BCS testing data
Nov 2, 2008, 12:17 AM
#1
BCS testing data
The BCS testing thread
June 29, 2010: Just realized that I never updated this post to include my testing of the Grimmspeed and OEM EGR 3-port solenoids. My findings are in this post: https://www.evolutionm.net/forums/7977427-post122.html
November 22, 2008: This post has been revised to include the AEM 30-2400 and much more detailed data on boost response to WGDC.
I finally got around to setting up all the equipment needed to do some realistic measurements of how well a BCS can relieve boost pressure at the WGA. The test setup is as follows:
- Air pressure source set to 29 psi.
- Water hammer arrestor used to simulate a WGA
- 30 psi pressure gauge in-line with the simulated WGA
- Tubing, fittings, etc.
- 14.0 V system voltage
- Variable resistor that can be set to anything from 1 to 10 ohms
- BCS wired in series with the resistor.
- 20 Hz square wave signal generator with adjustable duty from 20% to 80%.
I had four different BCS's to test:
- Stock BCS
- GM BCS
- Prodrive BCS
- AEM 30-2400 (MAC 35A)
The stock BCS was set to bleed mode while the 3-port solenoids were setup for interrupt mode. I tested each BCS with a 1-ohm resistor in series, and then I tested the GM BCS and the Prodrive BCS with a 5 ohm resistor and then a 10 ohm resistor in series. For each set of tests, I varied the duty cycle from 20% to 80%. I also measured the pressure with stock BCS at 100% duty. By design, the WGA pressure will be 0 psi at 100% duty for the 3-port BCS's.
Latency Measurements
I observed the BCS voltage waveform to measure the latency of each BCS. Here are the latency values:
Turns out that all of the 3-ports are way faster than the stock BCS. The AEM is the fastest. The Prodrive BCS is a bit slower than the GM BCS, especially with a 10 ohm resistor in series. I also observed that the MAC 35A and GM BCS latency were very consistent while the Prodrive BCS latency tended to bounce around a little. The higher latency of the Prodrive makes it less suitable to use at higher Seems that the $30 MAC 35A and GM BCS beat the $90 Prodrive BCS in these two tests.
Pressure Measurement Setup
I ran several variations in the hardware setup. Things that were varied:
- The use of a near-stock size pill at the "compressor".
- The use of a (Vishnu style) bleed hole. I tried several different locations for the bleed hole.
- Operation frequency of the BCS.
- The value of the resistor used in series.
Because of the on-off nature of solenoids, the pressure measured at the simulated WGA would bounce up and down, sometimes by very large amounts depending on duty cycle setting, operation frequency, and the particular solenoid. The highest flow solenoids had very large pressure variations due to the fact that these solenoids could bleed off a large amount of pressure during the "off" cycle. Most of the results are presented showing the upper and lower observed pressure at each duty cycle setting.
Effect of Series Resistor Value
Its become the standard to use a 10 ohm resistor in series with the GM solenoid to bring the circuit resistance up to the resistance of the stock solenoid. However, adding a resistor slows the rate at which a solenoid can open as shown in the latency results earlier. Pressure at the "WGA" was measured with the GM and Prodrive solenoids tested with 1 ohm, 5 ohm, and 10 ohm series resistance value at 20 Hz (stock operating frequency). A near-stock size pill was used at the "compressor" for these tests. The effect of resistance on the pressure vs duty cycle was minimal with the higher resistance causing higher pressure in the mid-duty cycle range due to the higher latency. The results are shown in comparison to the stock solenoid. At 30 Hz, a similar result was obtained except that the Prodrive solenoid was too slow to operate at 30 Hz when a 10 ohm resistor was used. My feeling is that a 5 ohm resistor with the GM and Prodrive solenoids is a better choice than a 10 ohm. 5 ohm still brings the overall circuit resistance to within stock specification range while keeping the latency lower.
Effect of Hardware Setup on WGA Pressure
There's a ton of data to present. These tests were all done with a 1 ohm resistor in series. I tried three different hardware configurations at 20 Hz and at 30 Hz. These configurations with their designations are:
- (N) - No pill at the compressor
- (1) - Near-stock size pill at the compressor
- (3) - Vishnu style bleed to atmosphere between BCS and WGA
The last hardware configuration was something I wanted to try because of the results I saw when I simulated the Vishnu hardware setup which uses a bleeder to atmosphere in combination with the stock BCS. The plot below compares the response of the stock BCS with and without a Vishnu bleeder. The effect of the bleeder is to lower the pressure curve and to reduce the pressure oscillations caused by the on-off nature of the solenoid. I tried the bleeder with the 3-port setups to see what effect it might have on pressure oscillations.
