Stock Honduh garbage, totally insufficient: http://www.sanken-ele.co.jp/en...e.pdf

Proposed replacement: http://www.onsemi.com/pub/Collateral/CS452-D.PDF

Not direct drop in, would require a very simple daughterboard. Honduh ECU's are already set up for ~1.2-1.4 amps worth of power through the existing saturated STA406A driver, which is approx what the beefier CS453 lays down during "hold" portion of operation.

Not entirely certain of stock ECU's ability to withstand 2.4-4.4 amp "peak" current load; this is easily fixed by running a jumper wire, more substantial than a trace, to ground. Doing so would likely **** USDM injector test circuit and click CEL 16. Would require an EMS that you can toggle injector test on/off with, such as Hyundaidata/Ubercrap/Clone. AFC hackfags and other useless wastrels need not apply.

Sad thing is, this is exactly what the AEM OMG WTF EMS needs. Too damn bad you lose your warranty if you open the case.

I expect a working prototype in 1-9 months, will share results. At that time I may offer to sell boards in kit form, or offer an installation service, but probably won't. I suck like that. Eat my shorts.

edit: Would be cakelike to design said daughterboard with provision for a second set of drivers... and possibly a hobbs switch... for staged injector setups or for water injection.


Modified by J. Davis at 2:31 AM 1/6/2005

 

Can't wait to see where this ends up. Great idea.

 

Go redneck!

Sonny

 

i just woke up and havent really thought about the caveats of doing this, but i am concerned that the CS452 appears to be discontinued?

 

yea, it says something about no reccommended for new design. But that doesn't mean you can't get ahold of them or something like it. Hope that he can find some atleast..

 

Only found a few thousand NOS. A little pricey at ~$4 per, might get better in large quantities.

National's LM1949 is a LOT cheaper, but requires you to assemble them into an array and insert a zener so the circuit doesn't reverse feed itself. CS452/453 are plug and play.

 

Wow, so much tech on honda-tech. I thought you only posted crap on this site?

Now I'm just a poor CompSci grad with a software focus, but I'm guessing the Ic-Vce of the stock NPN is what makes it crap, right? Compared to the CS452 output which has a 'jump' to kickstart larger injectors open very quickly and then levels off. It doesn't say how the CS452 responds to subzero temps, however, where the stock response levels off at above zero.

If you don't sell the boards, at least post the schematic... some of us aren't smart enough to make that ourselves

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by J. Davis &raquo;</TD></TR><TR><TD CLASS="quote">
I expect a working prototype in 1-9 months, will share results.
</TD></TR></TABLE>

Lol, giving yourself a little space to work, eh?!

I think this sounds like a great idea.

Would implementing the second set of drivers or a Hobbs switch into the daughterboard allow your ecu reprogramming software of choice to have control over this circuit. In other words, could tuning software have control over the secondary injectors, or water injection?

I hope that makes sense.

 

Very cool

Now we can control 1600cc's much better huh?

 

already have prototype boards using the National semiconductor IC and TIP122 darlingtons en route back to me...

 

J Davis,

Good idea. I have actually built this for years. We used to do this for the factory 911 Porsches when we modified the ECU's and added turbos much like the Honda crowd is doing these days with Hondata.

I've got the schematic already done and PCB's being made.

I used the CS453's with a 35V 5W Zener to clamp the kickback when you close the injector. You could have a smaller voltage Zener, but then it will take a lot longer for the coil to discharge when you close the injector which results in poor control of the injector closing, especially for low speed.

The catch is , what you need to do is run a pull-up resistor on the signal input
{pin 5} to keep it high all the time. This serves two purposes: 1 keeps the driver on all the time{I'll explain in a second}, two provides enough load for the injector drivers in the stock ECU to think the original injectors are still there so you don't get a CEL.

The deal is, the factory ECU is an open collector output {switch to ground}. The CS453 requires a "rising" signal going from low to high abouve about 2.5 volts or so.

Rather than use more circuitry to invert the factory Honda ECU signal, you can simply connect it to pin 2 {Control}

This pin was designed as a safety margin by Cherry and was normally connected to a voltage divider circuit between source and GND so that if the source voltage was to swing wildly high the driver turns off to prevent over current situations if the voltage on "Control" gets above about 1.5 volts or so.

What you do is connect a pull-up resistor to this pin so that it is always high. This leaves the driver normally disabled. When the Honda ECU signal pulls low on this pin the driver turns on and Viola...we have liftoff!

I can send you a board to play with if you like. Just PM me and I'd be happy to drop one in the mail.

-Ben

 

That's a really good idea Ben. Do you know how the price on those Cherry drivers compares to a TIP122/LM1949?

 

this thread has already exceeded all the tech i saw on honda-tech in all of 2004.

awesome stuff guys.

 

It also appears that the CS452 is an obsolete part... Couldn't get any availability data from their website. Digikey doesn't appear to even list them. Any ideas?

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by blundar &raquo;</TD></TR><TR><TD CLASS="quote">It also appears that the CS452 is an obsolete part... Couldn't get any availability data from their website. Digikey doesn't appear to even list them. Any ideas?</TD></TR></TABLE>

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by that big ****** J.Davis &raquo;</TD></TR><TR><TD CLASS="quote">Only found a few thousand NOS. A little pricey at ~$4 per, might get better in large quantities.</TD></TR></TABLE>

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Ibiza &raquo;</TD></TR><TR><TD CLASS="quote">Lol, giving yourself a little space to work, eh?!</TD></TR></TABLE>

Lazy/spacey/broke white trash. Being realistic.

