• Protip: Profile posts are public! Use Conversations to message other members privately. Everyone can see the content of a profile post.

Help diagnosing a voltage drop under high loads

Joined
23 May 2020
Messages
714
Location
SF Bay Area
Hey all, I recently discovered a voltage drop at high engine loads. I've replaced a few things but I'm still unsure what it might be or how to fix it. It might be kind of a rabbit hole as the car has so many aftermarket parts on it. Any help would be appreciated, please let me know!

The car

* Old whipple 1.6L CTSC at 9psi
* AEM series 1 ECU (tried two different V1 ECUs)
* RC 550 peak/hold injectors with stock resistor box
* K-series coils with foundry 3 harness
* Water meth injection system (WMI) coming on in boost
* Aftermarket fuel pump, unknown make, unknown wiring setup
* Fuel pressure at 46psi. I don't have the sensor running into the ECU, so not sure how it behaves during these voltage drops.
* Still using the stock fuel pump relay (it clicks at 4500 rpm)
* Prelude form-factor alternator cause CTSC (Tried stock 90A denso and Singer 180A)
* Brand new Antigravity ATX-30HD battery
* No accessories running during these situations (no lights, no stereo, no climate control)
* Been through all the grounds on the car a couple times for other electrical issues

The ECU is seeing a voltage drop at high loads. Basically it reads 13.8V under light loads, then it drops under high loads until it hits about 12.5V.

voltage-drop.jpg

My original thesis was that the 90A prelude alternator couldn't keep up with the addition of the WMI and maybe k-series coils. I bought a Singer high output alternator and installed it. It has both modifications _and_ runs a smaller pulley, but no help, still a volage drop. The data above is with the high output alternator.

I've been digging around in old data logs and did some experimentation last weekend. It seems like this voltage drop has been happening since I bought the car. Old logs show logged injector dead time indicating about 12.5V in boost, just like the logs above.

* Probably not the k-series coils, it happened with the stock coils
* Probably not WMI system, it happened without the system. I also had it turned off last weekend, note the air temps way high in logs above.
* I tried spinning the motor at highish RPM under light load, there was no voltage drop
* It happened with the previous AEM series 1 ECU and also with the current slightly-later-model series 1 ECU.

It seems like the drop is directly correlated with fuel demand. The more fuel necessary, the more the voltage drops. You can see in the logs that the voltage drops along with jumps in injector pulse width.

But even if the fuel pump or injectors have a high current demand, shouldn't the alternator cover it, especially a high output one?

* Belt slip? The belt is very tight already, but who knows. Maybe the current demand is super high and it's slipping? I plan to put the stock prelude pulley on the high output alternator for a little more leverage and maybe less slip.
* Would it be something to do with the factory dual-stage fuel pump wiring? Should I remove the fuel pump relay from the loop? Would my fuel pressure drop up high if I did? Should I get the SoS or some other pump wiring kit?
* Would the injectors cause this?
* Maybe the alt charging cable? Someone on an S2k forum replaced the charging cable and allegedly fixed it.

What am I missing? What else should I check?
 
How old is your main EFI relay? Assuming the wiring harness arrangement for the AEM is the same as the OEM ECU, the ECU and the injectors are supplied off of the same switched supply (yel/blk wire) out of the main EFI relay.

Main relay  and injectors.JPG

As you note, your ECU internal voltage is very closely correlated to fuel demand (injector pulse width). Increasing pulse width increases the overlap between injector pulses and the resulting higher current increases voltage drop in the supply system. I ask about the age of your main relay just because they do have a history of deterioration in the internal electrical connections. It is also possible that you could have deterioration in any of the connections in that same supply, including the connection from the main relay all the way back to the ignition switch. In the OEM ECU, the injector current is switched by power transistors internal to the ECU and the injector current is connected to ground through the ECUs common ground connection. If the AEM does the same, any deterioration in that ground connection results in voltage drop which shows up as reduced voltage inside the ECU.

You mention RC 550 peak/hold injectors. I assume that this means these are low impedance injectors? How, does the resistance of the RC injector coils compare to the OEM coils? If the resistance is lower than the OEM injectors the current flow will be higher and that will inherently increase the voltage drop in the power supply compared to OEM.

