the 12V PWM signal from the PCM through the solid-state fan relay 45. The EV fan clutch also has the Hall effect sensor which returns 5V square wave of variable frequency to the PCM; this signal is basicaly a measure of the fan speed. On my 04 Trailblazer, I tried to measure the frequency of this 5V return signal. When the engine was cold, the frequency was about 40-43 Hz. When the engine reached the operating temperature and with the A/C on, the frequency of this signal increased up to 77-78 Hz. However, it seems that the fan has not reached the maximal speed. I don't have a GM Tech 2 or similar device to command the speed of fan's rotation. Also, I did not find any valuable information on this subject. Could somebody tell me what is the maximal frequency of the 5V return signal from the Hall effect sensor of the EV clutch to the PCM?
 
Last edited:

Mooseman

Moderator
Dec 4, 2011
25,260
Ottawa, ON
Someone else had done exactly that but had problems emulating the desired vs actual speed readings and would eventually throw a fan speed code. I am unable to find that thread.
 

TJBaker57

Member
Aug 16, 2015
2,897
Colorado
I can direct you to information that will allow you to read what the PCM reports for actual fan speed, plus the PCM desired fan speed. It would require a $13 bluetooth OBD2 adapter and a smartphone app. This should allow you to compare your measured frequency with a PCM reported fan speed.

If you are interested I can also guide you in commanding a duty cycle for the cooling fan using a serial terminal app and that same OBD2 adapter. Would this help?
 

santon

Original poster
Member
Jun 3, 2020
94
Israel
I can direct you to information that will allow you to read what the PCM reports for actual fan speed, plus the PCM desired fan speed. It would require a $13 bluetooth OBD2 adapter and a smartphone app. This should allow you to compare your measured frequency with a PCM reported fan speed.

If you are interested I can also guide you in commanding a duty cycle for the cooling fan using a serial terminal app and that same OBD2 adapter. Would this help?
Thanks. Yes, the guide on how to use these tools would be great. I have some Bluetooth OBD2 adapter and I use it with Torque Lite application on Android phone. I believe another application will be needed. Can you command the duty cycle through Class 2 serial data? I would be glad to know how to do this; I have seen your thread on this subject and I wanted to learn it for another project (PCM and instrument cluster bench tester).

Meanwhile, it seems that 1Hz of the Hall effect sensor signal equals 10 rpm of fan rotation speed. I have seen some diagnostic guides on YouTube in which the fan was commanded 100% duty cycle and reached about 2000 rpm; the frequency of the Hall effect sensor signal should be about 200Hz under these conditions. If so, the bandwidth of of this signal is withing the range of 40 - 200Hz (0 - 100% duty cycle command). I just need to verify that these numbers are correct.
 

budwich

Member
Jun 16, 2013
2,027
kanata
Perhaps I don't understand the EV clutch but I don't think it controls the speed of the fan "directly". It controls how much "clutching" is required which then is "computed" into fan rpms after the speed of the engine is taken into account. Hence, 100% percent at an engine rpm of 800 is going to be XXX while at an engine rpm of 2200 rpm might be YYY.

The "duty cycle" for activation has more to do with how much "power" is sent to the clutch solenoid to control the amount of fluid moving in the clutch.
 

santon

Original poster
Member
Jun 3, 2020
94
Israel
Perhaps I don't understand the EV clutch but I don't think it controls the speed of the fan "directly". It controls how much "clutching" is required which then is "computed" into fan rpms after the speed of the engine is taken into account. Hence, 100% percent at an engine rpm of 800 is going to be XXX while at an engine rpm of 2200 rpm might be YYY.

