PWM Measurements

[z-ya]

Hear is what the injector current looks like when the PWM setting are correct:

 

 

As you can see, the injector driver allows the injector to draw full current for a period of time (peak), and then limits (hold) the current (rippled looking part of waveform).

 

The PWM time threshold is setting the amount time that the injector is allowed to draw full current to open quickly (peak). This should be set to a number slightly larger than your injector opening time to minimize overshoot, and injector overheating.

 

The PWM Current Limit % is the hold current percentage. So the nomenclature in MT is backwards if you ask me. When you set it to 100%, it is really allowing 100% current, not limiting it by 100%. So this should be set so the hold current meets the injector specification. Most peak and hold injectors have a hold current between 0.75 and 1A. For all the injectors I tested, a number between 30 and 35% will work fine.

 

For most applications, a injector opening time of 1ms should work fine. Keep the battery voltage correction at 0.2ms/V.

 

Hear is the current waveform when the PWM time threshold is too short:

 

 

You can see that the injector never fully opens, and the hold current never reaches 1A.

 

Hear is what it looks like if the PWM time threshold is set too long:

 

 

Notice that the injector opens, but the current spikes much higher than it needs too. This current spike or overshoot will cause the injector to heat up over time, and may cause intermittent heat related problems. Note that the hold current is also higher, but it eventually settles in on the proper hold current.

 

Hear are some injector opening times for a couple injectors:

 

280ZXT 26lb = 1.21ms

 

 

450cc/min DSM = 0.973ms

 

[eec564]

VERY nice, thankyou. You wouldn't have any N/A (183cc) injectors to test, would you?

[z-ya]

Yes I do. I'll try one over the next few days and post the results.

[eec564]

Now I'm wishing I had a better scope (not actually needed, but makes life easier) and a clamp on amp probe.

 

Also, do you think MS1s 0.1ms resolution for injector pulses is enough? I would have gotten one already, but after a lot of research I'm thinking MS2 once the Extra code comes to light.

[z-ya]

I'm an engineer at a scope company, so I have access to some pretty expensive equipment. Inexpensive current probes can be found on Ebay, but it not a measurement you will be making that often. You have to decide if you want to spend the money on a probe. Any scope will work at these frequencies. You just need to be sure that you calibrate the probe to the scope.

 

Yes, I think 0.1ms resolution is fine for most applications. A high boost application with large injectors (bigger than 600cc/min) may need more resolution. You would also need to run injectors with a very fast opening time.

[naviathan]

How big a scope do you need to take measurements like this? I have an old o-scope that a friend gave, but I think it's 100mhz at the most.

[z-ya]

100MHz is fine. You will just need a current probe that can measure in the 0-10A range.

[naviathan]

Hmmm, I'll have to check on that one.

[z-ya]

VERY nice, thankyou. You wouldn't have any N/A (183cc) injectors to test, would you?

 

I measured an OEM NA injector that is fresh from being cleaned and flow tested, and the opening time is around 1.2ms.

 

For OEM NA injectors I would set the PWM as follows:

 

PWM time threshold: 1.5ms

 

PWM Current Limit: 35%

[Zmanco]

I've been running with Ford 5.0L 437 cc injectors (280-150-400 / EOSE-A1A) with RESISTORS and have noticed that especially at idle, the AFR is very inconsistent from time to time. In other words, as MAT and load changes, the AFR moves a lot. No amount of tuning the VE table resolves this: I can get it nearly perfect one day, and then another day it's pretty far off. It dawned on me that if the opening times were not accurate, then VE wouldn't have a linear relationship to how much fuel is injected. So I ran the following test.

 

I drove the car, came back to the garage and let it idle to get the underhood temps stabilized. I allowed O2 correction to operate in the idle region so I could get a numerical value for how far off the mixture is. FYI target AFR at idle is 13:1. I bypassed the resistors and turned on PWM control and found that as I increased PWM Time Threshold the O2 correction would move negative - in other words, increasing time threshold and not changing anything else caused more fuel to be injected. It didn't level off until I got to 1.7 - 1.8 ms. I took this to mean that the injectors are slower than average.

 

With time threshold at 1.9 ms, I then experimented with changing the opening time. My theory was that when opening time is correct, then the change in VE should be linear with O2 correction to maintain a given AFR. I tested this by using the scaling feature in Tunerstudio to change all the cells around the idle point in the VE table by either 1.1 or .9 to cause a 10% change and then watched how much O2 correction actually changed. It turned out that it didn't correlate 1:1 until I shortened opening time to 0.900 - 0.950 ms.

 

So I have two questions:

 

1) Does my methodology for determining PWM Time Threshold and Opening Time without a scope make sense?

 

2) I'm confused by these times. 1.9 ms for time threshold seems long, and indeed the two injector driver transistors were noticeably warm to the touch - more than those around them. Also, it seems strange that the opening time is so much less than the time threshold.

 

 

 

Note: I debated starting another thread, but thought that this was relevant to how one might measure these times without a scope, so posted it here. I'll be glad to start another thread if this is too far off the original topic.