The alternator measures and controls the vehicle voltage and your log is the internal ECU voltage which is not the same thing. Your vehicle voltage may be rock steady and you could still have the internal voltage drop in the ECU because of voltage drop in the supply to the ECU. There is a connection between the ECU and the alternator; but, it does other control functions than setting the voltage monitoring point for the alternator so the alternator will not correct for low voltage in the ECU. You could watch the dash voltmeter; but, the little dip when the injector PW increases is so short that you are not going to see it. To confirm that the alternator / vehicle charging system is not the source of the voltage fluctuation problem you would need something like this to log the vehicle voltage.


Connect it up to a switched supply in the fuse box and use it to determine whether you are getting a voltage fluctuation originating in the charging / alternator system As an observation, your 'running' voltage of 13.8 volts is not great. My car is typically 14.3 - 14.6 volts. Check your vehicle voltage when the engine is running by measuring at the jump start terminal in the engine bay. If you have > 14 volts at that point and 13.8 volts in the ECU that is further indication that you have a voltage drop somewhere in the supply to the ECU. If your actual vehicle voltage is 13.8 volts that indicates marginal charging system performance. It will take much longer for a vehicle battery to recover the charge lost in starting if the vehicle voltage is only 13.8 volts. That is a different problem than the ECU voltage drop problem.

If your injectors have a significantly lower resistance / higher current than the OEM injectors you may have to consider separating the injector supply from the ECU supply via a separate relay with a separate fused source. However, this can cause other problems (read the next paragraph)

As an observation, an ECU supply voltage of 12.5 volts in not the end of the world. The actual ECU operates at 5 volts with an internal power supply that will typically tolerate input voltage drops down to about 7- 8 volts. The injectors may not be so happy. If you have accurately characterized and programmed in the injector offset change with supply voltage, the ECU will adjust the injector timing to correct for reasonable variations (13.8 to 12.5 volts would be reasonable) in injector voltage (that is why the ECU and the injectors share the same supply). If you separate the injector and ECU supply with a separate power relay you decouple the two and lose the ability to adjust injector timing based upon injector supply voltage (assuming the AEM allows you to enter injector offset versus voltage and that you are using this feature).
 
Last edited:
Thanks for the reply, I was hoping you would chime in.

I replaced the main relay, fuel pump relay, and fuel pump resistor with brand new parts a couple months ago, so the components themselves should be ok.

The AEM injector drivers work the same way a stock ECU would (ground trigger IIRC?). There is a battery offset / dead time table in the AEM and it is setup to their 'wizard' settings for the RC 550 PH injectors.

The RC injectors are 2.5 ohms. It looks like the stock injectors are 2.3 ohms, so should be all good there. I measured them and the injector resistor box when I had the misfire a while ago and everything was to spec.

I have a volts gauge in the car (that is reasonably fast acting) running off of IGN from the footwell fusebox, and I have a battery monitor (basically bluetooth voltmeter) on the battery terminals. When the car is running / cruising, both of those read 14.2-14.4V while the ECU is seeing 13.8V. I'm not sure how they move under high load, but getting data on that is the next move. You are right, it definitely is suspect that the ECU is showing a drop vs the IGN wire voltage.

I do have a little CAN sensor hub thing I could hook up and log voltage + fuel pressure under load. This weekend I will probably be ghetto, though, and just rig the gopro up so it can see all the bits at once: the navpod, the fuel press gauge, volt gauge, and the phone open to the battery monitor while under high load.

Based on your input I will check all the wiring to the injectors and the power & grounds to the ECU. Are there any specific grounds or connections in this system that are commonly problematic? Or any that I should check first? I suppose I will trace the schematic through the car. It's also possible there is still a bad connection from a T-tap at the ECU...
 
The main EFI relay gets what should be a pretty robust supply. It is fed off of the big ass #9 120 amp fuse which then splits off via the #18 20 amp fuse which connects to the main relay and one or two other things. With the engine running, check the voltage at the #9 fuse and #18 fuse points. For the #18 you will need to fabricate a stiff thin wire probe because you need to measure the voltage with the fuse in place (tricky). After the #18 fuse you want to check the voltage at the yel/blk in and yel /blk out wires on the main EFI relay by back probing the terminal plug on the relay (I don't have the pin numbers handy so you will have to consult the SM for that). I am assuming that the AEM must have a transition harness to interconnect with the NSX harness. You need to trace the yel/blk supply wire from the main relay to see how it gets into the AEM. Back probe plug on the AEM to get the final voltage entering the AEM.