First is the GM solenoid at 20 Hz. As expected, the GM solenoid allows for 0 pressure at the WGA at 100% duty. What turned out to be most important in my opinion were the size of the pressure oscillations. The GM clearly has larger pressure oscillations than either of the setups with the stock solenoid. The use of a bleed between the BCS and the WGA did help reduce the oscillations somewhat in the mid-duty cycle range. By having the Vishnu bleed in the system, it is possible to have finer control over the boost pressure. The reason is that the minimum boost value is increased, so the boost control range is narrower for the full range of duty cycle, so each increment in duty cycle results in a smaller increment in boost pressure change (finer control) compared to no bleeder. The effect of using a near-stock size restrictor at the "compressor" was to slightly lower the pressure at the WGA and slightly reduce the pressure oscillations. What I didn't show here was the smaller bleed also reduced the response time of the system to rapid increases in pressure at the compressor, e.g., during spool up. The slower response could cause the WGA to open a little late and cause some boost overshoot.
Next is the Prodrive at 20 Hz. The basic response is the same as the GM, except that the pressure oscillations are larger. I attribute this to a lower flow resistance of the Prodrive. The oscillations get quite large at around 40%-50% duty cycle.
The AEM MAC 35A at 20 Hz goes one step beyond the Prodrive with the pressure oscillations getting to be 7.5 psi at 40% duty cycle and still very large out at 80% duty where people are likely to be most of the time on boost.
Now onto the GM at 30 Hz. tephra has been experimenting with increasing the operating frequency, so I thought it would be interesting to test the GM and Prodrive at 30 Hz. I didn't bother with the AEM at 30 Hz because it is so far out of the optimum envelope compared to the GM and Prodrive. The main effect of the higher operating frequency is to reduce the pressure oscillations without significantly changing the pressure vs duty cycle curve.
Last is the Prodrive at 30 Hz. As with the GM, the higher operating frequency reduces the pressure oscillations. The oscillations are still larger than the GM. The Prodrive looks ok at 30 Hz except it needs a full 14 V with no resistor in series for it to even operate at <30% duty cycle.
Conclusions
The GM and Prodrive have fairly similar behavior, but the GM appears to have the best mix of characteristics: good response time, the right amount of bleed, and its affordable. Although the AEM is very fast, the pressure oscillations are pretty large, probably large enough to affect the boost behavior of the car.
June 29, 2010: Just realized that I never updated this post to include my testing of the Grimmspeed and OEM EGR 3-port solenoids. My findings are in this post: https://www.evolutionm.net/forums/7977427-post122.html
November 22, 2008: This post has been revised to include the AEM 30-2400 and much more detailed data on boost response to WGDC.
I finally got around to setting up all the equipment needed to do some realistic measurements of how well a BCS can relieve boost pressure at the WGA. The test setup is as follows:
- Air pressure source set to 29 psi.
- Water hammer arrestor used to simulate a WGA
- 30 psi pressure gauge in-line with the simulated WGA
- Tubing, fittings, etc.
- 14.0 V system voltage
- Variable resistor that can be set to anything from 1 to 10 ohms
- BCS wired in series with the resistor.
- 20 Hz square wave signal generator with adjustable duty from 20% to 80%.
I had four different BCS's to test:
- Stock BCS
- GM BCS
- Prodrive BCS
- AEM 30-2400 (MAC 35A)
The stock BCS was set to bleed mode while the 3-port solenoids were setup for interrupt mode. I tested each BCS with a 1-ohm resistor in series, and then I tested the GM BCS and the Prodrive BCS with a 5 ohm resistor and then a 10 ohm resistor in series. For each set of tests, I varied the duty cycle from 20% to 80%. I also measured the pressure with stock BCS at 100% duty. By design, the WGA pressure will be 0 psi at 100% duty for the 3-port BCS's.
Latency Measurements
I observed the BCS voltage waveform to measure the latency of each BCS. Here are the latency values:
Code:
1 ohm 5 ohm 10 ohm Stock BCS 10.3 ms Not Meas Not Meas GM BCS 6.0 ms 6.8 ms 7.9 ms Prodrive BCS 6.6 ms 7.4 ms 10.8 ms AEM MAC 35A 4.5 ms Not Meas Not Meas
Pressure Measurement Setup
I ran several variations in the hardware setup. Things that were varied:
- The use of a near-stock size pill at the "compressor".
- The use of a (Vishnu style) bleed hole. I tried several different locations for the bleed hole.
