IC Ignitor WIP

So, I'm going to be starting on a big project involving the electrical system of my XJ650, and while I don't have much to offer (and wont for a few months when my schedule opens up) I did have a little something that I wanted to share, a WIP diagram for the TID14-06 that my XJ uses.

 

here is some info that may help you

http://www.xjcd.org/TCI_repair

http://www.xjcd.org/TCI_transistors

is it your plan to reproduce the pcboard or map out the schmatic

 

Very cool. Keep us posted as you further develop this. TCI is something I personally would love to address build priorities permitting.

 

Thanks for the links, I suspect that those will be quite helpful.
I will absolutely be mapping it out, however, I may stop short of reproducing the board. At the very least I want to document what I can so that sharper minds and steadier hands will have the information to make things happen.

I will keep you posted, it will likely be a few months before I have more than this to share.

 

there is 1 other thread witch is a few years old where someone was doing something like what you are doing if you can find it it may have info for you

http://www.xjbikes.com/forums/threads/tci-upgrade-options.93295/

 

A little more progress. Should be decapping the HA1825P chip sometime in the next while

 

There's a place in the US...I'm sure there's more than one, that will make you custom PCB boards, depending on your needs.....

 

Some additional progress, but it looks like ill need to get some help from my good friend Nitric acid in order to read the die. Also, its the HA1825P chip that is imaged there

 

Nice. Someone's got an X-ray.

I suspect the HA1825 is a semi-custom done by Hitachi for use only in their own ignition stuff, so I doubt you'll find stock anywhere. Hitachi was big in that sort of thing and the TCI isn't the only example I've run across.

So what's your plan for the laser trimmed sub-assy on the ceramic substrate?

 

I finished the acid bath moments after you posted your picture, but it had already been a long day at work, so I didn't have the patience to hang around and image it under a microscope today.
Ill get to that on Monday I suspect, which should provide all the information needed on the custom chip.

I don't yet have a plan for the sub-assembly. I am tackling this one hurdle at a time, and I am almost ready to start grappling with that. As my field is Chemistry and not Electrical Engineering I am doing a lot of learning as i go here. I mean... I get how electrons work.... but that is not as useful in this instance.

That disassembly is a work of ART! especially compared to my hack-job.

If you have the time and the inclination, I would love to have verification on the values of the resistors and diodes, but only if it wont put you out.

 

Not trying to diminish or dissuade your efforts in any way, but I'm not sure what you hope to learn by looking at the HA1825 die under the microscope. Just because you can see the die with your own two eyes will lend little information as to what exactly it does, or where to find a replacement. I mean... Sure it would be cool, and I'd love to have a look with you, but I don't think it's going to help a whole lot in the big picture.

So when you say things like "(viewing under a microscope) should provide all the information needed on the custom chip", I'm not sure what it is you hope to gain.

I've seen a number of various die under microscopes, and even reversed engineered (simple) sections of circuits on one or two of them, but there's no way I or anyone I know would be able to make heads or tails of what you're going to see inside that HA1825 unless they had a whole lot of other supporting documentation about the design. I'd love to be proven wrong though, so good luck!!

And as for other info I can help with, etc, I'd be glad to help in any way I can. I can help with any of the macroscopic design details (like component values) you have questions about. Ask away when you have questions.

 

I don't so much have the goal of making a replacement, if I can actually construct something that will replace the original ignitor unit, I would be over the moon.
I have little intention of learning exactly what the chip does, or where to find a replacement, but if I stumbled on that, I would be thrilled.
As mentioned, I am a chemist and not an electrical engineer. My hope is that the image of the die allows those with expertise in that field the information that they need. Maybe that is naive of me to think, but hey, Ill follow this rabbit hole as far as I can if I think it will help those down the road.

I have one, and only one goal here: compile as much information as I can

and 15 years of riding has taught me that sometimes the journey is its own reward.

 

Preliminary image of die from HA1825P

 

Awesome. Love the pic of the die.

And I'm with ya for learning anything about the original design, even if it's just academic. I'm happy to help where I can.