From the #9 fuse to the power in terminal on the AEM there should be a small voltage drop. If you see a significant drop between any two measurement points then you may have a compromised connection somewhere between those two points.

Checking the ground should be easy. With the engine running, back probe the ground wire at the plug terminal on the AEM box and connect the other test lead of the voltmeter to a good clean chassis ground. You should measure a voltage of 0.1 volts or similar. Any significant voltage means you should check the condition of the ground path.

You will need a fairly good quality voltmeter. You are checking for hopefully small differences so 2-3 decimal places is an asset.

If the AEM uses a transition harness, that would be a good place to start looking. If the interconnection of the AEM relies on T taps, particularly these things or similar
self-strip-connect-automotive-22-18-f5460101-6962-4115-824f-590a56889d94.png


those would definitely be a candidate for a problem area.
 
Great, thanks for the detail, I will check all these things! My volt meter is pretty ok, almost certainly good enough to check this stuff. Worst case if I need the detail, I have a scope with a fancy voltmeter function on it.

FWIW, the old AEM unit doesn't have a transition harness, the stock plugs go right into the ECU. Eliminates one thing. And I have managed to remove all the dumb t-taps that were on there. The downside is that there looks to have been a lot of tapped wires, so quite a few of the wires at the ECU were in pretty rough shape, and a couple hanging on by only a few strands. I did fix what seemed like all the bad connections, but maybe there are more!
 
Out of curiosity, does your voltage dip when you are running lights, rear defogger, and the climate control unit?

My first thought is high RPM belt slip. I'm not familiar with the whipple, but IIRC, the popular belt for maybe the whipple or Basch was the Gatorback. Belt tension is always tricky but if it isn't a fresh belt I'd go ahead and replace it and temporarily put it near the tight range of tension and re-test.

Also, as Old Guy mentioned, I assume you've adequately compensated for voltage drops in your AEM Batt Offset Primary? You should have those values from RC to compare with whats in your cal file.

After stripping my car, my largest electrical loads are the power windows and the fuel pump. My power-hungry aftermarket Denso fuel pump is fed directly off of the alternator and goes through a dedicated 40A German relay and I've bypassed the OEM 4k RPM high speed switchover. I couldn't see from your plot above, but if your FP is still using some of the OEM stuff then that could also be part of the problem (especially if you're unsure what FP you're using), but I would start with belt slippage first.
 
Hey thanks for the response!

Out of curiosity, does your voltage dip when you are running lights, rear defogger, and the climate control unit?
I'm not sure, good call, I will check this weekend

My first thought is high RPM belt slip. I'm not familiar with the whipple, but IIRC, the popular belt for maybe the whipple or Basch was the Gatorback. Belt tension is always tricky but if it isn't a fresh belt I'd go ahead and replace it and temporarily put it near the tight range of tension and re-test.
Belt slip is definitely a possibility, I will tighten it a good bit to see if it fixes. I have had gatorbacks (now continental elite) on the car since I've owned it. With the new alternator's smaller pulley, I couldn't find a conti elite in the size I needed (68.5", 1740mm), so it now has a brand new conti OE technology on there now, which is a smooth belt.

I'm going to try to swap out the smaller pulley on the HO alternator with the stock pulley, then I can move back to a conti elite belt. With the small pulley, belt slip seems more likely. And also with the smaller pulley the alt will be spinning at 20k RPM at 8k engine RPM, which is too high from what I've read. Sounds like 16k or so is kind of the upper bound

Also, as Old Guy mentioned, I assume you've adequately compensated for voltage drops in your AEM Batt Offset Primary? You should have those values from RC to compare with whats in your cal file.
Yeah, the table is setup according to AEM's docs / wizard for the RC 550s

After stripping my car, my largest electrical loads are the power windows and the fuel pump. My power-hungry aftermarket Denso fuel pump is fed directly off of the alternator and goes through a dedicated 40A German relay and I've bypassed the OEM 4k RPM high speed switchover. I couldn't see from your plot above, but if your FP is still using some of the OEM stuff then that could also be part of the problem (especially if you're unsure what FP you're using), but I would start with belt slippage first.
Noted on the fuel pump being the major draw. I am running what I think is the stock fuel pump wiring, I'll dig into it to make sure. The FP relay is definitely connected and clicking. I'll look into bypassing it after checking to see what the pressure is at during these situations
 
The graphs in the first post are not scaled with units. Presumably this data is form a data log so I assume that you should be able to retrieve actual values. If you can retrieve the actual fuel PW values when the voltage is 13.8 volts and 12.5 volts you could calculate the change in injector overlap (when one or more injectors are firing at the same time. With the resistance value of the injector resistor and the injector this would allow you to calculate the change in total current for the injectors as the PW ramps up. The current change could indicate whether the voltage change is just due to voltage drop on the ECU / injector supply or whether it is being caused by a drop in the source voltage.