The "duty cycle" for activation has more to do with how much "power" is sent to the clutch solenoid to control the amount of fluid moving in the clutch.
Well, the EV clutch has two main separate components (please see attached): 1) the PCM-controlled valve for the viscous fluid, or the "clutch" itself; 2) the Hall effect sensor for reporting the actual speed. The duty cycle of the control signal (B, WHT) determines how much "clutching" is needed. However, the signal that reports the actual fan rpm (D, DK BLU) is not computed from the control duty cycle. This signal is rather measured with the help of the Hall effect sensor. Maybe the PCM is taking the engine rpm into consideration while it compares the actual vs. desired fan speed etc, but the operation of the fan clutch seems to be very simple: there is a relationship between the command "duty cycle" signal and the output signal of the Hall sensor. Thus, the more "power" is sent to the clutch the faster the fan rotates, and the higher the frequency of the output signal from the Hall sensor to the PCM. So, I wanted to create a simple controller that will emulate the EV clutch: measure the duty cycle of the command signal, and send another signal of the appropriate frequency to the PCM. This controller (or PCM "fooler") could be useful for eliminating the EV clutch when the electrical fans or thermal fan clutch from TB 2008-2008 are installed.
 

Attachments

  • TB 2004 fan circuit.pdf
    36.8 KB · Views: 8
Last edited:

Mooseman

Moderator
Dec 4, 2011
25,260
Ottawa, ON
Maybe the PCM is taking the engine rpm into consideration while it compares the actual vs. desired fan speed etc,
That's probably why the other guy that tried this it didn't work. Too may variables. That's why I just recommend getting a tune to turn off the codes.
 

budwich

Member
Jun 16, 2013
2,027
kanata
I think we aren't talking the same "words" ... I was questioning this statement that you made at the end of your post... "If so, the bandwidth of of this signal is withing the range of 40 - 200Hz (0 - 100% duty cycle command)"
As I was pointing out, you could have a 100% command and the rpm speed seen would / could be low ... ie. 100hz in your "math scenario" because the engine may be just above idle after a "hot run".
But if the engine rpm is higher and 3000 rpm, at 100% "command" then the sensor may see "3000rpm" depending on what the "gear ratio" inside the clutching is doing. The correlation between command requested (%) is not "1:X". Its somewhat a percentage of slip at a given engine rpm. I think you were saying that at 100% the sensor output would be "2000 RPM" (your example)... I don't think that follows. It will be some frequency (dependent on engine rpm and slip) that the PCM then "massages" to get an expected rpm result.

I think you are under the impression that power is driving the fan clutch. Power is driving the "slipability" of the fan clutch. The engine (rpm) is driving the fan and ultimately impacts its speed / rpm. The engine may be racing at 4000 rpm and IF the clutch is 100% commanded, then it will spin at 4000 rpm (and the hall sensor will see that) IF there is no slip and IF the internal gearing is 1:1 (it might actually be less).

your statement, "the operation of the fan clutch seems to be very simple: there is a relationship between the command "duty cycle" signal and the output signal of the Hall sensor" might not be true as you need to understand the relationship of the engine rpm to deal with the sensor output.
 
Last edited:

santon

Original poster
Member
Jun 3, 2020
94
Israel
The correlation between command requested (%) is not "1:X". Its somewhat a percentage of slip at a given engine rpm. I think you were saying that at 100% the sensor output would be "2000 RPM" (your example)... I don't think that follows. It will be some frequency (dependent on engine rpm and slip) that the PCM then "massages" to get an expected rpm result.
I see now. Well, I thought that I can do it by this simple "math" relationship. Thanks!
Too may variables. That's why I just recommend getting a tune to turn off the codes.
So, the tune will simply erase the DTC's, right? Maybe I should stick with the regular EV clutch... I have the Behr EV clutch, 3 years and 70K km (43K miles), still working good. Just wanted a small project...
 
  • Like
Reactions: Mooseman

TJBaker57

Member
Aug 16, 2015
2,897
Colorado
I have some Bluetooth OBD2 adapter and I use it with Torque Lite application on Android phone. I believe another application will be needed. Can you command the duty cycle through Class 2 serial data? I would be glad to know how to do this; I have seen your thread on this subject and I wanted to learn it for another project (PCM and instrument cluster bench tester).