So as far as replacing with a similar function device, I believe there are a couple aftermarket options at this point. The basic concept is pretty simple... Two ins, and two outs. The only remotely complicated hurdles are timing advance and dwell. I believe the original design did everything in the analog realm with calibration and tuning adjusted by the laser trimmed sub-board. Today, you would throw a little microcontroller at it instead.

 

I made TCI for XJ few years ago. I will cooperate as I can if you have what want to ask for me.

 

An electrical schematic of the TCI would be wonderful, although I know that is wishful thinking. Or, perhaps just a data sheet on the HA1825P Integrated Circuit, or one less a good educated guess to what it is and does - quad Op Amp for frequency to voltage conversion to adjust timing, or something else?

 

Yu Tanaka, I remember your work with your replacement TCI. Neat stuff.

http://xjbikes.com/forums/threads/original-ignitor-tci.95931/

 

Rooster53, I don't think you'll find a datasheet for the HA1825P unless you have a connection inside Hitachi.

I believe I have a schematic around here somewhere for the TCI, but it's kinda worthless without knowing more details about that HA1825 part. It's the only "interesting" part of the whole assembly.

 

How are all of these other guys (Hyperpak, the guys in Czechoslavakia, etc.) making usable replacement TCI units with the original specs?

 

Thanks everyone for offers of assistance and interest. Sadly, disaster struck and the die is now missing, so the images I have will have to do for now until I get my hands on another chip. On the plus side, I do have a few images of the die that ill try to clean and share once I have something worth sharing.

 

Bummer. Slipped with the tweezers and it flew across the room? BTDT.

Just keep telling yourself that in the grand scheme of things, it was an academic study.

 

Chacal,

I haven't seen any of the replacement units with my own two eyes, but I assume they are functionally equivalent from a "black box" perspective. In other words... They operate like the original one does, but there's no confines as to HOW they accomplish the task. Inside the box, the entire concept is probably completely different than the stock unit, but when viewed as an assembly from outside the box (black box), it appears to operate the same as the original.

I assume all the aftermarket units use a little microcontroller and do the advance and dwell controls with software, while the original one did it with stone knives and bearskins (capacitors, oscillators, comparators, etc).

The original was analog, while the new ones are probably digital. An offshoot of Moore's Law.

 

Rice, you are probably correct, but how is/are those dwell and advance curves determined by the software programmers (unless they had access to proprietary data?). The published advance curves in the shop manuals seems to have far too wide of a range to be acceptable for use in a modern (or even a stone knives) environment.........

 

Well I'm no ignition expert, but if the graph in the FSM is the ONLY piece of info at my disposal, I would just aim for the center of the range at a couple points and naively hope for the best. Put it on a dyno and hope I don't hole a piston? My experience with old Japanese documentation is that barring typos, it's usually pretty accurate.

Other ideas?

I would run a reality check on the numbers and do a little research to see if they were at least "reasonable". For example, according to the timing curve from the SECA 550 FSM:

Does it make sense that the advance is flat at 5 degrees advance up to 1200 RPM.
And from there, the advance increases to 7 degrees at 1850 RPM and ramps from there to 36 degrees at 3300 RPM.
And according to the graph, the timing is all in at 38 degrees by 4000 RPM.

I would try to find myself someone who knew a lot about ignition and try to find out... Does all that seem at least reasonable? Does that all sound reasonable for a 550 cc four cylinder engine with a redline of 10K?

I would test one of the factory units and characterize how it performed. Simulating the engine rotation, I have already run the stock unit up and down in RPM on the bench, and measured the advance at a couple points. But my setup was crude and not particularly accurate. I'd be happy to do a more thorough job, but I need a nice digital storage scope. My old analog scope just isn't up to the task.

You could measure the advance amounts with the engine running on a real bike as well. Noisier and hotter than doing it on the bench, but certainly an option.

I would research around a little and see if I heard reports of people blowing up their engines from detonation or ignition timing too advanced. I've been (mostly lurking) around here for a couple years and I don't think I've ever seen or heard about anyone holing a piston or grenading their engine from a detonation issue, so I think we're good there.