If your 550 injectors are so big that there is no injector overlap even at peak load then the voltage drop is not caused by the injectors and is a problem somewhere else on the car.
 
Yeah the chart is showing data from the ECU, the charts only show one Y scale, the bottom chart is showing the scale for ignition timing. The pulse width values at high boost are around 9.5 - 9.8ms, it looks like the voltage starts dropping over ~3ms pulsewidth. The AEM unit has a saturated driver which is ~1.5A per injector max, if it just held them all open, that'd be a 9A draw max. This doesn't seem like all that much, but it could be this plus extra load from the fuel pump? How much would the fuel pump draw change from low to high fuel load? I guess depends on the pump?

I put the stock pulley on the high output alternator and did some testing yesterday. All this was while it was warming up ~2 minutes after cold start, around 1500 rpm, 65 degF ambient

* Voltage from alt post to ground: 14.4v
* Voltage at gauge in the car (IGN): 14.2v - 14.3v
* Voltage from fuse #9 to firewall ground: 14.4v
* Voltage from fuse #9 to G101 ground: 14.2v - 14.3v
* Voltage from fuse #18 to firewall ground: 14.4v
* Voltage from fuse #18 to G101 ground: 14.2v - 14.3v
* Voltage at battery: 14.2v - 14.3v
* Voltage at fuse #33 underhood to frame rail ground: 14.2v - 14.3v
* Voltage backprobed A25 & C1 (ECU switched power from main relay) to ECU mounting point: 14.2v - 14.3v
* Voltage backprobed D1 (ECU constant power from fuse #33) to ECU mounting point: 14.2v - 14.3v
* Voltage backprobed A23 & A24 (ECU ground to G101) to ECU mounting point (also ground): 30mV - 32mV (is this too much?)
* Ohms from unplugged A ECU plug at A23 & A24 to ECU mounting point: 0.4 - 0.5 ohms

During all the measurements above, the ECU param I was logging earlier was showing 13.8 - 13.9v. Either something is weird internal to the ECU, OR there is a bug with the series 1 ECU in the conversion from raw 16bit to voltage. This old AEM has bugs, and I've chased my tail on stuff like this before only to find out it's just an issue inside the ECU. For example, this time I also logged the raw 12v ADC channel and it showed .1v higher than the processed one, i.e. 13.9 - 14.0v. On top of that, I sent the raw ADC value into my navpod via the CAN bus. Using the AEM-documented conversion factor in the navpod, it showed 14.6 - 14.7v during all this. Turns out the AEM pro logger uses a smaller conversion factor than is documented.........

All this tells me that voltage supply to the ECU is probably fine. Maybe the G101 ground could be improved, but it is clean and tight, and the voltage is above 14 at the ECU.

I also turned on the lights, climate control, and the radiator fan was running. At 975 idle, there is a big voltage drop, down to about 12.8 - 13.0v all over. Revving to 1500+ rpm brought it back up to 14-14.1v. This is probably why there was a smaller pulley on there. The singer alt guy did tell me that it wouldnt work as well at idle with the larger pulley. The larger pulley will spin it 700rpm slower at idle

I spent all my allotted time yesterday testing this stuff, but didnt have a chance to actually drive it. Today I drive with the bigger alt pulley and a super tight belt.
 