[z-ya]

I've been running with Ford 5.0L 437 cc injectors (280-150-400 / EOSE-A1A) with RESISTORS and have noticed that especially at idle, the AFR is very inconsistent from time to time. In other words, as MAT and load changes, the AFR moves a lot. No amount of tuning the VE table resolves this: I can get it nearly perfect one day, and then another day it's pretty far off. It dawned on me that if the opening times were not accurate, then VE wouldn't have a linear relationship to how much fuel is injected. So I ran the following test.

 

I drove the car, came back to the garage and let it idle to get the underhood temps stabilized. I allowed O2 correction to operate in the idle region so I could get a numerical value for how far off the mixture is. FYI target AFR at idle is 13:1. I bypassed the resistors and turned on PWM control and found that as I increased PWM Time Threshold the O2 correction would move negative - in other words, increasing time threshold and not changing anything else caused more fuel to be injected. It didn't level off until I got to 1.7 - 1.8 ms. I took this to mean that the injectors are slower than average.

 

With time threshold at 1.9 ms, I then experimented with changing the opening time. My theory was that when opening time is correct, then the change in VE should be linear with O2 correction to maintain a given AFR. I tested this by using the scaling feature in Tunerstudio to change all the cells around the idle point in the VE table by either 1.1 or .9 to cause a 10% change and then watched how much O2 correction actually changed. It turned out that it didn't correlate 1:1 until I shortened opening time to 0.900 - 0.950 ms.

 

So I have two questions:

 

1) Does my methodology for determining PWM Time Threshold and Opening Time without a scope make sense?

 

2) I'm confused by these times. 1.9 ms for time threshold seems long, and indeed the two injector driver transistors were noticeably warm to the touch - more than those around them. Also, it seems strange that the opening time is so much less than the time threshold.

 

 

 

Note: I debated starting another thread, but thought that this was relevant to how one might measure these times without a scope, so posted it here. I'll be glad to start another thread if this is too far off the original topic.

 

Since you are using closed loop mode, what is your target AFR table set to? All the PWM does is insure that the injector get the right amount of current required to open it. And then enough current to hold it open for the largest opening time in your map. If you are using resistors, just shut the PWM off and forget about it. Just so you know, injectors do not flow linear with respect to opening time. In fact two injectors with the same 80% duty cycle flow rate (made by two different manufacturers) can have different flow curves. So you really need to tune the base map, set the AFR target map, and then turn O2 correction (closed loop) on.

[Zmanco]

I know I could just fall back to using the resistors, but I still need to determine the opening time. Do you think the approach is valid, at least good enough for those of us without current probes for a scope?

 

Also, if my numbers are accurate, does that make sense? When I look at the values you measured in your other thread with the opening times spreadsheet, most have a pwm time threshold of just a few tenths of a ms more than the opening time. In my case, I'm seeing nearly 1 ms longer.

[z-ya]

I know I could just fall back to using the resistors, but I still need to determine the opening time. Do you think the approach is valid, at least good enough for those of us without current probes for a scope?

 

Also, if my numbers are accurate, does that make sense? When I look at the values you measured in your other thread with the opening times spreadsheet, most have a pwm time threshold of just a few tenths of a ms more than the opening time. In my case, I'm seeing nearly 1 ms longer.

 

Yes, the best way to find the correct PWM measurements is with a scope and current probe. But you can also use a voltage probe and get pretty close. Or use a low resistance power resistor (like .5 Ohms or so) and then float the scope and connect the voltage probe across the power resistor.

 

The problems I see with your measurement are:

 

- Finding the correct PWM settings is an electrical measurement related to the coil in the injector, not the fluid flow rate through it. The purposes is to find the safe electrical settings so that the injectors open completely without excess heat in the injector driver circuit and the injector itself.

- When you use your method, how do you know when you have increased the PWM time far enough beyond where you think the engine "sounds" good. What about the PWM current limit, how can you hear when the injector has enough current?

- The coil in the injector is non-linear, that is why flow rate increases in a non linear rate as opening time or duty cycle increases.

- You don't need to change the injector opening time for most all injectors. 1ms should be fine. Only adjust the PWM time and current limit.

 

The injectors you are using are Ford (Bosch) low impedance? If so, the DC resistance should be around 2.4 Ohms. I would use the Bosch settings from my spreadsheet:

 

http://forums.hybridz.org/index.php?/topic/55910-injector-opening-times-spread-sheet/

 

Turn O2 correction off until you have your VE table really close. You can't use the O2 correction to tune the engine. The VE table has to be pretty close to ideal before you can turn that on. Spend an hour or two on a dyno to get the heavy load tuning done. Then drive it around for a while using the O2 logging to get the drivability good. Then when you think you have it running really good, turn O2 correction on. Let face it, closed loop mode is only good for improving economy at cruising speeds and idle. When you have your foot into it, O2 correction is turned off anyway. Use your WBO2 for tuning, and only after it's running great should you think about turning O2 correction on.