I would also try to line up test mules and if I was planning to sell such devices, I would rent some dyno time. Not only could this be used for marketing fodder, but with the aid of a rectal O2 sensor and temp sensors on the exhaust pipes, they might also be able to determine how well the ignition was performing. Maybe.

 

No, you're no expert......! But you do know more than most of them.

 

Which way to the motorcycle forum? You guys are so over my head, it's like you're speaking a different language! Xrays, microscopes, chips flying around, this is the sci-fi channel!

 

What you're saying makes a lot of sense to me and I think your methodology is about the best way to go about it. But what do I know, I'm just a wrench turner.

 

this was to easy is this the part?
http://www.ic2ic.com/search.jsp?sSearchWord=HA1825P&prefix=H

or this
https://www.jotrin.com/product/parts/HA1825P

 

I had seen them for sale, but it is still difficult to tell if they are an exact match. The package (number of pins) might even be doubtful. And, the real interest is to determine what the part is, and not only is there not a valid link to a datasheet on those sites, but they don't even offer up a description of the part. I tried creating an account, but still no datasheet available. Looks like the account is mainly for purchasing the part.

 

thats why i thought it was to easy.

 

Correct, and reminding myself of that was the only way to stop the tears.

Yeah, just a question of if it is the same thing without any spec sheets. on the plus side, I have a reasonable idea now of what is on the inside, so I could at least decap one and give a preliminary report on how it appears to stack up.

 

I replaced mine with a B&G Microsquirt. I'm still using a stock pickup, but with a custom crank wheel. Nothing else between that pickup and the spark plugs is original, though.

However, since devices like the Microsquirt can be triggered by a stock crank pickup, I'm quite confident that the ignition logic, drivers, VR conditioner, etc, are all available that could be used to replace just a TCI as well (I'm doing fuel and ignition with the MS). The MS2Extra software is open source. The Microsquirt designs are also open source, and they have modules for some of them, if you don't want the whole assembly.

FWIW, I did replace the on-board VR conditioner in my Microsquirt with an NCV1124. That solved some high RPM problems that I think B&G finally fixed in a different way for themselves. Mine was one of the first batch of Microsquirts made, so there were some teething pains... Microsquirt uses VB921 drivers... at least mine does. I think these are NLA and have been replaced by BIP373. Probably the cheapest option would be if someone added an NCV1124, a pair of BIP373s, and a 7805 to one of these CHIP boards (https://getchip.com/pages/chip) and ported over the crank wheel and ignition logic from MS2Extra. Or maybe even use a Cypress PSOC as the MCU. Pretty sure those are still faster than the 68HC11 used in the Microsquirt.

Net net is that the single trigger that's used by the TCI should be OK to trigger a TCI replacement as well. My first design used a single tooth trigger. More teeth will allow the system to synch faster, (less cranking before the system knows what the engine speed is) and will allow better timing accuracy when engine speed is changing.

 

Haha! You give me way too much credit, my friend!

 

Macktion, Seriously... Been there, done that. Just the other day I was doing some very small electronics work, holding a part with the tweezers, and "snik" it was gone. I have no idea if it landed two feet away or twenty feet away. Looked a little bit and never found it.

So you're right. At least now we know what the die looks like, so if someone wants to try to buy some of those HA1825's and send you one... You can at least tell if they're legit!

 

SQLGuy, Sounds like a neat project. I know there are a couple aftermarket options out there like the microsquirt and jolt something.

I don't really know what I'm looking at, but back when Yu Tanaka did his first Tesla version, I took a quick look at the code, and it appeared that he was breaking up the RPM range into 500 RPM chunks (for example from 2000 to 2500) and using the range to adjust the timing conditions. At least that's how it appeared to me. Sort of quantizing the operating range into blocks to make the math easier?

Do you know if your system do something like that, or is it a continuous sliding calculation throughout the entire range? With a powerful enough processor, you could do more math.

 

It's based on tables that are usually 12x12 of RPM versus engine load, but it does interpolate intermediate values as well based on exact measurements. So, if you're at 1500 RPM, and your table points are 1200 and 2000, it will calculate an intermediate advance value. Factors like inlet air temperature can also be used to adjust advance.