Ok, I took some video today and learned a couple things:

* It looks like the voltage drop is just in or at the ECU. The volts gauge never drops below 14.0v. While cruising it fluctuates between 14.2 and 14.4v, then under high load, WMI on and all that, the gauge shows 14.0 - 14.1v while the ECU reports 12.5v
* Fuel pressure at idle / low load is 45 - 46psi, then at high load it seems to peak between 61 - 63psi
* With the really tight belt, the SC doesnt bleed any boost at all (9.25psi minimum!) and it feels super healthy. Gets loose rolling on in 2nd gear which it usually doesn't do. I guess the belt was still too loose previously

So now, it feels like there are one of two issues

1. The circuit feeding the ECU is seeing a voltage drop (is the fuel pump on that circuit?)
2. The AEM ECU is dumb and has some crappy circuits internally. If this is the case, it wouldn't be the first time. These early series1 ECUs apparently have a bunch of novice EE circuitry. For example many inputs are noisy because they pick up injector driver electrical noise.

Thinking I need to get power from A25 (+) A24 (-) onto the CAN bus or run the volts gauge from there or something. Happy to check other things if you guys have ideas.

Here's a representative screeshot of all the things. Small note that the boost formula in the navpod shows a little high, it's actually probably 9.5psi here

voltage-drop.jpg

Cue Honcho asking why I don't upgrade to the haltech already :p
 
If you use 5000 RPM as an example, the engine will go through a complete cycle (2 revolutions) every 24 ms. All 6 injectors will fire in a cycle. I don't know whether the C30 32 split crankpin is enough to completely even out the 90 V6 uneven firing pulses. Assuming it is, at 5000 RPM (and all lower RPMs) a 3 ms PW would have no overlap between injectors. However, at 5000 RPM a 9.5 ms PW will have significant overlap (6 x 9.5= 48 ms) which means 2 injectors will always be on during a complete injector cycle. Crank the RPM up to 8000 RPM and an engine cycle drops to 15 ms and with a 9.5 ms PW you have a minimum of 3 injectors firing at once (and possibility for a 4th)

The AEM saturated injector driver doesn't limit current unless that is an added feature. Saturated just means that it is not a peak and hold driver which has two current levels (and is real pain in the ass idea). The peak injector current will be limited by the combination of the NSX injector resistor and the resistance of the RC injector. Not having those values, lets assume that the peak RC injector currents are operating at the limit of the AEM drivers = 1.5 amps.

I will assume that your vehicle supply voltage is operating at 14.3 volts. With an injector PW of 3 ms at 5000 RPM the time averaged total injector current is a nudge less than 1.5 amp. With an internal ECU voltage of 13.8 volts, the inferred resistance of the ECU and injector supply is (14.3V - 13.8V)/1.5A = 0.33 ohms. If I assume that the source voltage stays at 14.3 volts; but, you increase the engine speed to 8000 RPM and the PW to 9.5 ms the time average injector current will be about 4.5 amps and the voltage in the ECU will be about 14.3V - 4.5Ax.33 ohms = 12.8 volts. That is not exactly what you are measuring; but, it is in the same ballpark and I made a bunch of assumptions, the key ones being I don't know what the actual injector current is and I don't know what the ECU load current is.

The take away is that the voltage drop that the ECU is recording might be perfectly normal because it is recording the ECU supply voltage which is not the same as the vehicle voltage. The best way to confirm whether you have a problem is to go back to previous data logs and try and find the same engine RPM and injector PW and look at the recorded ECU voltage. If under the same change in injector PW the ECU did not record a drop in internal supply voltage, then perhaps something has changed indicating that you need to investigate further. If the same operating conditions show the same voltage drop, then this might be perfectly normal and you need to stop looking.

If your AEM injector 'wizard' has accurately set both the injector offset and the change in injector offset versus operating voltage then the ECU will accurately adjust for the change in voltage to the injectors and everything should be just fine.

If you have a discrepancy between the voltages that the ECU is recording and what you measure on the input pin to the ECU, that could just be an ECU calibration error. The analog converter inputs on the AEM chip will be restricted to 5 volts. The 12v supply voltage to the ECU will typically be scaled using a simple resistive voltage divider. If the AEM uses a trim pot to calibrate the resistance divider the trim pot may have drifted in its calibration with age. You can confirm the calibration error by measuring the direct input voltage to the ECU and comparing it to what the log is showing - assuming you have a correctly calibrated multimeter.

Just like a man with two watches never knows what time it is, a man with two voltmeters is always unsure of the true voltage.
 
Interesting, ok an internal ECU resistance could explain the drop. Along those lines, maybe the 0.4 - 0.5ohms from chassis ground to A23 is an issue (maybe G101 could be better)? Could that also cause the same thing? The injectors are on this same circuit, + is the same, - comes out of the ECU, so maybe that little bit of resistance could be causing a drop on supply voltage to the ECU?