Torque Lite does not permit the addition of custom PIDs. At first look it appears to but in practice I have found it would not work. Torque Pro DOES allow for custom PIDs and has arguably the most powerful offering of equation functions but lacks in ease of data gathering and analysis.

The Car Scanner app does a good job of recording and graphing data parameters right within the app and can also export the recorded data to a few different formats. Attached is a file of the fan parameters and a couple more for import to Car Scanner. (Rename the file from a .txt file to a .csv file type)

Here is a short display of 7 minutes at startup of my 2002 4.2 Trailblazer. It can be seen that the desired fan speed never changes (stays at 496 rpm). At first the fan does the normal high rpm as the clutch has fluid accumulated in the working chamber. As the vehicle warms the idle and the resultant fan speed drop slowly. Before the idle comes down fully I rev the engine to approximately 2000 rpm and hold it so the fan clutch fluid will be pumped into the reservoir and out of the working chamber. I then release the throttle and the idle settles to about 600 and the fan speed settles down to about 720. After about a minute I again raise the throttle to about 2000 and it can be seen that the fan speed raises to about the same rpm it had settled out to previously 2000 engine rpm indicating most of the fluid has been retained in the reservoir.

Screenshot_20210218-180246.png


Yes, the fan clutch can be commanded to operate at various percentages of clutch applications through PWM instructions through the class 2 network. This is how the Tech 2 scantool does all its 'magic', through class 2 commands. However, being a fluid coupled device the control is quite limited. I have observed that while an increase in speed is quickly achieved, the opposite, a reduction in speed is quite slow as the fluid needs to pump out of the working chamber and back into the reservoir. I have also observed that lower levels of PWM have largely no effect. Since all my testing has been on used fan clutches, 3 so far, I suspect what I may be seeing is the result of the control valve either not fully seating or perhaps being held slightly open by detritus in the fluid. In this screenshot
I did not command the fan here in any way.

I currently have a PCM, a BCM, and a cluster on my workbench. Looks horrific but it is just for experimenting...

IMG_20201019_134034~2.jpg
 

Attachments

  • GMTN_CarScanner.txt
    396 bytes · Views: 3

budwich

Member
Jun 16, 2013
2,027
kanata
I am not sure I understand "I did not command the fan here in any way". OK, so what was the current level asked for by the system? Since you aren't feeding any inputs other than engine rpm (assumed) then I would expect the system's "command" to be some "percentage" which of operating the fluid valve... in this case, I would expect that to be rather low but have you or do know what that corresponds to... for instance, was the "system command" equal to say 20% (cold system)... which ultimately you "saw" as a desired rpm of 500ish. Further, if you now command the system to say 30 or 40 (ie. double based on this exercise), do you see a change in the desired rpm and is it just as stable (expected). Related, what was the engine temp value? sorry for the questions... they probably don't matter but interested in seeing more of this great work and jealous because I don't have the capability (both in body and on the bench). :smile:
 

santon

Original poster
Member
Jun 3, 2020
94
Israel
@TJBaker57, that's a great work indeed! It seems that once most of the fluid is retained in the reservoir, the relationship between the engine rpm and the fan rpm is quite constant (2000 rpm engine/1500 rpm fan). Also, the properties of the fluid could represent an additional variable affecting the above relationship but only initially, after the engine start up. It would be interesting to command the fan clutch to the maximal operation (100%) and observe the engine, fan rpm and the signal of the Hall sensor. I will probably try to do that using your code for the Car Scanner.
 

TJBaker57

Member
Aug 16, 2015
2,897
Colorado
I did not interfere with the normal operation of the cooling fan. I issued no commands to the system, just recorded the systems normal operation.