- Operation frequency of the BCS.
- The value of the resistor used in series.
Because of the on-off nature of solenoids, the pressure measured at the simulated WGA would bounce up and down, sometimes by very large amounts depending on duty cycle setting, operation frequency, and the particular solenoid. The highest flow solenoids had very large pressure variations due to the fact that these solenoids could bleed off a large amount of pressure during the "off" cycle. Most of the results are presented showing the upper and lower observed pressure at each duty cycle setting.
Effect of Series Resistor Value
Its become the standard to use a 10 ohm resistor in series with the GM solenoid to bring the circuit resistance up to the resistance of the stock solenoid. However, adding a resistor slows the rate at which a solenoid can open as shown in the latency results earlier. Pressure at the "WGA" was measured with the GM and Prodrive solenoids tested with 1 ohm, 5 ohm, and 10 ohm series resistance value at 20 Hz (stock operating frequency). A near-stock size pill was used at the "compressor" for these tests. The effect of resistance on the pressure vs duty cycle was minimal with the higher resistance causing higher pressure in the mid-duty cycle range due to the higher latency. The results are shown in comparison to the stock solenoid. At 30 Hz, a similar result was obtained except that the Prodrive solenoid was too slow to operate at 30 Hz when a 10 ohm resistor was used. My feeling is that a 5 ohm resistor with the GM and Prodrive solenoids is a better choice than a 10 ohm. 5 ohm still brings the overall circuit resistance to within stock specification range while keeping the latency lower.
Effect of Hardware Setup on WGA Pressure
There's a ton of data to present. These tests were all done with a 1 ohm resistor in series. I tried three different hardware configurations at 20 Hz and at 30 Hz. These configurations with their designations are:
- (N) - No pill at the compressor
- (1) - Near-stock size pill at the compressor
- (3) - Vishnu style bleed to atmosphere between BCS and WGA
The last hardware configuration was something I wanted to try because of the results I saw when I simulated the Vishnu hardware setup which uses a bleeder to atmosphere in combination with the stock BCS. The plot below compares the response of the stock BCS with and without a Vishnu bleeder. The effect of the bleeder is to lower the pressure curve and to reduce the pressure oscillations caused by the on-off nature of the solenoid. I tried the bleeder with the 3-port setups to see what effect it might have on pressure oscillations.
First is the GM solenoid at 20 Hz. As expected, the GM solenoid allows for 0 pressure at the WGA at 100% duty. What turned out to be most important in my opinion were the size of the pressure oscillations. The GM clearly has larger pressure oscillations than either of the setups with the stock solenoid. The use of a bleed between the BCS and the WGA did help reduce the oscillations somewhat in the mid-duty cycle range. By having the Vishnu bleed in the system, it is possible to have finer control over the boost pressure. The reason is that the minimum boost value is increased, so the boost control range is narrower for the full range of duty cycle, so each increment in duty cycle results in a smaller increment in boost pressure change (finer control) compared to no bleeder. The effect of using a near-stock size restrictor at the "compressor" was to slightly lower the pressure at the WGA and slightly reduce the pressure oscillations. What I didn't show here was the smaller bleed also reduced the response time of the system to rapid increases in pressure at the compressor, e.g., during spool up. The slower response could cause the WGA to open a little late and cause some boost overshoot.
Next is the Prodrive at 20 Hz. The basic response is the same as the GM, except that the pressure oscillations are larger. I attribute this to a lower flow resistance of the Prodrive. The oscillations get quite large at around 40%-50% duty cycle.
The AEM MAC 35A at 20 Hz goes one step beyond the Prodrive with the pressure oscillations getting to be 7.5 psi at 40% duty cycle and still very large out at 80% duty where people are likely to be most of the time on boost.
Now onto the GM at 30 Hz. tephra has been experimenting with increasing the operating frequency, so I thought it would be interesting to test the GM and Prodrive at 30 Hz. I didn't bother with the AEM at 30 Hz because it is so far out of the optimum envelope compared to the GM and Prodrive. The main effect of the higher operating frequency is to reduce the pressure oscillations without significantly changing the pressure vs duty cycle curve.
Last is the Prodrive at 30 Hz. As with the GM, the higher operating frequency reduces the pressure oscillations. The oscillations are still larger than the GM. The Prodrive looks ok at 30 Hz except it needs a full 14 V with no resistor in series for it to even operate at <30% duty cycle.
Conclusions
The GM and Prodrive have fairly similar behavior, but the GM appears to have the best mix of characteristics: good response time, the right amount of bleed, and its affordable. Although the AEM is very fast, the pressure oscillations are pretty large, probably large enough to affect the boost behavior of the car.