 

Hope this helps.

 

Pete

[Zmanco]

Yes, the best way to find the correct PWM measurements is with a scope and current probe. But you can also use a voltage probe and get pretty close. Or use a low resistance power resistor (like .5 Ohms or so) and then float the scope and connect the voltage probe across the power resistor.

 

The problems I see with your measurement are:

 

- Finding the correct PWM settings is an electrical measurement related to the coil in the injector, not the fluid flow rate through it. The purposes is to find the safe electrical settings so that the injectors open completely without excess heat in the injector driver circuit and the injector itself.

- When you use your method, how do you know when you have increased the PWM time far enough beyond where you think the engine "sounds" good. What about the PWM current limit, how can you hear when the injector has enough current?

- The coil in the injector is non-linear, that is why flow rate increases in a non linear rate as opening time or duty cycle increases.

- You don't need to change the injector opening time for most all injectors. 1ms should be fine. Only adjust the PWM time and current limit.

 

The injectors you are using are Ford (Bosch) low impedance? If so, the DC resistance should be around 2.4 Ohms. I would use the Bosch settings from my spreadsheet:

 

http://forums.hybrid...s-spread-sheet/

 

Pete, thanks for helping me see the flawed assumption I made: I forgot that the coil's current will saturate before the injector opens fully. So my procedure clearly won't be accurate at all, as my 1.9 ms measurement indicates.

 

Turn O2 correction off until you have your VE table really close. You can't use the O2 correction to tune the engine. The VE table has to be pretty close to ideal before you can turn that on. Spend an hour or two on a dyno to get the heavy load tuning done. Then drive it around for a while using the O2 logging to get the drivability good. Then when you think you have it running really good, turn O2 correction on. Let face it, closed loop mode is only good for improving economy at cruising speeds and idle. When you have your foot into it, O2 correction is turned off anyway. Use your WBO2 for tuning, and only after it's running great should you think about turning O2 correction on.

 

Hope this helps.

 

Pete

 

Actually, I've been operating with the resistors and no PWM control for some time now with a 1.500 ms opening time and the tune is very good, with the exception that at idle and other low MAP conditions, the amount of O2 correction required varies a lot from hour to hour, even minute to minute sometimes. FYI, normally I leave O2 off below 1100 rpm but turned it on only for these experiments.

 

At idle, PW is typically in the 2.2 - 2.3 ms range, so even changing the opening time by .1 ms makes a significant difference in how much fuel is "fixed" vs. controlled by MS calculations, hence my interest in getting a more accurate setting for these injectors. I'll try running with the values in your spreadsheet for the other Bosch injectors and see if I can get a more stable low MAP tune. The worst case is that I just go back to what I have now which all in all works pretty well.

[z-ya]

Did you check the DC resistance with a good meter? What did you measure? If they are high Z, it would explain a lot.

[Zmanco]

Did you check the DC resistance with a good meter? What did you measure? If they are high Z, it would explain a lot.

I measured 3 at a time in parallel at the MS relay board (much easier than trying fit my hands under the fuel rail) and saw 1.2 ohms for each bank which would suggest they're somewhere around 3.5 ohms individually.

 

Edit: I tried driving with threshold at 1.4 ms and pwm current limit as high as 40% last night and there were intermittent stutters. I bumped threshold up to the next step 1.6 ms (it increments in .128 ms chunks) and pwm current limit down to 30% and that solved it. But it didn't want to start reliably until I either bumped the threshold up further or increased battery voltage correction to 0.400.

 

I wish I had a spare injector to send to you Pete, I'm really curious to see some objective measurements for these. In the mean time I think I'm just going to run them with resistors as that's stable and reliable.

Edited July 7, 2010 by Zmanco

[z-ya]

I measured 3 at a time in parallel at the MS relay board (much easier than trying fit my hands under the fuel rail) and saw 1.2 ohms for each bank which would suggest they're somewhere around 3.5 ohms individually.

 

Edit: I tried driving with threshold at 1.4 ms and pwm current limit as high as 40% last night and there were intermittent stutters. I bumped threshold up to the next step 1.6 ms (it increments in .128 ms chunks) and pwm current limit down to 30% and that solved it. But it didn't want to start reliably until I either bumped the threshold up further or increased battery voltage correction to 0.400.

 

I wish I had a spare injector to send to you Pete, I'm really curious to see some objective measurements for these. In the mean time I think I'm just going to run them with resistors as that's stable and reliable.

 

I would pull the plug on one of injectors and measure the resistance at the injector without anything else in the circuit. You need a good quality DMM to measure low resistances accurately BTW. So you are better off measuring them individually. Most all Bosch injectors I have measured are around 2.4 Ohms.

 

What resistors are you using?

 

I don't think the problem you are seeing is related to the injector settings. It is probably related to some other setting in Megatune. What WBO2 are you using? Have you calibrated the output voltage from the WB02 with Megatune?

 

Pete