FWIW, when I had the misfire, I measured the injectors and the resistor box. Injectors were all 2.5ohms, and resistor box was 6ohms across the board. So, if there is no other resistor in the pathway there, it'd be 8.5 ohms total per injector (seems real low for a high impedance driver, right?), 14.3v / 8.5ohms = 1.68A. If I switched to EV14 bosch injectors, they'd be 12ohms, so 14.3 / 12 = 1.19A. I wonder if there'd be less voltage drop with them, in your formula, it'd be 14.3v - 3.57A * .33ohms = 13.12v

The best way to confirm whether you have a problem is to go back to previous data logs and try and find the same engine RPM and injector PW and look at the recorded ECU voltage.
The old logs also show the same voltage drop. I only have logged injector batt offset in the past, not directly battery V, but backing out from the offset based on the ECU's table gives me 12.5v at full load. So same drop.

If the same operating conditions show the same voltage drop, then this might be perfectly normal and you need to stop looking.
This is definitely very possible, it just seems so wrong! :)

The analog converter inputs on the AEM chip will be restricted to 5 volts. The 12v supply voltage to the ECU will typically be scaled using a simple resistive voltage divider. If the AEM uses a trim pot to calibrate the resistance divider the trim pot may have drifted in its calibration with age.
Interesting, that makes sense. I would have assumed they had some AD converter right off input voltage, but I'm certainly no EE. It does show a lower voltage all the time, even at idle when the injector current draw should be low

So my plan now is 2 things

1. Get a reading of voltage at the supply to the ECU under load. Does it drop? I will try this CAN expansion module thing that will emit its supply voltage to the CAN bus. I'll hook it directly to A25 and A23
2. See if I can improve G101 or figure out where that 0.5ohms is coming from

Bonus is to try the bosch injectors. I've been meaning to anyway. Newer tech, better atoms, you know? :p
 
Interesting, ok an internal ECU resistance could explain the drop. Along those lines, maybe the 0.4 - 0.5ohms from chassis ground to A23 is an issue (maybe G101 could be better)? Could that also cause the same thing? The injectors are on this same circuit, + is the same, - comes out of the ECU, so maybe that little bit of resistance could be causing a drop on supply voltage to the ECU?

It is very difficult / impossible to get an accurate measurement of a small resistance with a multimeter. That is why I suggested the voltage drop measurement. The value of 30 mV that you measured is reasonable. The easiest thing you can do to help is apply a little Ox-guard grease to the bolted connection at the chassis mounting point. I try to use Ox-guard grease on most of my bolted electrical connections on all of my vehicles.

FWIW, when I had the misfire, I measured the injectors and the resistor box. Injectors were all 2.5ohms, and resistor box was 6ohms across the board. So, if there is no other resistor in the pathway there, it'd be 8.5 ohms total per injector (seems real low for a high impedance driver, right?), 14.3v / 8.5ohms = 1.68A. If I switched to EV14 bosch injectors, they'd be 12ohms, so 14.3 / 12 = 1.19A. I wonder if there'd be less voltage drop with them, in your formula, it'd be 14.3v - 3.57A * .33ohms = 13.12v

The injector resistor value of 6 ohms and the injector resistance of 2.5 ohms are very typical for the low impedance injectors arrangements that I have seen on other cars. If you switched to the EV14 high impedance injectors that would reduce the injector current and the voltage drop in the circuit supplying the injectors and the ECU. However, there will still be voltage drop (just less) when you get 3 injectors firing together. Since you don't know that the voltage drop is actually creating a problem, I would not replace the injectors just to get rid of the injector problem. The only reason I see to change the injectors is the 1.5 amp nominal current rating of the injector switching transistors in the AEM. At 1.68 amps you are running 12% above the switching transistor rating. However, 1.5 amps may be a continuous transistor rating and the transistors to not operate with a 100% duty cycle so that could also be a non issue.

Interesting, that makes sense. I would have assumed they had some AD converter right off input voltage, but I'm certainly no EE. It does show a lower voltage all the time, even at idle when the injector current draw should be low

There are no families of integrated circuit devices that operate at 12 volts. Given the vintage of the AEM ECU, it will likely be 5 volt CMOS logic. More modern devices may be 3.3 volts - modern devices operate at lower voltages so they can pack stuff closer together without exceeding the internal voltage breakdown limits. There may be purpose built A-Ds with a 12 volt tolerance on the input; but, that is likely a custom chip or a 5 volt or 3.3 volt chip with a separate front end voltage reduction circuit.