A really simple version, though, could be done with a flat 12 bin table with no other factors feeding in... just advance based on RPM.

 

Gotcha. I believe you're talking about piecewise linearization (piecewise linear interpolation).

And now that I think about it, it's completely conceivable that the code I was looking at did that. I didn't really know what I was looking at in the first place, and didn't spend a lot of time trying to figure out exactly what was going on.

In any event, you can do that with just a small amount of processing power.

 

All modern microcontrollers have way more than enough power to do even full fuel and air with sequential injection, etc. Microsquirt has a 16-bit 24MHz CPU. The MPU in the Cypress PSoC 5 device I'm using for a completely different project right now is an 80Mhz 32-bit ARM device. And the development tools are really easy to work with. You can even do all sorts of modular digital and analog stuff on-chip, without CPU involvement.

 

You two need to get a private room together..............!

 

Chacal, is there a need/market for a TCI replacement? Something with adjustable timing tables and dwell? If so, what price point would it need to hit to be compelling?

 

Yeah, sorry.

The point of the latter part of this conversation is... There have been huge technological changes from when the original TCI was designed until now, and if you were to design an ignition module today, it would be microprocessor controlled. There are lots of choices for cheap, capable devices today that were not available then.

However, none of this should stop Macktion from his continued pursuit of understanding and exploration of the original module.

 

Would you mind sharing the schematic, I would like to look it over just as a learning experience of the 80's technology.

 

Still working on the Die, but this is what ive got so far for the board in the box

 

I will take a look and see what I can find. Been so long, I don't remember how complete (or incomplete) it was.

Let me see what I got.

 

Macktion, Nice work. Let me fill in a couple blanks, etc...

On the ceramic sub-board, everything you have identified as a capacitor, is correct. They are all what are known as "ceramic" capacitors.

However, the two things you have labeled as "varactor diodes" are also capacitors, even though they appear quite different. They are "electrolytic caps", and are probably "tantalum" capacitors. The reason they seem similar to diodes (with the polarity band) is that electrolytic caps like that, are in fact, polarized. The band probably indicates the positive terminal. One of them is a .47 uF cap, and the other is a .1 uF. Both of them carry a 35 volt maximum rating.

The L5 devices on that same side are most likely transistors. Sometimes they put diodes in that same package, and judging by the way the A3 and 62 devices are connected, I suspect that's what they are. They put little codes on them like A3 because the package was too small to put the real part number.

And on the other side of that ceramic board are a bunch of laser trimmed resistors. They silk screen the black rectangles down and then use a laser to burn across them until they reach the desired resistance. The more they burn away, the higher the resistance.

 

I did inquire of the Jotrin folk, and in spite of the quantity that they have listed as being in stock, according to their rep it is "unavailable from us". It might be the language barrier, but the wording makes it sound like it wont be available from them in the future either.

Splendid! I am not surprised at all that I mixed those up. Ill fix those for the future versions though.
(I kinda wish I could delete the old images though, since they might give someone the wrong idea).
I am looking forward to seeing your schematic, if you are able to find it.

 

I should ask my B in Law where he gets his printed circuit boards made ( he makes them for our chocolate machines runtime programs, but, who knows, ) ...I know there's a LOT more to it than just that, but........

 

hey, its not like it could hurt.

So, reached out to a few other sources, and still no data sheet (surprise surprise, but a guys gotta try). Havent made much progress on the die over the last week, but I am looking forward to some open time this week to dedicate to finishing it off.

 

OK, so some progress on the Die. Still working on it, but its been crazy at the lab lately, and I wanted to post some of the work

I am SOOOOO glad that I took that Xray, because it was very useful in orienting the chip as well as seeing which... nodes(?) go to which pins, and in turn to which pins on the board.

 

Call them "Pads" or "Bonding pads". They use thin gold wires ultrasonically (I think) welded at each end to connect the leadframe to the bonding pads on the die.

The bonding machine works similarly to a microscopic sewing machine. Stitching it's way around the die. One end - other end - cutoff. One end - other end - cutoff. One end - other end - cutoff.

Again, This may never lead to anything really useful in the long run, but I love it anyway. So, is the place you work at involved in microelectronics?