But, yeah, hobbs switch/opamp or other control (VTEC, IAB) can be used to trigger staged injectors.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by EFIGUY &raquo;</TD></TR><TR><TD CLASS="quote">I used the CS453's with a 35V 5W Zener to clamp the kickback when you close the injector. </TD></TR></TABLE>

I appreciate the experienced tips - don't think for a second I'm not saving yours to the .txt file I'm keeping all my notes on - but my ego forces me to remind you on days my blood sugar/insomnia/sequencing dysfunction don't trouble me too much I'm quite sharp.

Datasheet's proposed schematic was a 40V 10W Zener to clamp kickback.

Also:

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by CS452/453 datasheet &raquo;</TD></TR><TR><TD CLASS="quote">Automotive injectors at present time come in two types. The large throttle body injectors have an inductance of about 2.0 mH and an impedance of 1.2ohm and require the CS-453 driver. The smaller type, popular world-wide, has an inductance of 4.0 mH and an impedance of 2.4ohm and needs about a 2.0A pulse for good results, which can be met with the CS-452. Some designs are planned which employ two of the smaller types in parallel. The inductance of the injectors are much larger at low current, decreasing due to armature movement and core saturation to the
values above at rated current. </TD></TR></TABLE>

I don't have a Z-meter, but most P&H injectors I've seen are the 2.4 ohm units. Better suited to the CS452. Food for thought. Aside from old Ford 50# CFI injectors, I don't know offhand what injector you'd use that requires the CS453 that are topfeed aka applicable.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by blundar &raquo;</TD></TR><TR><TD CLASS="quote">It also appears that the CS452 is an obsolete part... Couldn't get any availability data from their website. Digikey doesn't appear to even list them. Any ideas?</TD></TR></TABLE>

Aeri, or 4 Star Electronics have listings for quite a number. I have inquiries in about pricing right now. A couple of the SE Asia based brokers list even more. GM bought these things in bulk, you can bet there is enough surplus for hobbyists scattered about the planet. Pricier than the National equivalent, but more convenient as previously stated. $8/per is a very steep price, have heard of $4/per in small quantities if you talk to the right person.

Cherry was purchased by ON Semi back in '00 (?) and made a run under that name as well. Cherry itself makes nothing more than switches these days. Chip comes in three packages, SIP5 (-GT5) or horizontal/vertical T0-220 (-GTAH5 and -GTAV5 respectively) with staggered pin layout like 5151's.

 

Son, that is good info.

J...still ****** with that stock **** to make it better aye...

 

Sure am. Since most OEM stuff is superior to 95% of the aftermarket stuff, it's a better platform to build from.

If you disagree, we can always line up some cars If you ignore the clutch, flywheel, and the stuff I made myself, my CRX is entirely "OEM"

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by J. Davis &raquo;</TD></TR><TR><TD CLASS="quote">Sure am. Since most OEM stuff is superior to 95% of the aftermarket stuff, it's a better platform to build from.

If you disagree, we can always line up some cars If you ignore the clutch, flywheel, and the stuff I made myself, my CRX is entirely "OEM"</TD></TR></TABLE>

Is this a challenge? If so, i hate to tell you, but i dont have anything to compare to your OEM kaa!

Ill hold that offer though, maybe in june or july ill come see you, since you are in walking distance

 

Sweet, we need to do the face to face + drink some beer thing. Ideally before the next millenium.

 

I dont drink beer, but as for the drink thing, ill break open a Spark with ya!

BTW, do you have any other "sources" for these Drivers?

 

Bump for the tech rich post of 2005!!

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by blundar &raquo;</TD></TR><TR><TD CLASS="quote">That's a really good idea Ben. Do you know how the price on those Cherry drivers compares to a TIP122/LM1949?</TD></TR></TABLE>

I haven't looked to closely at the pricing of the LM1949, but you can still source the CS453/452 family of stuff for around $4 which is kinda high but given the relatively low quantities required for this project I feel the simplicity of the CS devices outweigh the cost savings of the considerably more complex LM1949.

I haven't thoroughly investigated the LM stuff, but I know that for an all out stand alone application its very handy given that essentially it is "programmable" for current limiting various injector types and brands which helps in the case of trying to map out the control strategies for lots of different ones.

What I don't care for {in this particular project} though is that it takes a fair bit more planning to do it right and PWM the input properly to initiate the "hold" sequence where that logic is built into the CS453. Simplicity pays, especially when you are a subscriber to "Electronics for Dummies Monthly" like me!

According to the data sheet for the LM1949, to PWM control the driver cause a very high amount to RFI from the inductance created in the coil of the injector. This could create some problems if the RFI isn't shielded properly and that could really put a damper on someone's day at the track.

 

I'm inclined to agree. I'll probably end up testing out the LM1949 based driver board I have on my hatch and see how it goes... If you're interested Ben, take a look on PGMFI, OBD1 forum. There is a post with schematic snippets of the ECU, and the ECU already does signal inversion so you have an opportunity to grab both polarities of signal depending on where you intercept it.

I'm currently trying to find a couple dozen of the CS45x to play with, wish me luck. More than one way to skin the cat eh?