Bonus is to try the bosch injectors. I've been meaning to anyway. Newer tech, better atoms, you know?

If you want to try the Bosch injectors because 'you want to try the Bosch injectors' that is fine. Just make sure that you have all the data required to accurately characterize the opening time of the injectors
 
That was a good read and I'm totally not qualified to provide any potential solutions. I do wonder if you're still running the OEM fuel pump resistor?

And yes, by all means get rid of those peak/hold injectors and but my 550cc EV14s 😂😂😂 I can provide all of the dead time offsets at various voltages and RPMs
 
I do have the OEM fuel pump resistor, but pin A8 on the ecu is always high, so full speed, resistor bypass all the time. I did some experimentation this weekend with having it run low speed in vacuum, but there was no noticeable affect in fuel pressure, cruising afrs, or the ability to handle more electric load (e.g. lights on, same voltage drop) at idle or while cruising. I thought with the previous ecu it would run in low speed then switch to high under load, but after digging in, it’s always run full speed since I’ve owned the car.

I def want those EV14s!

I have some more data too I’ll share in a bit
 
I did some final tinkering with this thing over the last couple days. tl;dr: I'm back on the OG denso alternator and I think this voltage drop is probably fine. It feels dirty, but there doesn't seem to be any adverse affects, and I don't have a great idea of what else I can improve. Nothing is obviously broken, and maybe this also happens in stock form with the stock ECU? Not sure

I had 2 ideas

1. See if I can improve G101
2. Get a reading of voltage at the supply to the ECU under load. Does it drop?

The ECU grounds both go to G101 on the front of the engine near the front fuel rail. I figured this might have a bad connection. I was measuring 0.4 - 0.5 ohms from the A23 ECU pin to an ECU mounting point on the firewall. Could I get it down to 0? Unhooking G101 and measuring from the eyelet to the ECU pin I could get a 0.2ohm reading. So that seemed like the best case. I wire brushed the eyelet and the mounting point and was able to get the original reading points to show 0.2ohms. Cool:

vdrop-ground.jpg

Next step was getting voltage at the ECU on the CAN bus. I used some TE/AMP 070 connectors to patch into the voltage at the A23 / A25 pins. Those black/pink wires go into the CAN module

voltdrop-plug.jpg

Here's the AEM CAN bus expansion module (PN 30-2226). What a weird form factor:

vdrop-can-volts.jpg

Then the voltage into the navpod. The bottom right one is the voltage from the module. The ECU is reporting a little low for the same input voltage. The module voltage shows basically the same as the gauge. (I also got fuel pressure into the ECU + navpod for easy logging)

vdrop-navpod.jpg

So the question was: Where is the voltage drop? In the ECU? Or is it in the circuit feeding the ECU?

It is indeed the latter: the circuit feeding the ECU. Note that the CAN expansion module also shows a voltage drop here. It shows a higher voltage, but the drop still there, outside the ECU

voltage-drop-2a.jpg

So there you go. Probably the draw from the injectors causing the drop?

In the end I didnt solve it and I'm kind of back to where I started. The upside is that I learned a bunch about the car

* I'm back on the old Denso 80-90A alternator. It has way less voltage drop at idle with the lights on, there are no voltage spikes with RPM changes, it isn't noisy at all (HO alt has bearing howl), and it runs a whole lot cooler that the high output alternator.
* The belt was too tight. Yay for full boost all the way up, but it makes the supercharger run hot after ~15 minutes of the engine running. Check out the navpod in the pic above, it shows 183 deg F air temp. That is unusually HOT. And it will creep to 200 deg F just while cruising. With a looser belt AITs drop to 125F or so at 70F ambient while cruising. It started showing this hot behavior on the first drive after the tight belt. After taking it out yesterday, the gear case was super hot. My temp gun showed 180F on the gearcase and only ~150F over the rotors. I need to find the right tension balance between minimal boost fade but also putting minimal heat into the SC.
 