I have neither seen nor found a PID for the current PWM parameter of the fan speed control. The Tech 2 shows no such parameter, nor have I come across any. I would not really expect to stumble across such if it did exist, it would be difficult to recognize since I practically never see my fan commanded to a higher speed. My PID value for "desired fan speed" called out by a Tech2 as an rpm value has always stayed at 496 rpm unless I have issued a class 2 message to take control of that PCM output and drive it higher. And in those instances the desired fan speed does rise in response to my class 2 message.

I suspect that what I am seeing is a base level of fan speed with a commanded duty cycle of zero. That is, no output signal to the fluid valve. I think there may be such a minimum amount of clutch application going on, maybe by design, maybe not. Tech 2 gives us actual fan speed and desired fan speed but the only time when a PWM duty cycle is seen is when you are controlling the output manually. This is then controlled and displayed in increments of ten percent if memory serves. I think that display is not a PID value but simply a display of the duty cycle being currently sent over the class 2 network.

I may have already done this but I don't remember so might just do it again.... I could remove the solid state fan speed control module from the fuseblock and observe the fan speed again under the same cold start conditions. If it is the same then that might confirm my suspicion that in my case at least, the fan clutch is not receiving any directive to open the fluid valve and the resulting fan speed is just a minimum floor level for this particular fan clutch?

I have recently discovered an odd behaviour related to this... When dead cold (and perhaps elsewhere also) when one yries to spin the fan by hand there is much resistance and the fan stops dead when released. If I spin the fan briskly in the clockwise direction as viewed from the front of the vehicle (opposite the fans normal rotation but applying forct against the internals in the same manner force would be applied in operation) after maybe 30 to 45 seconds of repeated quick spins, the fan releases nearly entirely and will freely rotate. I'm not sure what to make of this. Anyone who has succesfully stopped their fan with the engine running may have witnessed the same phenomenon that when they release the fan it seems to be entirely free and takes a good while before the clutch re-engages.
 
Last edited:

TJBaker57

Member
Aug 16, 2015
2,897
Colorado
@TJBaker57, that's a great work indeed! It seems that once most of the fluid is retained in the reservoir, the relationship between the engine rpm and the fan rpm is quite constant (2000 rpm engine/1500 rpm fan). Also, the properties of the fluid could represent an additional variable affecting the above relationship but only initially, after the engine start up. It would be interesting to command the fan clutch to the maximal operation (100%) and observe the engine, fan rpm and the signal of the Hall sensor. I will probably try to do that using your code for the Car Scanner.


The relationship between fan speed vs engine rpm could be seen as the simple ratio of cranshaft pulley diameter vs fan/water pump pulley diameter. However, I suspect there to be some significant deviation at higher rpms where the horsepower required to drive the fan blade may exceed the clutches ability to maintain a fully locked state.
 

TJBaker57

Member
Aug 16, 2015
2,897
Colorado
I am not sure I understand "I did not command the fan here in any way". OK, so what was the current level asked for by the system? Since you aren't feeding any inputs other than engine rpm (assumed) then I would expect the system's "command" to be some "percentage" which of operating the fluid valve.

It only just now occured to me that I did not clearly state that the recorded graph was NOT an output from the benchtop testbed but was in fact generated from my vehicle. Is this why some of my post wasn't clear?
 

santon

Original poster
Member
Jun 3, 2020
94
Israel
I suspect that what I am seeing is a base level of fan speed with a commanded duty cycle of zero.
I think you are right; I tried to observe the command signal on the oscilloscope and haven't seen any voltage at idle on a cold engine. You can also see it in this video when the command signal is monitored with a bulb (at 7:18).
 

TJBaker57

Member
Aug 16, 2015
2,897
Colorado
It just occured to me that I should mention that my 2002 has a build date of 10\2001 and to the best of my knowledge my PCM still has the same original calibration. I think I checked that once and determined my PCM has never been updated. So my desired fan speed and I suspect quite a few other operating parameters are likely different than newer calibrations.
 

Forum Statistics

Threads
23,273
Posts
637,495
Members
18,472
Latest member
MissCrutcher

Members Online