Last edited by mrfred; Jun 29, 2010 at 10:20 AM.
Nov 2, 2008, 12:45 AM
#3
Account Disabled
iTrader: (7)
Join Date: Dec 2005
Location: Lansdale
Posts: 2,673
Likes: 0
Received 0 Likes
on
0 Posts
Hey mrfred, have you thought about testing the MAC Valve? I just finished reading a thread about dudical26 and fostytou using this for ECU controlled boost. I wonder how it would fair? I think they used a 10ohm 5watt resistor.
https://www.evolutionm.net/forums/sh...d.php?t=365191
https://www.evolutionm.net/forums/sh...d.php?t=365191
Nov 2, 2008, 02:26 AM
#4
NICE WORK!
I love your in-depth evaluation, another worthwhile contribution to our little(or maybe big :P) community!
So whats the verdict on the resistor, can we run without one?
If the GM3port is still the #1 choice then I guess I will get rid of my code that runs 4x25 loops and just use 2x50... less code/faster
I love your in-depth evaluation, another worthwhile contribution to our little(or maybe big :P) community!
So whats the verdict on the resistor, can we run without one?
If the GM3port is still the #1 choice then I guess I will get rid of my code that runs 4x25 loops and just use 2x50... less code/faster
Nov 2, 2008, 06:07 AM
#5
Newbie
Join Date: Sep 2006
Location: Italy
Posts: 41
Likes: 0
Received 0 Likes
on
0 Posts
GM BCS require 10 ohm resistor in series,bat Prodrive BCS require some resistor in series?I don't thinks,it is plug-and-play http://www.tactrix.com/product_info.php?products_id=47
Trending Topics
Nov 2, 2008, 08:40 AM
#8
Hey mrfred, have you thought about testing the MAC Valve? I just finished reading a thread about dudical26 and fostytou using this for ECU controlled boost. I wonder how it would fair? I think they used a 10ohm 5watt resistor.
https://www.evolutionm.net/forums/sh...d.php?t=365191
https://www.evolutionm.net/forums/sh...d.php?t=365191
NICE WORK!
I love your in-depth evaluation, another worthwhile contribution to our little(or maybe big :P) community!
So whats the verdict on the resistor, can we run without one?
If the GM3port is still the #1 choice then I guess I will get rid of my code that runs 4x25 loops and just use 2x50... less code/faster
I love your in-depth evaluation, another worthwhile contribution to our little(or maybe big :P) community!
So whats the verdict on the resistor, can we run without one?
If the GM3port is still the #1 choice then I guess I will get rid of my code that runs 4x25 loops and just use 2x50... less code/faster
I put together a spreadsheet that gives effective duty cycle as a function of operating frequency, applied WGDC, and BCS latency. 30 Hz looks like a good frequency for the GM BCS with a 5 ohm resistor or no resistor. Not too much loss in effective duty, and the BCS can still operate with 20% applied duty.
Hey tephra, when the WGDC is set to 100%, does the BCS still get turned off at the end of each duty cycle?
Bummer.
Last edited by mrfred; Nov 22, 2008 at 05:58 PM.
Nov 2, 2008, 08:46 AM
#9
Buy the GM 3 port from RockAuto.com, you can get it for roughly $40 shipped with the pigtail harness.
http://www.rockauto.com
GM 3 port: $17.83(AC Delco part number 214-474)
GM pigtails for $15.40 (AC Delco part number PT 374)
http://www.rockauto.com
GM 3 port: $17.83(AC Delco part number 214-474)
GM pigtails for $15.40 (AC Delco part number PT 374)
Last edited by Jack_of_Trades; Nov 2, 2008 at 08:53 AM.
Nov 2, 2008, 03:29 PM
#11
Are you sure that it never turns off when set to 100%?
Last edited by mrfred; Nov 2, 2008 at 03:36 PM.
Nov 2, 2008, 04:02 PM
#14
This is the algorithm in the stock system (sorry for the C pseudo code, i think its easier than proper pseudo code :P)
An example, say we have 75% wgdc in the map, this is 150 in the ROM
48*150 /200 = 36
So for the first 36 loops we turn the BCS on, and the last 12 (48-36) loops we turn it off.
Code:
max_loop = (48 * DC from WG table) / 200;
for (n = 0; n < 48; n++) {
if (n < max_loop) {
TURN ON BCS
} else {
TURN OFF BCS
}
}
48*150 /200 = 36
So for the first 36 loops we turn the BCS on, and the last 12 (48-36) loops we turn it off.