I struggled with CTSC belt tensioning. It was too tight I noticed the same issues you're seeing. At the end, I tighted the belt based on my boost curve. If it faded away it was too loose. I tightened it slowly until the boost fade went away. You have to have to have to use the gatorback V belt. It makes a big difference in traction of the belt on the pulley and doesn't require as much belt tension in my experience. For a time they were discontinued so hopefully you can still find them or they've been remade.
 
I struggled with CTSC belt tensioning. It was too tight I noticed the same issues you're seeing. At the end, I tighted the belt based on my boost curve. If it faded away it was too loose. I tightened it slowly until the boost fade went away. You have to have to have to use the gatorback V belt. It makes a big difference in traction of the belt on the pulley and doesn't require as much belt tension in my experience. For a time they were discontinued so hopefully you can still find them or they've been remade.
Good to know I'm not crazy. I loosened it way up but havent driven it since. I'm going to do exactly this process of creeping up on it until there is little to no boost fade. I got a tension gauge so when I get it right I can check tension and start there on belt / alt changes. I am using a gatorback belt, now they are called "continental elite". I think the conti elites are also discontinued, though, as I can only find them on ebay
 
Good to know I'm not crazy. I loosened it way up but havent driven it since. I'm going to do exactly this process of creeping up on it until there is little to no boost fade. I got a tension gauge so when I get it right I can check tension and start there on belt / alt changes. I am using a gatorback belt, now they are called "continental elite". I think the conti elites are also discontinued, though, as I can only find them on ebay
Just something I remembered when imagining what my CTSC boost curve looked like so long ago. Boost tended to peak around 6500-7000rpm and slowly dropped toward redline. It dropped about 0.5-0.75 psi IIRC towards redline. I think that's just natural efficiency losses of my 1.7L Autorotor. Your 1.6L may hit peak boost sooner and drop off sooner (i'm unsure).

My $0.02 - I'd loosen the belt to a point where you notice it starting to drop and tighten the belt slightly to it's final position.

Having an overtightened CTSC belt makes me sad for the upper crank bearing on #1 & #4. and the lower main bearings on #3 & #6.
 
Just something I remembered when imagining what my CTSC boost curve looked like so long ago. Boost tended to peak around 6500-7000rpm and slowly dropped toward redline. It dropped about 0.5-0.75 psi IIRC towards redline. I think that's just natural efficiency losses of my 1.7L Autorotor. Your 1.6L may hit peak boost sooner and drop off sooner (i'm unsure).

My $0.02 - I'd loosen the belt to a point where you notice it starting to drop and tighten the belt slightly to it's final position.

Having an overtightened CTSC belt makes me sad for the upper crank bearing on #1 & #4. and the lower main bearings on #3 & #6.
Yeah I have good data on how it behaves now. With it real tight it would only bleed .25psi near the top. Prior to that it was dropping about 1psi toward redline

Yeah will do on the belt tightening procedure. It really is hard to tell if it’s too tight. The difference between totally loose and super tight is like 3/16” difference at the tensioner

I know, same on the bearings, also the blower itself 😬 fortunately I didn’t drive it very much between tinkering sessions.
 
All of your voltage measurements appear to be engine running tests. Try a key on / engine off tests. Your voltages will be lower, probably around 12.1 - 12.2 volts if the battery is good; but, with the engine off there should be no current flowing through the injectors. The actual ECU current should be very low so the input to the ECU should read very close to the battery voltage. If that is correct, that would be consistent with the voltage drop to the ECU being caused by the injector currents.

The discrepancy between the voltage that the ECU reports versus what the CAN bus device reports could be caused by two things
- a calibration difference between the CAN device and the ECU measurement
- contact resistance on the supply side and ground side to the ECU plugs (looks like the CAN device is measuring the body harness side of the ECU supply.

Power up the ECU with the engine not running. If the ECU and CAN device report different voltages you likely have a calibration difference. When you start the engine up, watch for the difference between those two voltages. If the difference increases as you increase engine load (injector PW) then you likely have voltage drop across the contact resistance on the pins for the ground terminal. If the AEM ECU has a common ground connection for the ECU internals and sinking the injector current from the injector drivers, you could isolate the injector drivers on to a separate ground connection. Personally, unless you know that this is creating a problem I would be inclined to leave it alone.
 
Noted, I will do some more tests this weekend and report back. At this point I am inclined to leave it alone. It seems fine and the rest of the system is not seeing a voltage drop. I do wonder if there is similar behavior with a stock car.


Thanks for all your help on this!
 
Back
Top